Conflicts of interest among patients: do they exist and can they bias patient-reported outcome measures? A survey of shoulder and elbow patients.

Background: Potential conflicts of interest (COIs) are common among physicians and may bias physician-reported outcome assessment in orthopedic research. It is unclear whether patients have COIs and whether these COIs could affect patient-reported outcome measures (PROMs). The purposes of this study were to investigate (1) the existence of COIs among patients and (2) the potential of these COIs to bias PROMs with a pseudonymized survey among consecutive shoulder and elbow patients.

Materials And Methods: Between February and May 2021, 144 shoulder and elbow patients who underwent consultation at our outpatient clinic were included in the survey study. Of these patients, 79.2% (n = 114) completed the 11-item questionnaire. Variables including sex, existence of chronic diseases, level of education, and economic status were also assessed to investigate any association between these variables and patients' perceptions of COIs.

Results: Whereas 33.3% of respondents (n = 38) believe that COIs exist among patients and 28.1% (n = 32) believe that COIs could bias PROMs in general, fewer patients admit to personal COIs (24.6%, n = 28) and COIs biasing their self-assessment (23.7%, n = 27). Patients more frequently suggest COIs in their personal environment, such as among family members or friends (27.2%, n = 31), than in their own medical treatment. Financial factors such as sick pay (34.2%, n = 39) are thought to be the most likely reason for having COIs, followed by interpersonal reasons such as maintaining medical affection or bonds (29.8%, n = 34). Of respondents, 42.1% (n = 48) believe that sole usage of PROMs poses risks in treatment evaluation and only 21% (n = 24) consider PROMs reliable. In addition, 43% of respondents (n = 49) believe that patients should disclose COIs routinely in medical treatment. Multinomial logistic regression analysis revealed that wealthier patients are more likely to believe COIs could bias PROMs (odds ratio for poor vs. middle class, 0.23 [95% confidence interval, 0.053-0.963]; odds ratio for lower-middle class vs. middle class, 0.19 [95% confidence interval, 0.052-0.677]).

Conclusions: Although the majority of shoulder and elbow patients deny having self-experienced COIs and deny biased PROMs due to COIs in their own medical treatment, a considerable number of patients admit to having experienced both. Further studies might be justified to investigate the actual clinical relevance of patients' COIs and their impact on value-based health care.