In this study, count-based maternal mortality data obtained in the field of health were modeled using time series methods, and the performance of Classical Bootstrap and Moving Block Bootstrap methods were compared based on this model. Since the study is based on a methodological approach aimed at comparing the performance of Bootstrap methods in health data exhibiting time dependence, the analyses were conducted using a single health indicator. According to the findings, the Moving Block Bootstrap method was found to better preserve the dependence structure in health time series data, thereby producing more realistic standard errors and confidence intervals, and consequently, providing more realistic uncertainty estimates compared to the Poisson Autoregressive model and the Classical Bootstrap method. The results of the study demonstrate that the choice of statistical method used in the analysis of health data with time dependence is of critical importance for the accuracy and reliability of the resulting inferences.