HONK KONG PHYSIOTHERAPY JOURNAL

Editorial Hong Kong Physiotherapy Journal Vol. 38, No. 2 (2018) i–iii DOI: 10.1142/S1013702518010023 Hong Kong Physiother. J. 2018.38:i-iii. Downloaded from www.worldscientific.com Hong Kong Physiotherapy Journal by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. http://www.worldscientific.com/worldscinet/hkpj The importance of developing evidence-based clinical examinations for low back pain Arnold Wong Department of Rehabilitation Sciences The Hong Kong Polytechnic University, Hong Kong [email protected] Accepted 22 November 2018; Published 11 December 2018 Low back pain (LBP) is the number one cause of further studies should be conducted to evaluate the years lived with disability in the world.1 Approxi- clinimetric properties of the recommended assess- mately 80% of people experience LBP at least once ment criteria. Since some clinical examinations in their lifetime and many of them remain to have initially thought to be useful for di®erentiating LBP at older ages.1–3 Despite the high prevalence di®erent patient subgroups for treatment alloca- of LBP, approximately 90% of LBP are labelled as tions may display suboptimal clinical values in non-speci¯c LBP because no clear etiologies can be some patient populations later,8,9 any clinical found.1 Given that medical imaging has limited examinations derived from the Delphi process must values in diagnosing patients with LBP, lumbar be evaluated thoroughly before applying them in imaging is recommended only when serious pa- clinical practice. thologies (e.g., malignancy, fracture, infection) are suspected.4 In this issue of Hong Kong Physiotherapy Journal, Vongsirinavarat and co-workers10 con- In order to prescribe treatments for patients ducted a single-group, repeated measures reliabili- with LBP, physical therapists need to rely on pa- ty study to evaluate the agreement of two tient history and clinical examinations to inform experienced physical therapists in using nine as- clinical decisions. In the absence of robust evidence sessment criteria derived from a Delphi study to or knowledge on the assessment criteria for iden- diagnose patients with lumbar facet joint pain in a tifying patients with di®erent underlying causes of clinical setting.7 Speci¯cally, the assessment crite- non-speci¯c LBP, a Delphi process is commonly ria include three subjective assessments (i.e., used to solicit expert opinions regarding the most localized unilateral pain, referred pain above knee, appropriate assessment criteria for classifying and no radicular pain), three movement tests (i.e., patients into di®erent subgroups for treatment pain reduction in °exion, pain in extension, and allocations.5–7 Following the Delphi process, pain in extension with side °exion and rotation Copyright@2018, Hong Kong Physiotherapy Association. Published by World Scienti¯c Publishing Co Pte Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). i

Hong Kong Physiother. J. 2018.38:i-iii. Downloaded from www.worldscientific.com ii A. Wong References by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. toward the same side of the painful facet joint), 1. Balague F, Mannion AF, Pellise F, Cedraschi C. and three manual assessments (i.e., replication or Non-speci¯c low back pain. Lancet 2012;379 aggravation of pain by local pressure over a facet (9814):482–91. joint or a transverse process, localized muscle spasm over the lumbar facet joint, and reduced passive 2. Wong A, Karppinen J, Samartzis D. Low back pain range of movement or increased sti®ness on the in older adults: Risk factors, management options painful facet joint during palpation). The results and future directions. Scoliosis Spinal Disord revealed fair to substantial agreements between the 2017;12:14. two physical therapists in using the nine criteria to distinguish patients with and without lumbar facet 3. Ogunlana MO, Odole AC, Adejumo A, Odunaiya joint pain. The suboptimal agreement on some N. Catastrophising, pain, and disability in patients criteria may be attributed to poor reliability of the with nonspeci¯c low back pain. Hong Kong palpation-based assessments,11,12 and/or lack of Physiother J 2015;33:73–79. detailed de¯nitions and training on some assessment criteria prior to the commencement of study. As 4. Darlow B, Foster BB, O'Sullivan K, et al. It is time such, the authors provided speci¯c pragmatic to stop causing harm with inappropriate imaging solutions and suggestions for future research. for low back pain. Br J Sports Med 2017;51:414–5. While a Delphi consensus methodology has been 5. Tomkins-Lane C, Melloh M, Lune J, et al., ISSLS commonly used to solicit consensus from experts prize winner: Consensus on clinical diagnosis of regarding the diagnosis or classi¯cation of patient lumbar spinal stenosis: Results of an international with di®erent medical conditions,5,13 the results Delphi study. Spine 2016;41:1239–46. derived from these Delphi studies may not neces- sarily be adopted by clinicians. It is partly because 6. Wong A, Lauridsen H, Samartzis D, Macedo L, there is no standard methodology to validate the Ferreira P, Ferreira M. Global consensus from new knowledge and to put the validated knowledge clinicians regarding low back pain outcome indi- into practice. To facilitate this knowledge transla- cators for older adults: Pairwise wiki survey using tion, a multistage methodology used for the vali- crowdsourcing. JMIR Rehabil Assist Technol. dation of clinical prediction rules may be [accepted]. adopted.14,15 In brief, the validation process includes: (1) a narrow validation (i.e., validation of 7. Wilde VE, Ford JJ, McMeeken JM. Indicators of the diagnostic criteria in one or two clinics), (2) a lumbar zygapophyseal joint pain: Survey of an broad validation (i.e., validation of the criteria in expert panel with the Delphi technique. Physic separate populations), and (3) an impact analysis Ther 2007;87(10):1348–61. (i.e., evaluation of the usefulness of the diagnostic criteria in improving patient outcomes, changing 8. Mitchell U, Hurrell J. Clinical spinal stability: clinicians' behavior, and/or reducing resource 10 years since the derivation of a clinical prediction consumption). The ¯rst two stages validate the rule. A narrative literature review. J Back Muscu- psychometric properties of the experts derived loskelet Rehabil 2018;[Epub ahead of print]. diagnostic criteria in di®erent patient populations or di®erent healthcare settings. The diagnostic 9. Kendell M, Beales D, O'Sullivan P, Rabey M, Hill criteria can be re¯ned during these two stages. The J, Smith A. The predictive ability of the STarT third stage aims to evaluate the impact of imple- Back Tool was limited in people with chronic low menting the diagnostic criteria on the diagnostic back pain: A prospective cohort study. J Physi- accuracy, medical costs, and patient satisfaction in other 2018;64:107–13. clinical practice.15 Collectively, while a Delphi consensus process can be used as an initial step to 10. Vongsirinavarat M, Wahyuddin W, Adisaiphaopan derive diagnostic criteria for a disease,16 a standard R. Agreement of clinical examination for low back validation process should be followed to ensure the pain with facet joint origin. Hong Kong Physiother validity of the suggested diagnostic criteria and to J 2018;38(2):125–31. evaluate the bene¯ts of using those criteria in im- proving clinical practice. 11. Strender LE, Sjoblom A, Sundell K, Ludwig R, Taube A. Interexaminer reliability in physical examination of patients with low back pain. Spine 1997;22(7):814–20. 12. Wong A, Kawchuk G. The clinical value of lumbar posterioanterior segmental sti®ness: A narrative review. PM&R 2017;9:816–30. 13. INSITE Collaborations (INTernational Study group for Identi¯cation and Treatment of Endo¯- brosis). Diagnosis and management of iliac artery endo¯brosis: Results of a Delphi consensus study. Eur J Vasc Endovasc Surg 2016;52:90–8.

The importance of developing evidence-based clinical examinations for LBP iii 14. Falk G and Fahey T. Clinical prediction rule. BMJ 16. Nair R, Aggarwal R, Khanna D. Methods of formal 2009;339:b2899. consensus in classi¯cation/diagnostic criteria and guideline development. Semin Arthritis Rheum 15. Childs JD and Cleland JA. Development and ap- 2011;41:95–105. plication of clinical prediction rules to improve decision making in physical therapist practice. Phys Ther 2006;86:121–31. Hong Kong Physiother. J. 2018.38:i-iii. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles.

Review Hong Kong Physiotherapy Journal Vol. 38, No. 2 (2018) 77–90 DOI: 10.1142/S1013702518300010 Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com Hong Kong Physiotherapy Journal by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. http://www.worldscientific.com/worldscinet/hkpj Musculoskeletal disorder and pain associated with smartphone use: A systematic review of biomechanical evidence Aitthanatt Chachris Eitivipart1,2,*, Sirinya Viriyarojanakul3 and Lucy Redhead4 1Faculty of Allied Health Sciences Department of Physical Therapy Chulalongkorn University, Bangkok, Thailand 2Discipline of Exercise and Sport Science, Faculty of Health Sciences The University of Sydney, Sydney, Australia 3Faculty of Physical Therapy Saint Louis College, Bangkok, Thailand 4School of Health Sciences, University of Brighton, Brighton, UK *[email protected] Received 17 January 2017; Accepted 24 October 2017; Published 14 August 2018 The number of smartphone users is growing dramatically. Using the smartphone frequently forces the users to adopt an awkward posture leading to an increased risk of musculoskeletal disorders and pain. The objective of this study is to conduct a systematic review of studies that assess the e®ect of smartphone use on musculo- skeletal disorders and pain. A systematic literature search of AMED, CINAHL, PubMed, Proquest, Scien- ceDirect using speci¯c keywords relating to smartphone, musculoskeletal disorders and pain was conducted. Reference lists of related papers were searched for additional studies. Methodological quality was assessed by two independent reviewers using the modi¯ed Downs and Black checklist. From 639 reports identi¯ed from electronic databases, 11 were eligible to include in the review. One paper was found from the list of references and added to the review. The quality scores were rated as moderate. The results show that muscle activity of upper trapezius, erector spinae and the neck extensor muscles are increased as well as head °exion angle, head *Corresponding author. Copyright@2018, Hong Kong Physiotherapy Association. Published by World Scienti¯c Publishing Co Pte Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 77

78 A. C. Eitivipart, S. Viriyarojanakul & L. Redhead tilt angle and forward head shifting which increased during the smartphone use. Also, smartphone use in a sitting position seems to cause more shift in head–neck angle than in a standing position. Smartphone usage may contribute to musculoskeletal disorders. The ¯ndings of the included papers should be interpreted carefully in light of the issues highlighted by the moderate-quality assessment scores. Keywords: Smartphone; musculoskeletal disorders; pain. Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com Introduction experimental studies and may draw a de¯nite by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. conclusion regarding the research that focuses on Smartphones now have a signi¯cant role in people's the changes in musculoskeletal symptoms caused everyday lives as they are being used for commu- by smartphone usage. nication, internet browsing and gaming. In the past decade, the rate of smartphone usage, hours and Methods frequency of use, has been increased.1,2 A study in 2012 revealed that there were more than six billion A search of the Cochrane Library and the data- smartphone users worldwide.3 Additionally, re- bases included in this review revealed no equivalent search reported that over 65% of the owners in the systematic review. This systematic review was USA spent at least 1 h per day on their phone.4 A planned and accomplished based on the Preferred survey supported this trend by reporting that users Reporting Items for Systematic Review and Meta- spend more than 20 h weekly on texting, emailing, Analysis (PRISMA) statement for reporting sys- and using social network, representing the signi¯- tematic review.15 cant dependence on smartphones for connecting and communicating with others.5 Consequently, Literature Search the heavy reliance on the smartphone may con- tribute to musculoskeletal injuries in the users. A comprehensive search was performed in May Therefore, health professionals should be aware of 2016 by two independent researchers (AE and SV) the e®ect of smartphone use on physical health of the following databases: AMED, CINAHL, problems. Generally, the typical posture when PubMed, ProQuest and ScienceDirect. There was using smartphones (or other touchscreen handheld no date restriction. The combination of terms and devices) involves holding the tool with one or two keywords used were (smartphone OR mobile phone hands below the eye level, looking down at the OR texting OR typing) AND (musculoskeletal dis- device and using the thumb to touch the screen.6 order OR pain) AND (ergonomic OR human fac- This pattern of use forces the user to adopt an tor). Handsearching of the reference lists of all awkward posture such as forward neck °exion relevant papers was performed. Only papers written which is often maintained for long periods.6–9 The in English were included. The inclusion criteria were prolonged and frequent use of smartphones, as well the following: (1) the studies must be laboratory as the repeated movement of the upper extremities experimental studies (pre-post, quasi-experimental, in an awkward posture, have been shown to be or cross-sectional study) so that the actual data the main contributing factors to the incidence relating to the change in di®erent musculoskeletal of musculoskeletal symptoms.7–9 Musculoskeletal symptoms due to the use of smartphone could be symptoms, such as discomfort and pain, in smart- tracked in an objective way; (2) the outcome must phone users not only occur in the neck but also in contain at least one of the following aspects: pain, other areas of the body including shoulders, elbows, postural analysis or muscle activity; (3) the assess- arms, wrists, hands, thumbs and ¯ngers.1,6,10–14 ments of the subjects must focus on the upper ex- tremities including neck, shoulder, elbow, wrist, While some research has been conducted to hand, thumb, ¯ngers, and upper back; and (4) the study the e®ect of smartphone use on the muscu- e®ects of smartphone use must be the main focus loskeletal symptoms of the neck and upper ex- in the research. Studies were excluded if (1) the tremity, there has not been a systematic review evaluating this research. The purpose of this study is to systematically review the evidence from

Musculoskeletal disorder and pain associated with smartphone use 79 Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com research recruited subjects aged under 18; (2) the or explanation whether the number of subjects was by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. studies focused on the use of a tablet, computer, appropriate). and other visual display units; and (3) the primary outcome of the research was from survey or quali- Two reviewers (AE and SV) independently tative methods. scored the quality of each study. Disagreements were resolved by consensus or by a third reviewer (LR). In addition to the recruiting criteria, there is no The possible range of reporting quality summary clear and well-accepted diagnostic criteria for the scores was 0 to 28. There is no formal cut-o® point to term of \\musculoskeletal disorders and pain\". separate the level of quality scores in the modi¯ed Therefore, this review was speci¯cally designed to Downs and Black checklist. Therefore, as recom- include the relevant papers where the participants mended by the previous reviews,20 Quality scores were recruited based on one of the following indi- above 19 were considered as \\good,\" between 11 cations: the participants identi¯ed themselves as and 19 as \\moderate,\" and below 11 as \\poor\". having musculoskeletal disorders and pain, having participant screening processes that were able Results to identify those people who were symptomatic with musculoskeletal disorders and pain, having Selection of the study objective measurements that included but were not limited to electromyography (EMG), muscle The °owchart in Fig. 1 illustrates the selection pro- strength or cross-sectional area of muscles that cess of the included studies. 639 reports were iden- could detect change in musculoskeletal functions ti¯ed from the electronic databases (AMED ¼ 64, (either in comparison to base-line measurement or CINAHL ¼ 265, PubMed ¼ 153, ProQuest ¼ 70 while performing the assigned task). and ScienceDirect ¼ 87). Of these publications, 609 were excluded due to an irrelevant title and abstract. Data Extraction and Management Duplications were also excluded, leaving 28 studies. The selection criteria of this systematic review The papers were initially screened and analyzed on were then applied and 17 more studies were titles and abstracts by independent reviewers (AE excluded.6,12,15,21–34 Following this selection pro- and SV). Where there was any doubt, the full text cess, 11 papers were eligible to be included in the was read to determine if inclusion criteria were review.35–45 Additionally, a reference search was met. Studies that failed to meet the selection cri- conducted using the reference lists of relevant teria were excluded. The data extraction form was papers to retrieve any missing references. Conse- applied from the PECO questions on population, quently, a paper written by Akkaya et al.46 was exposure, comparison, and outcomes.16 added to the review. Therefore, the total number of studies included in the review was 12.35–46 Methodological Quality Study characteristics There appears no validated checklist or scale The main characteristics of the 12 studies are available to assess the methodological quality of presented in Table 1.35–46 All the included studies the cross-sectional experimental laboratory studies were cross-sectional experimental laboratory stud- in the literature.17 Therefore, the Downs and Black ies, which provided data collected form a total of checklist18 was modi¯ed based on the previous 755 subjects. When considering the inclusion cri- studies19,20 and used to assess methodological quality teria for the studies, four papers used the term of the included studies. The modi¯ed Downs and \\university students,\" ðn ¼ 406Þ,35,37,39,42 three Black checklist was developed that all items were papers used the term \\healthy (normal) adult\" scored 0 to 1, except the item number 5 with a score 0 ðn ¼ 214Þ,36,44,46 four papers used the term \\young to 2 and the item number 27 that the score was adult\" ðn ¼ 125Þ38,40,41,45 and one paper speci¯- changed from a scale of 0 to 5 (unclear wording and cally included only right-handed female subjects in di±cult to score) to a scale of 0 to 1 (where 1 their study ðn ¼ 10Þ.43 was scored if a power calculation or sample size calculation was present while 0 was scored if there Considering the inclusion criteria quoted in was no power calculation, sample size calculation the papers, seven studies failed to provide a clear list of inclusion criteria.35–38,43,44,46 Whereas, three

80 A. C. Eitivipart, S. Viriyarojanakul & L. Redhead 639 potentially relevant studies were identified and screened for retrieval (AMED = 64, CINAHL = 265, PubMed = 153, ProQuest = 70 and ScienceDirect = 87) 609 studies were excluded on screening abstracts and titles for inclusion criteria 30 studies were retrieved for full text screening 2 duplicate studies were removed 28 full text studies were read for more detailed application of the criteria for inclusion Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com 11 studies were included [35-45] 17 studies were excluded because they failed by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. to meet the inclusion criteria -Electromagnetic study (N = 1) [21] -Case report (N = 2) [15, 22] -Performance measurement (N = 3) [23-25] -Survey study (N = 4) [6, 26-28] -Observational study (N = 1) [29] -Epidemiology study (N = 1) [30] - Self-report study (N = 1) [31] -Recommendation for practice (N = 1) [32] -No study on smartphone (N = 2) [12, 33] -Age of participants below 18 (N = 1) [34] 1 study was included from the references searching [46] 12 studies were eligible for the review [35-46] Fig. 1. Flowchart for the selection of studies. studies indicated the amount of experience with a ultrasound to evaluate the diameter of the thumb touch screen smartphone,40,41,45 one study speci¯- tendons (APL, EPB: extensor pollicis brevis, EPL: extensor pollisis longus, and FPL: °exor pollicis cally included only participants aged between 18 to longus). Three studies focused on the e®ect of head 29 years,39 one study used the term \\use mobile and neck positioning during smartphone use in phone regularly\" as an inclusion criteria.42 Only di®erent positions; sitting position (lap and desk one study by Xie et al.45 demonstrated well- posture)36; standing position (using and without using smartphone)42; and sitting versus standing constructed inclusion criteria with an intention posture while using smartphone.40 Another study used EMG to assess the neck (UT: upper trapezius) to recruit participants with similar characteristics. and thumb muscle (EPL and AbP: abductor pol- licis) activity in sitting to compare the muscle ac- For the exclusion criteria, 10 studies excluded tivity between one and two hands smartphone use.43 Four studies had a comparison group and of participants with experience of injury, trauma, these; two studies compared the range of motion (ROM)41 and muscular activity45 in neck pain and deformity, surgery and/or any neurological condi- non-pain groups; the other two studies compared the ROM between frequent and infrequent smart- tion that a®ected head, neck, and upper phone users.44,46 The study by Inal et al.37 had limbs.36–39,41–46 However, participants who had three groups for comparison (non, low, and high any physical di±culty were excluded in Lee et al.,40 smartphone user) and used the ultrasonographic assessment of the FPL muscle and the median but this term was not de¯ned. There was one study nerve. Another study with three-group comparison which did not indicate any exclusion criteria.35 compared pain threshold and the muscle activity during smartphone use, computer use and in a Regarding the study intervention, six studies control group.38 had no comparison group.35,36,39,40,42,43 Of these, two studies focused on the thumb area.35,39 Xiong and Murasaki35 used EMG to assess thumb per- formance and muscular activity of the thumb (AdP: adductor pollicis, FPB: °exor pollicis brevis, APB: abductor pollicis brevis, APL: abductor pollicis longus, FDI: ¯rst dorsal interosseous, and ED: extensor digitorum) while Eapen et al.39 used

Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Table 1. Characteristic of the methodology, outcome measurement and ¯nding included in the review. References Participants ðnÞ Exposure PECO Outcome Recruitment criteria Reported ¯nding Akkaya Comparison et al. Healthy adults Using VAS to measure Frequent users VAS, pinch strength, I: N/A FPL tendons were larger and (2015)46 (149; 36 male thumb pain during versus grip strength, ROM, E: Experience of injury of had more pain on the Eapen C and 86 female) texting and using ROM infrequent users and ultrasonographic fracture to nerve, vessel and texting side in frequent et al. to measure MCP joint evaluation of FPL tendon and rheumatic disease. texter group. (2014)39 and IP joint of thumb, tendon Guan X et al. grip strength, pinch (2015)42 strength, and iNAL EE et al. ultrasonographic (2015)37 evaluation of FPL tendon in di®erent groups. Students (98) Using ultrasound N/A Finklestein test, pinch I: Age 18 to 29 years Participants express evaluation on APL, EPB, strength and APL, E: Past experience of in°a- musculoskeletal-related EPL, FPL and Thenar symptoms such as eminence. EPB, EPL, FPL, and mative and degenerative of tenderness on extensor Musculoskeletal disorder and pain associated with smartphone use 81 neuromuscular conditions of compartments, positive Thenar eminence Finklestein test, pain on thumb, hand, and UE due abduction and extension of diameter to other activity instead of smartphone use. thumb and increased °uid around dorsal compartment. University Using photographic analysis Standing while Head tilt angle, neck tilt I: Using smartphone regularly. Head tilt angle and forward students (186; to measure sagittal using angle, forward head E: Experience of craniofacial, head posture were 105 male and 81 posture of head tilt angle, smartphone shift, and gaze angle cervical, shoulder, thoracic, signi¯cantly increased female) neck tilt angle, forward versus standing during mobile phone use and spine pain. head shift, and gaze angle while not using whereas neck tilt angle was during smartphone use in smartphone decreased. di®erent standing conditions. University Using self-report hand Non-users versus VAS, grip strength, I: N/A High smartphone users had students (102; function questionnaire, low-users versus pinch strength, E: Experience of neuropathy, signi¯cantly larger median 30 male and 72 clinical evaluation, and high-users median nerve ration, radiculopathy, previous con- nerve CSA, less pinch female) ultrasonographic and FPL ratio tracture, and lateral or medial strength, and hand function assessment to measure epicondylitis. in dominant hand. hand performances in di®erent groups.

Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Table 1. (Continued ) 82 A. C. Eitivipart, S. Viriyarojanakul & L. Redhead References Participants ðnÞ Exposure PECO Outcome Recruitment criteria Reported ¯nding Jung SI Comparison et al. Healthy adults Using craniovertebral angle, Frequent users Craniovertebral angle, I: N/A Long duration of smartphone (2016)44 (50) scapular index, and versus scapular index, FVC, E: Past experience of pain, use negatively a®ected the Kim GY et al. respiratory function infrequent users FEV1, ratio of FEV1/ trauma, fracture or surgery to posture and respiratory (2012)38 assessment to measure cervical, thoracic, and ab- function, especially peak body functions in FVC, and peak expiratory °ow. Kim MS di®erent groups. dominal area, neurological (2015)41 expiratory °ow disorders, lung function Lee M restriction, unstable cardiac et al. (2015)43 conditions, recently smoking or smoker free within ¯ve years. Young adults (40; Using pain pressure Smartphone users Pressure pain threshold I: N/A Smartphone users showed 17 male and 23 threshold and EMG to versus and EMG E: Past experience of injury, statistically signi¯cant computer users surgery or deformity of spine di®erent in brachioradialis female) measure dominant UT, versus control and UE, visual problems, diz- muscle fatigue while ziness, vertigo, neurological computer user shows brachioradialis, FCU, disorders and using sedative statistically signi¯cant drug within 48 h. di®erent in UT muscle and APB in di®erent groups. fatigue. Both experimental groups showed signi¯cant reduction in pressure pain threshold of UT muscle. Young adults (27; Using ROM to measure Pain versus Upper and lower cervical I: At least one year experience Neck °exion angle was 12 male and 15 cervical angle during control ROM of using smartphone. increased with time during E: Past experience of neck smartphone use both on female) two-hand texting in pain, spinal trauma, cervical upper and lower cervical surgery, ¯bromyalgia, and spines. Neck pain group was sitting position. systematic or connective tissue found to have greater angle. disorder. Right-handed Using EMG and One-handed EMG and pressure pain I: N/A One-handed smartphone use female (10) dolorimeter to measure smartphone use threshold E: Past experience of UE showed higher muscular muscle activity and versus two- ROM limitation and ortho- activity in UT, AbP, and tenderness in UT, EPL, handed pedic problems. EPL. and AbP during di®erent smartphone use conditions of smartphone use on thigh in sitting position.

Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Table 1. (Continued ) References Participants ðnÞ Exposure PECO Outcome Recruitment criteria Reported ¯nding Comparison Lee S Young adults (18; Using ROM to measure head Sitting position Head °exion angle I: At least one year experience Head °exion angle was the et al. (2015)40 9 male and 9 °exion angle during text versus standing of using smartphone. highest during text female) messaging, web browsing, position E: Physical di±culties of messaging in sitting. using smartphone. and video watching in di®erent position. Shin H & Healthy adults Using VAS, EMG, and Desk posture Flexion relaxation ratio, I: N/A Sustained smartphone use in (15) Kim K ROM to measure cervical versus lap ROM, and VAS E: Past experience of neck lap posture could in°uence (2014)36 erector spinae during posture pain, spinal trauma, and neck pain. cervical surgery. smartphone use in di®erent posture. Xie Y Young adults (40; Using EMG, discomfort Pain versus non- EMG, discomfort score, I: Right-handed users with Participants with neck– and rate of perceived et al. 16 male and 24 score, and borg scale to pain exertion similar texting speed who shoulder pain showed higher Musculoskeletal disorder and pain associated with smartphone use 83 (2016)45 female) measure on cervical spent at least 2 h daily using muscle activity in cervical smartphone for the last six erector spinae and UT erector spinae, UT, LT, months. muscle during texting and E: Past experience of pain, typing tasks. Unilateral ECR, ED, FDS, and trauma, fracture or surgery texting showed higher to cervical and UE, neuro- APB during smartphone and computer use in di®erent groups. logical and systematic muscle loading in forearm disorders. muscles when compared to bilateral texting. Xiong J & Right-handed Using pressure sensor and Small button Thumb performance, I: N/A Smaller button negatively iEMG, contraction E: N/A a®ects thumb performance. Murasaki S university EMG to measure thumb versus large time and iEMG/s (2014)35 students (20; 10 performance and button male and 10 muscular activity during female) smartphone use in di®erent button size and speed. Notes: APB: abductor pollicis brevis; APL: abductor pollicis longus; ECR: extensor carpi radialis; ED: extensor digitorum; EMG: electromyography; EPB: extensor pollicis brevis; EPL: extensor pollicis longus; FCU: °exor carpi ulnaris; FDS: °exor digitorum super¯cialis; FEV1: force expiratory volume at 1 s; FPL: °exor pollicis longus; FVC: force vital capacity; iEMG: integrated electromyography; LT: lower trapezius; N/A: non-applicable; n: number; ROM: range of motion; SAS: smartphone addiction scale; UT: upper trapezius; VAS: visual analog scale.

Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com 84 A. C. Eitivipart, S. Viriyarojanakul & L. Redhead the control group, there was a statistically signi¯- by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. cant decrease in the median frequencies of the In this systematic review, it is not possible to brachioradialis muscle ðp < 0:05Þ. Lee et al.43 dis- perform a meta-analysis due to the heterogeneity covered that the muscular activity of the UT, ELP of the study designs and outcome measures. and AbP muscle was signi¯cantly higher when using the smartphone in one hand than in two Methodological Quality hands ðp < 0:05Þ. Xie et al.45 found that partici- pants with neck and shoulder pain had signi¯- Table 2 presents the methodological quality results cantly higher muscular activity in the cervical from the modi¯ed Downs and Black checklist. All erector spinae and UT muscles than non-symptomatic studies35–46 included in this review were rated as participants when performing a texting and \\moderate\" (ranged from 11 to 18). All studies35–46 typing task. Xie et al.45 also found that one- failed to provide information about representa- hand texting produced signi¯cantly more muscle tiveness of the population and the intervention as activity of the forearm muscles than two-hand well as adverse events, subjects recruiting periods, texting. blinding (both subjects and assessors) and ran- domization (allocation and concealment). The Range of motion study by Xiong and Murasaki35 did not provide information about the participants' characteristics. Five studies used ROM of the head and neck or the Six studies37,38,41,44–46 partially reported informa- thumb and hand as an assessment to evaluate the tion regarding principal confounders. One study40 change in posture during and after the smartphone failed to report the descriptive statistics from the use.36,40–42,44 Shin and Kim36 found an average raw data percentiles was reported but not the change of 44 Æ 4:31 in ROM of cervical °exion in mean and standard deviation of the measured the lap posture when compared to the baseline variable and also their main confounders were not measurements. Lee et al.40 concluded that the investigated. The actual p-value of the main out- cervical °exion angle was signi¯cantly larger when comes (0.05 rather than < 0:05) was reported in text messaging than when carrying out the other eight studies.35–39,42,45,46 Six studies36–38,40,43,46 had tasks (web browsing and video watching) ðp < no information about source of population and 0:05Þ and signi¯cantly larger in sitting than in their recruitment processes. Compliance with the standing ðp < 0:05Þ. When using the smartphone intervention was not mentioned in six stud- in a sitting position, one study41 discovered that ies.35,36,39,40,42,43 Only a study by Akkaya46 pro- the upper and lower cervical °exion angles were vided a statement of recruitment period. All signi¯cantly higher in the neck pain group than in studies with the exception of one39 failed to con- the control group ðp < 0:05Þ. In addition, another duct a power calculation. study42 compared the head and neck posture in standing with and without looking at the smart- Findings phone. They found that participants who were standing and looking at the smartphone had sig- The outcome of the studies can be divided into ni¯cantly increased the head tilt angle and forward seven categories: EMG, ROM, Pain, ¯nger and head shift ðp < 0:05Þ while signi¯cantly decreased hand performance, tendon diameter, and subjec- the neck tilt angle ðp < 0:05Þ. Jung et al.44 also tive measures of discomfort and exertion. found that frequent smartphone users have higher scapular index and craniovertebral angle ðp < Electromyography 0:05Þ compared to infrequent smartphone users. Four studies used EMG to assess muscular activi- Pain ty.35,38,43,45 Comparing between smaller buttons and larger buttons, Xiong and Muraki35 found that Measures of pain were presented in ¯ve stud- using smaller buttons signi¯cantly increased the ies.36–38,43,46 Shin and Kim36 presented the change muscle activity of the FDI muscle ðp < 0:01Þ and of mean value measured using a visual analog scale signi¯cantly decreased the muscle activity of the (VAS) after using a smartphone in a desk and lap APB muscle ðp < 0:01Þ. Kim et al.38 found that posture from 0 (baseline measurement) to 1.7 and after a smartphone typing task, when compared to

Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Table 2. An assessment of methodological quality of studies assessed by modi¯ed Downs & Black checklist. Check list 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 References Study aim Main outcome Subject characteristics Description of intervention Principal confounders Outcome data Range of results Adverse events Lost to follow-up Probability value (exact) Source population Representative of population Sta®, place, facility Subjects blind to intervention Blind assessors Data dredging Same length of follow-up Appropriate statistical tests Compliance with the intervention Accurate outcome measure Control recruited same Recruitment at same time Randomized allocation Concealed randomization Adjustment for confounders Subjects lost to follow-up Power Total Akkaya Y Y Y Y P Y Y N Y Y N N N N N Y Y Y Y Y N Y N N Y Y 0 17/28 et al. (2015)46 Eapen C Y Y Y Y N Y Y N Y Y Y N N N N Y Y Y N Y Y N N N Y Y 1 17/28 et al. (2014)39 Musculoskeletal disorder and pain associated with smartphone use 85 Guan X Y Y Y Y N Y Y N Y Y Y N N N N Y Y Y N Y Y N N N Y Y 0 16/28 et al. (2015)42 iNAL EE Y Y Y Y P Y Y N Y Y N N N N N Y Y Y Y Y Y N N N Y Y 0 17/28 et al. (2015)37 Jung SI Y Y Y Y P Y Y N Y N Y N N N N Y Y Y Y Y Y N N N Y Y 0 17/28 et al. (2016)44 Kim GY Y Y Y Y P Y Y N Y Y N N N N N Y Y Y Y Y N N N N Y Y 0 16/28 et al. (2012)38 Kim MS Y Y Y Y P Y Y N Y N Y N N N N Y Y Y Y Y Y N N N Y Y 0 17/28 (2015)41 Lee M Y Y Y Y N Y Y N Y N N N N N N Y Y Y N Y N N N N Y Y 0 13/28 et al. (2015)43 Lee S Y Y Y Y N Y N N Y N N N N N N Y Y Y N Y N N N N N Y 0 11/28 et al. (2015)40 Shin H & Y Y Y Y N Y Y N Y Y N N N N N Y Y Y N Y N N N N Y Y 0 14/28 Kim K (2014)36 Y Y Y Y P Y Y N Y Y Y N N N N Y Y Y Y Y Y N N N Y Y 0 18/28 Xie Y et al. (2016)45 Xiong J & Y Y N Y N Y Y N Y Y Y N N N N Y Y Y N Y Y N N N Y Y 0 15/28 Murasaki S (2014)35 Notes: *Items 1 to 27 of the modi¯ed Downs & Black checklist. \\Y\": the answer is yes; \\N\": the answer is no; \\U\": the answer is unable to determine; \\P\": the answer is partial. The question number 5 will assign a score of \\0\" if the answer is \\No\", \\1\" if the answer is \\Partial\", and \\2\" if the answer is \\Yes\". The question number 27 will assign a score of \\0\" if no power calculation is provided, and \\1\" if a power calculation is provided. All the questions except the question numbers 5 and 27 will assign a score of \\0\" if the answer is \\No\" or \\Unable to determine\", and \\1\" if the answer is \\Yes\". Total quality scores of studies: Less than 11 ¼ poor 11–19 ¼ moderate; Higher than 19 ¼ good.

Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com 86 A. C. Eitivipart, S. Viriyarojanakul & L. Redhead Discomfort and exertion level by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. 5.2, respectively. Inal et al.37 found that frequent Only two studies investigated the discomfort and smartphone users had signi¯cantly higher VAS exertion level.35,45 One reported45 a signi¯cant scores than the infrequent and non-user groups change in the discomfort scores ðp ¼ 0:008Þ as well ðp < 0:05Þ but found no di®erence between non- as the rate of perceived exertion ðp < 0:001Þ after users and infrequent users. Two studies38,43 con- performing the texting task. This e®ect was greater cluded that the pain threshold of the UT muscle in the symptomatic group than in the control decreased signi¯cantly after smartphone use group. Another study35 reported that smaller ðp < 0:01Þ. Lee et al.43 also found that one-hand button size leads to a signi¯cantly higher rating of smartphone use signi¯cantly increased muscle perceived exertion (using the Borg scale) of the tenderness compared to two-hand use ðp < 0:01Þ. FDI muscle in the tapping task. Moreover, they Akkaya et al.46 showed a statistically signi¯cant found a signi¯cant decrease of perceived exertion di®erence ðp ¼ 0:005Þ in the VAS scores between score of the APB and APL muscles and a signi¯- the texting side ð0:3 Æ 0:9Þ and the contralateral cant increase of perceived exertion score of the FDI side ð0:01 Æ 0:1Þ in a frequent texter group. muscle in the moving task. Thumb–¯nger–hand performance Discussion Four studies assessed the performance of the This systematic review has provided informa- thumb, ¯nger, and hand.35,37,39,45 Xiong and tion about the change37–39,46 and associations Muraki35 indicated that using a small button leads with musculoskeletal symptoms35,36,40–45 in the to signi¯cant shorter fatigue times than when using neck, the shoulder, the upper limb, the hands and the a large button ðp < 0:01Þ in a tapping task, while thumb associated with smartphone use. The ¯ndings the tapping speed found to be signi¯cantly slower of all studies emphasized that the use of smartphone in °exion–extension than in abduction–adduction may contribute to the musculoskeletal symptoms. of the thumb during a moving task ðp < 0:01Þ. Inal et al.36 presented a correlation between pinch Methodological Quality of Studies strength and smartphone addition scale (SAS) (p ¼ 0:022, r ¼ À0:281; negatively weak correla- The methodological quality of the studies included tion), pinch strength and duration of smartphone in this review was scored as moderate. This may be use (p ¼ 0:288, r ¼ 0:133; weak correlation), and due to the nature of cross-sectional experimental pinch strength with Duruoz hand index score laboratory studies where blinding and randomiza- (p ¼ 0:014, r ¼ À0:242; negatively weak correla- tion are hard to implement.47 In addition, more tion). Eapen et al.39 reported the signi¯cant than half of the included studies35–39,41–46 simu- reduction in tip ðp ¼ 0:002Þ and lateral ðp ¼ 0:02Þ lated the smartphone use conditions for partici- pinch grip strength in patients with thumb pain pants to perform in the laboratory setting. while text messaging when compared to the control Accordingly, these data may not represent the group. actual smartphone use in real life and therefore the studies have low external validity.48 Half of the Tendon–nerve diameter studies35,36,39,40,42,43 were lacking information re- garding confounding variables, source of popula- Three studies evaluated the thickness of the tendon tion and how they were recruited which, therefore, and nerve in symptomatic39 and non-symptomatic exposing to high risk of selection bias (low internal smartphone users.37,46 Eapen et al.39 applied validity). The presence of low internal and external ultrasound evaluation to the thumb area of the validity resulted in some concerns about the symptomatic subjects and found °uid around applicability of the study results.48 Moreover, half the thumb tendons at the wrist level (19%) and in of the studies included in this review35,36,39,40,42,43 the °exor muscles of the thumb (2%). Two did not provide su±cient information in order to studies37,46 discovered that the frequent smart- e®ectively assess the comparability of the interven- phone users had signi¯cantly larger FPL tendons tion and comparison groups. This notion made it ðp ¼ 0:001Þ46 and median nerves ðp < 0:001Þ37 than the infrequent smartphone users.

Musculoskeletal disorder and pain associated with smartphone use 87 Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com di±cult to analyze whether the change and asso- in the shoulder and forearm area when using a by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. ciations with musculoskeletal symptoms found in smartphone.38,43,45 This is because the increase in the study groups really originated from smart- muscle activity is associated directly with the rise of phone use, or from other factors. Moreover, almost muscle fatigue52,53 and the reduction of pain pres- all studies included in this review did not attempt sure threshold.54,55 The repeated upper limb move- to address potential sources of bias.35–38,40–46 ments during smartphone use activate a continuous Finally, only one study39 mentioned that their muscle contraction which may cause microscopic sample size was based on data from the pilot study damage to the muscle which is the risk factor for while the rest of the studies35–38,40–46 did not musculoskeletal disorders.38,43,56 mention a power calculation. For the hand–thumb region, this review also Consequently, the study quality scores were found that one-handed smartphone use may cause moderate. However, the issues identi¯ed above more musculoskeletal symptoms in the shoulder– must be taken into account when interpreting the arm and the hand–thumb areas than using two results of the studies included in this review. hands to operate a smartphone.43–45 The reason is that two-handed smartphone use allowed more Overall Findings e®ective cooperation between holding and con- ducting the smartphone tasks which resulted in The studies included in this review35–46 reported improving the task performance and variation in their ¯nding in three speci¯c body regions: the movements.25,55 Thus, less muscle activity was head–neck, shoulder–arm, and hand–thumb. found in two-hand smartphone use when compare to one-hand smartphone use (less stereotypical The ¯ndings of this review suggest that using and repetitive movements).25,43–45 Consequently, smartphone may induce musculoskeletal symptoms to reduce the risk of musculoskeletal problems, in the neck.36,40–42,44,45 During smartphone use, the using two hands to operate a smartphone is muscle activity of UT, erector spinae and the neck recommended.25,43 extensor muscles are increased,43,45 especially for those who already have pain in the neck region.45 Furthermore, this review also revealed that the Moreover, many studies found that neck °exion angle, head tilt angle and forward head shifting frequent smartphone users had reduced thumb were increased during the smartphone use36,40–42,44 and also increased with the duration of smartphone performance when compared to the infrequent use.40,41 Many studies suggested that people with users,37,39 especially, when performing sensitive pain in the neck region tended to adopt a more tasks or tapping on a small button.35 Additionally, °exed posture than those who have no pain41,44,45 which negatively a®ected the neck posture.44 This this study detected changes in the tendon, nerve could be explained by the theory that the motor control of the neck muscles was altered by pro- and space between muscular tissue in frequent longed poor neck posture during the use of smart- smart phone users.37,39 Practically, smartphone phones.49,50 In addition, the variation of the head– neck angle could possibly depend on the task, the users naturally adjust their hand and thumb pos- posture and the way of holding the smartphone.6,40 The recent review concluded that smartphone use tures to ¯t with the phone layout which may alter in a sitting position seems to cause more shift in head–neck angle than in a standing position.36,40 A their e±ciency of smartphone use. The prolonged possible explanation is that postural stability is associated with the head position and movement in altered static posture and repetitive use of the standing, since neck °exion or extension in an upright posture in standing can alter the postural stability.51 wrist and thumb during smartphone operation Therefore, when the smartphone is used in a standing position, the user tends to minimize the alternations may negatively impact the muscular and nervous in neck posture to avoid postural instability.40 tissue in the hand.57 Excessive repetitive or static For the shoulder–arm region, muscle activity use of wrist and thumb movements during the increased and the pain pressure threshold decreased smartphone use can increase the load on the joints,1,6,57 increase carpal tunnel pressure,58 and decrease the space available for the median nerve to move.59 Thus, leading to the acute trauma and causing the enlargement of the median nerve59–62 and muscular tendon (e.g., FPL tendon).46 Accordingly, the structural changes from frequent smartphone usage may aggravate pain36,37,43,46 which was also reported more frequently in the group of frequent smartphone users than the group of infrequent smartphone users.

Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com 88 A. C. Eitivipart, S. Viriyarojanakul & L. Redhead Conclusion by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Limitations of the Review This systematic review revealed that the use of smartphones may contribute to the occurrence of This review was based on a comprehensive search clinical and subclinical musculoskeletal changes as of all the evidence that relates to the research well as associated factors in the head–neck, shoulder– question and adheres to the inclusion and exclusion arm, and hand–thumb areas. While there is a criteria set. However, there were some limitations strong case presented in the ¯ndings of all the to the data found. studies reported in this review, the evidence must be considered in the light of the moderate scores This review only included publications that from the modi¯ed Downs and Black checklist. were published in English, leading to missing evi- dence that has been published in other languages. Con°ict of Interest There may be some possibility of publication bias because all reports presented more positive out- All authors declare that they have no con°ict of comes on musculoskeletal change than null results interest. which may indicate overestimation of the positive outcomes. In addition, the power calculations were Funding/Support not reported and the research design and outcome measures were di®erent between studies. There are No funding was received for this systematic review. some issues that lower the quality of the included studies. Most studies were done on university stu- Author Contributions dents or young healthy adults. Consequently, the research cannot be generalized to people of all ages. AE is the leading reviewer who contributed to the Furthermore, inclusion and exclusion criteria were conception and design of the study. AE and SV not explicit enough to recruit participants with contributed to the development of the search similar characteristics and did not mention existing strategy, conducted the systematic search, extrac- poor postures or personal habits that might a®ect ted the data, and performed the data analysis. All the association between the use of smartphone and authors assisted with the interpretation, prepared measured parameters. Additionally, the gender the manuscript, drafted and revised the ¯nal issue has not been addressed. The intervention and paper. LR contributed to the proof reading of the task simulations designed may not represent the whole manuscript. All authors approved the ¯nal use of smartphones in real life as it appears that submitted version of the manuscript. short duration tasks and standardized posture were used in the laboratory setting. The model of References smartphones used in each study were di®erent and, moreover, the role of examiners in all studies was 1. Jonsson P, Johnson PW, Hagberg M, Forsman M. not clearly described and intra- and inter-rater Thumb joint movement and muscular activity reliability were not reported. during mobile phone texting — A methodological study. J Electromyogr Kinesiol 2011;21(2):363–70. Implication for Further Research 2. Goggin G. Cell phone culture: Mobile technology in Future primary research should use publication everyday life. New York: Routledge, 2012. guidelines, for example, CONSORT or STROBE, to improve the reporting quality and study design. 3. International Telecommunication Union. Measuring Research planning should focus initially on the the Information Society, 2012. Available at http:// issue of study quality and study validity. More www.itu.int/en/ITU-D/Statistics/Documents/pub- clinical trials with comparison groups are needed lications/mis2012/MIS2012 without Annex 4.pdf. to further improve the strength of the evidence and Accessed September 2016. to identify the most suitable method of asses- sing the musculoskeletal changes due to the use of 4. Khalaf S. Mobile use grows 115% in 2013, propelled smartphones. by messaging apps. Flurry Analytics 2014.

Musculoskeletal disorder and pain associated with smartphone use 89 Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com 5. Madge C, Meek J, Wellens J, Hooley T. Facebook, 19. Eng J, Teasell R, Miller W, et al. Spinal cord injury by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. social integration and informal learning at univer- rehabilitation evidence: Method of the SCIRE sity: `It is more for socialising and talking to friends systematic review. Top Spinal Cord Inj Rehabil about work than for actually doing work'. Learn 2007;13(1):1–10. Media Technol 2009;34(2):141–55. 20. Methajarunon P, Eitivipart C, Diver CJ, Foong- 6. Berolo S, Wells RP, Amick BC. Musculoskeletal chomcheay A. Systematic review of published symptoms among mobile hand-held device users studies on aquatic exercise for balance in patients and their relationship to device use: A preliminary with multiple sclerosis, Parkinson's disease, and study in a Canadian university population. Appl hemiplegia. Hong Kong Physiother J 2016;35:12–20. Ergon 2011;42(2):371–8. 21. Vecsei Z, Csathó A , Thuróczy G, Hernadi I. E®ect 7. Gold JE, Driban JB, Yingling VR, Komaro® E. of a single 30 min UMTS mobile phone-like expo- Characterization of posture and comfort in laptop sure on the thermal pain threshold of young healthy users in non-desk settings. Appl Ergon 2012;43 volunteers. Bioelectromagnetics 2013;34(7):530–41. (2):392–9. 22. Gilman L, Cage DN, Horn A, Bishop F, Klam WP, 8. Maniwa H, Kotani K, Suzuki S, Asao T. Changes in Doan AP. Tendon rupture associated with exces- posture of the upper extremity through the use of sive smartphone gaming. JAMA Intern Med various sizes of tablets and characters. In: Int Conf 2015;175(6):1048–9. Human Interface and the Management of Infor- mation. Berlin, Heidelberg: Springer, 2013;89–96. 23. Kietrys DM, Gerg MJ, Dropkin J, Gold JE. Mobile input device type, texting style and screen size 9. Bababekova Y, Rosen¯eld M, Hue JE, Huang RR. in°uence upper extremity and trapezius muscle Font size and viewing distance of handheld smart activity, and cervical posture while texting. Appl phones. Optom Vis Sci 2011;88(7):795–7. Ergon 2015;50:98–104. 10. Kwon M, Lee JY, Won WY, et al. Development 24. Lee J, Seo K. The comparison of cervical reposi- and validation of a smartphone addiction scale tioning errors according to smartphone addiction (SAS). PloS One 2013;8(2):e56936. grades. J Phys Ther Sci 2014;26(4):595–8. 11. Ra®le AE, Mackenzie EF. Management of cervical 25. Trudeau MB, Asakawa DS, Jindrich DL, Denner- dyskaryosis. No easy answer. BMJ 1994;309(6949):270. lein JT. Two-handed grip on a mobile phone a®ords greater thumb motor performance, decreased vari- 12. Gustafsson E, Johnson PW, Hagberg M. Thumb ability, and a more extended thumb posture than a postures and physical loads during mobile phone one-handed grip. Appl Ergon 2016;52:24–8. use — A comparison of young adults with and without musculoskeletal symptoms. J Electro- 26. Stalin P, Abraham SB, Kanimozhy K, Prasad RV, myogr Kinesiol 2010;20(1):127–35. Singh Z, Purty AJ. Mobile phone usage and its health e®ects among adults in a semi-urban area of 13. Barr AE, Barbe MF, Clark BD. Work-related mus- Southern India. J Clin Diagn Res 2016;10(1):LC14. culoskeletal disorders of the hand and wrist: Epide- miology, pathophysiology, and sensorimotor changes. 27. Kim HJ, Kim JS. The relationship between J Orthop Sports Phys Ther 2004;34(10):610–27. smartphone use and subjective musculoskeletal symptoms and university students. J Phys Ther Sci 14. Ming Z, Pietikainen S, Hänninen O. Excessive 2015;27(3):575. texting in pathophysiology of ¯rst carpometacarpal joint arthritis. Pathophysiology 2006;13(4):269–70. 28. Shan Z, Deng G, Li J, Li Y, Zhang Y, Zhao Q. Correlational analysis of neck/shoulder pain and low 15. Moher D, Liberati A, Tetzla® J, Altman DG. back pain with the use of digital products, physical Preferred reporting items for systematic reviews activity and psychological status among adolescents and meta-analyses: The PRISMA statement. Ann in Shanghai. PloS One 2013;8(10):e78109. Intern Med 2009;151(4):264–9. 29. Sharan D, Ajeesh PS. Risk factors and clinical 16. Mattos CT, Ruellas AC. Systematic review and meta- features of text message injuries. Work 2012;41 analysis: What are the implications in the clinical (Supplement 1):1145–8. practice? Dental Press J Orthod 2015;20(1):17–9. 30. Oftedal G, Wilen J, Sandstr€om M, Mild KH. 17. Zeng X, Zhang Y, Kwong JS, et al. The method- Symptoms experienced in connection with mobile ological quality assessment tools for preclinical and phone use. Occup Med 2000;50(4):237–45. clinical studies, systematic review and meta-analysis, and clinical practice guideline: A systematic review. 31. Korpinen LH, Pääkk€onen RJ. Self-report of physi- J Evid-Based Med 2015;8(1):2–10. cal symptoms associated with using mobile phones and other electrical devices. Bioelectromagnetics 18. Downs SH, Black N. The feasibility of creating a 2009;30(6):431–7. checklist for the assessment of the methodological quality both of randomized and non-randomised 32. Gustafsson E. Ergonomic recommendations when studies of health care interventions. J Epidemiol texting on mobile phones. Work 2012; 41(Supplement 1): Community Health 1998;52:377e84. 5705–6.

Hong Kong Physiother. J. 2018.38:77-90. Downloaded from www.worldscientific.com 90 A. C. Eitivipart, S. Viriyarojanakul & L. Redhead workers performing monotonous keyboard work — by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. 1: Neck and shoulder muscle recruitment patterns. 33. Gustafsson E, Johnson PW, Lindegård A, Hagberg Man Ther 2005;10(4):270–80. M. Technique, muscle activity and kinematic dif- 50. Szeto GP, Straker LM, O'Sullivan PB. A comparison ferences in young adults texting on mobile phones. of symptomatic and asymptomatic o±ce workers Ergonomics 2011;54(5):477–87. performing monotonous keyboard work — 2: Neck and shoulder kinematics. Man Ther 2005;10(4):281–91. 34. Kee IK, Byun JS, Jung JK, Choi JK. The presence 51. Buckley JG, Anand V, Scally A, Elliott DB. Does of altered craniocervical posture and mobility in head extension and °exion increase postural insta- smartphone-addicted teenagers with temporoman- bility in elderly subjects when visual information is dibular disorders. J Phys Ther Sci 2016;28(2):339. kept constant? Gait Posture 2005;21(1):59–64. 52. Enoka RM, Stuart DG. Neurobiology of muscle 35. Xiong J, Muraki S. An ergonomics study of thumb fatigue. J Appl Physiol 1992;72(5):1631–48. movements on smartphone touch screen. Ergo- 53. Allen DG, Lannergren J, Westerblad H. Muscle cell nomics 2014;57(6):943–55. function during prolonged activity: Cellular mechan- isms of fatigue. Exp Physiol 1995;80(4):497–528. 36. Shin H, Kim K. E®ects of cervical °exion on the 54. Persson AL, Hansson GÅ, Kalliomäki J, Moritz U, °exion-relaxation ratio during smartphone use. J Sj€olund BH. Pressure pain thresholds and electro- Phys Ther Sci 2014;26(12):1899. myographically de¯ned muscular fatigue induced by a muscular endurance test in normal women. 37. İNal EE, Çetİntürk A, Akg€onül M, Savaş S. E®ects Clin J Pain 2000;16(2):155–63. of smartphone overuse on hand function, pinch 55. Cook DB, O'Connor PJ, Eubanks SA, Smith JC, strength, and the median nerve. Muscle Nerve Lee MI. Naturally occurring muscle pain during 2015;52(2):183–8. exercise: Assessment and experimental evidence. Med Sci Sports Exerc 1997;29(8):999–1012. 38. Kim GY, Ahn CS, Jeon HW, Lee CR. E®ects of the 56. Haaland KY, Mutha PK, Rinehart JK, Daniels M, use of smartphones on pain and muscle fatigue in the Cushnyr B, Adair JC. Relationship between arm upper extremity. J Phys Ther Sci 2012;24(12):1255–8. usage and instrumental activities of daily living after unilateral stroke. Arch Phys Med Rehabil 39. Eapen C, Kumar B, Bhat AK, Venugopal A. Ex- 2012;93(11):1957–62. tensor pollicis longus injury in addition to De 57. Ko K, Kim HS, Woo JH. The study of muscle Quervain's with text messaging on mobile phones. fatigue and risks of musculoskeletal system disorders J Clin Diagn Res 2014;8(11):LC01. from text inputting on a smartphone. J Ergon Soc Korea 2013;32(3):273–8. 40. Lee S, Kang H, Shin G. Head °exion angle while 58. Trudeau MB, Young JG, Jindrich DL, Dennerlein using a smartphone. Ergonomics 2015;58(2):220–6. JT. Thumb motor performance varies with thumb and wrist posture during single-handed mobile 41. Kim MS. In°uence of neck pain on cervical move- phone use. J Biomech 2012;45(14):2349–54. ment in the sagittal plane during smartphone use. J 59. Gelberman RH, Hergenroeder PT, Hargens AR, Phys Ther Sci 2015;27(1):15. Lundborg GN, Akeson WH. The carpal tunnel syndrome. A study of carpal canal pressures. J 42. Guan X, Fan G, Wu X, et al. Photographic mea- Bone Joint Surg Am 1981;63(3):380–3. surement of head and cervical posture when view- 60. Bower JA, Stanisz GJ, Keir PJ. An MRI evaluation ing mobile phone: A pilot study. Eur Spine J of carpal tunnel dimensions in healthy wrists: 2015;24(12):2892–8. Implications for carpal tunnel syndrome. Clin Biomech 2006;21(8):816–25. 43. Lee M, Hong Y, Lee S, et al. The e®ects of smart- 61. Wieslander G, Norbäck D, Go€the CJ, Juhlin L. phone use on upper extremity muscle activity and Carpal tunnel syndrome (CTS) and exposure to pain threshold. J Phys Ther Sci 2015;27(6):1743. vibration, repetitive wrist movements, and heavy manual work: A case-referent study. Br J Ind Med 44. Jung SI, Lee NK, Kang KW, Kim K, Lee DY. The 1989;46(1):43–7. e®ect of smartphone usage time on posture and re- 62. Harris-Adamson C, Eisen EA, Kapellusch J et al. spiratory function. J Phys Ther Sci 2016;28(1):186. Biomechanical risk factors for carpal tunnel syn- drome: A pooled study of 2474 workers. Occup 45. Xie Y, Szeto GP, Dai J, Madeleine P. A comparison Environ Med 2015;72(1). of muscle activity in using touchscreen smartphone among young people with and without chronic neck–shoulder pain. Ergonomics 2016;59(1):61–72. 46. Akkaya N, Dogu B, Ünlü Z, et al. Ultrasonographic evaluation of the °exor pollicis longus tendon in frequent mobile phone texters. Am J Phys Med Rehabil 2015;94(6):444–8. 47. Mann CJ. Observational research methods. Research design II: Cohort, cross-sectional, and case-control studies. Emerg Med J 2003;20(1):54–60. 48. Pannucci CJ, Wilkins EG. Identifying and avoiding bias in research. Plast Reconstr Surg 2010;126(2):619. 49. Szeto GP, Straker LM, O'Sullivan PB. A compar- ison of symptomatic and asymptomatic o±ce

Review and Meta-analysis Hong Kong Physiotherapy Journal Vol. 38, No. 2 (2018) 91–114 DOI: 10.1142/S1013702518500087 Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Hong Kong Physiotherapy Journal by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. http://www.worldscientific.com/worldscinet/hkpj E®ectiveness of surgical versus conservative treatment for carpal tunnel syndrome: A systematic review, meta-analysis and qualitative analysis Diony Klokkari1 and Ioannis Mamais2,3,* 1Private Physiotherapist, MSc, Nicosia, Cyprus 2Department of Hygiene, Epidemiology and Medical Statistics Medical School, National and Kapodistrian University of Athens, Greece 3Department of Health Science & Department of Life Science European University of Cyprus, Cyprus *[email protected] Received 9 October 2016; Accepted 6 October 2017; Published 2 July 2018 Background: Carpal tunnel syndrome (CTS) is the most common entrapment neuropathy of the upper limb. Treatment options include physiotherapy, splinting, steroid injections or surgery. Objective: To compare the e®ectiveness of surgical versus conservative treatment for CTS for symptom and functional improvement and improvement of neurophysiological parameters. Methods: Systematic searches of PubMed and EBSCO host were conducted to identify the studies published between 1990 and 2016, comparing any surgical treatment to any conservative treatment. Participants were adults with a diagnosis of CTS, with symptom duration ranging from 8 months to 3 years. A meta-analysis and a qualitative analysis were conducted to summarize the results of the included studies and establish any agreement between the two. Results: A total of 15 studies were included in the study and 10 were included in the meta-analysis, involving 1787 wrists. The qualitative and quantitative analyses were consistent with the results of both indicating that surgical treatment leads to a greater improvement of symptoms at six months (mean di®erence: 0.52, 95%CI 0.27 to 0.78) and a greater improvement of neurophysiological parameters [distal motor latency (mean di®erence: 0.31, 95%CI 0.06 to 0.56), sensory nerve conduction velocity (mean di®erence: 3.71 m/s, 95%CI 1.94 to 5.49)]. At 3 months and 12 months, the results were not signi¯cant in favor of surgery or conservative treatment. *Corresponding author. Copyright@2018, Hong Kong Physiotherapy Association. Published by World Scienti¯c Publishing Co Pte Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 91

92 D. Klokkari & I. Mamais Conclusion: Conservative treatment for CTS should be preferred for mild and short-term CTS. Surgery is more e®ective than conservative in CTS, and should be considered in persisting symptoms, taking into account the complications, which are more severe after surgery. Further research should focus on the ¯eld of manual therapy and compare it to surgical treatment for CTS. Keywords: Carpal tunnel syndrome; median nerve entrapment; surgical treatment; conservative treatment; systematic review; meta-analysis. Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Introduction parameters than conservative treatment, both short by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. term (3 months) and long term (12 months). An Carpal tunnel syndrome (CTS) remains more additional aim was to establish the extent of agree- common among non-computer-related jobs,1 de- ment between a qualitative analysis of the studies spite the increasing usage of computers in recent and a meta-analysis. years, a®ecting approximately 1–5% of the general population2–4 and approximately 34% of hospital Material and Methods workers.5 CTS is a result of compression of the median nerve at the wrist due to the con¯ned space Identi¯cation and selection of the carpal tunnel which restricts the movement of literature of the tissues. Any increase in internal or external canal pressure results in neurological impairment Systematic searches of PubMed and EBSCO host with numbness and tingling along the distribution were conducted to identify studies written in En- of the nerve. CTS often results from repetitive glish and published between 1990 and 2016 (see strain in manual jobs but it is also associated with Appendix A for full search strategy). Manual conditions like rheumatoid arthritis,6 pregnancy searches of reference lists from previous studies due to water retention,7 and diabetes mellitus, were also conducted to ensure that all relevant which increases the likelihood of a symptomatic studies were captured. Studies eligible for inclusion response in an already compressed nerve.8 Con- were randomized controlled trials, clinical trials servative treatment with physiotherapy, wrist (CTs), prospective and retrospective studies com- splints, corticosteroid injections, diuretics, vitamin paring any surgical intervention to any conserva- B6 is proposed initially, whereas surgical treatment tive intervention for CTS patients. The outcome is reserved for more severe cases with thenar measures short-term and long-term improvement muscle atrophy or after failure of conservative of symptoms, functional status and improvement treatment.9 of neurophysiological parameters. Case control studies were excluded. Studies comparing surgical Previous systematic reviews have reported an interventions and studies comparing conservative advantage of surgery over conservative treatment interventions were also excluded as the aim was for symptom functional improvement.10,11 These to compare the relative e®ectiveness of the two reviews were published in the last decade. More interventions. recent trials have been published since, some of which contradict the results suggested in these Two independent investigators (KD and MI) reviews. Given a substantial number of recent screened the titles and read the abstracts and the studies, an updated systematic review is justi¯ed to relevant papers were obtained in full text to assess incorporate data brought to light since 2011. Fur- further eligibility. Data extraction was performed thermore, this review includes both a meta-analysis by the two investigators independently. The fol- and a qualitative analysis to formulate conclusions lowing data were extracted: authors, year of regarding the relative e®ectiveness of the publication, study design, description of the sam- interventions. ple, description of the surgical and conservative intervention, duration of study, study outcomes, Therefore, this review sought to investigate assessment times, study results and study whether surgical treatment for CTS can lead to greater conclusions. symptom improvement, greater functional improve- ment and greater improvement of neurophysiological

E®ectiveness of surgical treatment versus conservative treatment in CTS 93 Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Quality assessment Participants by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. The methodological quality of the randomized Studies involving patients diagnosed with CTS controlled trials and CTs included in this system- irrespective of the cause, the way it was diagnosed, atic review were assessed using the CBRG meth- other associated conditions, the age or the sex of odological criteria scale proposed by van Tulder the person. Patient characteristics such as age, sex, et al.12 This scale consists of 11 items and is similar duration of symptoms were recorded in order to to the PEDro scale, which has good levels of va- assess heterogeneity between studies. lidity and reliability,13,14 but with clearer oper- ationalization criteria. It addresses the internal Intervention validity of the studies in order to minimize the risk of systematic bias (selection bias, performance The included studies compared any surgical inter- bias, attrition bias and performance bias). Items vention such as open carpal tunnel release (OCTR) relate to the adequacy of the randomization, or endoscopic carpal tunnel release (ECTR) to any treatment allocation concealment, baseline simi- conservative intervention such as steroid injec- larity of treatment groups, patient, treatment tions, wrist splints, physiotherapy with electro- provider and assessor blinding, similarity of co- therapy, exercise or manual therapy or a interventions, adequacy of compliance, adequacy combination of di®erent modalities. and description of the dropout rate, similar as- sessment timing across groups and analysis Outcome measures according to intention-to-treat. Each item is ac- companied by a strict list and was evaluated with a The primary outcome was the patient self-reported \\yes\", \\no\", or \\unclear\" (if it did not apply or if it improvement in symptoms and function measured was not mentioned). Each positive answer scored 1 using the symptom severity scale and functional point. The studies were regarded as high quality if status scale of the Boston questionnaire (BQ). the total score of positive answers in the criteria list Secondary outcome measures used to evaluate the was seven and above. In consequence, if the total e®ectiveness of the intervention were improvement score was below seven, studies were regarded as low of neurophysiological parameters measured using quality. electrodiagnostic studies, and side e®ects reported. The methodological quality of the prospective Data analysis and retrospective studies was assessed using the methodological criteria scale proposed by Moga Data extraction was performed by one investigator et al.15 using a modi¯ed Delphi technique. This (KD) and cross-checked by another (MI). Data scale is made up of 18 items addressing methodo- were documented on a customized table in order to logical quality (i.e., patient characteristics, ade- compare patient demographics, parameters of in- quacy of eligibility criteria, adequacy of tervention, duration and outcome measures of each intervention, similarity of co-interventions, rele- study post-intervention. Where su±cient informa- vancy and timing of outcome measures) and sta- tion was obtainable and the outcome measures tistical reporting (i.e., suitability of statistical were comparable, meta-analyses were performed, tests, length and loss of follow-up, random vari- allowing a quantitative analysis of the studies. The ability, adverse e®ect and competing interest pooled estimations regarding outcomes were reporting). Each item was accompanied by a strict expressed as dichotomous or as continuous vari- list and was evaluated with a \\yes\" or \\no\" and ables. These were calculated using a random e®ect each positive answer scored 1 point. The studies model or a ¯xed e®ect model. For dichotomous were regarded as high quality if the total score of data, the pooled odds ratio (OR) was calculated. positive answers was 14 and above. If the total The pooled mean di®erence was estimated to assess score was below 14, then the studies were regarded continuous data. Statistical analyses were per- as low quality. The methodological quality result formed using the Review Manager (RevMan) was used for the formulation of conclusions in the Version 5.0 software (The Nordice Cochrane qualitative analysis, which is described further Center, The Cochrane Collaboration, Copenhagen, below.

94 D. Klokkari & I. Mamais Box 1. Synthesis of results for the qualitative analysis. Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Denmark, 2008) and STATA Version 13. P < 0:05 Strong Consistent ¯ndings among two or by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. was considered signi¯cant. Moderate more, high quality RCTs Heterogeneity analysis Limited Consistent ¯ndings among one high quality RCT and one or more low The existence of statistical heterogeneity between Con°icting quality RCTs and/or CTs or one the included studies was assessed using the I 2 test. No evidence high quality observational study The heterogeneity was considered low, moderate or high if the I 2 was 25%, 50% or > 75%, respec- from studies Consistent ¯ndings from one high tively.16 If p-value was less than 0.05, the random quality RCT, or one low quality e®ect model was adopted or vice versa. The be- RCT or CT, or one high quality tween-trial heterogeneity was assessed using the Q observational study test and the I 2 statistic.17 Subgroup analyses by month were conducted in order to explore potential Inconsistent ¯ndings among multiple sources of the between-trial heterogeneity and studies (RCTs, CTs and/or potential e®ect modi¯ers in this study. observational studies) No studies found Publication bias assessment criteria and were eligible for data analysis. Of these, nine were Randomized CTs (RCTs), two To assess asymmetry, funnel plots were formulat- were CTs, two were prospective studies and two ed. The Begg's rank correlation and Egger's linear were retrospective studies. The meta-analysis only regression tests were used to detect potential publi- included 1017–26 of these studies with similar out- cation bias.18 A two-tailed p-value < 0:10 for Egger come measures, whereas all the studies were in- regression indicated the presence of publication bias. cluded in the qualitative analysis. The °ow of studies through the selection process is presented Sensitivity analyses in Fig. 1. (see Appendix B for excluded papers). A summary of the studies is presented in Table 1. The in°uence of individual studies, from which the meta-analysis estimates are derived, was examined Study characteristics by omitting low quality studies to see the extent to which inferences depend on a particular study or Quality group of studies (sensitivity analysis). The methodological quality of the eligible studies Qualitative analysis was low to moderate. There was a mean CBRG score of 5.1 out of 11 for the 9 RCTs and the 2 CTs. Based on the methodological quality score of each CTs had a lower quality score (2/11) due to lack of study, a qualitative analysis was also performed to randomization, whereas the maximum score in formulate conclusions thus allowing a wider inclu- RCTs was 8/11. Due to the nature of the inter- sion of studies. This was done using the Best Evi- vention (surgery), neither the patient nor the dence Synthesis,12 which was modi¯ed to include therapist could be blinded (criteria 4 and 5), so the observational studies. This method consists of ¯ve highest score that could be expected was 9/11. levels of scienti¯c evidence, presented in Box 1. Only 4 out of 11 trials had 80% retention rates for Consistency was de¯ned a priori at 60% (i.e., if short-term follow-up and 70% for long-tem follow- 60% or more of studies agreed in the same direction up and compliance was not clearly stated in the of results). studies with the exception of one.27 There was a mean score of 13 out of 18 of the criteria proposed Results by Moga et al.15 for 2 prospective and the 2 ret- rospective studies (range 11 to 17). Quality scores Selection of studies are presented in Tables 2 and 3. From the search strategy, 459 potentially relevant Participants studies were identi¯ed and 252 duplicate paper were removed after checking titles and abstracts. The review included 1787 wrists with a clinical di- Out of these studies, 15 ful¯lled the inclusion agnosis of CTS. The sample sizes of the 15 eligible

E®ectiveness of surgical treatment versus conservative treatment in CTS 95 PRISMA 2009 Flow Diagram IdenƟficaƟon Records idenƟfied through EBSCO AddiƟonal records idenƟfied (n = 387) through PubMed (n = 72) Records aŌer duplicates removed (n = 207) Screening Records screened Records excluded (Titles/ (n = 207) Abstracts (n = 186) Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Eligibility Full-text arƟcles assessed Full-text arƟcles excluded, with by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. for eligibility reasons (n = 21) (n = 6)* Studies included in Irrelevant outcome measures (carpal qualitaƟve synthesis tunnel size and cost-effecƟveness study) (n=4) (n = 15) Case control study (n=1) Studies included in quanƟtaƟve synthesis Study protocol- No outcome reported (n=1) (meta-analysis) Included (n = 10) Fig. 1. Selection of studies for inclusion in the systematic review. *Papers may have been excluded for failing to meet more than one inclusion criteria. studies ranged from 40 to 429 participants. Parti- which comprises of the Symptom Severity Scale (11 cipants were mostly females (79%) and one study28 items) and the Functional Status Scale (8 items). included only women. The weighted mean age in The BQ is a self-administered CTS-speci¯c tool the studies included was 48.8 years ranging from 20 measuring symptom severity and functional status years old19 to 88.5 years old29 although one study on a scale of 1 to 5, where 1 ¼ no symptoms or no concerned an elderly population,29 a®ecting the di±culty and 5 ¼ very severe symptoms or so dif- overall mean age. The duration of symptoms ran- ¯cult and could not do activity) and the overall ged from 8 months20 to 3 years.27 score is the mean of all items out of 5. High scores are indicative of more severe symptoms or func- Intervention tional limitation. The BQ's validity and reliability have been previously assessed.32 One study27 used All eligible studies compared surgery (OCTR or the CTS assessment questionnaire (CTSAQ) to ECTR) to a conservative intervention. Conserva- assess the symptoms (11 item scale) and functional tive interventions involved the use of steroid status (9 item scale) on a 1–5 scale similar to the injections in six studies,20–25 splinting in two BQ. Symptom severity was assessed with the studies,22,26 Low Level Laser Therapy (LLLT) in Global symptom score (GSS) in two studies.24,25 one study,30 manual physical therapy in one This scoring system rates symptoms on a scale of 0 study28 and multimodality in four studies.19,27,29,31 (no symptoms) to 10 (severe) in ¯ve categories: The treatment period in the CTs ranged from a pain, numbness, paresthesia, weakness/clumsiness single application to three months. and nocturnal awakening, and the result is the sum of the scores out of 50. Three publications of the Outcome measures same study20,33,34 assessed the functional status on a 100 mm Visual Analogue Scale (VAS), where Symptom severity and functional status were 0 ¼ no functional impairment and 100 ¼ the most assessed using the BQ in six studies,22,23,26,28,29

Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Table 1. Characteristics of included studies (n ¼ 15). 96 D. Klokkari & I. Mamais Study Design Participants Intervention Outcome measures Ref. 19 Retrospective Incl ¼ Clinical diagnosis of CTS con¯rmed by EDS A ¼ OCTR or ECTR (95 wrists/77 . Patient satisfaction ¼ Subjective Ref. 21 Prospective n ¼ 265 patients patients) . Repeat history, physical examination Ref. 31 Prospective Sex ¼ 43%F/57%M (114F/151M) Ref. 23 CT Age [ (range)] ¼ 45 (20–90) yrs; Gp A: 49 yrs; B ¼ Multimodality (patient education, and nerve conduction studies ¼ Objec- wrist splinting, vitamin B, NSAIDs, tive ¯ndings Gp B: 42 yrs steroid injections and job change or . Follow-up ¼ 3 to 9 month intervals modi¯cation) (188 patients) (for group B) Incl ¼ Clinical diagnosis of CTS con¯rmed by EDS n ¼ 125 wrists All groups ¼ educational videotape and . Improvement of neurophysiological Sex ¼ Gp A: 76%F/24%M; Gp B: 81%F/19%M; brochure, wrist splints, 100 mg Vit B6 parameters for median nerve: SNCV and 50 mg B2 daily and NSAIDs. GP C: 79%F/21%M . Follow-up ¼ Month: 1, 6 and 12 Age [median (range)] ¼ Gp A: 57.5 (30–88) yrs; A ¼ OCTR or ECTR (33 wrists) B ¼ 1–3 doses steroid injections . Symptoms ¼ SSS — 11 items (5-point Gp B: 58.6 (28–87) yrs; Gp C: 57.6 (28–87) yrs scale) Symptom duration ðÞ ¼ Gp A: 23.3 mo; (56 wrists) C ¼ No local treatment. NSAIDs or . Function ¼ FSS — 8 items (5-point Gp B: 22.9 mo; Gp C: 20.9 mo scale) vascular drugs (36 wrists) Incl ¼ Clinical diagnosis of CTS > 1 mo . Satisfaction ¼ satisfaction scale 7 items n ¼ 429 patients A1, A2 and A3 ¼ OCTR or ECTR (270 (5-point scale) Sex ¼ Gp A: 191F/79M; Gp B: 102F/23M patients) Age ( Æ SD) ¼ (3 subgroups) ( Æ SD): A1: . Health status ¼ SF-36: 36 questions B1, B2 and B3 ¼ Multimodality (125 assessing general health-related quality 68.0 Æ 9.1 yrs; A2: 42.0 Æ 7.3 yrs; A3: 39.0 Æ 8.1 patients) [NSAIDs (96), wrist splints of life yrs; B1: 64.0 Æ 7.0 yrs; B2: 41.0 Æ 8.9 yrs; B3: (115), physiotherapy (42), work 37.0 Æ 8.8 yrs modi¯cation (70), steroid injections . Follow-up ¼ Month: 6, 18 and 30 (48), Vitamin B6 (13)] Incl ¼ Clinical diagnosis of unilateral CTS > 6 mo, . Symptoms ¼ BQ — 11 items (5-point con¯rmed by EDS A ¼ OCTR (44 patients) scale) B ¼ 2 doses steroid injections 2 wks n ¼ 90 patients . Function ¼ BQ — 8 items (5-point scale) Sex ¼ Gp A: 38F/6M; Gp B: 42F/4M apart of 6.4 mg betamethasone (46 . Improvement of neurophysiological Age ( Æ SD) ¼ Gp A: 48.0 Æ 8.4 yrs; patients) 23-gauge needle at anterior wrist °exion crease, angulation 45 parameters for median and ulnar Gp B: 45.3 Æ 9.9 yrs distally and 45 radially nerves ¼ DML, MNCV, CMAP, DSL, SNCV, SNAP using the Nihon Kohden- Neuropack MEB 5504 K . Follow-up ¼ Month: 3 and 6

Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Table 1. (Continued ) Study Design Participants Intervention Outcome measures Ref. 26 RCT Incl ¼ Pain or paraisthesias in the median nerve A ¼ OCTR (87) . General improvement ¼ 6-point ordinal distribution, clinical diagnosis of CTS, con¯rmed B ¼ Splinting in neutral position of wrist transition scale (success ¼ \\completely Ref. 24 RCT by EDS, age > 18 yrs, ability to ¯ll out recovered\" or \\much improved\") E®ectiveness of surgical treatment versus conservative treatment in CTS 97 Ref. 20 RCT questionnaires in Dutch for 6 wks min at night and during the day as preferred (89) . No. of nights the patient woke due to n ¼ 176 symptoms in the last week Sex ¼ Gp A: 66F/11M; Gp B: 77F/2M A ¼ OCTR (25 patients) Age ( Æ SD) ¼ Gp A: 49 Æ 11; Gp B: 49 Æ 12 B ¼ 1 dose steroid injection 15 mg . Symptoms ¼ 11-point scale Symptom Duration [median (IQR)] ¼ Gp A: 40 . Symptoms ¼ SSS — 11 items (5-point methylprednisolone acetate with (16–104) wks; Gp B: 52 (24–104) wks 25-gauge needle 30 angle medial to scale) Characteristics ¼ small heterogeneity in sex and Palmaris longus tendon (25 patients) . Function ¼ FSS — 8 items (5-point symptom duration scale) . Severity of CTS-related complaints- Incl ¼ Clinical diagnosis of CTS > 3 mo but < 1 yr con¯rmed by EDS Physiotherapist assessment on 11-point scale n ¼ 50 patients . Improvement of neurophysiological Sex ¼ 48F/2M; Gp A: 24F/1M; Gp B: 24F/1M parameters for median nerve ¼ DML, Age ( Æ SD) ¼ Gp A: 50.8 Æ 11.6 yrs; Gp B: DSL, Median-Ulnar DSL di®erence . Follow-up ¼ Month: 3, 6 and 12 48.2 Æ 6.5 yrs Characteristics ¼ no large heterogeneity in age, . Symptoms ¼ GSS — 11-point scale for ¯ve symptom categories, total score/50 symptom severity, baseline measurements . Improvement of neurophysiological parameters for median nerve ¼ DML, SNCV . Grip strength ¼ JAMAR hydraulic hand dynamometer . Follow-up ¼ week: 6 and 20 Incl ¼ Age ! 18 yrs, Clinical diagnosis of CTS A ¼ OCTR (80 wrists) . Nocturnal paresthesias ¼ 100 mm VAS > 3 mo, con¯rmed by EDS, consecutive referral B ¼ 1 or 2 doses steroid injection 20 mg . Diurnal pain ¼ 100 mm VAS and unresponsiveness to ! 2 wks NSAIDs . Functional impairment ¼ 100 mm VAS treatment in 1 mL paramethasone acetonide . Follow-up ¼ month: 3, 6 and 12 with 22-gauge needle 45 angle n ¼ 163 wrists/101 patients Sex: 93F/8M distally 1–2 cm depth medial to Age ( Æ SD) ¼ Gp A: 50.53 Æ 10.87 yrs; Gp B: Palmaris longus tendon (83 wrists) 53.17 Æ 13.93 yrs (p ¼ 0:213) Symptom duration ( Æ SD): Gp A: 31.12 Æ 7.27 wks; Gp B: 33.25 Æ 8.17 wks (p ¼ 0:723)

Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Table 1. (Continued ) 98 D. Klokkari & I. Mamais Study Design Participants Intervention Outcome measures Ref. 29 Retrospective Incl ¼ Age ! 70 yrs, Recent clinical diagnosis of CTS, A ¼ OCTR or ECTR (47 wrists) . Symptoms ¼ BQ — 11 items (5-point Ref. 22 RCT with or without normal EDS, county residents B ¼ Multimodality (NSAIDs, splinting, scale) Ref. 30 CT n ¼ 96 wrists/60 patients steroid injections) (41 wrists) . Function ¼ BQ — 8 items (5-point scale) Sex ¼ 43F/27M . Patient expectations and satisfac- Age [ (range)]: 77.0 (70.2–88.5) yrs A ¼ OCTR (11) Characteristics ¼ no large heterogeneity in age, B ¼ Splinting (23) in neutral position of tion ¼ MODEMSQ — Musculoskeletal Outcome Data Evaluation and Manage- symptom duration or symptom severity the wrist for 3 mo night and day ment System Questionnaire — 5 items C ¼ Splinting & 1 dose steroid injection . Health status ¼ SF-36: 36 questions Incl ¼ Clinical diagnosis of mild to moderate CTS assessing general health-related quality > 6 mo con¯rmed by EDS (23) of life . Follow-up ¼ N/A n ¼ 57 wrists A ¼ OCTR (30 wrists) Sex ¼ 53F/4M B ¼ LLLT Neon (He–Ne) (632.8 nm, . Improvement of neurophysiological Age ( Æ SD) ¼ Gp A: 45.27 Æ 13.19 yrs; parameters for median nerve ¼ DML, Level Laser M300) continuous wave PML, MNCV, CMAP wrist and elbow, Gp B: 44.50 Æ 7.24 yrs; Gp C: 44.46 Æ 8.52 yrs (CW), ! 12 mW, 30 cm from skin, SNCV, SNAP using the Nihon-Cohden (p ¼ 0:976) 3 J/cm2, 2x/wk for 12 sessions Neuropack Symptom duration ( Æ SD): Gp A: 21 Æ 11 mo; (30 wrists) Gp B: 15.26 Æ 7.19 mo; Gp C: 19.13 Æ 13 mo . Symptoms ¼ BQ — 11 items (5-point (p ¼ 0:869) scale) Incl ¼ Clinical diagnosis of CTS . Function ¼ BQ — 8 items (5-point scale) n ¼ 60 wrists/54 patients . Satisfaction ¼ 5-point scale Sex ¼ Gp A: 23F/4M; Gp B: 25F/2M . Follow-up ¼ month: 3 and 6 Age ( Æ SD) ¼ Gp A: 42.65 Æ 8.0 yrs; . % patients with Gp B: 49.11 Æ 7.23 yrs . Symptom relief Symptom duration ( Æ SD) ¼ Gp A: 36.17 Æ 4.38 . Return to normal activities . Adverse e®ects from treatment mo; Gp B: 28.21 Æ 7.03 mo . Positive nerve conduction tests . Follow-up ¼ month: approximately 6

Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Table 1. (Continued ) Study Design Participants Intervention Outcome measures Ref. 27 RCT Incl ¼ Age ! 18 yrs, Clinical diagnosis of CTS A ¼ OCTR or ECTR (57 patients) . Function ¼ CTSAQ — 9 items > 2 wks con¯rmed by EDS, þve \\°ick test\" or B ¼ Multimodality [200 mg ibuprofen (5-point scale) nocturnal pain, failure of 2 wks conservative treatment 3x/day, 6 sessions hand therapy over . Symptoms ¼ CTSAQ — 11-items 6 wks, educational booklet, hand (5-point scale) n ¼ 116 patients exercises, splinting night and day, Sex ¼ Gp A: 28F/29M; Gp B: 34F/25M work modi¯cations] (59 patients) . Pain interference with work or Age ( Æ SD) ¼ Gp A: 50.2 Æ 10.3 yrs; activities ¼ 11-point scale E®ectiveness of surgical treatment versus conservative treatment in CTS 99 Gp B: 51.2 Æ 8.9 yrs (p ¼ 0:213) . Health status ¼ SF-36: 36 questions Symptom duration [median (IQR)] ¼ Gp A: 3.2 assessing general health-related quality of life (1.3–5.5) yrs; Gp B: 3.4 (1.0–8.7) yrs . Additional treatments ¼ Patient diary Incl ¼ Age ! 18 yrs, Clinical diagnosis of CTS . Follow-up ¼ month: 6 and 12 > 3 mo, con¯rmed by EDS, consecutive referral Ref. 33 RCT and unresponsiveness to ! 2 wks NSAIDs A ¼ OCTR (80 wrists) . Nocturnal paresthesias ¼ 100 mm VAS Ref. 34 RCT treatment B ¼ 1 or 2 doses steroid injection 20 mg . Diurnal pain ¼ 100 mm VAS . Functional impairment ¼ 100 mm VAS n ¼ 163 wrists/101 patients in 1 mL paramethasone acetonide . Follow-up ¼ month: 3, 6, 12 and 24 Sex: 93F/8M with 22-gauge needle 45 angle Age ( Æ SD) ¼ Gp A: 50.53 Æ 10.87 yrs; Gp B: distally 1–2 cm depth medial to Palmaris longus tendon (83 wrists) 53.17 Æ 13.93 yrs (p ¼ 0:213) Symptom duration ( Æ SD): Gp A: 31.12 Æ 7.27 A ¼ OCTR (80 wrists) . Nocturnal paresthesias ¼ 100 mm VAS B ¼ 1 or 2 doses steroid injection 20 mg . Diurnal pain ¼ 100 mm VAS wks; Gp B: 33.25 Æ 8.17 wks (p ¼ 0:723) . Functional impairment ¼ 100 mm VAS in 1 mL paramethasone acetonide . Follow-up ¼ month: 3, 6 and 12 Incl ¼ Age ! 18 yrs, Clinical diagnosis of CTS with 22-gauge needle 45 angle . Improvement of neurophysiological > 3 mo, con¯rmed by EDS, consecutive referral distally 1–2 cm depth medial to and unresponsiveness to ! 2 wks NSAIDs Palmaris longus tendon (83 wrists) parameters for median nerve ¼ DML, treatment MA, SNCV, SA . Follow-up ¼ month: 12 n ¼ 163 wrists/101 patients Sex: 93F/8M Age ( Æ SD) ¼ Gp A: 50.53 Æ 10.87 yrs; Gp B: 53.17 Æ 13.93 yrs (p ¼ 0:213) Symptom duration ( Æ SD): Gp A: 31.12 Æ 7.27 wks; Gp B: 33.25 Æ 8.17 wks (p ¼ 0:723)

Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Study Design Participants Table 1. (Continued ) Outcome measures 100 D. Klokkari & I. Mamais Intervention Ref. 25 RCT Incl ¼ Clinical diagnosis of CTS con¯rmed by EDS, A ¼ OCTR (20 patients) . Symptoms ¼ GSS — 11-point scale for Ref. 28 RT symptoms for > 3 mo B ¼ Steroid injection with 40 mg of ¯ve symptom categories, total score/50 n ¼ 40 patients methylprednisolone (20 patients) . Follow-up ¼ week: 2, 4 and 12 Sex ¼ 29F/11M (p ¼ 0:723) Age ( Æ SD) ¼ Gp A: 43.8 Æ 10.98 yrs; Gp B: A ¼ OCTR or ECTR (60 patients) . Pain intensity ¼ NPRS — 0–10 B ¼ Manual therapy 3 sessions of 30' (11-point scale where 0 = no pain 46.9 Æ 12.33 yrs (p ¼ 0:406) and 10 = worst possible pain) Symptom duration ( Æ SD) ¼ Gp A: 12.5 Æ 8.76 1x/wk (60 patients) . Symptoms ¼ BQ — 11 items (5-point mo; Gp B: 10.15 Æ 6.75 mo (p ¼ 0:348) scale) Incl ¼ Clinical diagnosis of CTS con¯rmed by EDS . Function ¼ BQ — 8 items (5-point scale) (sensory and motor de¯cit), symptoms for > 12 mo . Self-Perceived improvement ¼ GROC n ¼ 120 patients (from À7 (worse) to þ7 (better)) Sex ¼ 120F/0M . Follow-up ¼ Month: 1, 3, 6 and 12 Age ( Æ SD) ¼ Gp A: 46 Æ 9 yrs; Gp B: 47 Æ 10 yrs Symptom duration ( Æ SD) ¼ Gp A: 3.5 Æ 3.1 yrs; Gp B: 3.1 Æ 2.7 yrs Notes: Incl ¼ inclusion criteria, n ¼ number of patients randomized, Gp ¼ group, EDS ¼ electrodiagnostic studies, OCTR ¼ Open carpal tunnel release, ECTR ¼ Endoscopic carpal tunnel release, BQ ¼ Boston Questionnaire, SF-36 ¼ Short Form 36, DML ¼ Distal Motor Latency, MNCV ¼ Motor Nerve Conduction Velocity, CMAP ¼ Com- pound Muscle Action Potential, DSL ¼ Distal Sensory Latency, SNCV ¼ Sensory Nerve Conduction Velocity, SNAP ¼ Sensory Nerve Action Potential, SSS ¼ Symptom severity scale, FSS ¼ Functional status scale, GSS ¼ Global symptom score, VAS ¼ Visual analogue scale, MODEMSQ ¼ Musculoskeletal Outcome Data Evaluation and Management System Questionnaire, PML ¼ Proximal motor latency, CTSAQ ¼ Carpal tunnel syndrome assessment questionnaire, MA ¼ Motor amplitude, SA ¼ Sensory amplitude, RCT ¼ Randomized controlled trial, CT ¼ Clinical trial, RT ¼ Randomized trial, NPRS ¼ Numerical Pain Rating Scale, GROC ¼ Global rating of change.

Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Table 2. Quality scores for CTs (n ¼ 11). Randomization Allocation Baseline Participant Therapist Assessor Cointervention Dropout Intention to Total Study adequacy concealment comparability blinding blinding blinding avoidance Compliance rate < 20% Timing treat analysis (0 to 11) Ref. 23 N N Y N NN U U NY N 2 Ref. 26 Y Ref. 24 Y Y Y N NY Y N NY Y 7 Ref. 20 Y Ref. 22 Y Y Y N NY Y U YY N 7 Ref. 30 N Ref. 27 Y Y Y N NN U U YY Y 6 Ref. 33 Y Ref. 34 Y U Y N NN U U UY N 3 Ref. 25 Y Ref. 28 Y N Y N NN N U UY N 2 Y Y N NY U Y NY Y 7 Y Y N NN U U NY Y 5 E®ectiveness of surgical treatment versus conservative treatment in CTS 101 Y Y N NN U U NY N 4 Y Y N NU N U YY N 5 Y Y N NY Y U YY Y 8 Table 3. Quality scores for prospective and retrospective studies (n ¼ 4). Hypothesis stated Description of characteristics Multicenter study Appropriate eligibility criteria Consecutive recruitment Similar stage of condition Clearly described intervention Clearly described cointerventions Outcome measure description Outcome measure suitability Outcome measure timing Statistical test suitability Length of follow-up Loss to follow-up Random variability Adverse events Conclusion supported by result Competing interest Total (0 to 18) Study Ref. 19 Y Y N Y N N Y Y Y N Y N Y N N Y Y Y 11 Ref. 21 Y Y N Y N Y N Y Y Y Y Y Y N N N Y N 11 Ref. 31 Y Y Y Y Y Y N Y Y Y Y Y Y Y Y Y Y Y 17 Ref. 29 Y Y N Y N Y Y Y Y Y N Y Y Y Y N Y N 13

Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com 102 D. Klokkari & I. Mamais (MD 0.56, 95%CI 0.16 to 0.96, I 2 ¼ 89%, by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. p < 0:00001) (see Fig. C.1 in Appendix C). Meta- severe functional impairment. For meta-analysis analysis results at three months further increased purposes, the GSS scores and the VAS scale scores the total heterogeneity, therefore it was not astute were divided by 10 and 20, respectively, coming up to present them. Results for 18 or > 18 months with a common denominator of 5, in order to be could not be calculated due to insu±cient data coherent with the BQ and CTSAQ scores. The from the included studies. improvement of neurophysiological parameters was carried out by physicians and neurologists The qualitative analysis agreed with the meta- with nerve conduction studies. The most com- analysis, with strong evidence in favor of surgery monly assessed parameters were the Distal Motor for symptom improvement at 6 months, con°icting Latency (DML) and the Sensory Nerve Conduc- evidence at 12 months, moderate evidence at 18 tion Velocity (SNCV). months and limited evidence for the time period longer than 18 months (Table 4, see also E®ect of intervention on symptom Tables C.1–C.5 in Appendix C for details of in- improvement cluded studies). The meta-analysis pooled data from 6 studies with E®ect of intervention on functional a total of 805 wrists. One study26 was not included improvement in the pooled analysis as it presented the stan- dardized mean di®erence in contrast to the other The total meta-analysis results pooled data from 6 studies, which presented the mean score and studies with a total of 918 wrists. One study26 was standard deviation (SD). The results demonstrate not included in the pooled analysis for reasons that surgical treatment leads to a greater symptom explained previously. The results demonstrate that improvement at six months by 0.52 points lower surgery was superior to conservative treatment for score on a 5-point symptom severity scale com- functional improvement but the result was not pared to conservative treatment 95%CI (0.27 to signi¯cant (MD 0.06, 95%CI À0.10 to 0.22, 0.78). There was a statistically signi¯cant high I 2 ¼ 84%, p < 0:00001). Due to the high hetero- heterogeneity (I 2 ¼ 82%, p < 0:0001) (Fig. 2). The geneity, a subgroup analysis was ¯tting. At 3 results remain similar, if only high quality stud- months and 12 months, there was no statistically ies17,18,22,26 were included in the meta-analysis Fig. 2. MD (95%CI) of e®ect of surgical and conservative treatment on symptom improvement at six months of treatment by pooling data from six studies (n ¼ 805). Table 4. Qualitative analysis for symptom improvement. 3 months 6 months 12 months 18 months > 18 months Symptom Con°icting Strong evidence Con°icting Moderate evidence Limited improvement evidence fav. Surg evidence fav. Surg (2 studies) evidence fav. (6 studies) (8 studies) (4 studies) Surg (2 studies)

E®ectiveness of surgical treatment versus conservative treatment in CTS 103 Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Fig. 3. MD (95%CI) of e®ect of surgical and conservative treatment on functional improvement at 3, 6 and 12 months of by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. treatment by pooling data from six studies (n ¼ 918). signi¯cant di®erence between the surgical and (Table 5, see also Tables C.6–C.10 in Appendix C for details of included studies). conservative treatment for function. Surgery E®ect of intervention on proved more e®ective than conservative treatment improvement of neurophysiological parameters at six months (MD 0.20, 95%CI 0.05 to 0.36, I 2 ¼ 66%, p ¼ 0:01) (Fig. 3, but if only high The included studies presented results for di®erent quality studies17,18,26 were included, the results neurophysiological parameters. The meta-analysis was only applicable for the DML and SNCV. were not statistically signi¯cant at six months (MD Studies assessed the outcomes for the DML be- À0.22, 95%CI À0.55 to 0.12, I 2 ¼ 85%, p ¼ 0:001). tween 5 and 12 months. One study24 performed the The results remain similar at 12 months (MD follow-up measurement at 5 months, another34 at À0.23, 95%CI À0.72 to 0.26, I 2 ¼ 85%, p ¼ 0:01) 12 months and two more studies performed the with data from two studies17,18 (see Fig. C.2 in Appendix C). The qualitative analysis favored surgery with strong evidence at 6 and 12 months and limited evidence for the time period longer than 18 months Table 5. Qualitative analysis for functional improvement. 3 months 6 months 12 months 18 months > 18 months Functional Moderate evidence Strong evidence Strong evidence Con°icting Limited evidence improvement fav. Conserv. fav. Surg fav. Surg. evidence fav. Surg (5 studies) (8 studies) (4 studies) (2 studies) (2 studies)

104 D. Klokkari & I. Mamais Fig. 4. MD (95%CI) of e®ect of surgical and conservative treatment on improvement of DML between 5 and 12 months of treatment by pooling data from four studies (n ¼ 337). Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com follow-up at 6 months.22,23 Pooled data utilizing (p ¼ 0:08), which was a®ected by the results at 3 by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. 337 wrists demonstrated that surgery provides a months, so a subgroup analysis was appropriate greater improvement of the DML compared to (Fig. 5). At 6 months, the results were in favor of conservative treatment with a mean di®erence surgery (MD 3.71 m/s, 95%CI 1.94 to 5.49, 0.30 ms less time delay in the surgical group com- I 2 ¼ 0%, p ¼ 0:42). pared to the conservative (95%CI 0.09 to 0.51) and low to moderate non-signi¯cant heterogeneity In contrast to the meta-analysis, the qualitative (I2 ¼ 29%, p ¼ 0:24) (Fig. 4). analysis took into consideration the overall e®ec- tiveness from the nerve conduction studies repor- In the total meta-analysis results for the SNCV ted in the included studies. The data synthesis were in favor of surgery (MD 2.73 m/s, 95%CI 0.71 demonstrated that there was a bene¯t in favor of to 4.75) with a total of 80 wrists for the surgical surgery with moderate evidence at 6 and 12 and 94 for the conservative group. There was a months and insu±cient data to formulate conclu- moderate non-signi¯cant heterogeneity of 52% sions for the long-term e®ectiveness (Table 6, see Fig. 5. MD (95%CI) of e®ect of surgical and conservative treatment on improvement of SNCV at six months of treatment by pooling data from three studies (n ¼ 174). *Left ¼ results in favor of surgery, Right ¼ results in favor of conservative treatment. Table 6. Qualitative analysis for improvement of neurophysiological parameters. 3 months 6 months 12 months 18 months > 18 months Improvement of Con°icting Moderate evidence Moderate evidence N/A N/A neurophysiological evidence fav. Surg (3 fav. Surg parameters (2 studies) studies) Limited (3 studies) evidence fav Surg. for DML (1 study)

E®ectiveness of surgical treatment versus conservative treatment in CTS 105 Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Fig. 6. OR (95%CI) of e®ect of surgical and conservative treatment on complications reported by pooling data from eight by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. studies (n ¼ 759). also Tables C.11–C.13 in Appendix C for details of Sensitivity analyses included studies). Sensitivity analyses showed that the pooled esti- Side e®ect and complication mates of outcome measures did not vary substan- tially with the exclusion of low quality study. Eight of the trials included reported side e®ects from treatment. However, there was a wide range Qualitative analysis and of side e®ects and complications reported. Serious meta-analysis agreement complications were only reported in the surgery group and included re°ex sympathetic dystro- A comparison of a qualitative analysis and a phy25,26 and complex regional pain syndrome.22 quantitative analysis (meta-analysis) was only Mild side e®ects from both treatments included possible where su±cient data were available for a symptoms like pain, swelling, discomfort and were meta-analysis. It is worth noting that there was a reported in most CTs.20,22,24,26,30,33,34 The results considerable agreement between the qualitative from 370 wrists, which underwent surgery and and the quantitative analysis, despite the inclusion 389 wrists treated conservatively, showed that of di®erent studies in each type of analysis. In conservative treatment was more bene¯cial than particular, the outcomes with con°icting results in surgery with almost half the complications the qualitative analysis showed non-signi¯cant reported with conservative compared to surgical results in the meta-analysis at the same intervals treatment (OR ¼ 1:99, 95%CI 1.27 to 3.14, of reassessment. Furthermore, the outcomes dem- I 2 ¼ 0%, p ¼ 0:44) (Fig. 6). The results remain onstrating statistically signi¯cant results in the similar, if only high quality studies17,18,22,24 were meta-analysis had strong evidence in favor of a included in the meta-analysis (OR 2.07, 95%CI treatment. 1.20 to 3.57, I% ¼ 0%, p ¼ 0:47) (see Fig. C.3 in Appendix C). Discussion Publication bias Even with the inclusion of more recent studies, the results of this systematic review are consistent with There was symmetry in the funnel plots about the the previous systematic reviews with regards to the standard error. Therefore, no publication bias was direction of results. Surgical treatment outweighed identi¯ed because the funnel plots for symptom conservative treatment in all outcomes. Conser- improvement, functional improvement or im- vative treatment however caused fewer complica- provement of neurophysiological parameters, were tions than surgery. Both treatments were e®ective symmetric. No publication bias was noted on fun- in improving symptoms and function at six nel plots and Egger regression (p > 0:05). months.

Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com 106 D. Klokkari & I. Mamais vascular congestion35 around a median nerve which by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. often appears swollen when examined with ultra- There was some concordance in the results of sound.36 Injections can provide short-term analge- this systematic review compared to the latest sia but the mechanical compression persists,24 one,10 with statistically signi¯cant results at six resulting in a gradual recurrence of symptoms. months for symptom improvement. There was a disagreement regarding the results at 12 months, It was previously reported that 50% of the cases with the current meta-analysis showing no signi¯- treated with steroid injections are worse regarding cant di®erences in symptoms and function, in their clinical presentation and neurophysiological contrast to the previous systematic review, which studies over a period of six months.37 This is clearly supported surgery. These di®erences are re°ected in the studies using steroid injections used attributed to the inclusion of both additional and in this review.20,22,25,33,34 Ly-Pen et al.33 revealed di®erent studies in the meta-analysis. One trial26 that as the time passes from the injection until the was excluded for reasons previously mentioned, re-evaluation, the failures of the intervention in- and the most recent study28 showed no di®erence crease. They attributed the need for review of the between the two interventions at 12 months. original study to this event, and monitored the patients for another year after their initial inter- The exclusion of low quality trials, for meta- vention, where the results were clearly in favor of analysis purposes, was only possible for 6 months surgery. On the contrary, surgery provides a more symptom improvement, 6 and 12 months func- permanent solution, as the resection of the °exor tional improvement and complication and side retinaculum decompresses the median nerve. This e®ects. The results were only di®erentiated in fact, combined with the complications of surgery, functional improvement but this increased the the postoperative discomfort and the patient's re- heterogeneity by almost 20%. luctance to move after surgery to avoid irritation of the wound, results in a conservative treatment Since the heterogeneity of the studies included with injections appearing more e®ective than sur- in the meta-analysis was high, a qualitative anal- gery in the short term. ysis was carried out in addition to the meta-anal- ysis. The qualitative analysis allowed for a wider An additional reason for this discrepancy was inclusion of studies, upon which to draw conclu- the inclusion of the most recent study,28 which sions. In addition, it provided a classi¯cation of a®ected the overall result at 12 months as it studies according to their quality, with the highest showed no signi¯cant di®erences at 6 and 12 quality studies having a greater e®ect on the months. The unique feature of this study, which overall outcome. The qualitative analysis took into had the highest methodological score (8/11), was account the result of each study for each outcome, the nature of the intervention, which was not the methodological quality of each study and the concentrated on the wrist, unlike the other con- number of studies in favor of an intervention for servative interventions. This manual therapy in- each outcome. tervention included desensitization maneuvers, across the continuum of the median nerve from the In the results for symptom improvement, it was cervical spine to the wrist. Speci¯cally, treatment evident that surgery was superior to conservative targeted all possible locations of entrapment of the treatment at six months. Regarding functional median nerve along its path, prior to the applica- status at six months, the surgery group had greater tion of gliding movements of the nerve to improve functional improvement, which spread to the 12- its glide in relation to the adjacent tissues. This month re-evaluation (only in the qualitative anal- study only included female participants but it is ysis). However, the results for symptoms and unlikely that the results will be di®erent if parti- function at three months did not favor one of the cipants of both genders were included. interventions as the qualitative analysis showed con°icting evidence and the meta-analysis showed The improvement of neurophysiological para- a trend towards surgery, which was not statisti- meters was also superior with surgical treatment. cally signi¯cant. The meta-analysis data for DML were extracted from studies with di®erent time intervals, but this The reason for this discrepancy between the was the only way to group the studies together for short- and long-term e±cacy for symptoms and the meta-analysis. function is the use of steroid injections as conser- vative treatment intervention in the above studies. Steroid injections can reduce wrist joint e®usion and

E®ectiveness of surgical treatment versus conservative treatment in CTS 107 Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Complications and side e®ects were the only did not make such a reference. Instead, some pre- by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. occasion where conservative treatment outweighed sented baseline measurements for the severity of surgery with almost twice as many complications symptoms and functional status. Comparing these in the surgery group compared to the group of baseline values from the BQ showed that the se- conservative treatment. Severe complications oc- verity of symptoms and functional status were curred only with surgery. Most reported side e®ects comparable between studies, where the severity of were related to pain or tenderness at the incision CTS is mentioned (see Tables C.14 and C.15 in site, which are common signs and symptoms after Appendix C). Therefore, we can deduce that this open surgery. In addition, there was great hetero- systematic review refers mainly to people with mild geneity in the reported side e®ects. Only a few to moderate CTS. studies reported serious side e®ects, while others reported all adverse reactions regardless of severi- An additional issue of concern was the fact that ty. Because there was great variation in the se- the search strategy for this review was limited verity of complications reported in each study, it is to electronic databases and the gray literature not possible to verify the real advantage of con- (unpublished studies) was not searched. This could servative versus surgical treatment in this have a®ected the results of the publication outcome. analysis. Additional heterogeneity existed, regarding the From the studies included, many were limited chronicity and severity of symptoms, the period of by lack of randomization, lack of standardized re-evaluation, the outcome measures of each study, outcome measures and retrospective design, which and how these were measured. Analysis into sub- lacked information on patient baseline measure- groups according to the severity of symptoms or ments, so a comparison of the severity of initial the time of reassessment was not possible due to symptoms was not possible, nor was an estimate of the limited number of studies classifying the the improvement from baseline until reassessment. patients accordingly. Perhaps, this should be The incorporation of prospective and retrospective addressed in future trials comparing the two studies however allowed for an evaluation of the interventions. outcomes in the longer term. The severity of the symptoms could not be ac- In the absence of randomization, these obser- curately determined as there was great variation in vational studies are considered lower quality than the assessment methods. Moreover, some stud- RCTs and CTs for the collection of data on the ies29–31 did not con¯rm the diagnosis through e±cacy of an intervention because they can be af- electrodiagnostic studies, and as a result, some of fected by various types of bias such as selection, the patients included may not have been su®ering detection, performance, attrition, reporting and from CTS alone. The need for neurophysiological publication.15 However, their inclusion constituted studies however is not universally accepted. Some a strength of this review. In these studies, treat- studies, indicate that the nerve conduction studies ment was pre-determined according to the severity are not necessary for the diagnosis of CTS, since it of the condition. Patients with severe CTS under- can lead to false negative or false positive went surgery. Since the surgical intervention was results.38–40 Other studies indicate that there is no more e®ective than conservative, and in more se- correlation between clinical symptoms and results verely impaired patients, one can assume that the of neurophysiological studies in CTS. Researchers e®ect size of surgery might be higher than the one propose the use of neurophysiological studies as of the conservative. On the other hand, a long-term an additional independent tool for diagnosis and observational study42 presented evidence that in assessment of the severity of CTS.41 some cases, CTS may improve spontaneously, causing these patients to undergo unnecessary Regarding the severity of the condition, most surgery. In addition, severely impaired patients studies excluded patients with severe atrophy of might ¯nd it easier to score positively on subjective the thenar muscles, or previous surgery for CTS outcome measures like VAS and GSS. These excluding in this way severe cases of CTS. Six remarks make the formulation of a clear conclusion RCTs and CTs reported that they included people di±cult and future studies should address them. with mild to moderate CTS.20,23,24,26,33,34 One study included patients with mild, moderate or An evaluation of 331 hands identi¯ed 5 factors severe CTS28 and the remaining eight studies of poor prognosis and the need for surgery: age above 50 years, symptom for over 10 months,

108 D. Klokkari & I. Mamais Author Contributions consecutive symptoms of paraesthesia, °exor DK was responsible for the data collection, data tenosynovitis and a positive Phalen's test for less analysis, data interpretation and writing the than 30 s.43 It was reported that when these factors manuscript, IM was responsible for revising the did not exist, 2/3 of patients healed with conser- manuscript, statistical analysis and project man- vative treatment. When 4 or 5 of these factors agement. Both authors participated in editing and are present, they recommend surgical treatment. completion of the manuscript. Conversely, other researchers studied 45 hands conservatively with steroid injections and con- Appendix A. cluded that there is no correlation between signs and symptoms of CTS and the ¯nal result.44 They Summary of search strategy. argued that chronicity is the most signi¯cant factor Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com for the ¯nal result.44 Therefore, conservative PubMed by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. treatment is a feasible option for mild and short- term symptoms, but surgery can provide a more (Title/ EBSCO permanent solution to persisting symptoms. # Keywords Abstract) (Abstract) Conclusion #1 ((surgical*) OR 822.822 1.713.262 The results of this review demonstrate that surgery 1.579.763 leads to a greater improvement of symptoms and (surgical neurophysiological parameters at six months, 61.818 compared to conservative treatment. The decision intervention) OR 6.814.297 however, about the choice of treatment, needs careful consideration, taking into account the (open carpal tunnel 387 complications reported with surgical treatment and the fact that in some cases, CTS may be re- release) OR (OCTR) solved spontaneously. However, the conclusions derived from this review are based on a number of OR (endoscopic underpowered studies. Therefore, high quality prospective studies are needed in order to identify carpal tunnel release) the characteristics of individuals where CTS has promising path to avoid unnecessary surgery. In OR (ECTR)) addition, further research should focus on explor- ing the ¯eld of manual therapy and compare it to #2 ((conservative) OR 322.464 the surgical intervention for CTS. Research should also address the long-term e®ectiveness of the two (conservative interventions beyond 12 months. intervention) OR Con°ict of Interest (corticosteroid The authors declare that they have no con°ict of interest. injections) OR Funding/Support (steroid injections) This research did not receive any speci¯c grant OR (wrist splints) from funding agencies in the public, commercial or not-for-pro¯t sectors. OR (physiotherapy) OR (electrotherapy) OR (exercise) OR (manual therapy)) #3 ((Carpal tunnel 13.486 syndrome) OR (CTS) OR (median nerve entrapment) OR (nerve compression)) #4 ((RCT) OR (random*) 1.595.404 OR (randomized controlled trial) OR (controlled trial) OR (cohort study) OR (clinical trial) OR (controlled clinical trial) OR (retrospective) OR (prospective)) #5 #1 AND #2 AND 72 #3 AND #4

E®ectiveness of surgical treatment versus conservative treatment in CTS 109 Appendix B. List of excluded papers Study Year Title Reason for exclusion Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Martin et al. 2005 RCT of surgery versus conservative therapy Study protocol by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Schrijver et al. 2005 for CTS Study comparing nerve Korthals-de 2006 Correlating nerve conduction studies and conduction and clinical Bos et al. clinical outcome measures on CTS: Lessons improvement using the data 2009 from a randomized controlled trial from the study of Ref. 26 Pomerance 2010 et al. Surgery is more cost-e®ective than splinting for Cost-e®ectiveness study using 2013 CTS in the Netherlands: Results of an the data from Ref. 26 Vogelin et al. economic evaluation alongside a randomized controlled trial Cost-e®ectiveness study Onuma et al. The cost-e®ectiveness of non-surgical versus Di®erent outcome measures. surgical treatment for CTS Study measuring the size of the carpal tunnel after the Sonographic follow-up of patients with CTS intervention. undergoing surgical or non-surgical treatment: Prospective cohort study Case-control study Bilateral CTS due to gouty tophi: Conservative and surgical treatment in di®erent hands of the same patient Appendix C. Fig. C.1. Detailed meta-analysis results for symptom improvement if only high quality studies17,18,22,26 were included.

110 D. Klokkari & I. Mamais Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Fig. C.2. Detailed meta-analysis results for functional improvement at 6 and 12 months if only high quality studies17,18,26 were by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. included. Fig. C.3. OR (95%CI) of e®ect of surgical and conservative treatment on complications reported if only high quality stud- ies17,18,26 were included. Table C.1. Qualitative analysis (symptom improve- Table C.3. Qualitative analysis (symptom improve- ment — 3 months). ment — 12 months). Favors No Favors No Favors conservative Favors surgery di®erence conservative surgery di®erence High quality Ref. 26 Ref. 28 High quality Refs. 26 and 27 Ref. 28 Low quality Ref. 25 Ref. 26 Refs. 20 and 22 Low quality Ref. 20 Table C.2. Qualitative analysis (symptom improve- Table C.4. Qualitative analysis (symptom improve- ment — 6 months). ment — 18 months). No Favors No Favors Favors surgery di®erence conservative Favors surgery di®erence conservative High quality Refs. 24, 26 and 27 Ref. 28 High quality Refs. 26 and 31* Low quality Refs. 22, 26 and 30 Ref. 20 Low quality Note: *Observational study.

E®ectiveness of surgical treatment versus conservative treatment in CTS 111 Table C.5. Qualitative analysis (symptom improve- Table C.10. Qualitative analysis (functional improve- ment — > 18 months). ment — > 18 months). No Favors No Favors Favors surgery di®erence conservative Favors surgery di®erence conservative High quality Ref. 31* High quality Ref. 31 Low quality Refs. 29 and 33 Ref. 21 Low quality Refs. 29* and 33 Note: *Observational studies. Note: *Observational studies. Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Table C.6. Qualitative analysis (functional im- by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. provement — 3 months). Favors No Favors Table C.11. Qualitative analysis (improvement of neurophysiological parameters — 3 months). surgery di®erence conservative High quality Ref. 26 Ref. 28 Favors Favors Low quality surgery conservative Ref. 23 Refs. 20 and 22 No di®erence High quality Ref. 23 (apart from Low quality SNCV and DSL) and Ref. 22 Table C.7. Qualitative analysis (functional improve- ment — 6 months). No Favors Favors surgery di®erence conservative High Refs. 24, 26, Ref. 28 Table C.12. Qualitative analysis (improvement of neuro- quality 27 and 31* physiological parameters — 6 months). Refs. 20 Low Ref. 23 and 22 quality Favors Favors surgery No di®erence conservative Note: *Observational study. High quality Ref. 24 (apart Low quality from SNCV) Table C.8. Qualitative analysis (functional improve- Ref. 23 (apart ment — 12 months). from SNCV and DSL), No Favors Refs. 22 and 30 Favors surgery di®erence conservative High quality Refs. 26 and 27 Ref. 28 Low quality Ref. 20 Table C.13. Qualitative analysis (improvement of neu- rophysiological parameters — 12 months). Table C.9. Qualitative analysis (functional improve- Favors ment — 18 months). Favors surgery No di®erence conservative Favors No Favors High Ref. 26 (apart Ref. 26 (apart surgery di®erence conservative quality from DSL) from DML) High quality Ref. 31* Ref. 26 Low Refs. 21* and 34 Low quality quality Note: *Observational study. Note: *Observational study.

112 D. Klokkari & I. Mamais Table C.14. Symptom severity baseline measurements. Study Measurement tool Intervention Baseline measurement Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com Ref. 31 Symptom severity scale Surg: 3.2 Æ 0.8 by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Ref. 23 BQ Cons: 2.6 Æ 0.8 Ref. 26 Surg: 3.4 Æ 0.7 Ref. 24 Symptom severity scale Cons: 3.3 Æ 0.7 Ref. 22 GSS/10 Surg: 2.5 (1.9–3.1) BQ Cons: 2.4 (1.8–2.9) Ref. 27 Surg: 2.86 Æ 1.10 Ref. 25 CTSAQ Cons: 2.52 Æ 1.05 Ref. 28 GSS/10 Surg: 3.09 Æ 0.5 Cons (Splinting): 2.66 Æ 0.35 BQ Cons (Splint þ injection): 2.79 Æ 0.63 Surg: 2.95 Æ 0.77 Cons: 3.01 Æ 0.64 Surg: 3.545 Æ 0.74 Cons: 3.48 Æ 0.81 Surg: 2.7 Æ 0.6 Cons: 2.5 Æ 0.7 Table C.15. Functional status baseline measurements. Study Measurement tool Intervention Baseline measurement Ref. 31 Functional status scale Surg: 2.7 Æ 0.9 Cons: 2.1 Æ 0.9 Ref. 23 BQ Surg: 3.3 Æ 1.0 Cons: 3.0 Æ 0.8 Ref. 26 Functional status scale Surg: 2.3 (1.5–3.0) Cons: 2.0 (1.5–2.9) Ref. 20 VAS scale/20 Surg: 1.95 Æ 1.40 Cons: 1.895 Æ 1.318 Ref. 22 BQ Surg: 2.7 Æ 0.62 Cons (Splinting): 2.47 Æ 0.65 Ref. 27 CTSAQ Cons (Splint þ injection): 2.19 Æ 0.51 Ref. 28 BQ Surg: 2.42 Æ 0.82 Cons: 2.53 Æ 0.82 Surg: 2.4 Æ 0.6 Cons: 2.3 Æ 0.5 References syndrome: Prevalence in the general population. J Clin Epidemiol 1992;45(4):373–6. 1. Mediouni Z, Bodin J, Dale AM, et al. Carpal tunnel 4. Atroshi I, Gummesson C, Johnsson R, Ornstein E, syndrome and computer exposure at work in two Ranstam J, Rosen I. Prevalence of carpal tunnel large complementary cohorts. BMJ Open 2015; syndrome in a general population. Jama 1999; 5(9):e008156. 282(2):153–8. 5. Castro Ado A, Skare TL, Nassif PA, Sakuma AK, 2. Tanaka S, Wild DK, Seligman PJ, Behrens V, Barros WH. Sonographic diagnosis of carpal tunnel Cameron L, Putz-Anderson V. The US prevalence syndrome: A study in 200 hospital workers. Radiol of self-reported carpal tunnel syndrome: 1988 Bras 2015;48(5):287–91. National Health Interview Survey data. Am J 6. Kostopoulos D. Treatment of carpal tunnel Public Health 1994;84(11):1846–8. syndrome: A review of the non-surgical approaches 3. de Krom MC, Knipschild PG, Kester AD, Thijs CT, Boekkooi PF, Spaans F. Carpal tunnel

E®ectiveness of surgical treatment versus conservative treatment in CTS 113 Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com with emphasis in neural mobilization. J Bodyw 21. Seror P. Nerve conduction studies after treatment by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Mov Ther 2004;8(1):2–8. for carpal tunnel syndrome. J Hand Surg 1992;17 7. Ibrahim I, Khan WS, Goddard N, Smitham P. (6):641–5. Carpal tunnel syndrome: A review of the recent literature. Open Orthop J 2012;6:69–76. 22. Ucan H, Yagci I, Yilmaz L, Yagmurlu F, Keskin D, 8. Rempel D, Dahlin L, Lundborg G. Pathophysiolo- Bodur H. Comparison of splinting, splinting plus gy of nerve compression syndromes: Response of local steroid injection and open carpal tunnel peripheral nerves to loading. J Bone Joint Surg Am release outcomes in idiopathic carpal tunnel 1999;81(11):1600–10. syndrome. Rheumatol Int 2006;27(1):45–51. 9. Bland JD, Rudolfer SM. Clinical surveillance of carpal tunnel syndrome in two areas of the United 23. Demirci S, Kutluhan S, Koyuncuoglu R, et al. Kingdom, 1991–2001. J Neurol Neurosurg Psychi- Comparison of open carpal tunnel release and atry 2003;74(12):1674–79. local steroid treatment outcomes in idiopathic 10. Shi Q, MacDermid JC. Is surgical intervention carpal tunnel syndrome. Rheumatol Int 2002; more e®ective than non-surgical treatment for 22(1):33–7. carpal tunnel syndrome? A systematic review. J Orthop Surg Res 2011;6:17. 24. Hui AC, Wong S, Leung CH, et al. A randomized 11. Verdugo RJ, Salinas RA, Castillo JL, Cea JG. controlled trial of surgery vs steroid injection for Surgical versus non-surgical treatment for carpal carpal tunnel syndrome. Neurology 2005;64 tunnel syndrome. Cochrane Database Syst Rev (12):2074–8. 2008(4):CD001552. 12. van Tulder M, Furlan A, Bombardier C, Bouter L. 25. Ullah I. Local steroid injection or carpal tunnel Editorial Board of the Cochrane Collaboration release for carpal tunnel syndrome — which is Back Review Group. Updated method guidelines more e®ective? J Postgrad Med Instit (Peshawar– for systematic reviews in the cochrane collaboration Pakistan), North America, 27, Mar. 2013. back review group. Spine 2003;28(12):1290–9. 13. Maher CG, Sherrington C, Herbert RD, Moseley 26. Gerritsen AA, de Vet HC, Scholten RJ, Bertels- AM, Elkins M. Reliability of the PEDro scale for mann FW, de Krom MC, Bouter LM. Splinting vs rating quality of randomized controlled trials. Phys surgery in the treatment of carpal tunnel syn- Ther 2003;83(8):713–21. drome: A randomized controlled trial. Jama 14. Macedo LG, Elkins MR, Maher CG, Moseley AM, 2002;288(10):1245–51. Herbert RD, Sherrington C. There was evidence of convergent and construct validity of Physiotherapy 27. Jarvik JG, Comstock BA, Kliot M, et al. Surgery Evidence Database quality scale for physiotherapy versus non-surgical therapy for carpal tunnel syn- trials. J Clin Epidemiol 2010;63(8):920–5. drome: A randomised parallel-group trial. Lancet 15. Moga C, Guo B, Schop°ocher D, Harstall C. De- 2009;374(9695):1074–81. velopment of a quality appraisal tool for case series studies using a modi¯ed Delphi technique. Institute 28. Fernandez-de-Las Penas C, Ortega-Santiago R, de of Health Economics: Edmonton AB; 2012. la Llave-Rincon AI, et al. Manual physical therapy 16. Sedgwick P. Meta-analyses: Heterogeneity and versus surgery for carpal tunnel syndrome: A ran- subgroup analysis. Bmj 2013;346(2):f4040. domized parallel-group trial. J Pain 2015; 17. Higgins JP, Thompson SG. Quantifying heteroge- 16(11):1087–94. neity in a meta-analysis. Stat Med 2002;21 (11):1539–58. 29. Ettema AM, Amadio PC, Cha SS, Harrington JR, 18. Egger M, Davey Smith G, Schneider M, Minder C. Harris AM, O®ord KP. Surgery versus conservative Bias in meta-analysis detected by a simple, graph- therapy in carpal tunnel syndrome in people aged ical test. Bmj 1997;315(7109):629–34. 70 years and older. Plast Reconstr Surg 19. Harter BT, Jr., McKiernan JE, Jr., Kirzinger SS, 2006;118(4):947–58; discussion 959–60. Archer FW, Peters CK, Harter KC. Carpal tunnel syndrome: Surgical and nonsurgical treatment. 30. Elwakil TF, Elazzazi A, Shokeir H. Treatment J Hand Surg 1993;18(4):734–9. of carpal tunnel syndrome by low-level laser 20. Ly-Pen D, Andreu JL, de Blas G, Sanchez-Olaso A, versus open carpal tunnel release. Lasers Med Sci Millan I. Surgical decompression versus local ste- 2007;22(4):265–70. roid injection in carpal tunnel syndrome: A one- year, prospective, randomized, open, controlled 31. Katz JN, Keller RB, Simmons BP, et al. Maine clinical trial. Arthritis Rheum 2005;52(2):612–9. carpal tunnel study: Outcomes of operative and nonoperative therapy for carpal tunnel syndrome in a community-based cohort. J Hand Surgery 1998;23(4):697–710. 32. Leite JC, Jerosch-Herold C, Song F. A systematic review of the psychometric properties of the boston carpal tunnel questionnaire. BMC Musculoskelet Disord 2006;7:78. 33. Ly-Pen D, Andreu JL, Millan I, de Blas G, Sanchez-Olaso A. Comparison of surgical decom- pression and local steroid injection in the treatment

Hong Kong Physiother. J. 2018.38:91-114. Downloaded from www.worldscientific.com 114 D. Klokkari & I. Mamais 39. Braun RM, Jackson WJ. Electrical studies as a by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. prognostic factor in the surgical treatment of carpal of carpal tunnel syndrome: 2-year clinical results tunnel syndrome. J Hand Surg 1994;19(6):893–900. from a randomized trial. Rheumatology 2012;51(8): 1447–54. 40. Redmond MD, Rivner MH. False positive electro- 34. Andreu JL, Ly-Pen D, Millan I, de Blas G, San- diagnostic tests in carpal tunnel syndrome. Muscle chez-Olaso A. Local injection versus surgery in Nerve 1988;11(5):511–8. carpal tunnel syndrome: Neurophysiologic out- comes of a randomized clinical trial. Clin Neuro- 41. Glowacki KA, Breen CJ, Sachar K, Weiss AP. physiol 2014;125(7):1479–84. Electrodiagnostic testing and carpal tunnel release 35. Werner RA, Andary M. Electrodiagnostic evalua- outcome. J Hand Surg 1996;21(1):117–21. tion of carpal tunnel syndrome. Muscle Nerve 2011;44(4):597–607. 42. Padua L, Padua R, Aprile I, Pasqualetti P, Tonali 36. Wong SM, Gri±th JF, Hui AC, Tang A, Wong KS. P, Italian CTSSGCts. Multiperspective follow-up Discriminatory sonographic criteria for the diag- of untreated carpal tunnel syndrome: A multicenter nosis of carpal tunnel syndrome. Arthritis Rheum study. Neurology 2001;56(11):1459–66. 2002;46(7):1914–21. 37. Girlanda P, Dattola R, Venuto C, Mangiapane R, 43. Kaplan SJ, Glickel SZ, Eaton RG. Predictive fac- Nicolosi C, Messina C. Local steroid treatment in tors in the non-surgical treatment of carpal tunnel idiopathic carpal tunnel syndrome: Short- and syndrome. J Hand Surg 1990;15(1):106–8. long-term e±cacy. J Neurol 1993;240(3):187–90. 38. Finsen V, Russwurm H. Neurophysiology not re- 44. Irwin LR, Beckett R, Suman RK. Steroid injection quired before surgery for typical carpal tunnel for carpal tunnel syndrome. J Hand Surg 1996; syndrome. J Hand Surg 2001;26(1):61–4. 21(3):355–7.

Research Paper Hong Kong Physiotherapy Journal Vol. 38, No. 2 (2018) 115–123 DOI: 10.1142/S1013702518500099 Hong Kong Physiother. J. 2018.38:115-123. Downloaded from www.worldscientific.com Hong Kong Physiotherapy Journal by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. http://www.worldscientific.com/worldscinet/hkpj Reliability and validity of transfer assessment instrument version 3.0 in individuals with acute spinal cord injury in early rehabilitation phase Preeti Baghel1, Shefali Walia1 and Majumi M Noohu2,* 1ISIC Institute of Rehabilitation Sciences, New Delhi, India 2Centre for Physiotherapy and Rehabilitation Sciences, Jamia Millia Islamia, New Delhi, India *[email protected] Received 9 January 2017; Accepted 28 July 2017; Published 14 August 2018 Background: Transfers are very important in functional activities of subjects with spinal cord injury (SCI). The transfer assessment instrument (TAI) was the ¯rst tool to standardize the assessment of transfer technique. Objective: The purpose of this study was to establish the reliability and validity of TAI 3.0 in people with SCI in early rehabilitation phase. Methods: Thirty subjects with acute traumatic SCI were recruited from a tertiary care center for SCI management. Four raters assessed the quality of transfer using TAI 3.0 and a ¯fth rater used global as- sessment of transfer scale (VAS). TAI 3.0's intraclass correlation coe±cient (ICC) for intrarater and inter- rater reliability, standard error of measurement (SEM), minimal detectable change (MDC), limits of agreement and concurrent validity was determined. Results: The intrarater ICC was 0.93 to 0.98 and interrater ICC was 0.99, indicating high levels of reliability. The SEMs among the raters for TAI 3.0 total was from 0.23 to 0.28. The MDC among the raters TAI 3.0 total was from 0.54 to 0.86. Correlation for di®erent raters between the TAI 3.0 and VAS ranged between 0.88 and 0.90. Conclusion: TAI 3.0 is a reliable and valid tool to assess the transfer skill in individuals with SCI in early rehabilitation phase. Keywords: Spinal cord injury; transfers; activities of daily living. *Corresponding author. Copyright@2018, Hong Kong Physiotherapy Association. Published by World Scienti¯c Publishing Co Pte Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 115

Hong Kong Physiother. J. 2018.38:115-123. Downloaded from www.worldscientific.com 116 P. Baghel, S. Walia & M. M. Noohu and validity of TAI 3.0 in people with SCI in early by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. stage of rehabilitation. Introduction Methods Wheelchairs are the primary mode of mobility in individuals with spinal cord injury (SCI). The Sample transfers are vital in performing activities of daily living in wheelchair users. A transfer is considered A sample of convenience of 30 subjects with acute as a movement of oneself from one surface to an- traumatic SCI who met the inclusion criteria and other in one direction. If a full-time wheelchair user were willing to participate in the study was in- is unable to transfer, or transfers become more cluded. The sample was selected from rehabilita- di±cult, the individual's quality of life may be tion department of Indian spinal injuries center, further a®ected.1 New Delhi. The subjects were in the early reha- bilitation phase. The transfer training was started Transfers often cause excessive loading of the one week before the date of evaluation. The design arms. This may lead to upper extremity pain and was methodological research-repeated measure injuries, such as rotator cu® tears, elbow pain, and study. The study was approved by institutional carpal tunnel syndrome.2 The transfers may also ethical committee where the study was carried out. place joints of upper limb in extreme positions and expose to high internal joint forces. Therefore, it Sample size seems that prevention may be critical in reducing upper limb pain and overuse injuries. Learning the A sample size of 30 subjects with 2 observations per transfer methods that reduces forces and awkward subject achieves 91% power to detect an intraclass joint motions is an important strategy for prevent- correlation of 0.70 under the alternative hypothesis ing impairments in upper limb of wheelchair users.3,4 when the intraclass correlation under the null hy- pothesis is 0.30 using an F-test with a signi¯cance There is a wide variation in the amount and the level of 0.05 was required for reliabilty testing. A type of transfer training and no uniform way to sample size of 25 achieves 81% power with the al- evaluate transfer quality in early stage of rehabil- ternative hypothesis correlation of 0.60 by using a itation after SCI. Observation by a therapist and two-sided hypothesis test with a signi¯cance level qualitative assessment was the standard method of of 0.05 which was needed for validity testing.7,8 evaluating transfers. Therefore, McClure et al. de- The sample size was calculated using PASS 2008 veloped a tool called transfer assessment instru- software. ment (TAI) to assess the quality of transfer in full- time wheelchair users. It is a safe and easy tool to Subjects included were ¯rst time manual wheel- administer outcome measure to assess transfers in chair users who were potential full-time wheelchair wheelchair users. It was found to have an accept- users with tetraplegia, high paraplegia (T2–T7) or able interrater and intrarater reliability.5 Tsai low paraplegia (T8–L4). They were able to sit with et al., introduced the re¯ned TAI version 3.0. It or without hand support for 30 s. Subjects with fol- was found to have high reliability (0.74 to 0.88) lowing problems were excluded, such as, unhealed among raters of di®erent clinical backgrounds and pressure ulcers, > 19 score on Beck depression in- experience in people using wheelchairs for more ventory-II,9,10 wheelchair user's shoulder pain index than a year.6 score > 8:5,11 weight relief raises, musculoskeletal deformities of upper extremities, unstable medical In early stage of rehabilitation following SCI, condition (e.g., angina, seizures), respiratory dis- transfer evaluations are not done objectively and in tress, cardiovascular, emotional or psychiatric pro- a consistent manner. The transfer assessment can blems and with signi¯cant visual impairments. be in°uenced by the subjective experience of the therapists. It may lead to less accurate evaluations Procedure and variability in transfer skill assessment. TAI was the ¯rst tool to standardize the evaluation of The subjects were recruited by convenient sam- transfer technique. The items included in the TAI pling method who met the inclusion criteria. The were based on clinical practice guidelines, available information on literature, and best clinical prac- tices for transfers.5 There is a lack of valid assess- ment tool to evaluate transfers in early stage of rehabilitation following SCI. Therefore, the purpose of this study was to establish the reliability

Psychometric properties of TAI 3.0 in early rehabilitation phase of SCI 117 Hong Kong Physiother. J. 2018.38:115-123. Downloaded from www.worldscientific.com purpose of the study was explained to the subjects 0 points or \\not applicable\" (N/A), which means a by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. who gave the consent to participate in the study. removed item. Part 1 is completed after each Then their age (years), gender, height (cm), weight transfer and item scores are averaged to produce a (kg), level of injury and American spinal injury single representative item score. The part 1 is the association (ASIA) impairment level,12 area of summation of each item's score multiplied by 10 body a®ected, spinal cord independence-III mea- and then divided by the number of applicable sure (SCIM) score,13 modi¯ed functional reach test items, ranging from 0 to 10. The items in part 2 are score,14 type of transfer, duration since injury completed after all transfers have been performed. (months), body mass index (BMI) were collected. The 12 items in part 2 are scored on a likert scale ranging from 0 (strongly disagree) to 4 (strongly Five physiotherapists who had an experience of agree). The part 2 score is the summation of each 3–4 years in rehabilitation of the people with spinal item's score multiplied by 2.5 and then divided by cord injury (SCI) were the raters for transfers done the number of applicable items, resulting in a range by the participants. Four raters evaluated the of scores from 0 to 10. The ¯nal score of TAI is transfer skill of participants on TAI 3.0 while the the average of the part 1 and part 2. The items of ¯fth therapist evaluated the transfer on global as- the instrument and what is evaluated during a sessment of transfer scale (VAS). Before the test- transfer are given in Table 1. All the recruited ing, the raters were explained how to administer subjects completed both the sessions (session 1 and TAI 3.0 using the text instruction manual of the session 2) of the study.6 assessment tool. The instructions contained the details and how to score each item. In addition, Global assessment of transfer scale general recommendations were provided including where to stand when observing di®erent aspects of This is a likert scale which rates the overall transfer transfer. The raters were also told about the quality on a 10-point scale. The participants' instructions to be given to participants during transfer from a wheelchair was evaluated and transfers. The study investigator ensured that each graded from poor (0) to excellent (10). The criteria rater was trained properly using TAI 3.0.6 The for rating a transfer as poor were if the individual raters got individualized instruction on items with does not make use of equipment when needed, do di±culty in scoring. not make transfer easy and safe and inappropriate placement of the hand and feet. A transfer was Participants were asked to perform four transfers rated excellent if the transfer was appropriately each in two sessions (sessions 1 and 2). The transfers done without transfer devices, easy and safe, were done from their own wheelchair, to and fro from placing the hand and feet on right places and using a wheelchair level surface bed or a height adjustable human assistance when need. During session 1, hospital bed with their usual way of transfer. The while four raters evaluated the transfer skill of the height of the hospital bed was adjusted, depending participants, the ¯fth therapist who has not seen on the individual's preference. An attendant was the TAI rated the study participants transfer skill there with subject to prevent any fall or provide on a VAS. The VAS evaluation was done only in assistance during transfers. If participants needed session 1.5 assistance for transfers, they were permitted to use transfer device (e.g., transfer board or lift) or the Data Analysis attendant to provide assistance. In session 1, while participants performed transfer, four raters used Data analysis was performed using SPSS 21.0 TAI 3.0 to score and evaluate their transfer skills. software. Descriptive statistics were calculated For session 2, all participants were asked to return for the subject's demographic data including after 72 h to perform the transfer again. Then the age, gender, type of transfer, type of disability, same four raters evaluated them for second time BMI. The intraclass correlation coe±cients (ICCs) using TAI 3.0. Each session lasted for 30 to 45 min within each rater and between raters in part 1, part 2 per subject. and ¯nal TAI 3.0 scores were calculated to assess reliability. The limits of agreement (LOA) analysis TAI scoring (TAI 3.0) were done by plotting Bland and Altman (B&A) The TAI contains two parts. Part 1 comprises of 15 items and is scored as follows: \\yes\", 1 point, \\no\

,"118 P. Baghel, S. Walia & M. M. Noohu Table 1. The items and evaluation component of the TAI 3.0. Item no. What is being evaluated Part 1 1 Distance between the wheelchair and object to which he/she is transferring on to. The subject's wheelchair is within 3 inches of the object to which he is transferring on to. 2 The angle between the subject's wheelchair and the surface to which he is transferring. 3 Whether the subject attempts to place his chair to perform the transfer forward of the rear wheel. 4 If possible, the subject removes his armrest independently or with assistance. 5 Level or downhill transfer. 6 Placement of feet in a stable position. 7 Scoots to the front edge of the wheelchair seat before he transfers. 8 Hands' position. Hong Kong Physiother. J. 2018.38:115-123. Downloaded from www.worldscientific.com 9 Handgrip of the leading arm. by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. 10 Handgrip of the trailing arm. 11 Control over °ight. 12 Head–hip relationship. 13 Positioning of the lead arm. 14 The landing phase of the transfer. 15 The assistant supporting the subject's arms during the transfer. Part 2 Weight bearing 1 The lead arm position. arm position Set-up phase 2 Sets up for a safe and easy transfer. 3 Change the height of the object he is transferring to/from to make the transfer level. 4 Gets close to the object that he is transferring on to. 5 Uses handgrips when necessary. Conservation 6 Uses a transfer device when necessary. 7 Alternate the leading/trailing arm over the course of the assessment. Quality 8 Transfer is smooth and well controlled. 9 Clearly communicate his needs in transfer. 10 Does not allow the assistant to pull on his arms during the transfer. 11 The subject corrects the assistant. 12 The subject is able to correctly direct his care in an assertive and polite manner. plots with graph pad prism software (Prism version and 0.59 were moderate, and lower than 0.40 were 6.00).15 To determine variability of TAI 3.0 scores, weak.16 standard error of measurement (SEM) and mini- mal detectable change (MDC) were analyzed. Standard error of measurement Statistical signi¯cance was set at p 0:05 with con¯dence interval of 95%. To establish convergent The SEM provides a value for measurement error validity, Pearson correlation coe±cients were cal- in the same units as the measurement itself, it is a culated for each rater to evaluate the correlation of measure of absolute reliability.17 The SEM was TAI 3.0 scores (total) with global assessment of calculated for part 1, part 2 and ¯nal TAI scores transfer scores. using the formula: SEM ¼ SD Â ð1 À rÞ1=2, where SD is the standard deviation of the dataset and r is Reliability Testing the reliability coe±cient or ICC value. The SEM was calculated for individual rater based on ICC Interrater and intrarater reliability values in case of intrarater reliability analysis. For calculating intrarater and interrater reliability, Minimal detectable change ICC coe±cient value (ICC 3,1 : two-way mixed e®ect and consistency) was calculated separately The MDC was calculated for individual rater for for part 1, part 2 and total score of TAI3.0. ICCs intrarater reliability analysis. It is an estimate of higher than 0.80 were considered strong, between the smallest change in score that can be detected 0.60 and 0.79 were acceptable, between 0.40

Psychometric properties of TAI 3.0 in early rehabilitation phase of SCI 119 objectively for a subject, it is the amount by which of body a®ected, ASIA impairment level, gender, type of transfer, and duration since injury are a subject's score needs to change and ensure that shown in Table 2. The mean Æ SD of part 1 (TAI 1), the change is greater than measurement error.18 part 2 (TAI 2) and total score (TAI total) of TAI at two di®erent time points for all four raters is MDC was analyzed based on 95% con¯dence in- given in Table 3. The mean Æ SD global assessment terval (MDC ¼ 1:96 Â 21=2Â SEM) for part 1, of transfer scores was 7.9 Æ 1.48. part 2 and ¯nal TAI scores. Hong Kong Physiother. J. 2018.38:115-123. Downloaded from www.worldscientific.com Limits of agreement Reliability by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. LOA between session 1 and 2 were determined by The intrarater ICCs ranged from 0.93 to 0.95 for B&A method.19 TAI 1, 0.97 to 0.98 for TAI 2 and TAI total, sug- gesting high levels of reliability (Table 4). The Validity Testing interrater ICC of TAI 1 at ¯rst time was 0.98 and Convergent validity for TAI 2 and TAI total was 0.99 and for the sec- ond time, ICC of TAI 1, TAI 2 and TAI total was Pearson correlation coe±cient was calculated for 0.99 (Table 5). each rater by correlating their ¯nal TAI 3.0 scores with global assessment of transfer scores (VAS) of SEM and MDC session 1. The SEM among the raters for TAI 1 was from Results 0.34 to 0.43, TAI 2 was from 0.23 to 0.27 and for TAI total was from 0.23 to 0.28. The MDC among Demographic characteristic of the subjects, the raters for TAI 1 ranged from 1.19 to 0.94, TAI 2 mean Æ SD, such as age (years), height (cm), was from 0.69 to 0.86 and for TAI total was from weight (kg), BMI, SCIM score, modi¯ed functional 0.54 to 0.86 (Table 3). Between the raters, SEM reach test score, and frequency distribution of area and MDC is given in Table 5. Table 2. Demographic characteristics of the sample. Variable Mean Æ SD/n (%) Age (years) Male 31.9 Æ 12.3 Height (cm) Female 163.78 Æ 9.64 Weight (kg) 62.61 Æ 13.49 BMI (kg/m2) Tetraplegia 23.33 Æ 2.04 High paraplegia (T2–T7) 1.33 Æ 0.47 Duration(months) Low paraplegia (T8–L4) 40.32 Æ 11.64 SCIM-III score 10.27 Æ 3.05 mFR (cm) A Gender B 25 (83.3) C 5 (16.7) Area of body a®ected D 6 (20) ASIA impairment level Independent sitting pivot 5 (16.7) Assisted sitting pivot 19 (63.3) Type of transfer 11 (36.7) 11 (36.7) 5 (16.7) 3 (10) 20 (66.7) 10 (33.3) Notes: SCIM: spinal cord independence measure; mFR: modi¯ed functional reach; BMI: body mass index; SD: Standard deviation, n: number.

120 P. Baghel, S. Walia & M. M. Noohu Table 3. Mean Æ SD of TAI at two di®erent time points. Raters TAI 1 Time 1 TAI total TAI 1 Time 2 TAI total TAI 2 TAI 2 1 8.42 Æ 1.56 7.62 Æ 1.90 7.91 Æ 1.68 8.30 Æ 1.45 7.764 Æ 1.89 8.03 Æ 1.61 2 8.05 Æ 1.57 7.571 Æ 1.81 7.81 Æ 1.65 8.23 Æ 1.43 7.77 Æ 1.85 8.00 Æ 1.60 3 8.21 Æ 1.45 7.63 Æ 1.89 7.92 Æ 1.61 8.23 Æ 1.52 7.83 Æ 1.80 8.03 Æ 1.61 4 8.12 Æ 1.69 7.71 Æ 1.87 7.91 Æ 1.71 8.16 Æ 1.52 7.96 Æ 1.78 8.06 Æ 1.60 Notes: TAI 1: Part 1 of TAI; TAI 2: Part 2 of TAI; TAI total: TAI total score. Hong Kong Physiother. J. 2018.38:115-123. Downloaded from www.worldscientific.com Table 4. Intrarater reliability analysis for TAI 3.0. by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. Raters TAI 1 (95 CI) Difference Bewteeen Sessions 1 and 2 ICC (95 CI) MDC SEM TAI 2 (95 CI) TAI total (95 CI) TAI 2 TAI total TAI 2 TAI total 1 0.95 (0.89–0.97) 0.98 (0.94–0.99) 0.98 (0.94–0.98) 0.94 0.75 0.64 0.34 0.27 0.23 2 0.94 (0.88–0.97) 0.98 (0.96–0.99) 0.98 (0.94–0.98) 1.02 0.69 0.40 0.37 0.25 0.23 3 0.94 (0.86–0.96) 0.98 (0.96–0.99) 0.98 (0.94–98) 1.02 0.72 0.86 0.37 0.26 0.23 4 0.93 (0.85–0.96) 0.97 (0.94–0.98) 0.97 (0.93–0.98) 1.19 0.86 0.78 0.43 0.23 0.28 Notes: TAI 1: Part 1 of TAI; TAI 2: Part 2 of TAI; TAI total: TAI total score; ICC Intraclass correlation coe±cient; MDC: minimum detectable change; SEM: standard error of measurement; CI: con¯dence interval. Table 5. Interrater reliability analysis for TAI 3.0. ICC (95CI) MDC SEM Item Time 1 Time 2 Time 1 Time 2 Time 1 Time 2 Part 1 0.98 (0.97–0.99) 0.99 (0.94–0.99) 0.44 0.42 0.16 0.15 Part 2 0.99 (0.98–0.99) 0.99 (0.94–0.99) 0.53 0.50 0.19 0.18 Total 0.99 (0.96–0.99) 0.99 (0.96–0.99) 0.47 0.44 0.17 0.16 1. 5 1.0 + 1 .9 6 S D 0 .9 9 0. 5 Mean 0.0 -0 .1 5 - 0.5 -1 .9 6 SD -0 .6 8 - 1.0 0 5 10 15 Mean of Part 1 Session 1-Mean of Part 1 Session 2 Fig. 1. B&A plot of agreement between sessions 1 and 2 for part 1 score. The ¯gure reveals that only one data point lies outside Æ 1.96 SD.

Psychometric properties of TAI 3.0 in early rehabilitation phase of SCI 121Difference Bewteeen Sessions 1 and 2 Hong Kong Physiother. J. 2018.38:115-123. Downloaded from www.worldscientific.com 2 by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. + 1.96 SD 1 0.99 Mean 0.05 0 -1 .9 6 SD -1 -0.89 0 5 10 15 Mean of Part 2 Session 1-Mean of Part 2 Session 2 Fig. 2. B&A plot of agreement between sessions 1 and 2 for part 2 score. The ¯gure reveals that only one data point lies outside Æ 1.96 SD. Difference Bewteeen Sessions 1 and 2 1. 5 1. 0 + 1.96 SD 0 .7 6 0. 5 Me an 0 .1 0 0. 0 - 0.5 -1 .9 6 SD -0 .5 6 - 1.0 0 5 10 15 Mean of Total Score: Sessions 1 and 2 Fig. 3. B&A plot of agreement between sessions 1 and 2 for total score. The ¯gure reveals that only two data points lie outside Æ 1.96 SD. LOA Plots Convergent validity The LOA plot showed that there was 1 data Pearson correlation coe±cients ranged from 0.88 point for part 1 and part 2 outside þ1.96 SD to 0.90 among the raters with the VAS score (Figs. 1 and 2) and for total score, there was 2 data (p ¼ 0:001) as given in Table 6. points outside þ1.96 SD (Fig. 3). Discussion Table 6. Correlation of total TAI score with the criterion standard (VAS). Measurements of functional outcomes are an inte- gral part of any goal-orientated, interdisciplinary Rater Pearson correlation p rehabilitation program. It is important for quan- coe±cient tifying the success of rehabilitation program. A good clinical assessment tool should be both reli- 1 0.89 0.001 able and valid.20,21 The study results showed that 2 0.89 0.001 TAI 3.0 has higher levels of intrarater (ICCs ran- 3 0.88 0.001 ged from 0.93 to 0.98) and interrater reliability 4 0.90 0.001 (ICC — 0.99). For convergent validity, correla- tions ranged between 0.88 and 0.90. Results

Hong Kong Physiother. J. 2018.38:115-123. Downloaded from www.worldscientific.com 122 P. Baghel, S. Walia & M. M. Noohu uniform throughout the scale, but the measure- by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. ment error can vary at di®erent points in the showed signi¯cant correlations between TAI and scale.25 The raters were given handouts with an global assessment of transfer skills. The ¯ndings explanation of each item, a description of di®erent are similar, but psychometric properties appear scoring scenarios and a short practice session. better than the previous reliability and validity Instructions administering TAI 3.0 might have analysis done in wheelchair users with various improved the consistency among raters.6 disabilities with the same scale.5,6 The reason may be that all the subjects were with SCI and from the Currently, no other outcome measure exists to same rehabilitation care setup. This might have assess transfer quality in population with SCI made the sample more homogenous and another population in early rehabilitation phase. Hence, component might be the uniformity in transfer global assessment (VAS) was used to evaluate training. convergent validity. Using a non-validated tool to evaluate convergent validity is not only a preferred The reason for choosing ICCs to analyze the option, but also cannot be avoided because of a reliability of the TAI is because the ICC measures lack of a comparable criterion standard. The the association and agreement. The ICC can be results showed an excellent convergent validity. used to assess reliability for more than two raters The VAS was also previously used to establish the and can be used to analyze ordinal type of data.6 concurrent validity of TAI.5 The mean di®erence determined by LOA analysis was very small. Only total of four data points (one The objective evaluation of transfers may help each for parts 1 and 2 and 2 for total score) were the clinicians to improve the transfer training, outside the 95% con¯dence interval limits. Hence, identify and correct improper transfer techniques. there was an agreement with the two sessions of The identi¯cation of improper transfer techniques measurements.22 B&A plot is the quanti¯cation of may prevent musculoskeletal injuries and pain in the agreement between two measurements by the upper extremities. The evaluation was done plotting it graphically, the mean di®erence and only on participants in early phase of rehabilita- constructing LOA. The di®erence of the two paired tion, who performed independent or assisted sitting measurements is plotted against the average of the pivot transfers, so results cannot be generalized to two measurements. The LOA recommended by all full-time wheelchair users. Future researches B&A is that 95% of the data points should lie should be done to ¯nd out the e®ect of transfer within Æ 2SD of the mean di®erence. The part 1 training program on changes in TAI score in people and part 2 scores meet the LOA criteria whereas with SCI. total score is slightly less at 93.33%.19,23 Conclusion Scores may vary, given expected variability of individual performance and measurement error. A The TAI is a reliable and valid tool which can be measure of absolute variability provides useful in- used as an outcome measure to evaluate transfer formation to delineate the \\expected\" changes quality in people with acute SCI in early rehabili- from \\true\" changes in performance. Statistically, tation phase. absolute reliability is determined by the SEM. Clinically useful mechanism for looking at absolute Con°ict of Interest reliability is the MDC score.24 Results showed that SEMs ranged from 0.23 to 0.28 within raters We hereby declare that there is no con°ict of in- and 0.15 to 0.19 between raters. The smaller the terest involved in this study in terms of monetary SEM, the more accurate are the assessments that bene¯ts or in any other form. are being made. The smaller SEM in this study further indicates the accuracy of measurements Funding/Support with TAI 3.0.25 This study found that the MDC ranged from 0.64 to 0.86 within raters and 0.47 to The study did not receive any funds and it was 0.44 between raters. Minimum of 0.86 point change supported by Indian Spinal Injuries Center, New would be needed to identify a true di®erence in Delhi, India, by providing space to carry out the transfer skills that is not a measurement error. The study. We acknowledge the support of Chitra MDC is relatively easy to calculate which provides clinically relevant information. The limitation of MDC is that it assumes that detectable changes are

Psychometric properties of TAI 3.0 in early rehabilitation phase of SCI 123 Hong Kong Physiother. J. 2018.38:115-123. Downloaded from www.worldscientific.com Kataria, Principal, and all the supporting sta® of the First 6 Months after the Spinal Cord Injury. by 27.58.229.138 on 05/28/22. Re-use and distribution is strictly not permitted, except for Open Access articles. ISIC institute of rehabilitation sciences, New Delhi, Ann Rehabil Med 2012;36:119–25. India for this study. 11. Curtis KA, Roach KE, Applegate EB et al., Reli- ability and validity of the Wheelchair User's Shoul- Author Contributions der Pain Index (WUSPI). Paraplegia 1995;33: 595–601. The conception and design of the study, or acqui- 12. Maynard FM Jr, Bracken MB, Creasey G et al., sition of data were made by Preeti Baghel, Shefali International standards for neurological and func- Walia and Majumi. The analysis and interpreta- tional classi¯cation of spinal cord injury. Spinal tion of data were carried out by Shefali Walia and Cord 1997;35:266–74. Majumi. Drafting the paper or revising it critically 13. Itzkovich M, Gelernter I, Biering-Sorensen F et al., for important intellectual content was carried out The Spinal Cord Independence Measure (SCIM) by Preeti Baghel, Shefali Walia and Majumi. All version III: Reliability and validity in a multi-cen- authors approved the ¯nal submitted version of the ter international study. Disabil Rehabil 2007;29: manuscript. 1926–33. 14. Lynch SM, Leahy P, Barker SP. Reliability of References measurements obtained with a modi¯ed functional reach test in subjects with spinal cord injury. Phys 1. Rice LA, Smith I, Kelleher AR, Greenwald K, Ther 1998;78:128–33. Hoelmer C, Boninger ML. Impact of the clinical 15. Prism version 6.00 for Windows, GraphPad Soft- practice guideline for preservation of upper limb ware, La Jolla California USA, www.graphpad.com. function on transfer skills of persons with acute 16. Krebs DE. Declare your ICC type. Phys Ther spinal cord injury. Arch Phys Med Rehabil 1986;66:1431. 2013;94:1230–46. 17. Harvill LM. Standard error of measurement. Educ Meas 1991;10:33–41. 2. Preservation of Upper Limb Function Following 18. Finch E, Brooks D, Stratford PW, Mayo NE (eds.) Spinal Cord Injury: A Clinical Practice Guideline Physical rehabilitation outcome measures: A guide for Health-Care Professionals. Washington, DC: to enhanced clinical decision making. 2nd ed. Paralyzed Veterans of America, 2005. Hamilton: Canadian Physiotherapy Association, BC Decker Inc, 2002. 3. Tsai, CY, Hogaboom NS, Boninger ML, Koontz 19. Bland JM, Altman D. Statistical methods for AM. The relationship between independent trans- assessing agreement between two methods of clin- fer skills and upper limb kinetics in wheelchair ical measurement. Lancet 1986;1:307–10. users. BioMed Res Int 2014:984526. 20. Emerich L, Parsons KC, Stein A. Competent care for persons with SCI and dysfuction in acute in- 4. Mohammed K, Dunn JA. Shoulder pain in tetra- patient rehabilitation. Top Spinal Cord Inj Rehabil plegia. Orthop Trauma 2014:28:27–32. 2012;18:149–66. 21. Sullivan GM. A primer on the validity of assess- 5. McClure LA, Boninger ML, Ozawa H, Koontz AM. ment instruments. J Grad Med Edu 2011;3:119–20. Reliability and validity analysis of the transfer as- 22. Rankin G, Stokes M. Reliability of assessment sessment instrument. Arch Phys Med Rehabil tools in rehabilitation: An illustration of appro- 2011;92:499–508. priate statistical analyses. Clin Rehabil 1998;12: 187–99. 6. Tsai CY, Rice LA, Hoelmer C, Boninger ML, 23. Giavarina D. Understanding Bland Altman analy- Koontz AM. Basic psychometric properties of the sis. Biochem Med 2015;25:141–51. transfer assessment instrument (version 3.0). Arch 24. Donoghue D. How much change is true change? Phys Med Rehabil 2013;94:2456–64. The minimum detectable change of the berg bal- ance scale in elderly people. J Rehabil Med 7. Walter SD, Eliasziw M, Donner A. Sample size and 2009;41:343–6. optimal designs for reliability studies. Stat Med 25. Stratford PW, Binkley J, Solomon P, Finch E, Gill 1998;17:101–10. C, Moreland J. De¯ning the minimum level of de- tectable change for the Roland–Morris question- 8. Zar JH, Biostatistical Analysis. 2nd ed. Engelwood naire. Phys Ther 1996;76:359–65. Cli®s, NJ:Prentice-Hall, 1984. 9. Arnau RC, Meagher MW, Norris MP, Bramson R. Psychometric evaluation of the beck depression inventory-II with primary care medical patients. Health Psychol 2001;20:112–9. 10. Shin JC, Goo HR, Yu SJ, Kim DH, Yoon SY. Depression and Quality of Life in Patients within