Machine Learning and Health Care: Potential Benefits and Issues.

We discuss the potential for machine learning (ML) and artificial intelligence (AI) to improve health care, while detailing caveats and important considerations to ensure unbiased and equitable implementation. If disparities exist in the data used to train ML algorithms, they must be recognized and accounted for, so they do not bias performance accuracy or are not interpreted by the algorithm as simply a lack of need. We pay particular attention to an area in which bias in data composition is particularly striking, that is in large-scale genetics databases, as people of European descent are vastly overrepresented in the existing resources.

Improving Quality Measurement: Design Principles for Quality Measures.

The use of quality measures to adjust health care payments and to rank providers is growing rapidly, but there are many problems with the quality measures that are currently being used. This article discusses some of these problems and then lays out some principles and procedures that should be used in the development and combination of quality measures. Many of the problems with existing quality measures would have been avoided had these principles been applied as they were developed.

Problematic Risk Adjustment in National Healthcare Safety Network Measures.

This article reviews the risk-adjustment models underpinning the National Healthcare Safety Network (NHSN) standardized infection ratios. After first describing the models, the authors focus on hospital intensive care unit (ICU) designation as a variable employed across the various risk models. The risk-adjusted frequency with which ICU services are reported in Medicare fee-for-service claims data was compared as a proxy for determining whether reporting of ICU days is similar across hospitals.

Indications of biased risk adjustment in the hospital readmission reduction program.

The Centers for Medicare and Medicaid Services hospital readmission reduction program administers substantial penalties to hospitals with excess readmissions. In our analysis of Medicare claims data, we find hospitals with the highest percentages of patients with several chronic conditions and advanced age have excess readmission ratios that are overstated because of inadequate risk adjustment. The distribution of chronic conditions and age is sufficient to cause 4.

Conceptual errors in the CMS refusal to make socioeconomic adjustments in readmission and other quality measures.

The Centers for Medicaid & Medicare Services has made a policy decision that socioeconomic factors should not be adjusted for in its various quality measures and point both to arguments made by the National Quality Forum and to analysis of the distributions of quality results to support this view. We present counterarguments to this viewpoint and use the results reported by the Centers for Medicaid & Medicare Services to support its position to demonstrate that adjustments are necessary. We further argue that the incentives for providers to improve performance would not be weakened by including socioeconomic factor adjustments.

Hospital readmission rates: the impacts of age, payer, and mental health diagnoses.

We examine impacts of age, payer, and mental health conditions upon hospital readmissions and the comparability of same-hospital and multiple-hospital readmission rates. Medicaid primary payment and extreme age are associated with significantly higher readmission rates. We find low correlation between same-hospital and multiple-hospital readmission rates and identify urban hospitals with high proportions of Medicaid patients and mental health admissions as factors driving the use of multiple hospitals within readmission chains.

State hospital rate-setting revisited.

In an attempt to control rapid growth in hospital costs, beginning in the mid-1970s several states implemented rate-setting programs to regulate hospital payments. In seven states, rate-setting was in effect for a substantial period of time (14 years or more). While most of these programs were discontinued by the mid-1990s, two are still active.

Provider-initiated pay-for-performance in a clinically integrated hospital network.

Long Island Health Network developed a provider-initiated pay-for-performance (PI-PFP) program beginning in 2004 and operated by 10 clinically integrated hospitals. The PI-PFP administrative processes, length of stay, patient satisfaction, and Hospital Quality Alliance measures are elaborated. PI-PFP has evolved over time and supports best practice sharing.

The use of ambulatory patient groups for regulation of hospital ambulatory surgery revenue in Maryland.

The Maryland Health Services Cost Review Commission (HSCRC or the commission) is a government agency with the authority to establish rates for both inpatient and outpatient services for all general acute care hospitals in the state. By law and consistent with the state's unique Medicare waiver, all payers (including Medicare and Medicaid) must pay hospitals on the basis of these rates. The HSCRC has used diagnosis related groups to set case-mix-adjusted limits on the revenue per discharge for inpatient services (similar to Medicare inpatient prospective payment nationally) yet, the Maryland rate-setting system for outpatient services has not embodied incentives to control utilization of services.