Area Variations of Cesarean Section Rates, Factors Influencing Physician’s Cesarean Section Rates and Association between Cesarean Section Rates and Perinatal Morbidity

In the past two decades, most investigators have focused on overall CSRs and have used individual level data to demonstrate non-clinical associated factors of the CSR. The units of analysis in this study focused on branches of the National Health Insurance, institutions and physicians. Aims of this study were: 1. Evaluate variations in CSRs at different NHI branches and changes of indications for cesarean section; 2. Investigate factors influencing physicians’ CSRs from the physician and institutional levels; and 3. Investigate relationships between CSRs and perinatal morbidity in the, institutions. The in-patient expenditures by admissions (DD) for other research usage were employed to confirm the trend of CSRs from 1998 to 2010 in a pilot study. The specific subject datasets for annual childbirth in Taiwan were used throughout these three areas of research. This dataset includes registry for contracted medical facilities (HOSB), registry for medical personnel (PER), inpatient expenditures by admissions (DD), details of inpatient orders (DO), ambulatory care expenditures by visits (CD) and details of ambulatory care orders (OO). The order codes in the NHI claim data were used to determine modes of delivery. The operational definition of CSR was annual cesarean delivery divided by all annual live births (claims of live births by cesarean delivery and vaginal delivery) in Taiwan, for each branch, institution and physician. In the third subject area, the in-patient expenditures by admissions (DD) in 2002, 2006 and 2010 were joined with the specific subject datasets for annual childbirth and were used to determine the perinatal admission rate in each institution. In the first subject area, the differences in CSRs between the highest and lowest branch were 9.53% in 1998, 12.5% in 2002, 8.44% in 2006 and 10.84% in 2010. These findings revealed that the CSR in the highest branch was ranging from 129.16% to 146.52% of the CSR in the lowest branch. Distributions of indications for cesarean section in different branches were divergent. Subjective indications such as failure to progress in labor and fetal distress fluctuated only slightly in these four years. Previous cesarean sections gradually decreased. However, malpresentation and malposition increased significantly. High risk indications increased in 2006 at the NHI branch and institutional levels as well as in 2010. In the second subject area, physicians with lower volume of services performed significantly more cesarean sections than their counterparts with high volume of services when controlling institutional factors (0.0237, 95% CI: 0.0041~0.0434, p =0.018 for less than 104 per year; 0.0252, 95% CI: 0.0075~0.0429, p=0.0053 for ≧104-<364 per year). Institutional factors such as policy or preference in institutions (0.026, 95% C.I.：0.0068~0.0452, p=0.00781 for the institutions with medium variation in cesarean section ratios; 0.0332, 95% C.I.: 0.0069~0.0595, p=0.0135 for the institutions with high variation in cesarean section ratios) and accreditation level, not ownerships, played an important role in influencing the physician’s choice of mode of delivery. In the third subject area, higher institutional CSRs correlated to higher institutional perinatal admission rates for respiratory distress syndrome or other respiratory conditions, such as intrauterine hypoxia and birth asphyxia or infections of the fetus and newborn after controlling indications for cesarean section and proportion of older women(≧35 years old) in institutions (β= 0.21797, p=0.0221). This study showed that variations of CSRs in the NHI branches did exist and the CSRs in Taiwan would be less than 32% if the NHI bureau could control up-coding or code creep by institutions or physicians. Physicians’ CSRs were influenced by institutional policy or preference for modes of delivery and accreditation level as well as service volume of physicians. Higher CSRs could result in a negative externality in prenatal newborns.