Abstract
Objective. To assess the quality of nonpharmacological care received by people with knee and/or hip osteoarthritis (OA) in the community and to assess the associated factors.
Methods. We conducted a postal survey to evaluate 4 OA quality-of-care indicators for knee/hip OA: (1) advice to exercise; (2) advice to lose weight; (3) assessment for ambulatory function; and (4) assessment for nonambulatory function, including dressing, grooming, and arising from a seated position. Eligible participants were identified from the administrative database of British Columbia between 1992 and 2006.
Results. In total, 1349 participants reported knee and/or hip OA [knee only = 700 (51.9%); hip only = 261 (19.3%); knee and hip = 388 (28.8%)]. Their mean age was 67.1 years (SD 11.1); 816 (60.5%) were women, and 921 (68.3%) were diagnosed with OA for 6 years or longer. The overall pass rate of the 4 quality indicators was 22.4% (95% CI 20.5, 24.3). The pass rate for the individual quality indicator ranged from 6.9% for assessment of nonambulatory function to 29.2% for receiving assessment of ambulatory function. Receiving exercise advice was associated with having a university degree (vs high school diploma; OR 3.10, 95% CI 2.00, 4.80). Receiving weight-loss advice was associated with being female (OR 2.64, 95% CI 1.71, 4.08), being aged 55–64 years (compared to being aged 75 and over; OR 1.96, 95% CI 1.02, 3.76), and having higher Western Ontario and McMaster Universities Osteoarthritis Index scores (for every 10-point increment; OR 1.14, 95% CI 1.02, 1.26). On the other hand, having less than a high school education reduced the odds of weight-loss advice (OR 0.52, 95% CI 0.30, 0.88).
Conclusion. The quality of nonpharmacological care for people with knee/hip OA in the community is suboptimal. Advice on exercise and weight management may not be provided equally across sex, age, disability, and formal education levels.
Osteoarthritis (OA) is the most common form of joint disease, affecting about 10% of the population1. The prevalence of OA is increasing rapidly with the aging of the population. In Canada, the prevalence of arthritis is projected to be 20% by 2031, representing an increase of 38% from 2007, and with the majority of these people having OA2. The knee and hip are commonly affected, and ample evidence supports the use of pain medication, exercise, and weight management as the first-line treatment3,4,5. Previous studies suggested that 33% to 63% of patients with knee or hip OA had received a recommendation from a health professional to exercise6,7,8, 31% had been told to lose weight, and 27% had been recommended to use mobility aids7. Patients might also try these interventions on their own initiative, but some might require guidance from a health professional to achieve therapeutic benefits9. It was unclear whether those who did not receive a recommendation for these interventions really did not need it or if they simply did not receive a required intervention.
In 2004, the Arthritis Foundation supported the Rand Group to develop 14 quality indicators covering the assessment, treatment, and followup for patients with OA10. These indicators were developed using a comprehensive literature synthesis and were reviewed by a multidisciplinary expert panel. They represent the basic care that is appropriate for all clinical practice and can be used to evaluate quality of care. In our study, we applied these quality indicators to assess the nonpharmacological, nonsurgical care received by people with knee and/or hip OA in British Columbia, Canada. In addition, we explored the sociodemographic factors associated with nonpharmacological care for knee and/or hip OA.
MATERIALS AND METHODS
Study sample
Our study was 1 part of the 2007 British Columbia Osteoarthritis Survey, which aimed to evaluate the use of, and barriers to, health services and treatments by patients with OA11. A random sample of 6000 patients was identified using administrative databases collected by the Ministry of Health of British Columbia, for the purpose of reimbursements for outpatient physician visits (Medical Services Plan Fee-For-Service Database) and hospitalizations (Discharge Abstract Database). The sample was stratified by the 5 regional authorities that oversee healthcare delivery. Eligible participants were those who (1) met the case definitions for OA of any joints or hip/knee replacement surgeries based on the International Classification of Diseases 9/10 codes between April 1, 1992, and March 31, 200612; (2) had at least 2 medical visits for OA or 1 hospitalization within a 365-day period; (3) were aged 19 years or older on March 31, 2006; (4) were living in British Columbia; and (5) were alive (i.e., no date of death recorded on the administrative databases at the time of sampling).
The case date was defined as the first date by which the case definition was met. To ensure confidentiality, the Ministry of Health used a computer-generated algorithm to randomly select individuals for the survey. Those selected were then asked if they had ever been told by a health professional that they had arthritis. Only those who responded positively were invited to complete the full questionnaire.
Questionnaire development and administration
The British Columbia OA Survey questionnaire covered areas related to the use of health services and the health of people with OA, including (1) use of services from health professionals and alternative therapy practitioners, treatments, and community services for managing arthritis; (2) general health and comorbid conditions, including diabetes, kidney and/or bladder problems, fibromyalgia, high blood pressure, lung problems, osteoporosis, heart problems, intestinal or stomach ulcers, cancer, liver problems, bowel disorder, depression; and (3) physical function, assessed by standardized outcome measures such as the Western Ontario and McMaster Universities OA Index (WOMAC).
The questionnaire, consisting of 62 questions, was pilot-tested on 200 people with OA and 50 who had hip or knee replacement surgery because of OA. We used the same 5 eligibility criteria to identify participants for the pilot testing (response rate = 49.2%). Fifteen questions were found to have > 10% missing data. They were subsequently reviewed by the research team and the wording was modified unless the question was from a standardized questionnaire.
The survey comprised 3 mailings between June and July 2007. In the first mailing, all participants received a survey package from the Ministry of Health, including an information letter, a questionnaire booklet, and a stamped return envelope. Reminder cards were sent at 2 weeks and 4 weeks. To protect confidentiality, the ministry assigned an identification number to all eligible individuals and the researchers had no access to any personal contact information. This procedure complied with British Columbia’s Freedom of Information and Protection of Privacy Act. The study protocol was approved by the University of British Columbia Behavioural Research Ethics Board (application number H04-80289).
Quality of nonpharmacological care for OA
To assess the quality of nonpharmacological care, we applied 4 of the Arthritis Foundation Quality Indicators for OA10 to the British Columbia OA Survey data. The quality indicators included (1) advice to exercise, (2) advice to lose weight, (3) assessment for ambulatory function, and (4) assessment for nonambulatory function. The latter includes self-care activities such as dressing and grooming, as well as arising from sitting to standing. A typical quality indicator includes 2 components: The “IF” statement determines the eligibility for the care process in question; the “THEN” statement specifies what care process should be performed. The pass rate of each quality indicator was calculated by dividing the number of people who received the care (i.e., those who passed the “THEN” statement) by the number of those eligible for the care (i.e., those who passed the “IF” statement). Table 1 presents the criteria for passing the 4 quality indicators and the survey questions used.
Criteria for meeting a quality indicator for osteoarthritis and the corresponding items on the questionnaire.
Statistical analysis
Participant characteristics and health status were summarized in frequencies, means (SD) or medians [interquartile ranges (IQR)], depending on the measure. The pass rate of each of the 4 quality indicators was calculated. In addition, we calculated each participant’s summary score of nonpharmacological care by dividing the number of times the individual passed a quality indicator and the total number of quality indicators for which he/she was eligible (ranging from 0 to 4). We also calculated the mean and 95% CI of the overall pass rate for nonpharmacological care.
After controlling for the aggregate WOMAC score (0–96), logistic regression with backward elimination was used to assess the association between each of the 4 quality indicators and the following independent variables: (1) sex; (2) age group (24–54, 55–64, 65–74, 75+); (3) education (university degree, trade/vocational certificate, high school diploma, < high school); and (4) employment (employed, retired for medical reasons, unemployed/retired not because of medical reasons). We began with models that included number of comorbidities (maximum 12), plus all 2-way interactions between age group, sex, employment, and education. Backward elimination was used to drop nonsignificant interaction terms, and number of comorbidities if nonsignificant. In the Assessment of Ambulatory Activities model, education was omitted for model convergence. The level of statistical significance was set at p ≤ 0.05.
RESULTS
Of the 6000 records, 501 had invalid addresses, leaving 5499 eligible individuals. A total of 2259 individuals (41.1%) responded; 1713 of those confirmed that they had received a diagnosis of arthritis from a health professional, 442 had never been told that they had arthritis or OA, 96 returned an empty questionnaire, and 8 were deceased. Of the 1713 participants, 1349 reported knee and/or hip OA [knee only = 700 (51.9%), hip only = 261 (19.3%), knee and hip = 388 (28.8%); Table 2]. The participants’ mean age was 67.1 years (SD 11.1), 816 (60.5%) were women, and 921 (68.3%) were diagnosed with OA 6 years ago or longer. They had a median of 1 comorbid condition (IQR 1.0, 3.0). The mean WOMAC aggregate score was 32.9 (SD 19.4). Most of the participants (n = 783; 58.0%) rated their health excellent, very good, or good; and the vast majority met the eligibility (i.e., fulfilling the criteria for the “if” statement) of at least 1 quality indicator (n = 1304; 96.7%).
Sociodemographic characteristics and health status of survey participants (n = 1349).
The overall pass rate of the 4 quality indicators was 22.4% (95% CI 20.5, 24.3). The pass rates of individual quality indicators ranged from 29.2% for receiving assessment of ambulatory function to 6.9% for receiving assessment for nonambulatory function (Table 3). The pass rates were similar in both sexes, except for “advice to lose weight,” in which 32.3% of women vs 14.4% of men who were overweight or obese received this advice (Figure 1).
Quality of nonpharmacological care for osteoarthritis: women compared to men. *p < 0.05.
Quality of nonpharmacological care for osteoarthritis (n = 1349).
None of the interaction terms or the number of comorbidities was found to be statistically significant. They were subsequently dropped from all logistic regression modes. The OR (Table 4) revealed that receiving exercise advice was associated with having a university degree (vs a high school diploma; OR 3.10, 95% CI 2.00, 4.80). Receiving weight-loss advice was associated with being female (OR 2.64, 95% CI 1.71, 4.08), the age range of 55 to 64 years (compared to age 75 and over; OR 1.96, 95% CI 1.02, 3.76), and higher WOMAC scores (for every 10-point increment; OR 1.14, 95% CI 1.02, 1.26). However, less education (i.e., < high school) had a reverse effect on receiving weight-loss advice (OR 0.52; 95% CI 0.30, 0.88). For those who met the quality indicator for assessment of nonambulatory activities, retirement because of medical reasons (compared to retirement/unemployment because of other reasons) emerged as a significant factor for receiving care (OR 3.66, 95% CI 1.22, 10.97). No statistically significant association was found between the demographic/health status variables and the quality indicator on assessment of ambulatory activities.
Logistic regression models for the quality indicators.
DISCUSSION
The majority of those who completed the survey perceived themselves as having good health, but this study also identifies a substantial gap in nonpharmacological care. Of the participants, 22.4% received the recommended nonpharmacological, nonsurgical care for knee/hip OA, based on 4 quality indicators. This overall pass rate echoes the findings in previous studies, including the 57% reported by Ganz, et al13 in the medical treatment for community-dwelling older adults with knee/hip OA, and by McGlynn, et al14 in the general management of OA among adults of all ages. In addition, only 1 in 4 of those who needed advice to exercise or lose weight received the necessary care and < 7% of those with severe difficulties in self-care were assessed for nonambulatory activities. These findings add to the observation of the suboptimal care received by people with OA elsewhere6,13,15.
Previous studies on quality of care in OA focused on the use of exercise, but not other nonpharmacological interventions. McGlynn, et al14 used an earlier version of the quality indicator for exercise, which states, “Providers caring for patients with symptoms of hip or knee OA should recommend exercise programs at least once in 2 years.” The pass rate for this quality indicator was 29.7%. In contrast, the current version has narrowed the timeframe to once per year, which may be a reason for the difference in pass rate (25.2%). The exercise quality indicator in our study is similar to that used by Ganz, et al except that they limited the age to 75 years and over. The pass rate of this quality indicator was 44.0%. It should be noted that their sample was drawn from the ACOVE-2 study, in which some of the participants had received an intervention for cognitive impairment, falls/mobility problems, and/or incontinence16. This intervention also included information on specific exercises and community-based exercise groups, although it was not specifically designed for OA management16.
The observed gaps in nonpharmacological care are alarming, especially because physical activity and exercise have been shown to reduce pain and improve quality of life4,17,18, and have the potential to reduce the progression of joint damage19. In addition, a high body mass index (BMI) is associated with an increased likelihood of deterioration in the functional status in the first 3 years following an OA diagnosis20. In a case-control study, Wendelboe, et al21 found that severely obese women (BMI > 40) were 4 times more likely to have hip replacement surgery and 19 times more likely to have knee replacement surgery when compared to normal-weight people. Among those with OA who are overweight, even a moderate weight loss of 5% can significantly reduce physical disability22. Further, a decrease in the BMI is associated with people’s adherence to exercise and the improvement in function23, suggesting that it may be beneficial for people who are obese to lose weight even before starting an exercise program. Our findings suggest that these effective interventions have not reached the community level, as evidenced by the low pass rates for weight management and exercise.
Our study also revealed that the majority of patients with severe functional limitation had not been assessed by a physical therapist or an occupational therapist in the past year. Little is known about the use of physical therapy and occupational therapy by patients with OA in the community, but a recent study in Ontario found that about 43% of these patients had been referred for physical therapy within a 4-year period8. Given that only 29% and 7% of our participants with severe functional disability had seen a physical therapist or an occupational therapist in the past year, respectively, future efforts should be directed toward supporting the appropriate use of rehabilitation services.
Our study also identified a few sociodemographic factors associated with OA care. For example, participants with a university education were more likely to have passed the quality indicator on exercise advice compared to those with lower education levels. The findings echo an earlier study by Dexter, et al24 that reported that patients with lower education levels received less comprehensive instruction and monitoring of exercise. Also, among participants who were overweight or obese, those with less than a high school education were less likely to meet the weight-management quality indicator. Interestingly, women were more likely than men to have joined a weight-loss program or have seen a dietician, as were the middle-aged group compared to those over the age of 75 years, and those with more severe symptoms. We could not determine whether this was associated with patients’ help-seeking decisions, health professionals’ prescribing practices, or both. Nonetheless, with the prevalence and severity of OA being higher among those with lower education and socioeconomic status25,26,27, our results add to the growing literature that urges immediate action to address the inequity in OA care28. It should be noted that in a 2011 systemic review, Borkhoff, et al29 identified 10 studies evaluating interventions that aimed to reduce the disparity in care for disadvantaged populations with OA, but none of these interventions targeted health professional practices. Our findings lend further support for research in this area.
Our findings, however, should be interpreted with caution because of limitations of the eligibility definitions for the quality indicators (Table 1), which might have resulted in underreporting or overreporting of the pass rates. Also, the return rate of this survey was low (41.1%); hence, the findings are subject to response bias where those who responded may be systematically different from those who did not, thus affecting the generalizability. A recent study has shown that more than 80% of people aged 50 years or older and with knee pain had undiagnosed knee OA, and many of them had only mild radiographic joint damage30. Thus many with undiagnosed early OA are not included in the study. These individuals could benefit from early intervention, but without a diagnosis they would be unlikely to receive the appropriate advice and followup. Our findings, therefore, may underestimate the care gaps in OA. Also, this study relied on patient self-reported data, which is an inherent limitation in most surveys examining chronic illnesses and treatment use. To address this limitation, a few recent studies have advocated for the use of electronic medical records to assess quality of care31,32. Finally, the percentage of English-speaking participants was higher than that of the British Columbia population33 (71%, vs 96% in the survey) because the questionnaire was in English only. Hence, our data are insufficient to estimate the health services use and care gaps in the non-English-speaking population, particularly new immigrants.
Proponents of quality indicators suggest that judicious application of these tools can optimize healthcare because it provides the means to measure the process of care and feedback to health professionals34,35,36,37. The results from our province-wide survey indicate that nonpharmacological care for people with knee/hip OA is not consistent with the current treatment guidelines. Further, there may be inequity in the provision of advice on exercise and weight management. Our findings indicate the need to identify barriers to providing advice on exercise and healthy weight and referring patients for functional assessment when needed. In addition, we have argued for a need to invest in interventions, specifically those that target health professionals, to address the disparity in care for patients with OA due to their sociodemographic status.
- Accepted for publication May 19, 2011.
REFERENCES
In this issue
