Browsing by Subject "Proportional Hazards Models"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Open Access
    A comparison of the conditional inference survival forest model to random survival forests based on a simulation study as well as on two applications with time-to-event data
    (2017) Nasejje, Justine B; Mwambi, Henry; Sabur, Natasha F; Lesosky, Maia
    Abstract Background Random survival forest (RSF) models have been identified as alternative methods to the Cox proportional hazards model in analysing time-to-event data. These methods, however, have been criticised for the bias that results from favouring covariates with many split-points and hence conditional inference forests for time-to-event data have been suggested. Conditional inference forests (CIF) are known to correct the bias in RSF models by separating the procedure for the best covariate to split on from that of the best split point search for the selected covariate. Methods In this study, we compare the random survival forest model to the conditional inference model (CIF) using twenty-two simulated time-to-event datasets. We also analysed two real time-to-event datasets. The first dataset is based on the survival of children under-five years of age in Uganda and it consists of categorical covariates with most of them having more than two levels (many split-points). The second dataset is based on the survival of patients with extremely drug resistant tuberculosis (XDR TB) which consists of mainly categorical covariates with two levels (few split-points). Results The study findings indicate that the conditional inference forest model is superior to random survival forest models in analysing time-to-event data that consists of covariates with many split-points based on the values of the bootstrap cross-validated estimates for integrated Brier scores. However, conditional inference forests perform comparably similar to random survival forests models in analysing time-to-event data consisting of covariates with fewer split-points. Conclusion Although survival forests are promising methods in analysing time-to-event data, it is important to identify the best forest model for analysis based on the nature of covariates of the dataset in question.
  • Thumbnail Image
    Item
    Open Access
    Risk factors for unstructured treatment interruptions and association with survival in low to middle income countries
    (2016) McMahon, James H; Spelman, Tim; Ford, Nathan; Greig, Jane; Mesic, Anita; Ssonko, Charles; Casas, Esther C; O’Brien, Daniel P
    Abstract Background Antiretroviral therapy (ART) treatment interruptions lead to poor clinical outcomes with unplanned or unstructured TIs (uTIs) likely to be underreported. This study describes; uTIs, their risk factors and association with survival. Methods Analysis of ART programmatic data from 11 countries across Asia and Africa between 2003 and 2013 where an uTI was defined as a ≥90-day patient initiated break from ART calculated from the last day the previous ART prescription would have run out until the date of the next ART prescription. Factors predicting uTI were assessed with a conditional risk-set multiple failure time-to-event model to account for repeated events per subject. Association between uTI and mortality was assessed using Cox proportional hazards, with a competing risks extension to test for the influence of lost to follow-up (LTFU). Results 40,632 patients were included from 11 countries across 33 sites (17 Africa, 16 Asia). Median duration of follow-up was 1.61 years (IQR 0.54–3.31 years), 3386 (8.3 %) patients died, and 3453 (8.5 %) were LTFU. There were 14,817 uTIs, with 10,162 (25 %) patients having more than one uTI. In the adjusted model males were at lower risk of uTI (aHR 0.94, p < 0.01, and age 20–59 was protective compared to <20 years (20–39 years aHR 0.87, p < 0.01; 40–59 years aHR 0.86, p < 0.01). Preserved immune function, as measured by higher CD4 cell count, was associated with a reduced rate of uTI compared to CD4 <200 cells/μL (CD4 200–350 cells/μL aHR 0.89, p < 0.01; CD4 >350 cells/μL aHR 0.87, p < 0.01), whereas advanced clinical disease was associated with increased uTI rate (WHO stage 3 aHR 1.10, p < 0.01; WHO stage 4 aHR 1.21, p < 0.01). There was no relationship between uTI and mortality after adjusting for disease status and considering LTFU as a competing risk. Conclusions uTIs were frequent in people in ART programs in low-middle income countries and associated with younger age, female gender and advanced HIV. uTI did not predict survival when loss to follow-up was considered a competing risk. Further evaluation of uTI predictors and interventions to reduce their occurrence is warranted.