The severity of tennis elbow is not related to health-seeking behaviour and treatment outcomes are not influenced by patient characteristics

Abstract

Purpose: The aims of these post hoc analyses were 1) to determine whether or not baseline patient characteristics in a tennis elbow cohort were different between patients who sought primary care intervention and those who had not, and 2) to determine if subgroups of patients with tennis elbow responded differently to treatment. Relevence: It is unknown if patients who have a more severe condition exhibit health-seeking behaviour such as primary care consultation. In addition, whilst clinical trials provide insight into the efficacy of interventions for the average population, it is unclear if patient characteristics affect these treatment outcomes. Participants: 198 participants with a mean age (SD) of 47.6 (7.8) years were sub-grouped according to whether or not they had consulted a primary care practitioner for their current episode of tennis elbow. In addition, we pooled individual patient data (N=383) from two randomised controlled trials (RCTs) that investigated a wait-and-see policy, corticosteroid injections and physiotherapy in patients with tennis elbow. Methods: Post-hoc sub-group analyses of a tennis elbow cohort (Australia) were performed according to health-seeking behaviour. Outcome measures were baseline pain, sex, age, duration of condition, previous history, dominant side affected and employment, assessed using analysis of covariance for continuous measures and logistic regression for dichotomous measures (p<0.05). Secondly, post-hoc sub-group analyses of individual patient data pooled from two RCTs (Australia and Dutch) were performed, with baseline characteristics of previous history, baseline pain, duration of condition and employment identified for inclusion a priori. Outcome measures common and homogeneous to both trials were pain severity and global improvement at six and 52 weeks (p<0.05). Results: Within the Australian trial, significant differences were found between the sub-groups of primary care and community participants for the number of women, previous history of elbow complaints and age (p<0.05). There were no differences between the sub-groups for pain severity, duration of condition, pain-free grip strength or functional status. Based on pooled patient data from both trials, subgroup effects were scarce and small. Patients with higher baseline pain did not respond as well to physiotherapy at 6 weeks follow up as patients with lower baseline pain. It also appeared that non-manual workers did not respond as well to corticosteroid injections at 52 weeks compared to manual and no workers. The treatment outcomes were largely similar between trials and not different between most subgroups studied. Conclusion: This study showed that based on the characteristics of pain and function, participants who exhibited health-seeking behaviour prior to enrolment into the clinical trial were no different to those who did not seek primary care consultation. In tennis elbow, it would appear that treatment outcomes are only minimally dependent on baseline patient characteristics, which supports the generalisability of individual trial results. Clinicians can be confident of the applicability of the findings of these clinical trials to their clinical practice.