Dataset Information

Generalizing observational study results: applying propensity score methods to complex surveys.

ABSTRACT:

Objective

To provide a tutorial for using propensity score methods with complex survey data.

Data sources

Simulated data and the 2008 Medical Expenditure Panel Survey.

Study design

Using simulation, we compared the following methods for estimating the treatment effect: a naïve estimate (ignoring both survey weights and propensity scores), survey weighting, propensity score methods (nearest neighbor matching, weighting, and subclassification), and propensity score methods in combination with survey weighting. Methods are compared in terms of bias and 95 percent confidence interval coverage. In Example 2, we used these methods to estimate the effect on health care spending of having a generalist versus a specialist as a usual source of care.

Principal findings

In general, combining a propensity score method and survey weighting is necessary to achieve unbiased treatment effect estimates that are generalizable to the original survey target population.

Conclusions

Propensity score methods are an essential tool for addressing confounding in observational studies. Ignoring survey weights may lead to results that are not generalizable to the survey target population. This paper clarifies the appropriate inferences for different propensity score methods and suggests guidelines for selecting an appropriate propensity score method based on a researcher's goal.

SUBMITTER: Dugoff EH

PROVIDER: S-EPMC3894255 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundThe association between preoperative aspirin use and postoperative acute kidney injury (AKI) in cardiovascular surgery is unclear. We sought to evaluate the effect of preoperative aspirin use on postoperative AKI in cardiac surgery.MethodsA total of 770 patients who underwent cardiovascular surgery under cardiopulmonary bypass were reviewed. Perioperative clinical parameters including preoperative aspirin administration were retrieved. We matched 108 patients who took preoperative aspirin continuously with patients who stopped aspirin more than 7 days or did not take aspirin for the month before surgery. The parameters used in the matching included variables related to surgery type, patient's demographics, underlying medical conditions and preoperative medications.ResultsIn the first seven postoperative days, 399 patients (51.8%) developed AKI, as defined by the Kidney Disease Improving Global Outcomes (KDIGO) criteria and 128 patients (16.6%) required hemodialysis. Most patients took aspirin 100 mg once daily (n = 195, 96.5%) and the remaining 75 mg once daily. Multivariable analysis showed that preoperative maintenance of aspirin was independently associated with decreased incidence of postoperative AKI (odds ratio [OR] 0.46, 95% confidence interval [CI] 0.21-0.98, P = 0.048; after propensity score matching: OR 0.39, 95% CI 0.22-0.67, P = 0.001). Preoperative maintenance of aspirin was associated with less incidence of AKI defined by KDIGO both in the entire and matched cohort (n = 44 [40.7%] vs. 69 [63.9%] in aspirin and non-aspirin group, respectively in matched sample, relative risk [RR] 0.64, 95% CI 0.49, 0.83, P = 0.001). Preoperative aspirin was associated with decreased postoperative hospital stay after matching (12 [9-18] days vs. 16 [10-25] in aspirin and non-aspirin group, respectively, P = 0.038). Intraoperative estimated or calculated blood loss using hematocrit difference and estimated total blood volume showed no difference according to aspirin administration in both entire and matched cohort.ConclusionsPreoperative low dose aspirin administration without discontinuation was protective against postoperative AKI defined by KDIGO criteria independently in both entire and matched cohort. Preoperative aspirin was also associated with decreased hemodialysis requirements and decreased postoperative hospital stay without increasing bleeding. However, differences in AKI and hospital stay were not associated with in-hospital mortality.