Summary of the Article
The article by Ng et al. (2012) is aimed at identifying the health outcomes of children who had pediatric liver transplantation and survived for 10 years after that operation. Thus, the article investigates the presence or absence of an event, which is common for public health research (Telke & Eberly, 2011). The authors conducted a cross-sectional study and utilized a variety of statistical methods, such as chi-square, Wilcoxon log-rank test, Student t-test, and independent samples t-tests for estimating differences between groups, Kaplan-Meier curves to show “patient survival, graft survival, and acute cellular rejection,” etc. (Ng et al., 2012, pp. 2-3). The researchers also described an “ideal” survivor, a person whose health outcome was optimal after the transplantation. As for Kaplan-Meier method, two graphs (related to a) probability of patient survival over 10 years and b) probability of ACR over 10 years) are provided.
Advantages of Using Kaplan-Meier Method
There are a number of advantages to using the Kaplan-Meier method in order to estimate the probability of survival and of acute cellular rejection in the study by Ng et al. (2012). For instance, such an estimation is based on actual times of survival (instead of, for instance, using time intervals), which allows for a rather precise estimation (Forthofer, Lee, & Hernandez, 2007, p. 306). Second, the method provides a visual estimation of the survival function, which permits for its easier interpretation.
Disadvantages of Using Kaplan-Meier Method
There are also some disadvantages to utilizing the Kaplan-Meier method in the study. For instance, Kaplan-Meier does not handle the censored cases in the optimal way (Costella, 2010). That is, for censored cases, the even in question practically does not happen during the time of observation. Another problem is related to the confusing effect of the visual estimation of the survival function. This estimation has certain points at which the graph “drops” or “rises,” which denotes the occurrence of the event (in this case, a death of a patient or a case of acute cellular rejection). This makes it look as if there are certain points of time during which there are greater risks of the event occurring, whereas, it is possible that the probability of the event remains the same over the interval.
Are the Authors’ Conclusions Supported by Their Analysis?
In fact, the authors draw no specific conclusions from the results of the Kaplan-Meier analysis (Ng et al., 2012); instead, the results of the approximation of the function of survival are provided, and the authors also supply additional details related to the incidence of events; for instance, the approximations depict the total levels of acute rejection, whereas in the text, the authors provide more information (first or second case of rejection, etc.). On the whole, it should be noted that the authors tend to report the results of the study rather than boldly draw conclusions from them.
An Alternative Method
As an alternative to the Kaplan-Meier method, it also would be possible to use the life tables in order to estimate the chances of an event (death or acute cellular rejection) among the patients (Forthofer et al., 2007). Since the results of the Kaplan-Meier analysis are summarized only in the graphs approximating the survival functions, it would also be possible to use the life tables for examining the survival rates. In addition, it is apparent that Ng et al. (2007) do not use any factors for the Kaplan-Meier analysis, for all the children had the same intervention – liver transplantation.
References
Costella, J. P. (2010). A simple alternative to Kaplan–Meier for survival curves. Web.
Forthofer, R. N., Lee, E. S., & Hernandez, M. (2007). Biostatistics: A guide to design, analysis, and discovery (2nd ed.). Burlington, MA: Elsevier Academic Press.
Ng, V. L., Alonso, E. M., Bucuvalas, J. C., Cohen, G., Limbers, C. A. Varni, J. W.,…Anand, R. (2012). Health status of children alive 10 years after pediatric liver transplantation performed in the US and Canada: Report of the studies of pediatric liver transplantation experience. Journal of Pediatrics, 160(5), 820-826, e3. Web.
Telke, S. E., & Eberly, L. E. (2011). Statistical hypothesis testing: Associating patient characteristics with an incident condition: Kaplan-Meier curves, hazard ratios, and Cox proportional hazards regression. Journal of Wound, Ostomy, and Continence Nursing, 38(6), 621-626.