Polypharmacy and the impact of antidepressants, antipsychotics and benzodiazepines on falls and fractures risks in older adults: a triangulation approach towards causal inference for the study of drug combinations

I aim to use evidence from different causal inference study designs that rely on different assumptions and have differing strengths and limitations to evaluate combination use of antidepressants, antipsychotics and benzodiazepines on risks of falls and fractures in older adults. This will include use of the target trial framework, an appropriate self-controlled method (e.g., case cross-over or self-controlled case series) and mendelian randomization. The target trial approach can clarify the causal question of interest and account for biases introduced by analysis decisions such as immortal time bias, whereas the self-controlled design can better control for time-invariant confounding because it is a within-individual design. Mendelian randomization can allow assessment of the on-target and off-target effects of drugs and their corresponding indications on falls and fractures. The population of interest is older adults aged 40 years and above who have prescription records of an antidepressant, antipsychotic and/or benzodiazepine. Confounding and selection bias will be accounted for using appropriate analysis methods. I anticipate that results from this study will inform clinical guidelines, improve geriatric and psychiatric care, and help provide patients with informed choice about medications.