Joint frailty analysis of recurrent hospital-acquired infections highlights the importance of accounting for competing terminal outcomes

Patients in intensive care units (ICUs) patients frequently experience recurrent infections associated with prolonged hospitalization and increased mortality. Conventional analyses treating death or discharge as simple censoring may bias recurrence estimates, whereas joint modeling appropriately accounts for recurrent and terminal events. This study discusses models to investigate recurrent ICU-acquired infections using a joint frailty model accounting for terminal events and unobserved heterogeneity, quantifying infection burden and assessing risk factors while adjusting for competing risks. We analyzed a publicly available prospective single-center neuro-ICU cohort using a joint frailty model to assess recurrent hospital-acquired infections and competing terminal events (death and discharge), while evaluating their interdependence. We employed recurrent event analysis approaches, notably total time scales and gap time scales for sensitivity analysis. The model was first implemented with a single composite terminal event and later extended to treat death and discharge as discrete, unique outcomes. All analyses were performed in R version 4.5.1 using the frailtypack package (version 3.7.0). We found individual heterogeneity, frailty variance θ = 0.070, and a high positive dependency on the terminal outcome and recurrent infections α = 4.40 using a joint frailty model with combined terminal events. We examined a significant individual heterogeneity θ = 0.286 in relation to joint modeling for separate terminal outcomes. Similarly, we discovered a high positive association between recurring infections and death (α₁ = 6.86) and recurrent infections and discharge (α₂ = 5.81). There were no statistically significant effects of sex or Charlson Comorbidity Index on recurrent infections in the model with separate terminal events. However, under the composite terminal event model, female patients had a significantly lower hazard of recurrent infections compared with males (HR = 0.90, 95% CI: 0.82–0.98; p = 0.028). However, the covariate impacts varied according to the terminal outcome. For a separate terminal events, a higher Charlson Comorbidity Index was associated with an increased hazard of death (HR = 1.11 [95% CI: 1.05–1.18, p < 0.001]) but a decreased hazard of discharge (HR = 0.93 [95% CI: 0.90–0.97, p < 0.0001]). The present study reveals a strong link between recurrent infections and terminal outcomes, with moderate heterogeneity among patients. It suggests that neglecting terminal events leads to an overestimation of recurrence burden and reduces the observed effects.