Study population
A cross-sectional study was designed in the area of Hospital del Mar in Barcelona, Spain (a public teaching hospital), and its primary care program for sexual and reproductive health in the Barcelona districts of Ciutat Vella and San Martí, representing a catchment area of 320,000 inhabitants.
All the deliveries in Hospital del Mar from October 2006 to September 2007 were included. We gathered all the public healthcare and economic information on the pregnancy (starting at 9 months before delivery), the delivery itself (including associated maternal and neonatal costs), and data on the mother and child up to 3 months after delivery. As a public service, professionals involved in the childbirth healthcare process provided services to women regardless of their origin or any other characteristic. Foreign-born women were from South America, Asia and the North of Africa, being Morocco, Pakistan and Ecuador the most represented countries.
During the study period there were 1,475 childbirths in Hospital del Mar. A total of 449 deliveries were excluded: 385 (26.1%) because the mother's residence fell outside the catchment area and 64 (4.3%) because data could not be linked among the information systems. Finally, 1,026 childbirths were analyzed.
All the information was treated confidentially. The ethical principles for medical research in human beings defined in the Declaration of Helsinki of 1964 and revised by the World Medical Organization in Edinburgh 2000 and Organic Law 15/1999 of Data Protection in Spain were followed. The study was approved by the ethics comitee of IMIM-Hospital del Mar.
Cost assessment
The overall cost of a delivery was calculated by taking into account the cost of the activities related to the process including the 9 months of pregnancy prior to delivery, delivery itself and the 3-month postnatal period for both the mother and the newborn. The information sources were the activity records of the Program for Sexual and Reproductive Health, the health care information system and the cost accounting system of Hospital del Mar.
Hospital del Mar uses a hospital cost accounting system based on full-costing allocation. In the present study cost estimation was based on a full-costing accounting system and on the criteria of clinical Activity Based Costing (ABC) methods to obtain maximal sensitivity in the assessment of clinical activity variability. This system ensures that the hospital's total costs are distributed among the patients. Allocation is based on directly assigning the cost of the following services to the patient: laboratory, pharmacy, radiology, nuclear medicine, pathology, prostheses and other diagnostic and therapeutic tests. The information system contains exhaustive data on human resources and their activity: storage, admissions planning, ambulatory and emergency care, operating rooms, diagnostic and complementary tests, and inter-hospital consultations. This information creates and automatically updates the cost for overheads. Fixed cost derived from surgical procedures, hospitalization, and intensive care unit stay, is distributed based on routine criteria: duration of the intervention or the number of stays among the different hospitalization units. The cost predictor variables were age, maternal origin (Spanish-born versus foreign-born), prenatal care (yes or no), delivery type [according to the ICD-9: vaginal non-instrumented (73.4 and 73.59); instrumented (72.XX, 73.1, 73.2, 73.3 and from 73.80 to 73.99) and cesarean section (74.XX)], multiple pregnancy and maternal and neonatal severity (measured by the APR-DRG [20]) severity index. The APR-DRG severity index was based on the comorbidites included in the secondary diagnostics used to calculate the APR-DRGs. For newborns, it was available for admitted newborns only. All newborns are registered after birth, but only those newborns needing tests or clinical assistance are admitted to the hospital.
Statistical analysis
The unit of analysis was maternal admission. A Chi-square test was used to compare categorical variables. The Kruskal-Wallis and Mann-Whitney U tests were used to compare continuous variables. Partial costs of the process and the number of acts were analyzed overall and by maternal origin and delivery type. Overall cost was analyzed by maternal age, maternal origin, prenatal care, delivery type, postnatal care, maternal and neonatal severity and multiple pregnancy. Overall cost was normalized through the log and the inverse transformations, lending the inverse transformation a closer approximation to normality. Multivariate analyses of hospital cost were performed using generalized linear models with inverse-transformed costs. All the variables significantly related to the outcome of interest (p < 0.20) were included in the multivariate model. Final multivariate models were adjusted by maternal age, maternal origin, pregnancy follow-up, delivery type, maternal and neonatal severity score (according to APR-DRG severity of illness subclass [20]), and multiple delivery. After fitting the model, predictions of the overall inverse-transformed cost were calculated for each category of the predictor variables, while the remaining variables were kept at the reference level (15 for age, Spanish-born woman for origin, yes for prenatal care, vaginal non-instrumented for delivery type, mild for maternal severity, non-admitted to hospital for neonatal severity and single for multiple birth). To interpret the results of the model, these predictions were transformed back to euros. The value of the constant of the model corresponded to the cost of a vaginal non-instrumented delivery of a 15-year-old Spanish-born woman with prenatal care and low severity, who delivered a singleton not admitted to hospital. By using the estimated overall cost, the incremental cost and its weight (%) for each category of the predictor variables was calculated.
The level of statistical significance was set at 0.05 and all tests were two-tailed. Analyses were conducted using SPSS, version 13.0 (SPSS, Chicago, IL, USA, 2003).