Career decision regret and its associations with depressive symptoms and Suicide Probability Scale scores among medical students at a Turkish university: a cross-sectional study

Abstract Background Career decision regret may reflect a mismatch between expectations of medicine and experiences during training. We examined whether regret was associated with depressive symptoms and Suicide Probability Scale (SPS) scores independently of key demographic and clinical covariates. Methods This single-centre cross-sectional survey used a voluntary convenience sample of 208 first-, fourth-, and sixth-year medical students. Participants completed a sociodemographic questionnaire, Beck Depression Inventory (BDI), SPS, and Career Decision Regret Scale (CDRS). Two multivariable linear regression models with HC3 robust standard errors were fitted with BDI and SPS total scores as the respective outcomes. The SPS model included BDI; both models adjusted for previous suicide attempt, psychiatric diagnosis, sex, year, and socioeconomic status. Results Of 208 participants, 149 reported career regret (sample proportion 71.6%, Wilson 95% confidence interval [CI] 65.2%–77.3%). CDRS correlated with BDI (Spearman’s ρ = 0.421, bootstrap 95% CI 0.292–0.536) and SPS total scores (ρ = 0.415, 95% CI 0.285–0.527; Holm-adjusted p < 0.001 for both). In adjusted analyses, each 10-point higher CDRS score was associated with a 1.98-point higher BDI score (95% CI 1.40–2.56; standardized β = 0.430; p < 0.001; ΔR² = 0.164) and, after adjustment for BDI and other covariates, a 1.43-point higher SPS total score (95% CI 0.56–2.30; standardized β = 0.173; Holm-adjusted p = 0.001; ΔR² = 0.021). Students directly reporting regret had a 10.94-point higher mean SPS score (95% CI 6.08–15.60; Hedges’ g = 0.67, 95% CI 0.36–0.96). Conclusions Higher career regret remained associated with greater depressive and suicide-related symptom scores after adjustment for measured covariates, although its incremental association with SPS scores beyond depressive symptoms was modest. The cross-sectional design, unknown response rate, and non-representative sampling preclude causal, prevalence, or individual suicide-risk interpretations.