¶ … inferential statistic tests used in study. What were these tests typically used for? Why were they chosen here?
The objective of the study was to analyze the true costs of hypertensions. The researchers did this by analyzing the data of four patient groups using propensity score matching to control for possible bias in cost estimates. The regression model that followed estimated for costs of hypertension by controlling for sex, length of hospital stay, Charlson comorbidity index, region of residence, and urbanization of residence.
Researchers used the 2005 MarketScan CCAE database, that contained information about hospitalized patients who belong to more than 100 health insurance plans offered by about 40 employers, in order to estimate hypertension associated hospitalization costs for patients with hypertension as a secondary diagnosis.
The problem was that since more than 95% of the hypertensive patients in the CCAE study had hypertension as secondary rather than primary diagnosis, hypertension-related costs would be dramatically underestimated if they were based only on costs that were incurred by the patients who had a primary diagnosis of hypertension. On the other hand, if cost estimates were based on total costs for all hypertension patients without complications (such as age and gender and other comorbidities included), the costs for hypertension might become grossly exaggerated. For this reason, the researchers used propensity score matching.
Since researchers wanted the true costs of hypertension but since their results after screening the data from the CCAE database resulted in skewed characteristics between patients with hypertension and patients without hypertension, they employed propensity score matching. Examples of difference in distribution of characteristics include the fact that the average age for patients with hypertension...
More significantly, whilst 48% of the patients with hypertension were men, only 30% of patients without hypertension were men. These differences could create potential bias in cost estimates (since differences in age and gender may effect differences in medical treatment received, accordingly skewing medical costs). Therefore a propensity score-matching (PSM) test was used. PSM tests are often used, particularly in observational studies when randomization is difficult or impossible. In some studies such as in this one researchers must rely on previous data (or naturalistic studies) from which they infer the impact (or cost as in this case) of different types or levels of treatment form differences among individuals receiving different care or having different treatment or disease (as in this case no hypertension or hypertension as primary). Ordinarily, without these a priori confounding variables and when randomization can be used, a simple regression model would be performed to test for differences. Here, however, the already innate differences between the two groups can interfere in the regression model and confound data. Statistical matching could be used instead where cases are grouped to minimize differences between matched cases, and poor matches are excluded, but statistical matching is unwieldy and clumsy when used for data that is complex and when many characteristics (such s in this case) are involved.
PSM that essentially uses probability in order to predict probability that an individual receives the treatment of interest. Consistent research has indicated that the PSM provides a reliable measure for capturing all of the variance in the covariates that are needed for forming and adjusting between-group comparisons. Using the PSP, therefore, one can simply match the two or more groups based on…
