Discussing depression in patients with visual impairment differs across countries: Validation of a prediction model in healthcare providers

Purpose: Healthcare providers often experience difficulties in discussing depression with adults with visual impairment (VI), obstructing timely referral. The purpose of this study was to examine predictors of routine discussions of depression with adults with VI from the perspective of different healthcare providers from different countries. Methods: Cross-sectional survey data from Welsh (N = 122), Australian (N = 94) and Dutch (N = 100) healthcare providers, that is eye care practitioners (ECPs) and low-vision care providers (LVCPs), were analysed. Multivariable logistic regression analysis was performed in the Welsh sample to determine predictors for discussing depression. Internal validation was conducted by using a bootstrap method, and the recalibrated model was externally validated in the Australian and Dutch sample. Results: Work experience in eye care services (OR 0.95; 95% confidence interval (CI) 0.92 to 0.99) and perceived barriers (OR 0.95; 95% CI 0.92 to 0.98) was found to predict discussing depression with patients. The area under the curve (AUC) of 0.73 reflected good discrimination of the model. The model showed a slightly better fit in the Australian sample (AUC = 0.77), but a poor fit in the Dutch sample. Conclusion: The final prediction model was not generalizable to Dutch healthcare providers. They perceived less barriers in depression management than Welsh and Australian healthcare providers. This could be explained by differences in ECPs and LVCPs roles and responsibilities, increased attention on mental health and differences in organizing health care. Differences between healthcare providers' responsibilities and support needs should be taken into account while creating a facilitating environment to discuss depression.