Missing Data Analysis Paper

1140 Words3 Pages

The study design was appropriate to solve the research questions. The sample size was sufficient. However, eleven percent of cases (36,839) were removed due to missing data on one or multiple factors, which was a threat to the generalizability of the findings. I think the method used to deal with the missing data was not appropriate. An analysis of the pattern of the missing data was necessary and methods of managing the missing data should be based on an overall consideration of the pattern of missing data, how much was missing, and why it was missing. I am interested in the conceptual framework developed in this study. This conceptual framework can be used to guide my research after modifications. Overall it is a good, well-written article …show more content…

Survey and administrative data were used to include 4,540 NH residents in 290 NYS NHs who died in 2007. The hospice use rate was 18.0% with an average length of stay of 93 days. NH residents’ hospice utilization was not only correlated with individual and facility characteristics but also with NHs’ self-reported EOL care practices styles. A thorough review of the literature was done, which provided direction for the formation of the two research questions. The clearly defined covariates were identified based on literature review and opinions from clinical experts. The rationale of the choice of the statistical analysis method was clearly described. Implications of the findings for future study were discussed at length. However, the findings were not quite consistent with the research questions. The research questions for this study were too narrow and ignored the value of individual factors and facility …show more content…

The sample size was large. Study facilities and all NHs in NYS were compared to evaluate possible selection bias. This study was conducted by the same researchers as the last article (Zheng, Mukamel, Caprio, Cai, & Temkin-Greener, 2011). Some issues I found in the last article were addressed by this article, for example, the data screening before the analysis and the use of prior available full assessment. The data linkage the researchers used to examine the effect of EOL care practices is of interest to me in relation to my research question because some factors I am interested in examining are not included in the MDS assessment and can be obtained by data linkage. Overall, it is a well-written and thought-provoking study. It would be interesting and important to know whether the facility self-reported EOL care practices affect residents’ hospice

Open Document