Statistical Groupings of Mental Health and Osteoarthritis Severity Correlate With 10-year Trajectories of Levels of Capability and Comfort Among People With Hip Pain: A Nationwide Prospective Cohort Study (CHECK).

BACKGROUND

Among patients seeking care for osteoarthritis (OA), there is evidence that mindsets contribute more to variation in levels of musculoskeletal discomfort and incapability than does radiographic severity. While the importance of mindsets among those seeking specialty care is well established, less is known about the relationship of thoughts and feelings with levels of joint-related discomfort and incapability in the general population, many of whom experience symptoms of aging joints and are accommodating these symptoms. By studying individuals with hip pain who have not sought specialty care, using a statistical technique that can account for the interrelationship of psychological and social variables (cluster analysis) and accounting for levels of OA, we can better study the association of mindset with levels of discomfort and incapability relative to the grade of OA while limiting or avoiding the potential for distortion of linear and logistic regression by even relatively lower levels of collinearity among mental health variables.

QUESTIONS/PURPOSES: Using a Dutch population-based cohort we asked: (1) Are there distinct 10-year trajectories of comfort and capability by statistical grouping based on mental health measures and grade of OA in a population-based cohort of individuals with hip pain? (2) Are groupings with less healthy mindsets associated with worse 10-year trajectories for both comfort and capability compared with groups defined by healthier mindsets?

METHODS

We analyzed data from a prospective longitudinal cohort from the Dutch general population from between October 2002 and September 2005. Individuals age 45 to 65 years were included if they experienced new onset of hip or knee pain or stiffness and either had not sought care for their symptoms or their first general practitioner consultation for symptoms was within 6 months of enrollment. Originally, a total of 1002 participants were included, of which 74% (n = 740, 79% [584 of 740] women, mean ± SD age 56 ± 5 years) met our inclusion criteria and were included in this study. Each year for 10 years, participants completed the SF-36 (measuring general capability and mental health, with higher scores indicating better overall health); the WOMAC questionnaire (measuring capability specifically for hip and knee OA, with lower scores indicating greater capability); the EuroQol-5D-3L (EQ-5D-3L) questionnaire, which measures five domains of health (mobility, self-care, usual activities, pain, and anxiety/depression); and the VAS for pain intensity. All 740 included participants completed the WOMAC, the SF-36 physical component summary, and the Numeric Rating Scale for pain at 10 years. Cluster analysis identified statistical groupings of participants with similar scores on the SF-36 mental component summary (MCS), EQ-5D-3L anxiety/depression item, and Kellgren-Lawrence grade of radiographic hip OA. We then constructed a conditional growth model, which is a statistical technique that analyzes the average rate of decline in capability and increase in pain intensity over 10 years between groups of people with similar baseline mental health and radiographic arthritis severity. The conditional growth model quantifies the relationship between mindset and grade of OA at baseline and 10-year trajectories of levels of comfort and capability.

RESULTS

The cluster analysis identified four statistical groups of participants with similar mean grades of radiographic OA at the 10-year evaluation and variation in mental health scores. All groups had a mean Kellgren-Lawrence Grade of 2. Group 1, summarized as "accommodative mindset" (44% [326 of 740] of patients), had high mean scores on the SF-36 MCS (a mean of 60, which represents an SD better than the population mean of 50) and low mean scores on the EQ-5D-3L anxiety/depression questions, representing minimal symptoms of anxiety and depression. Group 2, summarized as "neutral mindset" (37% [276 of 740] of patients), had scores near the population mean of 50 on the SF-36 MCS and slightly higher mean scores on the EQ-5D-3L anxiety/depression questions, representing relatively neutral mental health. Group 3, summarized as "less healthy mindsets" (12% [89 of 740] of patients), had mean scores of 1 SD below the population mean on the SF-36 MCS and higher mean scores on the EQ-5D-3L anxiety/depression questions, representing relatively greater symptoms of anxiety and depression. Group 4, summarized as "least healthy mindset" (6% [45 of 740] of patients), had mean scores of 2 SD below the population mean on the SF-36 MCS and high mean scores on the EQ-5D-3L anxiety/depression questions, representing notable symptoms of anxiety and depression. Patients with less healthy mindsets experienced more rapid decline in comfort and capability over a decade.

CONCLUSION

The finding among the general population of people with hip pain-many of whom have sensations from hip arthritis but are not seeking care-that worse mental health accounted for worse 10-year trajectories of hip symptoms independent of the radiographic severity of hip arthritis is further evidence that strategies for enhancing musculoskeletal health must account for symptoms of anxiety and depression. Seeking care for hip pain from relatively mild arthritis might, in part, signal higher levels of emotional distress. Timely diagnosis and effective treatment of unhealthy levels of anxiety and depression have the potential to improve a patient's hip health and their health in general.

LEVEL OF EVIDENCE

Level II, prognostic study.