224 Andrew et al: Physiotherapy service delivery, people with severe and persistent mental illness Table 1 Participants Subtheme: Lack of understanding of the role of physiotherapists in the Characteristics of focus group and interview participants. (n = 31) treatment of people with SPMI Characteristics 32 (11) When asked about their role in managing people with SPMI, many 21 (68) participants focused on patients presenting with pain and how that Age (yr), mean (SD) 10 (10) may be influenced by or may influence their mental illness. Partici- Gender, n female (%) pants also focused on the role of exercise in reducing symptoms of Clinical experience (yr), mean (SD) 31 (100) anxiety and fatigue. Overall, participants were uncertain of the role State or territory of residence, n (%) they could play in treating people with SPMI. 1 (3) Western Australia 1 (3) Theme: Barriers relating to people with SPMI Primary area of professional practice, n (%) 1 (3) 11 (35) Despite the majority of participating physiotherapists having cardiorespiratory 9 (29) limited or no contact with people with SPMI, a number of barriers disability 1 (3) that may interfere with physiotherapy management were identified. intensive care 1 (3) musculoskeletal 1 (3) Subtheme: Characteristics and behaviours of people with SPMI neurological 4 (13) Participants reported that obstacles in providing effective treat- non-clinical 1 (3) orthopaedics ment to people with SPMI could include motivation issues, mani- rehabilitation 12 (39) festing in low attendance rates (especially for follow-up hospital ward rotations 18 (58) appointments), and poor adherence to home exercise programs. women’s health Participants also found communication with this population Primary area of clinical work, n (%) 1 (3) extremely hard, making it difficult to establish a rapport. Other beliefs inpatient were that people with SPMI might be ‘unpredictable’ or ‘aggressive’. outpatient 18 (58) not applicable 9 (29) Subtheme: Limited knowledge of the role of physiotherapy in physical Place of primary employment, n (%) 3 (10) health amongst people with SPMI private practice (non-hospital) 1 (3) public hospital (non-tertiary) Participants believed that people with SPMI might have limited public hospital (tertiary) 15 (48) insight into their physical health conditions and limited awareness university 16 (52) about how physiotherapy could be helpful. Participants identified this Location of primary employment, n (%) as having an impact on both access to physiotherapy services, as well regional as the clients’ commitment to physiotherapy goals, and therefore urban treatment and management effectiveness. confirm key themes that were identified and discuss conflicts where Theme: Health system barriers necessary. Key themes were further refined once consensus was established. Qualitative participant characteristics and all quantitative Structural barriers within the healthcare system identified by results were reported using descriptive statistics. participants were categorised into three subthemes. Results Subtheme: Referrals The limited services provided to people with SPMI was Qualitative study perceived to be influenced by limited referrals from other health Thirty-one physiotherapists participated in either one of three professionals. General practitioners (GPs) were identified as the main focus groups (n = 20) or one of seven interviews with one or two source of referrals. However, physiotherapists questioned whether participants (n = 11). Participant characteristics are presented in GPs understood the relevance of physiotherapy intervention in this Table 1. The mean age of participants was 32 years and 68% were population. female. Areas of professional practice were diverse, including: musculoskeletal (36%) and neurological (29%) clinical areas; private Subtheme: Time and model of service delivery practice (58%) and public hospital (29%) settings; and urban (52%) The need for longer appointment times was identified because of and regional (48%) locations. the difficulties in building a rapport and communication with people No amendments to the focus group and interview transcripts were with SPMI. A rigorous structure of appointments was identified as requested by participants and four key themes were identified. unsuitable for this population due to the commonality of missed Themes and supporting quotes are summarised in Table 2. appointments. Participants also reported that a likely reliance on public transport by people with SPMI is also a challenge to rigid Theme: Physiotherapist-specific barriers appointment times. In general, the physiotherapists who were interviewed expressed Subtheme: Funding little confidence in their capacity to competently manage people with Participants identified those people with SPMI as being more SPMI. Personal barriers limiting treatment of this population were classified into four subthemes. likely to be in a low socio-economic group. Therefore, while it may be possible to build the capacity to manage people with SPMI within the Subtheme: Concerns about safety private sector, participants felt that the cost was likely to be prohib- Participants felt the behaviour of people with SPMI may pose itive for them. There is limited availability for physiotherapy in the public sector due to limited funding. Limited funding was also re- some physical risks and compromise their safety or that of their ported to have resulted in a decrease in physiotherapy staff on mental colleagues and other patients. This theme was primarily a concern health wards. voiced by physiotherapists working in outpatient settings. Theme: Enablers to improving access to physiotherapy services Subtheme: Limited knowledge and skill to manage people with SPMI Participants acknowledged a lack of both theoretical knowledge Subtheme: Education of physiotherapists Participants expressed the need for education in order to increase and practical skills to manage people with SPMI. In particular, they identified a gap in knowledge regarding the disorders encompassed their capacity to treat people with SPMI. Increased knowledge about within SPMI, and appropriate management and communication strategies. This theme was attributed to a lack of training at both undergraduate and postgraduate levels.
Research 225 Table 2 Themes, subthemes, and examples of supporting quotes derived from interviews and focus groups. Theme: Perceived personal barriers attributed to physiotherapists Subtheme: Concerns about safety of the practitioner and other patients I have immense fear for my staff safety (P21) I think there would be situations where some other clients might not feel as safe (P24) You feel uncomfortable or scared to address them [people with SPMI] (P6) Subtheme: Limited knowledge and skill to manage people with SPMI There is a bit of a gap in knowing how to increase their [people with SPMI] insight. and like how to get them to understand why it’s important to engage and participate (P4) There’s a distinct lack of training and our understanding of that [how to treat people with SPMI] and also our ability to communicate beyond what people have a natural affinity for doing (P28) Subtheme: Lack of understanding of the role of physiotherapists in the treatment of people with SPMI I don’t understand a lot about our role to be honest (P4) Theme: Perceived personal barriers attributed to people with SPMI by the physiotherapists Subtheme: Characteristics and behaviour of people with SPMI They can be quite difficult to treat, quite hard to motivate sometimes I find (P11) Difficult in terms of communication and developing rapport (P5) [People with SPMI have] unpredictable behaviour (P16) [People with SPMI are] people who are mumbling to themselves like almost that stereotypical kind of crazy (P7) Subtheme: Limited knowledge of the role of physiotherapy in physical health amongst people with SPMI. I guess they [people with SPMI] just don’t really know or can’t seem to comprehend or connect our role I suppose and they just lack insight into what we are actually doing (P6) Theme: Health system barriers Subtheme: Referrals It probably won’t even cross their [GPs] mind to refer to physio (P3) We don’t know where we could send them and we don’t ever get asked to refer them anywhere (P2) Subtheme: Time and model of service delivery And time you know keeping appointments, health department again you miss your appointment and then you’re off. and that sometimes doesn’t work for these clients (P17) There might be transport issues as well coming to appointments because they are relying on public transport and it doesn’t arrive on time for that appointment at 10 o’clock (P19) Subtheme: Funding [People with SPMI] financially may not be able to afford the rates for physio (P7) Theme: Enablers to improving access to physiotherapy services Subtheme: Education of physiotherapists Definitely better education and training in terms of what to expect from certain disorders and even just one or two things in terms of how to manage them or how to deal with them (P7) Subtheme: Funding models Maybe in the mental health care plan there should be acknowledgement that physiotherapy has a role. that would be a small token but it could help (P30) Subtheme: Service models Someone who knows the person well, and the person’s family and their condition well and they are the key contact person for that person so we keep going back to this one person yeah, I think that would help (P8) Subtheme: Referral pathways We should know where patients need to go and what service we should be referring them to (P11) SPMI = severe and persistent mental illness. the range of clinical presentations and symptoms, and information Quantitative study about more effective communication were perceived as critical in facilitating treatment. There was a consensus that education needs to Fifty-nine respondents completed the survey. Two respondents be embedded in the undergraduate course and made accessible to were excluded as they self-identified as allied health assistants. Sur- practising physiotherapists. vey respondent characteristics are presented in Table 3 and their clinical experience in mental health is summarised in Table 4. Forty- Subtheme: Funding models nine (86%) participants were female and mean age was 38 years. Shortcomings of current funding models were identified and Thirteen (23%) respondents had experience working in a mental health setting. Depression and anxiety were the mental health con- discussed. More funding was proposed to assist people with SPMI in ditions most commonly encountered (53% of respondents each). In managing the cost of private practice physiotherapy services. In people with SPMI, the most commonly managed physical health addition, participants felt that funding in the public sector was conditions were chronic pain (40%) and musculoskeletal issues (26%). necessary to ensure that physiotherapists had capacity to treat people When physiotherapists treated people with mental illness for their with SPMI. mental illness, they mostly administered exercise (65%) as well as lifestyle or motivation programs (58%). General practitioners were the Subtheme: Service models most common referral source (28%) followed by other specialist staff A multi-disciplinary team approach was thought to be favourable (25%). both in the outpatient and inpatient settings for these patients. Case Fifty-five (93%) respondents agreed that physiotherapists should management, and better communication pathways with occupational play a significant role in managing the physical health problems of therapists, psychologists, social workers and physiotherapists was people with SPMI. Thirty-three (58%) respondents felt confident identified as necessary to deliver effective treatment. Longer managing cardiovascular diseases, chronic pain, respiratory-related appointment times were also thought to be highly desirable to fully disorders, metabolic syndrome, and diabetes in the general popula- address the range of conditions in this population. tion. However, only 10 (18%) respondents felt confident managing these conditions in people with SPMI (Figure 1, with numerical data Subtheme: Referral pathways presented in Appendix 2 on the eAddenda). Receiving more referrals and having someone or somewhere to Only 16% of respondents felt that their workplace was thoroughly refer people with SPMI was important to the majority of participants. well equipped to manage the physical health needs of people with Participants felt that there was a need for community outpatient mental illness (Figure 2, with numerical data presented in Appendix 2 services with suitably trained physiotherapists to which they could on the eAddenda). Fifty-one (90%) respondents reported needing refer people with SPMI.
226 Andrew et al: Physiotherapy service delivery, people with severe and persistent mental illness Table 3 Table 4 Characteristics of survey respondents. Experience of survey respondents in mental health. Characteristics Participants Characteristics Participants (n = 57) (n = 57) Age (yr), mean (SD), range Gender, n female (%) 38 (13), 21 to 63 Experience working in a mental health setting, n yes (%) 13 (23) Clinical experience (yr), mean (SD), range 49 (86) Type of SPMI among patients seen, n (%) a,b State or territory of residence, n (%) 30 (53) 14 (12), 1 to 38 depression 30 (53) New South Wales 12 (21) Queensland 9 (16) anxiety 12 (21) South Australia 6 (11) schizophrenia Tasmania 3 (5) bipolar disorder 23 (40) Victoria 1 (2) Common physical health conditions among patients with SPMI, n (%) b 15 (26) Western Australia 11 (19) chronic pain 6 (11) Primary area of professional practice, n (%) 27 (47) musculoskeletal cardiorespiratory poor cardiovascular fitness 1 (2) gender health 4 (7) obesity 7 (12) gerontology 2 (4) other c intensive care 7 (12) Modalities physiotherapists used to treat people with SPMI, n (%) a,b 37 (65) mental health 2 (4) exercise (supervised) 38 (67) musculoskeletal 1 (2) exercise (other) 19 (33) neurological 20 (35) general physical activity advice 20 (35) non-clinical 4 (7) education 33 (58) paediatrics 1 (2) other d sports 3 (5) Referral sources to physiotherapists for people with SPMI, n (%) a,b 16 (28) chronic disease 2 (4) 6 (11) community health 1 (2) general practitioners 8 (14) orthopaedics 3 (5) self-referred 7 (12) pain 3 (5) other physiotherapists 14 (25) rehabilitation 1 (2) mental health professionals surgical 1 (2) other specialist staff e women’s health 1 (2) Primary area of clinical work, n (%) 1 (2) SPMI = severe and persistent mental illness. inpatient a This question allowed multiple responses. outpatient 17 (30) b This question was not offered to participants who answered ‘none’ to the question: not applicable 37 (65) Place of primary employment, n (%) What proportion of your current caseload includes people with a mental health illness that community health organisation 3 (5) private hospital is contributing to their presenting complaint? (eg, chronic pain, obesity, diabetes, smoking, private practice (non-hospital) 9 (16) public hospital (non-tertiary) 3 (5) etc.). public hospital (tertiary) 15 (26) c For example, medicinal overdose or functional neurological disorders. university 3 (5) d For example, lifestyle programs or motivational programs. department of education 19 (33) e For example, specialists or nursing staff. public rural health and private practice 4 (7) nursing home 1 (2) Discussion Location of primary place of employment, n (%) 1 (2) rural 2 (4) It is believed that this study is the first to examine physiothera- urban pists’ perspectives about their involvement in managing people with Secondary area of professional practice, n (%) 13 (23) mental health conditions. The majority of physiotherapists in the cardiorespiratory 44 (77) study considered treatment of people with SPMI to be part of their gender health job and that physiotherapists can play a role in managing the physical gerontology 3 (5) wellbeing of people with SPMI. This finding complements the large musculoskeletal 1 (2) body of literature supporting the positive impact of physiotherapists neurological 7 (12) in the treatment and management of physical health in the SPMI non-clinical 8 (14) population.9,14–17 However, personal and systemic barriers were sports 5 (9) identified to explain why physiotherapy services are not more sys- not applicable 10 (18) tematically accessible to people with SPMI. clinical Pilates 3 (5) rehabilitation 18 (32) Limited confidence in treating this population was a major Secondary area of clinical work, n (%) 1 (2) barrier identified by physiotherapists in both components of our inpatient 1 (2) study. This limited confidence was largely attributed to a lack of outpatient training at both undergraduate and postgraduate levels. In addi- not applicable 10 (18) tion, despite it being defined in both components of the study, it Place of secondary employment, n (%) 17 (30) appeared that many physiotherapists considered depression and community health organisation 30 (53) anxiety as forming the bulk of SPMI, reflecting a limited under- private hospital standing of the broad range of existing mental health conditions. private practice (non-hospital) 5 (9) These results are consistent with the findings of Connaughton and public hospital (non-tertiary) 3 (5) Gibson,18 who identified a gap in foundational knowledge about public hospital (tertiary) 8 (14) mental illness in undergraduate physiotherapy students due to university 2 (4) insufficiencies in physiotherapy curricula across Australia and New not applicable 9 (16) Zealand. These results also confirm the perception held by mental aged care 7 (12) health professionals and those with SPMI that physiotherapists do 21 (37) not have the necessary training to effectively manage people with 2 (4) these conditions.11 One implication of this finding is that access to mental health education for physiotherapists at an undergraduate education about how to effectively manage the physical health of and/or postgraduate level would greatly improve understanding of people with SPMI (Figure 3, with numerical data presented in mental illness, with the potential to substantially enhance work- Appendix 2 on the eAddenda). Further, 48 (84%) felt that they needed force capacity. information about how to communicate with people with SPMI appropriately (Figure 1). Consistent with previous literature,18–20 physiotherapists in our study recognised that theoretical and practical education is required to ensure that physiotherapists are better equipped to treat people with SPMI. This education may include
Research Percentage of participants 227 0% 20% 40% 60% 80% 100% Do you feel confident in managing comorbidities in the general population? Do you feel confident in managing comorbidities in people with mental illness? Do you feel confident in interacting and communicating with people with mental illness? Confident Somewhat confident Neutral Somewhat unconfident Not confident at all Unable to judge Figure 1. Percentage of survey responses for questions about confidence in managing people with severe and persistent mental illness (n = 57). Comorbidities include cardiovascular diseases, chronic pain, respiratory-related disorders, metabolic syndrome and diabetes. communication strategies, symptoms and clinical presentation of this study and Lee et al.11 Importantly, both studies question whether mental illness, antipsychotic medication side-effects and the role GPs know about physiotherapists’ relevance in this clinical area. This physiotherapy plays in the treatment of physical conditions in is of significance when discussing systemic barriers to physiotherapy people with SPMI.18,19,21 Some training opportunities and resources access. Importantly, these findings provide physiotherapists with the are offered to physiotherapists through the APA Mental Health opportunity to educate GPs about their willingness to receive Group, and to all health professionals through organisations such referrals for people with SPMI who present with the common, health- as Black Dog, Beyond Blue and Australian Red Cross. However, the related co-morbidities or sequelae of SPMI, for which physiothera- exact approach to training physiotherapists to treat this population pists have effective interventions. adequately needs to be further explored. Referrals of people with SPMI both to and from physiotherapists The gap identified in mental health training for physiotherapists was identified as being lacking or limited by both groups of physio- may be a major factor in other reported barriers, including stigma- therapists. However, it was still reported that GPs were the main tisation of those with SPMI and uncertainty of the role that physio- source of referrals to physiotherapy services, directly reflecting Lee therapists play. et al.11 Therefore, to ensure appropriate referrals to improve physio- therapy service access, it is critical that GPs are fully cognisant of the Participants in both components of our study expressed concerns role that physiotherapists can play in managing the physical health of about staff safety and the perceived bizarre and unpredictable people with SPMI. behaviours of people with SPMI. Such beliefs have been previously reported amongst other health professionals, including physiothera- Physiotherapists recognised that the structure and cost of the pists.22–24 Concerns about risk of violence were notable. Negative healthcare system was inadequate to accommodate people with perceptions about risks and behaviours in people with SPMI can be SPMI. High costs of private physiotherapy services was identified problematic because such perceptions have a detrimental effect on as a barrier for people with SPMI and has been previously help-seeking behaviours for mental health problems generally,25 and highlighted in the literature.12 Furthermore, the rigorous might influence access to physiotherapy services by people with appointment schedules that are typical in the Australian health- SPMI. While stereotypes about mental health are held by society in care system were reported to be ill-suited to this population. The general,26 and are therefore difficult to address, it has been previously commonality of missed healthcare appointments in people with reported that specific initiatives such as direct contact and education SPMI has also been previously represented in the literature.11,31,32 about mental illness in the health sector might work to improve As a result, fitting people with SPMI into the healthcare system negative prejudice.27–29 The results also highlight a need for better can be likened to fitting a square peg in a round hole. Alter- staff education about assessment of risk, and relationships with other ations to the healthcare system to make accommodations for this factors such as substance abuse, insight and illness severity, and how population is an essential step towards enabling greater access to to manage aggression.30 physiotherapy services. Uncertainty about the role that physiotherapists play in managing A major strength of this study was the mixed methods design. the physical health of those with SPMI is a common thread between Qualitative data collected from focus groups and interviews provided Percentage of participants 0% 20% 40% 60% 80% 100% Do you feel that your current workplace is well equipped to manage the physical health needs of people with mental illness? Well equipped Somewhat well equipped Neutral Somewhat ill equipped Not well eqipped at all Unable to judge Figure 2. Percentage of survey responses for question about workplace readiness for treating people with severe and persistent mental illness (n = 57).
228 Andrew et al: Physiotherapy service delivery, people with severe and persistent mental illness Percentage of participants 0% 20% 40% 60% 80% 100% Do you feel a need for more information and/or training about managing the physical health needs of people with mental illness? Definitely no Somewhat no Neutral Somewhat yes Definitely yes Unable to judge Figure 3. Percentage of survey responses for question about training requirements for treating people with severe and persistent mental illness (n = 57). in-depth information about the multifaceted issues, and survey data participants gave written informed consent before data collection about a sensitive topic could be provided with complete anonymity.33 began. Results from both study methods were complementary. Another strength was representation of both regional and urban, and private Competing interests: Nil. practice and hospital-based physiotherapists in both components of Source(s) of support: This study was funded by the School of our study to ensure that a variety of geographical locations and a Physiotherapy and Exercise Science at Curtin University. range of clinical experience were represented. Acknowledgements: The authors would like to acknowledge Ms Novia Minaee for her guidance in using SPSS software. This study had some limitations. The survey was originally Provenance: Not invited. Peer reviewed. circulated through APA networks. Members of the APA arguably hold Correspondence: Dr Robyn Fary, School of Physiotherapy and Ex- more contemporary, evidence-based views and this may have intro- ercise Science, Curtin University, Perth, Australia. Email: duced bias. The use of word of mouth and snowballing as a recruit- [email protected] ment strategy for the qualitative component may have limited capacity to generate a representative sample, and all qualitative data References were collected from Western Australian physiotherapists. However, the qualitative findings are reflected in the results of the survey, 1. Hardy S, White J, Deane K, Gray R. Educating healthcare professionals to act on the which were collected from across Australia, suggesting that there can physical health needs of people with serious mental illness: a systematic search for be confidence generalising results within Australia. evidence. J Psychiatric Mental Health Nurs. 2011;18:721–727. Physiotherapists are ideally poised to become leaders in the 2. Grubaugh AL, Clapp JD, Frueh BC, Tuerk PW, Knapp RG, Egede LE. Open trial of treatment and management of the physical health in people with exposure therapy for PTSD among patients with severe and persistent mental SPMI. However, a number of factors limit physiotherapists’ ability to illness. Behav Res Ther. 2016;78:1–12. provide adequate healthcare services to this population. Of impor- tance were the identification of limited education of physiotherapists 3. Alexandratos K, Barnett F, Thomas Y. The impact of exercise on the mental health regarding the treatment of people with SPMI; negative perceptions and quality of life of people with severe mental illness: a critical review. Br J Occup and beliefs held by physiotherapists about people with SPMI; and Ther. 2012;75:48–60. inadequate healthcare system structures. Physiotherapists interested in this area of healthcare are encouraged to seek out currently 4. Knapik GP, Graor CH. Engaging persons with severe persistent mental illness into available generic professional development opportunities. Further- primary care. J Nurse Pract. 2013;9:283–287. more, better undergraduate education of physiotherapists, physiotherapist-specific postgraduate training, and flexibility in 5. Robson D, Gray R. Serious mental illness and physical health problems: A discus- health service delivery are believed to be essential steps towards sion paper. Int J Nurs Stud. 2007;44:457–466. improving the physical health in this important yet overlooked population. 6. Dickerson F, Origoni A, Schroeder J, Schweinfurth LA, Stallings C, Savage CL, et al. Mortality in schizophrenia and bipolar disorder: clinical and serological predictors. What was already known on this topic: People with severe Schizophr Res. 2016;170:177–183. and persistent mental illness typically have poor physical health and comorbid conditions such as obesity, diabetes, and cardio- 7. Pina-Camacho L, Díaz-Caneja CM, Saiz PA, Bobes J, Corripio I, Grasa E, et al. Phar- vascular and respiratory disease. Physiotherapists are well macogenetic study of second-generation antipsychotic long-term treatment trained in delivering interventions that can assist with these metabolic side effects (the SLiM Study): rationale, objectives, design and sample comorbidities and reduce the risk of future health problems. description. Rev Psiquiatr Salud Ment. 2014;7:166–178. Physiotherapy has limited presence and advocacy within the multidisciplinary mental health team. 8. Bishop MD, Torres-Cueco R, Gay CW, Lluch-Girbes E, Beneciuk JM, Bialosky JE. What this study adds: Despite having effective interventions What effect can manual therapy have on a patient’s pain experience? Pain Manag. to offer people with mental illness, physiotherapists may not 2015;5:455–464. engage with this population because of insufficient education and confidence in this clinical area, health system structure, and 9. Mental health and physiotherapy. Australian Physiotherapy Association Web Site. stigmatisation of people with mental illness. Physiotherapists https://www.physiotherapy.asn.au/DocumentsFolder/Advocacy_Position_Mental_ should seek existing training opportunities in mental health, in- Health_2011.pdf. Accessed July 1, 2018. crease undergraduate and postgraduate education in mental health, and address health system barriers. 10. Morgan VA, Waterreus A, Jablensky A, Mackinnon A, McGrath JJ, Carr V, et al. People living with psychotic illness in 2010: the second Australian national survey eAddenda: Appendices 1 and 2 can be found online at https://doi. of psychosis. Aust N Z J Psychiatry. 2012;46:735–752. org/10.1016/j.jphys.2019.08.001. 11. Lee S, Waters F, Briffa K, Fary RE. Limited interface between physiotherapy primary Ethics approval: The Curtin University Human Research Ethics care and people with severe mental illness: a qualitative study. J Physiother. Committee (HREC) has approved this study (HRE2017-0866). All 2017;63:168–174. 12. Pope C, Mays N. Qualitative research: reaching the parts other methods cannot reach: an introduction to qualitative methods in health and health services research. BMJ. 1995;311:42–45. 13. Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32 item checklist for interviews and focus groups. Int J Qual Health Care. 2007;19:349–357. 14. Vancampfort D, Knapen J, Probst M, Scheewe T, Remans S, De Hert M. A systematic review of correlates of physical activity in patients with schizophrenia. Acta Psy- chiatr Scand. 2012;125:352–362. 15. Stubbs B, Soundy A, Probst M, De Hert M, De Herdt A, Vancampfort D. Under- standing the role of physiotherapists in schizophrenia: an international perspec- tive from members of the International Organisation of Physical Therapists in Mental Health (IOPTMH). J Ment Health. 2014;23:125–129. 16. Stubbs B, Soundy A, Probst M, Parker A, Skjaerven LH, Gyllensten AL, et al. Addressing the disparity in physical health provision for people with schizophrenia: an important role for physiotherapists. Physiother. 2014;100:185–186. 17. Stubbs B, Probst M, Soundy A, Parker A, De Herdt A, De Hert M, et al. Physio- therapists can help implement physical activity programmes in clinical practice. Br J Psych. 2014;204:164.
Research 229 18. Connaughton J, Gibson W. Physiotherapy students’ attitudes toward psychiatry and 26. Schulze B, Angermeyer MC. Subjective experiences of stigma. A focus group study mental health: a cross-sectional study. Physiother Can. 2016;68:172–178. of schizophrenic patients, their relatives and mental health professionals. Soc Sci Med. 2003;56:299–312. 19. Pope C. Recovering mind and body: a framework for the role of physiotherapy in mental health and well-being. J Public Ment Health. 2009;8:36–39. 27. Phelan M, Stradins L, Morrison S. Physical health of people with severe mental illness: can be improved if primary care and mental health professionals pay 20. Vancampfort D, Rosenbaum S, Probst M, Connaughton J, du Plessis C, Yamamoto T, attention to it. Br Med J. 2001;322:443–444. et al. Top 10 research questions to promote physical activity in bipolar disorders: a consensus statement from the International Organization of Physical Therapists in 28. Rüsch N, Angermeyer MC, Corrigan PW. Mental illness stigma: concepts, conse- Mental Health. J Affect Disord. 2016;195:82–87. quences, and initiatives to reduce stigma. Eur Psychiatry. 2005;20:529–539. 21. Vancampfort D, De Hert M, Skjerven LH, Gyllensten AL, Parker A, Mulders N, et al. 29. Stubbs A. Reducing mental illness stigma in health care students and professionals: International Organization of Physical Therapy in Mental Health consensus on a review of the literature. Australas Psychiatry. 2014;22:579–584. physical activity within multidisciplinary rehabilitation programmes for mini- mizing cardio-metabolic risk in patients with schizophrenia. Disabil Rehabil. 30. Fazel S, Långström N, Hjern A, Grann M, Lichtenstein P. Schizophrenia, substance 2012;34:1–12. abuse, and violent crime. JAMA. 2009;301:2016–2023. 22. Peer J, Warnecke A, Baum C, Goreczny A. stigmatization of people with schizo- 31. Pastore P, Griswold K, Homish G, Watkins R. Family practice enhancements for phrenia: perspectives of graduate students in various healthcare fields. Intl J Ment patients with severe mental illness. Community Mentl Health J. 2013;49:172–177. Health. 2015;44:186. 32. Coodin S, Staley D, Cortens B, Desrochers R, McLandress S. Patient factors associ- 23. Rao H, Mahadevappa H, Pillay P, Sessay M, Abraham A, Luty J. A study of stigma- ated with missed appointments in persons with schizophrenia. Can J Psychiatr. tized attitudes towards people with mental health problems among health pro- 2004;49:145–148. fessionals. J Psychiatr Ment Health Nurs. 2009;16:279–284. 33. Huan-Niemi E, Rikkonen P, Niemi J, Wuori O, Niemi J. Combining quantitative and 24. Setchell J. What has stigma got to do with physiotherapy? Physiother Can. qualitative research methods to foresee the changes in the Finnish agri-food sector. 2017;69:1–5. Futures. 2016;83:88–99. 25. Clement S, Schauman O, Graham T, Maggioni F, Evans-Lacko S, Bezborodovs N, Websites et al. What is the impact of mental health-related stigma on help-seeking? A systematic review of quantitative and qualitative studies. Psychol Med. www.qualtrics.com 2014;45:11–27.
Journal of Physiotherapy 65 (2019) 243–245 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Appraisal Research Note: Prognostic model research: overfitting, validation and application Introduction modelling (Table 1), increasing the risk of overfitting. Better ap- proaches are imputation methods,10 where missing values may be In physiotherapy, many prognostic models have been developed substituted with the mean or the mode with single imputation, and m to predict future outcomes after musculoskeletal conditions, completed data sets are created with multiple imputation procedures. including neck pain.1 Prognostic models combine several character- Multiple imputation is recommended, because single imputation ig- istics to predict the risk of an outcome for individual patients and nores potential correlation of predictors and leads to an underesti- may enable personalised prevention and care. In practice, they can be mation of variability of predictor values among subjects.11 This may used to inform patients and relatives on prognosis, and to support lead to an overestimation of the precision of regression coefficients. clinical decision-making. Moreover, models may be useful to stratify Imputation methods are widely available through modern statistical patients for clinical trials. Prediction models are increasingly being software. published, including 99 prognostic models for neck pain alone, pre- dicting recovery (pain reduction, reduced disability, and perceived It is difficult to select the most promising predictors. Selection of recovery).2 Although guidelines for developing and reporting prog- candidate predictors based on literature and expert knowledge is nostic models have been proposed,3,4 a recently proposed assessment often preferred over selection based on a relatively limited dataset.10 tool found that many prognostic models in physiotherapy are prone Also, some related predictors can sometimes be combined in simple to risks of bias.2,5 scores. For example, comorbid conditions are often combined in a comorbidity score,12 and frailty in the elderly can be scored according Various limitations have been noted regarding design and ana- to various characteristics.13 After selection of candidate predictors, lyses, which make models at risk of overfitting.2 Overfitting relates to the set of predictors may be reduced; this can be done using uni- the notion of asking too much from the available data, which results variate analysis and/or stepwise methods. However, both approaches in overly optimistic estimates of model predictive performance; re- do not truly reduce the problem of statistical overfitting, since the sults that cannot be validated in underlying or related populations.6 model specification is driven by findings in the data. Univariate Consequently, the model may predict poorly, with serious limita- analysis is common as a first step to select the most potent risk fac- tions when the model is applied in clinical practice: it does not tors, which are then used in multivariable analysis. This approach was separate low-risk from high-risk patients (poor discrimination), and followed in the development of the OMPQ (Table 2). A common may give unreliable or even misleading risk estimates (poor alternative is to use backward stepwise selection from a model that calibration). includes all candidate predictors, as was done by Schellingerhout to develop a model to predict non-specific neck pain (Table 2). Stepwise We aim to describe a number of challenges related to the design selection procedures are known to result in biased regression coef- and analysis in different stages of prognostic model research, and ficient estimates (testimation bias).6 A modern approach to reduce opportunities to reduce overfitting (summarised in Table 1). We such testimation bias and overfitting is by shrinkage of regression emphasise validation before the application of prediction models is coefficients towards zero.10 A key example of this approach is the considered in clinical practice. For illustration, we consider the Öre- Least Absolute Shrinkage and Selection Operator, which penalises for bro Musculoskeletal Pain Screening Questionnaire (OMPQ) (Table 2).7 the absolute values of the regression coefficients. It shrinks some The model has extensively been validated, and its use is recom- coefficients to zero, which means that predictors are dropped from mended by clinical guidelines.8 We also consider the Schellingerhout the model. non-specific neck pain model predicting recovery after six months (Table 2),9 which was indicated as one of the few externally validated Validation: apparent, internal and external performance models with a low risk of bias.2 Model development The aim of prognostic models is to provide accurate risk pre- dictions for new patients. Therefore, validation of prognostic models The development of a prognostic model involves a number of is crucial. Three types of validation can be distinguished: apparent, steps. These include handling of missing data, selection and coding of internal and external validation. predictor variables, choosing between alternative statistical models, and estimating model parameters.10 Prognostic models are usually Apparent validation entails the assessment of model performance developed with multivariable regression techniques on data from directly in the derivation cohort. Because the regression coefficients (prospective) cohort studies, while machine learning techniques are are optimised for the derivation cohort, this provides optimistic es- gaining increased attention. timates of the model’s performance (overfitting). To correct for overfitting, several internal validation procedures are available. Missing data is common in prognostic research. A complete case Bootstrap resampling and cross-validation provide stable estimates analysis is often conducted (ie, the exclusion of participants that have with low bias and are therefore recommended.10 missing data on one or multiple predictor variables, resulting in smaller sample size). As a consequence, the number of events per Before a prognostic model can be applied in practice it is crucial variable may drop below the number deemed necessary for reliable to explore how the model performs outside the setting in which it was developed, preferably across a range of settings. External https://doi.org/10.1016/j.jphys.2019.08.009 1836-9553/© 2019 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
244 Appraisal Research Note Table 1 Overview of challenges and opportunities categorised by the stage of prognostic model research in which they occur, and illustrated with two prediction models.7,9 Stage of Challenges Opportunities Örebro Schellingerhout prognostic Musculoskeletal non-specific model research Pain Screening Questionnaire neck pain model Design Insufficient sample size Collaborative efforts to reach No information on EPV Restricted to 17 predictors . 10 EPV, cross-validate Complete case analysis based on EPV (10) Development Inappropriate handling across setting Univariate analysis Multiple imputation with of missing data; Multiple imputation methods 5 repetitions Development complete case analysis Apparent validation Backward stepwise selection Selection of predictors Shrinkage and penalisation in Yes Internal based on univariate multivariable Externally validated; AUC, Apparent validation validation analysis or stepwise analysis but no calibration plot External selection procedures Yes validation Apparent validation or Bootstrap resampling or Externally validated; AUC and External inefficient internal cross-validation calibration plot validation validation procedures Full model equation is Present full model equation not presented No external validation Validation of models in cohort other than development cohort through collaborative research EPV = events per variable; AUC = area under the receiver operating characteristic curve. validity relates to the generalisability of the prognostic model to high-risk and low-risk patients. It can be measured by the concor- another population.10 A cross-validation across different non- dance statistic (C-statistic or area under the receiver operating random parts of the development data gives an indication of characteristic curve: AUC). The AUC ranges between 0.50 (no external validity.14 Heterogeneity in predictor effects across set- discrimination) and 1.00 (perfect discrimination). For instance, the tings indicates that the model should be calibrated to each specific OMPQ was validated in an observational study of patients with acute setting, to achieve robust model performance across settings. To back pain in Australia.17 At external validation of the OMPQ the AUC enable external validation of the model the full model equation was 0.80 (95% CI 0.66 to 0.93) for absenteeism at 6 months (Table 2).17 should be presented in the paper (Table 1). The OMPQ has been The discriminative ability of the Schellingerhout non-specific neck extensively validated in international cohorts,15 while such pain model was lower: AUC 0.66 (95% CI 0.61 to 0.71) at development, external validation is rare for other prognostic models for muscu- and validation cohort AUC 0.65 (95% CI 0.59 to 0.71).9 loskeletal conditions.2,16 Calibration refers to the agreement between predicted and Performance measures observed probabilities. This agreement can be illustrated with a calibration graph. Ideally, the plot shows a 45-deg line with calibra- Model performance at internal and external validation is tion slope 1 and intercept 0. Calibration is more informative at commonly expressed with discrimination and calibration. Discrimi- external than internal validation because a model is expected to nation indicates the ability of the model to differentiate between provide correct predictions for the derivation cohort it is fitted on. At external validation, the Schellingerhout non-specific neck pain score chart showed reasonable calibration (Figure 1); it slightly Table 2 Overview of prognostic model characteristics of the Örebro Musculoskeletal Pain Screening Questionnaire and the Schellingerhout non-specific neck pain model. Örebro Musculoskeletal Schellingerhout non-specific Pain Screening neck pain model Questionnaire Development n = 137; adult patients; acute/subacute back pain; Sweden7 n = 468; adult patients (18 to 70 yrs); Patient population non-specific neck pain; primary care; of development Accumulated sick leave; 6 months follow-up The Netherlands9 cohort Global perceived recovery; dichotomised into Outcome 21 predictors: physical functioning, fear-avoidance beliefs, ‘recovered or much improved’ versus ‘persistent the experience of pain, work, and reactions to the pain complaints’; 6 months follow-up Predictors 9 predictors: age, pain intensity, previous neck complaints, radiation of pain, accompanying low back pain, accompanying headache, employment status, health status, and cause of complaints External validation External validation n = 106; adult patients; acute/subacute low back pain; n = 346; adult patients (18 to 70 yrs); Model performance workers’ compensation and medical practitioner referral; non-specific neck pain; primary care; randomised observational study; Australia17 controlled trial; PANTHER trail; United Kingdom9 AUC 0.80 (CI 95% 0.66 to 0.93); no calibration plot AUC 0.65 (CI 95% 0.59 to 0.71); calibration plot Application Recommended in clinical guidelines as screening instrument,8 Score chart9 Practical application and used to select trial participants19
Appraisal Research Note 245 0.8 Summary Observed risk for persistent complaints 0.6 The aim of prognostic models for predicting future outcomes after musculoskeletal conditions is to provide accurate and patient-specific 0.4 estimates of the risk of relevant clinical outcomes such as delayed recovery. These models may be applied in primary care to identify 0.2 patients likely to have poor outcomes. Most models in physiotherapy have been judged to be at moderate to high risk of bias.2 Approaches 0.0 0.2 0.4 0.6 0.8 to reduce overfitting should be better utilised. These include appro- 0.0 Predicted risk for persistent complaints priate handling of missing data, careful selection of predictors with domain knowledge, and internal and external validation (Table 1). Figure 1. Calibration of the Schellingerhout non-specific neck pain score chart in Assessment of performance across a range of settings may show external validation cohort. suboptimal results, specifically with respect to calibration of pre- dictions. Such suboptimal performance may motivate updating of a = deciles of risk model before it can be considered for application in a specific —— = Perfect calibration setting.10 Furthermore, clinical impact studies are recommended to Adapted from Schellingerhout et al.9 assess the (cost-) effectiveness of a prognostic model in clinical practice. The presentation format of a prognostic model is also important, as this can facilitate implementation of prognostic models in clinical practice to improve decision-making and outcome by personalised medicine. Competing interests: Nil. Sources of support: Nil. Acknowledgements: None. Provenance: Invited. Not peer reviewed. Correspondence: Isabel RA Retel Helmrich, Public Health, Center for Medical Decision Making, Erasmus MC-University Medical Center Rotterdam, The Netherlands. Email: [email protected] overestimated the risk of persistent complaints in adult patients Isabel RA Retel Helmricha, David van Klaverena,b and presenting with non-specific neck pain.9 More severe miscalibration Ewout W Steyerberga,c is common for prediction models.18 aDepartment of Public Health, Center for Medical Decision Making/ Application of prognostic models in practice Erasmus MC-University Medical Center Rotterdam, The Netherlands bPredictive Analytics and Comparative Effectiveness Center, Institute for A prognostic model is more likely to be applicable for imple- mentation in practice if the model was developed with high-quality Clinical Research and Health Policy Studies/Tufts Medical Center, data from an appropriate study design, and with careful statistical Boston, USA analysis.10 Even better is when the model is externally validated in the setting where it is to be used.14 For instance, the OMPQ is recom- cDepartment of Biomedical Data Sciences, Leiden University Medical mended in clinical guidelines to be applied in screening to predict Center, Leiden, The Netherlands delayed recovery,8 and was used to select trial participants,19 likely motivated by the extensive and positive external validation studies References across multiple settings. When a prognostic model is deemed appro- priate for implementation, the impact (clinical effectiveness and costs) 1. Kelly J, et al. Musculoskelet Sci Pract. 2017:155–164. of the use of the model in clinical practice should be studied.4 Although 2. Wingbermühle RW, et al. J Physiother. 2018;1:16–23. recommended, these clinical impact studies are scarce, and some 3. Collins GS, et al. BMC Med. 2015;1:1. prediction models have been recommended to be used in clinical 4. Steyerberg EW, et al. PLoS Med. 2013;2:e1001381. practice without adequate evaluation of their (cost-)effectiveness. 5. Wolff RF, et al. Ann Intern Med. 2019;1:51–58. 6. Babyak MA. Clin J Pain. 1998;3:209–215. The presentation of clinical prediction models is important to 7. Linton SJ, et al. Clin J Pain. 1998;3:209–215. facilitate implementation of prognostic models in practice. The Schel- 8. ACC. New Zealand acute low back pain guide. https://www.acc.co.nz/assets/ lingerhout model was presented as a score chart that can readily be used by physicians. Although the score chart may be easy to use, pre- provider/ff758d0d69/acc1038-lower-back-pain-guide.pdf. Accessed 14 June, 2019. dictions of risks are only approximate because continuous predictors 9. Schellingerhout JM, et al. Spine J. 2010;17:E827–E835. are categorised and regression coefficients are rounded. The score chart 10. Steyerberg EW. Stat Methods Med Res. 2007;3:277–298. should ideally be externally validated across various settings before it 11. Ambler G, et al. Stat Methods Med Res. 2007;3:277–298. can be considered for use in broader practice. Other common formats 12. Charlson ME, et al. J Chronic Dis Manag. 1987;5:373–383. include web-based calculators and apps for mobile devices.10,20 13. Searle SD, et al. BMC Geriatr. 2008;1:24. 14. Steyerberg EW, et al. J Clin Epidemiol. 2016:245–247. 15. Hockings RL, et al. Spine J. 2008;15:E494–E500. 16. van Oort L, et al. J Clin Epidemiol. 2012;12:1257–1266. 17. Linton SJ, et al. Clin J Pain. 2003;2:80–86. 18. Riley RD, et al. BMJ. 2016:i3140. 19. Schmidt CO, et al. BMC Musculoskelet Disord. 2010;1:5. 20. Bonnett LJ, et al. BMJ. 2019:l737.
Journal of Physiotherapy 65 (2019) 215–221 j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j p hy s Research Short-term cryotherapy did not substantially reduce pain and had unclear effects on physical function and quality of life in people with knee osteoarthritis: a randomised trial Lucas Ogura Dantas a, Carolina Carreira Breda a, Paula Regina Mendes da Silva Serrao a, Francisco Aburquerque-Sendín b,c, Ana Elisa Serafim Jorge a, Jonathan Emanuel Cunha a, Germanna Medeiros Barbosa a, Joao Luiz Quagliotti Durigan d, Tania de Fatima Salvini a a Physical Therapy Department, Federal University of São Carlos, Brazil; b Departamento de Ciencias sociosanitarias, Radiología y Medicina física, Universidad de Córdoba, Spain; c Instituto Maimónides de Investigación Biomédica de Córdoba, Spain; d Physical Therapy Division, University of Brasilia, Distrito Federal, Brazil KEY WORDS ABSTRACT Osteoarthritis Objective: Does short-term cryotherapy improve pain, function and quality of life in people with knee Cryotherapy osteoarthritis (OA)? Design: Randomised controlled trial with concealed allocation, blinded assessment of Randomised trial some outcomes, and intention-to-treat analysis. Participants: People living in the community with knee OA. Knee Interventions: The experimental group received cryotherapy, delivered as packs of crushed ice applied to the Physical therapy knee with mild compression. The control group received the same regimen but with sham packs filled with sand. The interventions were applied once a day for 4 consecutive days. Outcome measures: Participants were assessed at baseline and on the day after the 4-day intervention period. The primary outcome was pain intensity according to a visual analogue scale. Secondary outcomes were baseline to post-intervention changes according to the Western Ontario and McMaster Universities Osteoarthritis, Knee injury and Oste- oarthritis Outcome; Timed Up and Go test; and 30-Second Chair to Stand test. Results: Sixty participants were randomised into the experimental group (n = 30) or the control group (n = 30). Twenty-nine partici- pants from each group completed the trial. The mean between-group difference in change in pain severity was 20.8 cm (95% CI 21.6 to 0.1), where negative values favour the experimental group. This result did not reach the nominated smallest worthwhile effect of 1.75 cm. The secondary outcomes had less-precise esti- mates, with confidence intervals that spanned worthwhile, trivial and mildly harmful effects. Conclusion: Short-term cryotherapy was not superior to a sham intervention in terms of relieving pain or improving function and quality of life in people with knee OA. Although cryotherapy is considered to be a widely used resource in clinical practice, this study does not suggest that it has an important short-term effect, when compared with a sham control, as a non-pharmacological treatment for people with knee osteoarthritis. Registration: NCT02725047. [Dantas LO, Breda CC, da Silva Serrao PRM, Aburquerque-Sendín F, Serafim Jorge AE, Cunha JE, Barbosa GM, Durigan JLQ, Salvini TdF (2019) Short-term cryotherapy did not sub- stantially reduce pain and had unclear effects on physical function and quality of life in people with knee osteoarthritis: a randomised trial. Journal of Physiotherapy 65:215–221] © 2019 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Introduction inexpensive, and easy to administer for healthcare professionals and Knee osteoarthritis (OA) is a prevalent and costly chronic patients. Moreover, it can be prescribed in isolation or in conjunction musculoskeletal condition associated with pain and disability.1 Clin- with other therapies,5 and seems to be well accepted by people with ical guidelines recommend a combination of non-pharmacological knee OA.7,8 treatments2 – including patient education, exercise and some other physiotherapy interventions – together with pharmacological treat- Some international knee OA guidelines recommend cryotherapy ments3 to improve pain and symptoms. as a treatment option,9,10 but others have found insufficient evidence to recommend it.11–13 Recent and relevant systematic reviews Cryotherapy is a non-pharmacological intervention that is widely used in various rheumatic joint diseases4,5 for its effects on pain, conclude that further trials are needed to evaluate the isolated effects inflammation and oedema.6 It is considered to be relatively safe, of cryotherapy on pain, function and quality of life in people with knee OA.14–16 The most recent of these reviews16 identified five randomised trials, almost all of which scored only 4 out of 10 on the https://doi.org/10.1016/j.jphys.2019.08.004 1836-9553/© 2019 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
216 Dantas et al: Cryotherapy for knee osteoarthritis PEDro scale. When the authors applied the Grading of Recommen- knees to determine whether they had knee OA based on the clinical dations, Assessment, Development and Evaluation (GRADE) approach and radiographic criteria of the American College of Rheumatology.20 to the evidence, they concluded that the review identified only low- Participants were required to have a symptomatic and radiographic level evidence regarding cryotherapy for pain management, knee grade (Kellgren-Lawrence scale) of 2 (at least mild radiographic OA) stiffness, knee range of motion and physical function.16 in at least one knee compartment.17 To be included in the study prospective participants also needed to: be aged between 40 and 75 Some consistent methodological limitations among the existing years; be engaged in , 45 minutes/week in accumulated physical studies were: unconcealed allocation, lack of blinding, poor baseline activity of at least moderate intensity;21 have a body mass index 35 comparability of the groups, lack of confirmation of OA grade of the kg/m2; and have pain intensity in the previous week of 4 cm on a participants, and analysis not according to the principle of intention 10-cm visual analogue scale.17 Exclusion criteria were: physiotherapy to treat.16 Future studies should therefore seek to achieve these in the previous 3 months; intra-articular knee injections in the pre- methodological criteria, where possible. In this way, the generated vious 6 months; medical restrictions such as cardiorespiratory, information is likely to contribute to strategies for targeted knee OA neurological or any other rheumatology dysfunctions; previous hip, rehabilitation, and improve pain management and overall quality of knee or ankle surgery; and any other chronic condition that leads to life of people with knee OA.5,14,16 pain. The aim of this study was to determine the effects of short-term Intervention cryotherapy in people with knee OA. We hypothesised that cryo- therapy would relieve pain and improve function and quality of life Experimental group when compared with a sham intervention. The experimental intervention consisted of four cryotherapy ses- Therefore, the study question for this randomised trial was: sions performed by a trained physiotherapist over 4 consecutive days. Cryotherapy was only applied to the more-affected knee, in an air- Does short-term cryotherapy improve pain, function and quality conditioned room controlled at 21 C (6 2). The therapist explained of life in people with knee osteoarthritis? that the intervention consisted of cryotherapy applied to the more- affected knee for 20 minutes. Participants were positioned in dorsal Methods decubitus with both legs extended and relaxed. To protect the skin from possible frostbite, the entire knee surface was covered with a Design moist surgical gauze (45 3 50 3 0.01 cm). Next, two plastic bags (24 3 34 3 0.08 cm), each containing 1 kg of crushed ice, were placed This study was a randomised sham-controlled trial carried out on the knee, covering the anterior, posterior, medial and lateral sur- over a period of 6 consecutive days. A baseline assessment was per- faces. A comfortable, non-painful compression was applied over the formed on the first day, followed by 4 days of intervention and a post- ice packs by wrapping an elastic bandage around them, and the intervention assessment on the final (sixth) day. Each participant was therapy was left in situ for 20 uninterrupted minutes. The main assessed in the same period of the day (morning or afternoon) in a purposes of compression were to maintain the ice packs in position physiotherapy research laboratory by the same assessor. To reduce on the knee22 and to enhance the effects of the cryotherapy.23 To bias, both the therapist responsible for applying the intervention and allow participants to mimic the usual clinical setting treatment, they the outcomes assessor followed standardised scripts to give expla- were provided all the necessary materials (plastic bags, elastic nations regarding the general objective of the study.17 Moreover, the bandage and surgical gauze) to use cryotherapy at home whenever therapist responsible for the intervention participated in a 10-hour they felt in pain or discomfort. Moreover, they received a booklet with training module before the start of the study, which consisted of illustrated pictures and all the instructions needed for cryotherapy scientific information and clinical training regarding cryotherapy application. effects and application for people with knee OA. Intervention adherence, medication intake and adverse events were tracked with a Control group 4-day assessment diary given to the participants at the baseline For the control intervention, the bags were filled with 1 kg of dry assessment. All the participants were told to not practise any kind of physical exercises/activities during the intervention week. sand instead of ice. The sand bags were applied according to the same regimen in the same locations. The therapist’s explanation about the Participants were randomly allocated into two groups of 30: an intervention was changed to mention ‘application of sand packs’ experimental group that received cryotherapy, and a control group instead of ‘cryotherapy application’. The sand packs were applied that received a sham intervention. Random allocations were deter- with the same gauze underneath and the same bandage for mined by a computer-generated random numbers program and compression. Participants in the control group were provided with matched for gender (15 men and 15 women in each group). Alloca- the sand bags and other materials for application at home whenever tion was concealed by placing the random allocations in opaque they felt in pain or discomfort. The booklet was modified to refer to sealed envelopes that were locked in a central location. Each partic- the application of sand bags. ipant’s random allocation was revealed just before the intervention was commenced.17 Outcome measures Verbal and written explanations of the objectives and methodol- All the outcomes were measured by the same blinded assessor ogy of the study were provided to the patients, and those who were before and after intervention. Table 1 describes the main outcome willing to participate signed a written informed consent form measures included in this study and the recommended estimate of approved by the local ethics committee. The study was reported ac- the minimum clinically important difference for each outcome cording to the Consolidated Standards of Reporting Trials (CONSORT) measure. Pain intensity, knee subjective and objective physical Statement for Randomised Trials of Nonpharmacologic Treatments18 function, and quality of life were measured. and the Template for Intervention Description and Replication checklist (TIDieR).19 Primary outcome The primary outcome was pain intensity assessed using a visual Participants analogue scale. This self-reported pain score is a valid and reliable Participants were recruited through public announcements and measure among people with OA.24 The visual analogue scale was waiting lists from local orthopaedic and rheumatology outpatient administered at baseline and on the final assessment day. clinics. People who expressed interest in participating in the study underwent lateral, anteroposterior and axial radiography of both
Research 217 Table 1 Detailed description of the study’s outcome measures. Outcome measure Description of the test Scoring Minimum clinically important difference Visual analogue scale The scale is placed in front of the patient who is The scale ranges from 0 (no pain) to A pain reduction of 1.75 cm asked to rate their pain intensity in the previous 10 cm (maximum pain intensity). is recommended in OA research.33 Western Ontario & week.24 Each question is scored from 0 to 4. An improvement of 12% from baseline McMaster Universities This self-report questionnaire assesses the The maximum score is 96. is recommended in OA research.34 Osteoarthritis questionnaire problems experienced by people with lower High scores indicate worse status. limb OA in the previous 72 hrs. It contains A difference of 8 to 10 in the total Knee Injury and 24 questions in three domains: pain, stiffness The answers are standardised and score is recommended in OA research.35 Osteoarthritis and physical function. scored from 0 to 4. The total score is Outcome Score This self-report questionnaire assesses the problems 168. High scores indicate worse status. A reduction of 0.8 to 1.4 s is experienced by people with lower limb OA in the recommended in OA research.36 Timed Up and Go test previous week, by measuring quality of life and knee Total time to complete the test. function. It contains 42 questions in five domains: An increase of 2 to 3 repetitions is 30-Second Chair to Stand test pain; other symptoms; function in daily life; Total number of repetitions recommended in OA research.36 sports-related function and recreation; and within 30 s. knee-related quality of life. This test assesses: balance moving from sitting to standing, stability in walking, and gait course changes without using compensatory strategies. The participant is asked to stand up from a chair, walk 3 m, turn around, return and sit back in the chair. A chair with no arms is placed against a wall to prevent oscillations. Patients sit in the middle of the chair, with their back straight and feet resting on the floor in line with their shoulders. The participant is asked to rise from sitting to standing as many times as possible in 30 s. OA = osteoarthritis. Secondary outcomes radiographic screening. Of these, 60 participants matched the eligi- The Western Ontario and McMaster Universities Osteoarthritis bility criteria and were randomised, of whom 58 completed the intervention. The baseline demographic characteristics of each group (WOMAC) questionnaire was used to assess knee function and asso- of participants are presented in Table 2. The baseline scores on the ciated problems. The Knee Injury and Osteoarthritis Outcome Score outcome measures are presented in the first two columns of Table 3. (KOOS) was used to assess knee function and quality of life. Two objective physical functional tests were also used: the Timed Up and Adherence to the study protocol Go test and 30-Second Chair to Stand test. All registered outcome measures are reported in this manuscript. To collect preliminary data to support future randomised One participant in each group did not complete the four scheduled controlled trials, pressure pain thresholds (algometry) and knee skin intervention sessions. These participants also declined to attend the temperature (thermography) were measured. The data from these post-intervention assessments, so their data were imputed as outcomes are presented in Appendix 1 on the eAddenda. described above. Data analysis In the experimental group, the adherence diary showed that 12 (41%) of the participants used the cryotherapy intervention at home. A blinded biostatistician performed all analyses using commercial Of these, 10 participants used it between one and three times and two softwarea. The Kolmogorov-Smirnov test was applied to evaluate data participants used it more than three times. distribution and all variables showed p . 0.05. A two-factor analysis of variance was conducted for the primary outcome (visual analogue In the control group, the adherence diary showed that 19 (66%) of scale for pain) and secondary outcomes, with time (baseline and the participants used the sham intervention at home. Of these, five post-intervention) as the within-subject factor and group (experi- participants used it between one and three times and 14 participants mental or control) as the between-subject factor. Tukey’s test was used it more than three times. used for post-hoc analysis when necessary and an intention-to-treat analysis was performed for all randomised participants. Missing data Effect of intervention were replaced using the expectation maximisation method. Between- group differences and their 95% CIs were reported and interpreted Primary outcome against the nominated thresholds for minimum clinically important The data about pain severity measured using the 10-cm visual difference. For the algometry and thermography data, where mini- mum clinically important differences were not nominated, Cohen’s analogue scale are presented in Table 3. The individual participant d coefficient was calculated to aid interpretation. An effect size . 0.8 data are presented in Tables 4 and 5 on the eAddenda. was considered large, around 0.5 moderate, and 0.2 small.25 The mean between-group difference in change in pain severity Sample size was based on a significance level of 0.05 and power of was 20.8 cm (95% CI 21.6 to 0.1). That is, although pain severity 0.90 to detect a difference of 1.75 cm on the visual analogue scale, reduced in both groups, the mean between-group difference favoured assuming a standard deviation of 2.00 cm.26 Based on these criteria, the experimental group by indicating 0.8 cm greater reduction in pain 29 participants with knee OA were required in each group. To allow severity than in the control group. Neither that mean estimate nor for possible dropouts during the intervention period, 30 participants the 95% CI reached the nominated threshold (ie, a reduction in pain were recruited per group. severity of 1.75 cm) for the minimum clinically important difference. Therefore, the data in this study are consistent with a range of Results possible effects on pain severity that are not beneficial enough to make undertaking the intervention worthwhile. Flow of participants through the study Secondary outcomes Figure 1 shows the design of the trial and flow of participants Pre-intervention and post-intervention results of the WOMAC, through the trial. Of the 188 volunteers, 83 attended the physical and KOOS, Timed Up and Go test, and 30-Second Chair to Stand test are
218 Dantas et al: Cryotherapy for knee osteoarthritis People with knee osteoarthritis assessed for eligibility by telephone (n = 188) Excluded (n = 105) . did not meet inclusion criteria (n = 90) . declined to participate (n = 11) . other (n = 4) People assessed by physical screening and radiography (n = 83) Excluded (n = 23) .. did not meet inclusion criteria (n = 20) .. declined to participate (n = 0) .. other (n = 3) Day 1 Measured pain on the visual analogue scale, Western Ontario & McMaster Universities Osteoarthritis questionnaire, Knee Injury and Osteoarthritis Outcome Score, Timed Up and Go test, and 30-second Chair to Stand test Randomised (n = 60) (n = 30) (n = 30) Loss to follow-up ..Experimental group . . Control group Loss to follow-up cryotherapy sham cryotherapy .(n = 1) 20 minutes per day × 20 minutes per day × .(n = 1) time restriction 4 days, plus additional 4 days, plus additional time restriction (n = 1) home use if pain (n = 1) home use if pain Measured pain on the visual analogue scale, Western Ontario & McMaster Universities Osteoarthritis questionnaire, Knee Injury and Osteoarthritis Outcome Score, Timed Up and Go test, and 30-second Chair to Stand test Day 6 (n = 29) (n = 29) Analysed (imputation of missing data via expectation maximisation method) (n = 30) (n = 30) Figure 1. Design and flow of participants through the trial. shown in Table 3. The individual participant data are presented in reduction (ie, improvement) in the WOMAC score than in the control Tables 4 and 5 on the eAddenda. group. In order to compare this result with the minimum clinically important difference, the score can be converted into percentage The mean between-group difference in change in WOMAC score change: 28% (95% CI 221 to 5). The mean estimate of 28% does not was 26 points (95% CI 213 to 2). That is, the mean difference fav- reach the nominated threshold (ie, a change of 212% from baseline), oured the experimental group by indicating 6 points greater whereas the 95% CI spans the smallest worthwhile effect and also spans no effect. Therefore, the WOMAC data in this study do not Table 2 provide a very precise estimate; the true average effect of cryotherapy Baseline characteristics of the participants. may be worthwhile (221 to 212%), too small to be worthwhile (211 to 0%), or mildly harmful (1 to 5%). Characteristic Exp Con (n = 30) (n = 30) The results for the remaining secondary outcome measures are similar to the results for WOMAC (Table 3). For each outcome mea- Age (yr), mean (SD) 60 (7) 60 (7) sure, the 95% CI spans the minimum clinically important difference Gender, n (%) female 15 (50) 15 (50) and no effect, so the study data are consistent with worthwhile, Weight (kg), mean (SD) 85.0 (18.4) 83.5 (19.6) trivial and mildly harmful effects, as detailed below. Height (cm), mean (SD) 166 (9) 164 (9) Body mass index (kg/m2), mean (SD) 30.7 (5.8) 31.0 (7.6) Con = control group, exp = experimental group.
Research 219 Table 3 Mean (SD) of groups and mean (95% CI) within and between-group differences. Outcome Groups Difference within groups Difference between groups Day 6 minus Day 1 Day 6 minus Day 1 Day 1 Day 6 Exp Con Exp Con Exp Con Exp minus Con (n = 30) (n = 30) (n = 30) (n = 30) Pain visual analogue scale (0 to 10) 6.8 6.8 1.6 2.3 25.2 24.4 20.8 WOMAC (0 to 96) (1.8) (1.2) (1.7) (1.9) (2.5) (2.3) (21.6 to 0.1) KOOS (0 to 168) Timed Up and Go test (s) 44 41 12 15 232 226 26 30-s Chair to Stand test (repetitions)a (15) (18) (12) (14) (19) (22) (213 to 2) 89 84 27 30 262 254 (23) (25) (18) (23) (29) (33) 28 8.6 9.0 7.5 8.3 21.1 20.8 (220 to 4) (1.7) (2.1) (1.2) (1.5) (2.2) (6.0) 10 10 13 12 20.3 (2) (3) (3) (4) 3 2 (21.1 to 0.4) (3) (3) 1 (21 to 2) Shaded row = primary outcome. Small anomalies in subtraction are due to the effects of rounding. Negative between-group differences favour the experimental group, except where indicated. Con = control group, exp = experimental group, KOOS = Knee Injury and Osteoarthritis Outcome Score, WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index. a A positive between-group difference favours the experimental group for this outcome. The mean between-group difference in change in KOOS was 28 consider that the effect on pain severity justifies the use of cryo- points (95% CI 220 to 4). That is, the mean difference favoured the therapy on a short-term basis. The estimates of the effect of short- experimental group by indicating 8 points greater reduction (ie, term cryotherapy on the secondary outcome measures were less improvement) in the KOOS than in the control group. However, the precise. For each secondary outcome, the confidence interval ranged confidence interval spanned from 220 points (ie, a worthwhile from worthwhile to mildly harmful effects. Therefore, this study reduction in KOOS because it is greater than the minimum clinically does not generate any clear implications about whether or not important difference) to 4 points (ie, a mildly harmful effect). short-term cryotherapy should be recommended to improve WOMAC, KOOS, Timed Up and Go test or the 30-Second Chair to The mean between-group difference in change in the Timed Up Stand test in people with knee OA. and Go test was 20.3 seconds (95% CI 21.1 to 0.4). That is, the mean difference favoured the experimental group by indicating a 0.3- As a natural analgesic and anti-inflammatory,6 cryotherapy is second greater reduction (ie, improvement) in the Timed Up and widely used in clinical practice to reduce pain and thereby improve Go test than in the control group. However, the confidence interval function and quality of life. However, our findings do not support shows that the effect of cryotherapy on the Timed Up and Go test may such widespread use. Our results also do not support the tentative be worthwhile (95% CI 21.1 to 20.8), trivial (20.7 to 0.0), or mildly observation (based on low-quality evidence) from a previous sys- harmful (0.1 to 0.4). tematic review that cryotherapy was more effective than control groups using untuned short-wave diathermy and electrodes with no The mean between-group difference in change in the 30-Second electrical current, respectively, for pain, stiffness, knee range of mo- Chair to Stand test was 1 repetition (95% CI 21 to 2). That is, the tion, and physical performance improvements.16 Our results indicate mean difference favoured the experimental group by indicating that any beneficial effect of cryotherapy on pain is meagre, and would greater improvement by 1 repetition than in the control group. not be considered worthwhile, which seems consistent with a study However, the confidence interval shows that the effect of cryotherapy that suggests that people with knee OA prefer heat rather than on the 30-Second Chair to Stand test may be worthwhile (2 repeti- cryotherapy.8 However, that result does not seem consistent with a tions), trivial (1 repetition), or mildly harmful (21 repetition). previous study where cryotherapy was found to be as effective as transcutaneous electrical nerve stimulation for quadriceps activation Medication use in people with knee OA.27 In total, 27 (93%) of the participants in the cryotherapy group did Regarding the secondary outcomes, the between-group differ- not use any analgesic or anti-inflammatory drugs during the inter- ences that were observed were smaller than the smallest worthwhile vention period; two participants (7%) used medication only once for effects recommended for use in OA research, but the 95% CIs were other symptoms but not knee pain. In the control group, 29 (100%) of wide enough to be unclear about whether the effect was or was not of the participants did not use any analgesic or anti-inflammatory drugs clinical importance. While this suggests that further research into the during the intervention period. effect of cryotherapy on the secondary outcomes could be under- taken, worthwhile effects seem unlikely because benefits on the Adverse events secondary outcomes would presumably occur via reducing pain, and All 29 patients in each group who were followed up at the end the effect on pain was small. of the study reported no adverse events during the intervention The differences observed within groups cannot be taken as an period. indication of the effect of the intervention, since there are a number of factors that can explain what happened. These include Discussion regression to the mean,28 where participants are more likely to improve after a consultation regardless of intervention, due to We believe that this is the first study to assess the isolated ef- symptom fluctuation;29 polite patients effect,29 where patients do fects of short-term cryotherapy compared with a sham control in not want to fail the therapist treating them; and the placebo ef- people with knee OA. The results showed that the effects of short- fect,30–32 where the knowledge of receiving a topical treatment, term cryotherapy application were not sufficiently superior to a with the considerable attention given to all the participants, their sham control to make the intervention worthwhile. The study’s expectations concerning the upcoming therapy, and bio- estimate of the effect on pain severity was small enough that the psychosocial elements could all have indirectly influenced the re- confidence interval did not exceed the nominated smallest worth- sponses to the interventions applied. while effect. Therefore, although this study indicates that the effect of short-term cryotherapy on pain severity is beneficial, the effect is The recommendations of current clinical guidelines vary small enough that people with knee OA typically would not regarding cryotherapy for knee OA.2 The American College of
220 Dantas et al: Cryotherapy for knee osteoarthritis Rheumatology and the National Institute for Health and Care Excel- References lence conditionally recommend cryotherapy as a complementary treatment option for people with knee OA.9,10 However, the Osteo- 1. Hunter DJ, Bierma-Zeinstra S. Osteoarthritis. Lancet. 2019;393:1745–1759. arthritis Research Society International, the European League Against 2. Collins NJ, Hart HF, Mills KAG. OARSI year in review 2018: rehabilitation and Rheumatism, and the Ottawa Panel did not achieve expert panel consensus and failed to report cryotherapy in their final recommen- outcomes. Osteoarthr Cartil. 2019;27:378–391. dations.11–13 Our results agree with these latter guidelines. However, 3. Mandl LA. Osteoarthritis year in review 2018: clinical. Osteoarthr Cartil. 2018;27:1–6. the most appropriate interpretation of the total body of evidence will 4. Demoulin C, Vanderthommen M. Cryotherapy in rheumatic diseases. Jt Bone Spine. come when our study is incorporated into a high-quality systematic review on this topic. 2012;79:117–118. 5. Guillot X, Tordi N, Mourot L, Demougeot C, Dugué B, Prati C, et al. Cryotherapy in It is important to underscore that well-established reporting guidelines were followed to improve the evidence synthesis inflammatory rheumatic diseases: A systematic review. Expert Rev Clin Immunol. regarding this topic.18,19 The methodology of this study was designed 2014;10:281–294. to minimise potential for bias by including concealed treatment 6. Bleakley CM, Davison GW. Cryotherapy and inflammation: evidence beyond the allocation, and blinding of the outcome assessor and biostatistician. cardinal signs. Phys Ther Rev. 2010;15:430–435. Participants presented radiographically confirmed knee OA and a 7. Porcheret M, Jordan K, Jinks C. Primary care treatment of knee pain - A survey in sufficient level of pain to ensure ample scope for improvement. older adults. Rheumatology. 2007;46:1694–1700. Moreover, the amount of missing outcome data was relatively small 8. Denegar CR, Dougherty DR, Friedman JE, Schimizzi ME, Clark JE, Comstock BA, et al. for a randomised controlled trial. Despite these strengths, the study Preferences for heat, cold, or contrast in patients with knee osteoarthritis affect also exhibited some limitations. The therapist who delivered cryo- treatment response. Clin Interv Aging. 2010;5:199–206. therapy or the placebo intervention, and the patients were not blin- 9. Chae KJ, Choi MJ, Kim KY, Ajayi FF, Chang IS, Kim IS. National Institute for Health and ded. Although this may have resulted in bias, given the study design, Care Excellence. Osteoarthritis: Care and Management. December 2014. a method of blinding patients and therapists to thermal agents has 10. Hochberg MC, Altman RD, April KT, Benkhalti M, Guyatt G, McGowan J, et al. yet to be established. In addition, there was not a no-treatment group American College of Rheumatology 2012 recommendations for the use of non- and there was not a follow-up period to evaluate the residual effect of pharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and each intervention. knee. Arthritis Care Res. 2012;64:465–474. 11. McAlindon TE, Bannuru RR, Sullivan MC, Arden NK, Berenbaum F, Bierma- In conclusion, in people with symptomatic knee OA, the short- Zeinstra SM, et al. OARSI guidelines for the non-surgical management of knee term use of cryotherapy was not substantially superior to a sham osteoarthritis. Osteoarthr Cartil. 2014;22:363–388. control in terms of relieving pain, and it had uncertain effects on 12. Fernandes L, Hagen KB, Bijlsma JWJ, Andreassen O, Christensen P, Conaghan PG, function and quality of life. Although cryotherapy is widely used in et al. EULAR recommendations for the non-pharmacological core management of clinical practice, and recommended by some treatment guidelines, hip and knee osteoarthritis. Ann Rheum Dis. 2013;72:1125–1135. the estimates of its effects in this study suggest that it might not have 13. Brosseau L, Taki J, Desjardins B, Thevenot O, Fransen M, Wells GA, et al. The Ottawa a worthwhile effect as a treatment for knee OA. panel clinical practice guidelines for the management of knee osteoarthritis. Part two: Strengthening exercise programs. Clin Rehabil. 2017;31:596–611. What was already known on this topic: People with 14. Brosseau L, Yonge KA, Robinson V. Thermotherapy for treatment of osteoarthritis. symptomatic knee osteoarthritis experience pain and disability. Cochrane Database Syst Rev. 2003;4:CD004522. Cryotherapy is sometimes used to treat knee osteoarthritis. 15. Brosseau L, Rahman P, Toupin-April K, Poitras S, King J, De Angelis G, et al. Clinical guidelines for osteoarthritis differ in their recommenda- A systematic critical appraisal for non-pharmacological management of osteoar- tions about cryotherapy. Recent systematic reviews conclude thritis using the appraisal of guidelines research and evaluation II instrument. PLoS that further evidence about cryotherapy for knee osteoarthritis is One. 2014;9:e82986. needed. 16. Dantas LO, Moreira R de FC, Norde FM, Mendes Silva Serrao PR, Alburquerque- What this study adds: In people with symptomatic knee Sendín F, Salvini TF. The effects of cryotherapy on pain and function in individuals osteoarthritis, any beneficial effect of short-term application of with knee osteoarthritis: a systematic review of randomized controlled trials. Clin cryotherapy on pain was small enough that most people would Rehabil. 2019:026921551984040. not consider it to be worthwhile. The effects of short-term 17. McAlindon TE, Driban JB, Henrotin Y, Henrotin Y, Hunter DJ, Jiang GL, et al. OARSI cryotherapy on function and quality of life in people with Clinical Trials Recommendations: Design, conduct, and reporting of clinical trials symptomatic knee osteoarthritis were unclear. for knee osteoarthritis. Osteoarthr Cartil. 2015;23:747–760. 18. Barbour V, Bhui K, Chescheir N, Schulz KF, Ravaud P. CONSORT Statement for Footnotes: a SPSS 24.0 software, SPSS Inc, Chicago, USA. randomized trials of nonpharmacologic treatments: A 2017 update and a CONSORT eAddenda: Tables 4 and 5 and Appendix 1 can be found online at extension for nonpharmacologic trial abstracts. Ann Intern Med. 2017;167:40–47. https://doi.org/10.1016/j.jphys.2019.08.004. 19. Hoffmann TC, Glasziou PP, Boutron I, Milne R, Perera R, Moher D, et al. Better Ethics approval: This study was approved by the Institutional reporting of interventions: template for intervention description and replication Ethics Committee of Federal University of Sao Carlos, São Paulo, (TIDieR) checklist and guide. BMJ. 2014;348:g1687. Brazil. Registration approval number: (CAEE: 65966617.9.0000.5504). 20. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, et al. Development of All participants gave written informed consent before data collection criteria for the classification and reporting of osteoarthritis. Classification of began. The trial was conducted according to the Helsinki Statement. osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the Competing interests: All authors have completed a uniform American Rheumatism Association. Arthritis Rheum. 1986;29:1039–1049. disclosure form and declare that they have no conflicts of interest. 21. Dunlop DD, Song J, Lee J, Gilbert AL, Semanik PA, Ehrlich-Jones L, et al. Physical Sources of support: This study was supported by Sao Paulo activity minimum threshold predicting improved function in adults with lower- Research Foundation (FAPESP # 2015/21422-6; 2011/22122-5) and extremity symptoms. Arthritis Care Res. 2017;69:475–483. Conselho Nacional de Desenvolvimento Científico e Tecnológico 22. Silva P, Lott R, Wickrama S, Mota J, Welk G. Elastic bandaging for orthopedic and (CNPQ # 401333/2016-7). sports injuries prevention and rehabilitation: a systematic review. Int J Sport Nutr Acknowledgements: The authors thank O J Neto for performing Exerc Metab. 2016;32:1–44. the RX exams at the UFSCar University Hospital. 23. Song M, Sun X, Tian X, Zhang X, Shi T, Sun R, et al. Compressive cryotherapy versus Provenance: Not invited. Peer reviewed. cryotherapy alone in patients undergoing knee surgery: a meta-analysis. Spring- Correspondence: Professor Tania de Fatima Salvini, Physical erplus. 2016;5:1–12. Therapy Department, Federal University of São Carlos, São Carlos, 24. Hawker G, Mian S, Kendzerska T, French M. Measures of adult pain: Visual Analog Brazil. Email: [email protected] Scale for Pain (VAS Pain), Numeric Rating Scale for Pain (NRS Pain), McGill Pain Questionnaire (MPQ), Short-Form McGill Pain Questionnaire (SF-MPQ), Chronic Pain Grade Scale (CPGS), Short Form-36 Bodily Pain Scale (SF-36 BPS). Arthritis Care Res. 2011;63:240–252. 25. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale NJ, USA: L. Erlbaum Associates; 1998. 26. Hinman RS, Heywood SE, Day AR. Aquatic physical therapy for hip and knee osteoarthritis: results of a single-blind randomized controlled trial. Phys Ther. 2007;87:32–43. 27. Pietrosimone BG, Hart JM, Saliba SA, Hertel J, Ingersoll CD. Immediate effects of transcutaneous electrical nerve stimulation and focal knee joint cooling on quadriceps activation. Med Sci Sports Exerc. 2009;41:1175–1181. 28. Bland JM, Altman DG. Statistics Notes: Some examples of regression towards the mean. BMJ. 1994;309:780. 29. Kamper SJ. Engaging with research: linking evidence with practice. J Orthop Sport Phys Ther. 2018;48:512–513. 30. Bannuru R, McAlindon T, Sullivan M, Wong J, Kent D, Schmid C. Effectiveness and implications of alternative placebo treatments: a systematic review and network meta-analysis of osteoarthritis trials. Ann Intern Med. 2015;163:365–372. 31. Zhang W, Robertson J, Jones C, Dieppe P, Doherty M. The placebo effect and its determinants in osteoarthritis: meta-analysis of randomised controlled trials. Ann Rheum Dis. 2008;67:1716–1723. 32. Dieppe P, Goldingay S, Greville-Harris M. The power and value of placebo and nocebo in painful osteoarthritis. Osteoarthr Cartil. 2016;24:1850–1857.
Research 221 33. Bellamy N, Carette S, Ford P, Kean WF, Lussier A, Wells GA, et al. Osteoarthritis 35. Roos EM, Toksvig-Larsen S. Knee injury and Osteoarthritis Outcome Score (KOOS) - Antirheumatic Drug Trials. III. Setting the Delta for Clinical Trials–Results of a validation and comparison to the WOMAC in total knee replacement. Health Qual Consensus Development (Delphi) Exercise. J Rheumatol. 1992;19:451–457. Life Outcomes. 2003;1:17. 34. Angst F, Aeschlimann A, Stucki G. Smallest detectable and minimal clinically impor- 36. Dobson F, Hinman RS, Roos EM, Abbott JH, Stratford P, Davis AM, et al. OARSI tant differences of rehabilitation intervention with their implications for required recommended performance-based tests to assess physical function in people sample sizes using WOMAC and SF-36 quality of life measurement instruments in diagnosed with hip or knee osteoarthritis. Osteoarthr Cartil. 2013;21:1042– patients with osteoarthritis of the lower ex. Arthritis Rheum. 2001;45:384–391. 1052.
Search
