A carregar...
Publicação
Competing risks modeling in survival analysis: valuation in peritoneal dialysis
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 6.33 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
A análise de sobrevivência é uma pedra angular na pesquisa médica. A
estimativa de Kaplan-Meier é o método estatístico mais amplamente utilizado, mas a
presença de riscos competitivos viola o seu pressuposto fundamental que afirma que o
mecanismo de censura é independente do tempo de sobrevida. Isto leva à sobrestimação
da probabilidade cumulativa de falha de risco específico. Assim, nessas circunstâncias a
estimativa da função de incidência cumulativa e análise de riscos competitivos são
métodos preferíveis. O objetivo deste estudo foi comparar os métodos de análise de
sobrevivência diferentes: estimativas de Kaplan-Meier e função de incidência
cumulativa numa coorte de doentes em Diálise Peritoneal (DP).
Avaliou-se a sobrevivência de 123 pacientes incidentes em DP num hospital
universitário. Foram estimadas as curvas de Kaplan-Meier, função de incidência
cumulativa, risco específico e subdistribuição dos riscos.
O grupo PD first apresentou sobrevivência superior, quando comparado com os
doentes transferidos para DP, a probabilidade de morte foi sobrestimada pelo método de
Kaplan-Meier. O risco específico revelou como factores associados a maior
sobrevivência a opção PD first, níveis de albumina e função renal residual superiores e
menor índice de Charlson. Considerando a análise de riscos competitivos, na
subdistribuição dos riscos a opção PD first, a função renal residual e índice de Charlson
permaneceram significativos, mas não os níveis séricos de albumina.
É primordial reconhecer a presença de riscos competitivos e adoptar um método
de risco competitivos em estudos com vários outcomes, como em estudos de Diálise
Peritoneal, para estimar a incidência cumulativa e produzir resultados mais precisos.
Survival analysis is a cornerstone in medical research. For that purpose Kaplan- Meier estimate is the most widely used statistical test, but the presence of competing risks violates the fundamental assumption that states the censoring mechanism is independent of survival time. This leads to overestimation of the cumulative probability of cause-specific failure. Cumulative incidence estimate and competing risks analysis are preferred. The purpose of this study was to compare different survival analysis methods: Kaplan-Meier and cumulative incidence function estimates in a cohort of Peritoneal Dialysis (PD) patients. The survival of 123 incident patients on PD in a university hospital was evaluated. Kaplan-Meier, cumulative incidence function, cause-specific and subdistribution hazards were performed. PD first group had a better survival when compared with PD transfer patients, the probability of death was overestimated by the Kaplan-Meier method. The causespecific hazard revealed factors associated with survival were PD first option, higher albumin levels, higher renal residual function and lower Charlson Index. Taking competing risks into account in the subdistribution hazard model PD first option, renal residual function and Charlson Index remained significant, but not albumin levels. It is primordial to recognize the presence of competing risks in order to adopt a competing risk method in studies with multiple outcomes, as in Peritoneal Dialysis studies, to estimate cumulative incidence and yield more accurate results.
Survival analysis is a cornerstone in medical research. For that purpose Kaplan- Meier estimate is the most widely used statistical test, but the presence of competing risks violates the fundamental assumption that states the censoring mechanism is independent of survival time. This leads to overestimation of the cumulative probability of cause-specific failure. Cumulative incidence estimate and competing risks analysis are preferred. The purpose of this study was to compare different survival analysis methods: Kaplan-Meier and cumulative incidence function estimates in a cohort of Peritoneal Dialysis (PD) patients. The survival of 123 incident patients on PD in a university hospital was evaluated. Kaplan-Meier, cumulative incidence function, cause-specific and subdistribution hazards were performed. PD first group had a better survival when compared with PD transfer patients, the probability of death was overestimated by the Kaplan-Meier method. The causespecific hazard revealed factors associated with survival were PD first option, higher albumin levels, higher renal residual function and lower Charlson Index. Taking competing risks into account in the subdistribution hazard model PD first option, renal residual function and Charlson Index remained significant, but not albumin levels. It is primordial to recognize the presence of competing risks in order to adopt a competing risk method in studies with multiple outcomes, as in Peritoneal Dialysis studies, to estimate cumulative incidence and yield more accurate results.
Descrição
Palavras-chave
Saúde pública Estatística para a saúde