Baylor

The Alchemy Problem.

This is a pilot study designed as a preparatory project to the principal investigator’s application for a mentored career award. The goal of the pilot study is to identify whether high performance in one quality measure of neonatal intensive care translates into high performance in other measures. The applicant’s career goal is to develop valid and reliable quality measures for neonatal intensive care.

The Institute of Medicine's (IOM) reports on quality of care have sparked a renewed focus on health care quality. The IOM defines quality of care as "the degree to which health care services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge." High quality care requires excellence in various domains of quality including safety, effectiveness, efficiency, patient-centeredness, timeliness, and equity. Currently the health system is not meeting these standards. McGlynn and colleagues reported that only slightly more than half of adults receive care consistent with evidence based guidelines.

It is reasonable to assume that neonatal intensive care suffers from similar quality problems. Donabedian provided a framework for quality of care using structure, process, and outcomes. In this framework the physical and organizational make-up of a hospital and its NICU influences the quality of provider-patient interactions, which in turn influence patient outcomes. The neonatal literature supports the validity of Donabedian's framework in demonstrating how wide variability in terms of structure and processes of care eventually lead to differences in morbidity and mortality.

Awareness of such performance differentials has led to creative attempts at improving the quality of neonatal intensive care. However, projects usually focus on a single measure of care, not systemic improvement. It is therefore not clear whether high NICU performance or improvement in one measure is a signal for system-wide performance or whether there are trade-offs in other areas of care.

The focus on quality has been intensified as ever increasing health care and insurance premium costs are becoming a heavy financial burden for the US government and a competitive disadvantage for US industry. As a result there has been a growing conviction on the side of purchasers of health care that financial incentives may play a central role in improving health care quality. A number of “pay-for-performance” programs are being implemented in industrialized countries. In theory, such incentives will induce competition among providers, improve efficiency of care and reduce waste. Financial incentives for quality should work if we can demonstrate the existence of NICUs that provide superior or "gold" quality in not only one domain but consistently across various domains of care. Else, pay-for-performance may simply shift quality between domains as NICUs scramble to receive the reward in the measured behaviors but neglect unmeasured ones.

In summary, the central question we are posing is whether we can identify gold performers across multiple dimensions of quality among NICUs or if gold performance in one area of quality is accompanied by average performance in another.

This research project seeks to retrospectively identify the existence of a high performing or “gold” NICU using de-identified data from the California Perinatal Quality Care Collaborative (CPQCC). A gold NICU will achieve consistently high rankings across various NICU quality domains.

The specific aims of this project are:

Aim 1: To identify the existence of high performing, “gold” NICUs by investigating the correlation of NICU rankings on several quality measures across different domains.

Hypothesis 1.1: When the eight measures are used to classify the NICUs into the upper 20% and the remaining 80%, the agreement among the eight measures will be significant at the 0.05 level.

Aim 2: To investigate the degree of correlation of NICU rankings on several quality measures across different domains in other than high performing NICUs.

Hypothesis 2.1: When the eight measures are used to classify the NICUs into the lower 20% and the remaining 80%, the agreement among the eight measures will be significant at the 0.05 level.

Hypothesis 2.2: When the eight measures are used to classify the NICUs into the middle 60% and the remaining 40%, the agreement among the eight measures will be significant at the 0.05 level.

Database Review

None

While there will not be any direct benefit to the patients whose charts have been abstracted, we anticipate that future generations of premature newborns will benefit from our improved understanding of quality of neonatal intensive care.