|
Abstract:
|
In the United States the number of health systems that own practices or hospitals have increased in number and complexity. This has led to interest in assessing the relationship between health organization factors and health outcomes. To meet this goal, national surveys of United States health care systems, hospitals and (physician) practices were developed. The existence of multiple types of organizations combined with the nesting of some hospitals and practices within health systems and the nesting of some health systems within larger health systems generates multiple questions of interest and analytic objectives. Targets of inference range from the nationwide prevalence of health systems with certain characteristics to regression coefficients in hierarchical models that represent the relationship between specific health system, hospital and practice level predictors measured in the respective surveys and quality or other outcomes. Because the optimal design for one objective may be far from optimal for another (e.g., because statistical precision at the different levels of the model depends on different elements of the design), compromise is necessary. An objective function that explicitly weighs all objectives is theoretically appealing but becomes unwieldy and ad hoc as the number of objectives increases. To overcome this problem, we consider an alternative approach based on constraining the sampling design to satisfy desired statistical properties. For example, to support evaluations of the comparative importance of factors measured in different surveys on health system performance, a constraint requiring that at least one organization of each type (corporate owner, hospital, practice) is sampled whenever any component of a system is sampled may be enforced. Multiple such constraints define a nonlinear system of equations whose solution yields the sample inclusion probabilities for each organization. In addition to developing novel sampling design methodology, another important contribution of this paper is the Monte Carlo algorithm that simultaneously solves the system of equations to estimate the sampling probabilities and to extract the survey samples. We use this algorithm to illustrate the virtues of “coupled sampling” by comparing the proportion of eligible systems for whom the corporate owner and both a hospital and a practice are expected to be sampled to that obtained under alternative sampling designs. Comparative and descriptive analyses that illustrate features of the sampling design are also presented.
|
