Representation and outcomes of individuals with major depression in routine care who are ineligible for randomized controlled trials: a nationwide register-based study

Randomized controlled trials (RCTs) are the foundation of current clinical treatment guidelines. However, they may not reflect real-world populations, due to strict eligibility criteria. We determined the proportion of individuals with major depressive disorder (MDD) receiving maintenance antidepressant treatment in routine care who would be ineligible for RCTs, and compared their outcomes with those who were eligible. Utilizing specialized health care registers in Finland (2004-2018) and Sweden (2006-2021), we identified adults diagnosed with non-psychotic MDD (ICD-10: F32, F33) who were stabilized on maintenance antidepressant monotherapy. Through multidisciplinary expert consensus on latest meta-analytic evidence, we derived a standardized list of RCT inclusion and exclusion criteria. These criteria were systematically applied to classify individuals as RCT-eligible or RCT-ineligible. We then used Cox proportional models to derive hazard ratios (HRs) of a composite primary outcome of hospitalization due to any psychiatric reason or suicide attempt, and all-cause mortality, during a 6-month follow-up. Secondary outcomes were treatment changes (i.e., discontinuation, switch or augmentation) and psychiatric sick leave ≥2 weeks. A total of 73,720 individuals in Finland and 135,092 in Sweden were included. More than one third of patients with MDD (33.5% in Finland and 35.3% in Sweden) were found to be ineligible for RCTs. The most common reasons for ineligibility were comorbidities (serious somatic disease, other psychiatric disorders, or substance use disorder). RCT-ineligible individuals had more than twice the risk of the composite primary outcome compared to eligible individuals (HR=2.44, 95% CI: 2.15-2.76 in Finland; HR=2.61, 95% CI: 2.37-2.87 in Sweden). Approximately one third of the composite primary outcome was attributable to RCT-ineligibility factors (32.5%, 95% CI: 27.9-37.1 for Finland; 36.2%, 95% CI: 32.6-39.8 for Sweden). Risk of treatment change was slightly but significantly higher in ineligible individuals. Findings were consistent in a wide range of sensitivity analyses. We conclude that more inclusive eligibility criteria for RCTs, and their integration with real-world data, are needed to improve the generalizability of antidepressant trial evidence and MDD clinical treatment guidelines.