Cost-effectiveness of Alternative NICU Designs

Abstract

The design of the physical space of the neonatal intensive care unit (NICU) has been proposed as a target for intervention to improve neonatal outcomes. Most NICUs are currently designed so that infants are cared for in a shared space which is commonly referred to as an open bay design. The purpose of this study was to identify the cost-effective NICU design from a Canadian public payer perspective using a lifetime time-horizon. The study used a decision model approach informed by individual participant data for estimating the baseline history of disease and the impact of morbidities on length of stay. Efficacy parameters were informed through analysis from an ongoing randomized controlled trial (RCT) of single-family room design and a systematic and targeted literature search. Meta-analysis of efficacy parameters was conducted using a multivariate network meta-analysis. A stochastic multi-criteria acceptability analysis (SMAA) was conducted to provide an additional perspective limited to clinical outcomes of alternative room designs in addition to an ordinal ranking of their degree of family centeredness. The network meta-analysis included a total of 25 studies (2 RCTs, 23 observational studies) evaluating old open bay, new open bay, half-wall, private room, combined single-family room and open bay, and single-family room only designs. When using a life-time horizon all designs had a higher expected value of net monetary benefit than old open bay designs although there was substantial uncertainty in relative ranking and the probability that any given design was cost-effective. Half-wall designs had the highest mean expected value of net monetary benefit over a wide range of values of willingness to pay. Conclusions were robust to sensitivity analyses, including the results of the SMAA. Results were limited by the strong assumptions required to create a connected network that assessed all outcomes of interest. The decision to undertake new construction and the decision on which design aspects to integrate should reflect the complex and multi-factorial nature of the decision problem.