Impact of alternatives to pure fee-for-services payment within Accountable Care Organizations

Rationale: The way to achieve health care delivery efficiency gains in order to support neither the maintenance nor the extension of the level of health care coverage is a key issue for the United States. In that prospect, policies have been implemented to reinforce primary health care delivery notably by supporting primary care teams (PCT) and new organizational forms. The American health care systems have faced similar barriers to ensuring that payment structures foster care coordination and meet the needs of particular patient groups – including high-need, high-cost and other vulnerable groups. Within a context of controversy on the optimal structure of payment in primary care, new organizational forms have emerged as alternatives to the pure fee-for-service payment models: the accountable care organizations (ACOs). Few researches evaluate the impact of these advanced models on performance depending on the regional and local context, on the organizational characteristics and the level of financial risk bearing, and finally, on the characteristics of the patients.

Objective: This research aims to analyze the impact of these alternative remuneration schemes on two efficiency of health care delivery dimension: the technical efficiency (i.e. maximization of outputs with a set of inputs) and the balance of care and services (e.g. specialist vs. primary care physician’s) delivered and “used” by patients. The research hypothesis is that a mixed payment system which combines prospective payments and/or saving models, additionally to fee-for-services payment, is effective in stimulating team working and efficiency gains.

Methods: The research framework and design combine three main steps. First, an explanatory design, with factor and cluster analysis of mixed data based on the National Survey of Accountable Care Organizations (NSACO), to categorize ACOs and to determine which ACO look like another ACO and on what characteristics it is based with a focus on how are sharing risk and savings internally. Second, a qualitative design, based on few visits and interviews of ACO’s executives and/or PCTs, to confirm (or not) the above hypotheses. Third, a quantitative design, with econometric analysis of panel data based on Dartmouth claims database, to estimate if different risk sharing schemes have differential impact on efficiency taking into account of the results of the two previous steps. Stochastic production frontier models will be implemented to estimate technical efficiency using production outputs of patients encountered, listed, visits and acts delivered. Sample selection or hurdle models will be implemented to estimate ambulatory care and services expenditures by category conditional to the utilization.

Data sources: Wave one to three of the National Survey of Accountable Care Organizations (NSACO), from The Dartmouth Institute for Health Policy and Clinical Practice, Hanover, New Hampshire, and The School of Public Health, University of Berkeley, California. 347 ACOs complete the full survey that will be merged for their beneficiaries with Dartmouth claims database, which is one of the largest nonfederal repositories of Medicare in USA.