Predictors of cost-related outcomes of neonatal intensive care

Abstract

In “Acuity Scores as Predictors of Cost-Related Outcomes of Neonatal Intensive Care” (J PEDIATR 1995;126:88-93), the predictive power of the Iameter Acuity Index Method (AIM) was compared with that of the Medicus Patient Classification System (PCS). My fundamental concern is that the comparison was made at all. The two systems are entirely different in their intention, uses, and data requirements. Concluding that the AIM score could not be validated as an appropriate method for assessing cost-related outcomes after newborn care is misleading, inappropriate, and damaging. The AIM statistically predicts hospital inpatient charges, length of stay, and mortality by use of the widely available Uniform Hospital Discharge Data Set and is intended for retrospective analysis of these outcomes relative to expected. The elimination of labor-intensive data capture and predictions relative to large federal and state normative data sets are AIM's main strengths. AIM has been documented in several rigorous independent reviews to perform similarly to other systems based on similar data.1Thomas JW Ashcraft MLF. Measuring severity of illness: six severity systems and their ability to explain cost variation.Inquiry. 1991; 28: 39-55PubMed Google Scholar, 2MacKenzie TA William AR Lichter J et al.Patient classification systems: an evaluation of the state of the art.. Case Mix Research, Queen's University,, Kingston, Ontario, Canada:1991Google Scholar The PCS is based on the daily accumulation of copious amounts of detailed concurrent data. It is neither surprising nor useful to point out that such detailed data are better at predicting charges and length of stay than the retrospectively acquired Uniform Hospital Discharge Data Set information. The appropriateness of testing, vital signs, and length of stay is not addressed. A patient receiving the intensive testing documented by PCS is automatically classified as being more severely ill. The higher charges and length of stay are justified by service-related activity rather than comparisons with a predicted level based on independent variables. This circular logic is what has engendered the intense scrutiny by purchasers that medicine now endures. Systems like PCS are useful in areas of focused clinical analysis. However, when comparisons to national norms are needed and until more economically justifiable detailed predictive methods become available, AIM and similar methods are the best alternative. 9/35/66834