Preterm birth incidence has risen globally and remains a major cause of neonatal mortality despite improved survival. Demand and cost of initial hospitalization has also increased. This study assessed the cost of preterm birth during initial hospitalization from care provider perspective in neonatal intensive care units (NICU) of two hospitals in the state of Kedah, Malaysia. It utilized universal sampling and prospectively followed up preterm infants till discharge. Care provider cost was assessed using mixed method of top down approach and activity based costing. A total of 112 preterm infants were recruited from intensive care (93 infants) and minimal care (19 infants) units. Majority were from the moderate (23%) and late (36%) preterm groups followed by very preterm (32%) and extreme preterm (9%). Median cost per infant increased with level of care and degree of prematurity. Cost was dominated by overhead (fixed) costs for general (hospital), intermediate (clinical support services) and final (NICU) cost centers where it constituted at least three quarters of admission cost per infant while the remainder was consumables (variable) cost. Breakdown of overhead cost showed NICU specific overhead contributing at least two thirds of admission cost per infant. Personnel salary made up three quarters of NICU specific overhead. Laboratory investigation was the cost driver for consumables. Gender, birth weight and length of stay were significant factors and cost prediction was developed with these variables. This study demonstrated the inverse relation between resource utilization, cost and prematurity and identified personnel salary as the cost driver. Cost estimates and prediction provide in-depth understanding of provider cost and are applicable for further economic evaluations. Since gender is non-modifiable and reducing LOS alone is not effective, birth weight as a cost predictive factor in this study can be addressed through measures to prevent or delay preterm birth.
* Title and MeSH Headings from MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine.