Recent documentation shows that rates of cesarean delivery are again on the rise for both women who have not had a cesarean delivery before (primary) and for women who have undergone a previous cesarean delivery (repeat).1 Primary cesarean deliveries are an important target for reduction, because they lead to an increased risk for a repeat cesarean delivery. Of particular interest are the cesarean deliveries that are elective, although the clinical use and implications of the term elective requires clarification. Elective cesarean deliveries can include medically and obstetrically indicated procedures that generally occur before labor. Elective cesarean deliveries can also include procedures for which there is no clear medical or obstetric indication. There is a growing concern that there is a rising rate of the latter. These cesarean deliveries are referred to as “maternal-choice” cesarean deliveries and are performed for convenience, pelvic preservation, and reduction of neonatal morbidity. Maternal-choice elective primary cesarean deliveries generate both clinical and ethical controversy and concern.2
As yet, information on trends in the use of elective primary cesarean delivery to support or dispel concerns about either quantity or appropriateness of use has been lacking. This is due to the fact that one cannot easily identify elective primary cesarean deliveries from typical data sources. Birth certificates and hospital discharge data record whether the cesarean delivery is primary or repeat but they do not include information about whether the woman labored before the operative procedure or what specific indication(s) led to the cesarean delivery.
International Classification of Diseases, 9th Revision (ICD-9) algorithms to identify labor and indications for cesarean delivery have been established.3–5 These algorithms are based on a comparison of hospital discharge data and chart review by obstetricians. In these algorithms, elective primary cesarean delivery was defined as a procedure that occurred before labor and without a previous history of cesarean delivery.
Tracking of national trends in the use of elective primary cesarean delivery is important for informing obstetric practice, and the use of ICD-9 algorithms allows for this type of investigation. The purpose of our study was to evaluate national trends in the elective primary cesarean delivery rate, defined as occurring before labor, from 1994 to 2001 and to comment on shifts in indications using previously reported methodology. We also sought to describe differences in maternal and hospital characteristics for the group identified as having undergone elective primary cesarean deliveries between 1994 and 2001.
MATERIALS AND METHODS
We used data without identifiers from the Nationwide Inpatient Sample (NIS) of the Healthcare Cost and Utilization Project (HCUP) for 1994 to 2001. The Healthcare Cost and Utilization Project is built through a partnership between the state data organizations and the Agency for Healthcare Research and Quality in which the state partners contribute their unique statewide hospital discharge data to HCUP. The data are then subjected to internal consistency and edit checks. Data elements that are similar across the states are recoded into a uniform coding scheme and data elements unique to individual states are retained if they are useful for research purposes. These uniformly formatted data sets are the core of the HCUP. The number of participating states has increased from 17 in 1994 to 33 in 2001. The 33 states contributing data to HCUP in 2001 captured approximately 85% of all hospital discharges in the United States.
Once the states contribute their data to HCUP, the Nationwide Inpatient Sample is created. The NIS is a stratified probability sample of hospitals in the United States. Hospitals are selected on the basis of a sampling frame that uses five strata: rural or urban location, number of beds, region, teaching status, and ownership. All discharges are retained for each sampled hospital. The hospital universe is defined as all hospitals located in the U.S. open during any part of the calendar year and designated as community hospitals in the American Hospital Association (AHA) Annual Survey.6
The 2001 NIS includes information on 7.1 million discharges from 986 nonfederal community hospitals located in the 33 states, which when weighted, provide estimates representing the total number of inpatient hospital discharges in the United States. The NIS provides a research database for conducting national and regional studies of inpatient care delivered in the United States.7 This study was exempt from internal review board approval.
The ICD-9 algorithms developed by Gregory et al3 and Henry et al4 were applied to the NIS to identify cases of previous cesarean delivery, labor, and indications. Henry et al4 validated by chart review 1,885 cesarean deliveries at Cedars Sinai Medical Center in 1992 and then modified a hierarchical classification scheme developed by Anderson and Lomas8 to identify repeat cesarean procedures, indications, and labor status for cesarean deliveries. Gregory et al3 refined the Henry et al4 classification algorithm with recursive modeling techniques in conjunction with obstetric review of appropriateness of indications and applied the resulting algorithm to the 1995 California state hospital discharge data to arrive at 12 indications for elective cesarean deliveries that comprised 92.9% of elective cesarean procedures, whereas 7.1% had unclear or unspecified indications.3 When testing the algorithms on the HCUP California data, we obtained the same results.
Cases of cesarean delivery for numerators and deliveries for denominators using International Classification of Diseases, 9th Revision, Clinical Modification primary and secondary procedure codes are listed in Table 1. Cases of repeat cesarean delivery and laboring women, as defined in the Henry algorithm and identified by codes listed in Table 1 were removed.4 The remaining cases that had not labored and had not had a previous cesarean delivery were then assigned a primary indication for cesarean delivery according to the hierarchy of ICD-9 codes described by Gregory et al.3 Once the first indication was identified, those cases associated with it were removed from the pool and the next indication in the hierarchy was applied until all 12 indications were exhausted. The remaining cases were labeled unspecified.
All cesarean delivery rates and primary and elective primary cesarean delivery rates were calculated per 100 deliveries per year and graphed from 1994 to 2001 except for those with a code for fetal heart rate abnormality, which was only available from 1998 to 2001. To account for the complex sampling design and sample discharge weights in the NIS, SUDAAN 8.02 (Research Triangle Institute, Research Triangle Park, NC) was used to generate appropriate variances, standard errors, and confidence intervals. Trends were also assessed in SUDAAN 8.02 by testing polynomial contrasts for goodness of fit for linear, quadratic, and cubic trends. Because of the increase in the number of cesarean deliveries over the study period, rates per 1,000 deliveries were calculated to evaluate differences in use of coded indications.
Demographic and hospital data for elective primary cesarean deliveries between 1994 and 2001 are presented in Table 2. There were more women aged older than 40 years in 2001 and a shift from nonteaching to teaching hospitals.
The total cesarean delivery, the primary cesarean delivery, and the elective primary cesarean delivery rates decreased from 1994 to 1997 and then showed an increase (Fig. 1). The rate for all cesarean deliveries increased approximately 14% from 1998 to 2001; the rate for primary cesarean deliveries increased approximately 13%, and the rate for elective primary cesarean deliveries increased approximately 53%. Elective primary cesarean deliveries rose from 19.7% of all cesarean deliveries in 1994 to 28.3% in 2001. The trend tests show a quadratic trend over the whole study period at the P < .05 level.
There was a considerable rise in the elective primary cesarean delivery rate after 1997, yet minimal changes in the use of the 12 indications we identified. The exception to this observation was unspecified procedures, which rose from 3.5 per 1,000 (95% confidence interval [CI] 3.1–3.9) deliveries in 1997 to 16.6 per 1,000 (95% CI 15.8–17.4) deliveries in 2001. Because the largest coding shift was an increase in the unspecified procedures and because the hierarchical approach of assigning the indication resulted in grouping cesarean deliveries with new codes to the unspecified category, we investigated coding changes in the unspecified category during the study period. We identified a new International Classification of Diseases, 9th Revision, Clinical Modification code for abnormality in fetal heart rate or rhythm, 659.71, which was introduced in October 1998 and rapidly adopted during 1998–2001 as the ICD-9 code that accounted for the rise in the unspecified group and the leading cause for the rise in the overall total of elective cesarean procedures. Using the hierarchical approach for indications for 2001 data, fetal heart rate abnormalities accounted for 16.8% (95% CI 16.0–17.6%) of elective primary cesarean deliveries or 47,325 (95% CI 43,060–51,591) procedures. The 12 indications in hierarchical order per the Gregory model and their respective rate of use per 1,000 deliveries are presented for 1997 and 2001 NIS data in Table 3.
We have presented a national estimate of the elective primary cesarean delivery rate with attention to indications using a previously derived methodology. The overall elective primary cesarean delivery rate, that is, primary procedures occurring before labor, is rising. Within the group of procedures with unspecified indications, use of a new code for fetal heart rate abnormalities accounts for the largest shift in coding.
A recent study from Nova Scotia reported very similar rates of elective primary cesarean delivery, also defined as cesarean delivery performed before labor, despite a much smaller sample size and differences in proportions of indications. The authors report an initial rate of cesarean delivery without labor of 2.9% in 1988, which rose to 7.6% in 2001.9
Some indications had increased use. That is, malpresentation, antepartum bleed, hypertension and severe hypertension, macrosomia, unengaged head, preterm gestation, and maternal soft tissue disorder were coded more frequently over the study period. Changes in the use of the codes could be due to an increased incidence or severity of the disease or practice pattern changes. Practice pattern changes in this case refer to the change in choice of mode of delivery by the physician for the same indication. For example, there was an increase in the use of coding for malpresentation. This could be a reflection of a decrease in the use of external cephalic version, an intervention during pregnancy that can reduce the number of nonvertex presentations.
Although there was clear nationwide use of the ICD-9 code for fetal heart rate abnormalities, it is unclear how fetal heart rate abnormality is diagnosed, why there is a rapid rise in the use of this new code, and whether there is a relationship between this new code and maternal-choice cesarean delivery or cesarean delivery for fear of medical liability. Whether to accept this new code as a clinically appropriate indication is an important question. Critical reviews of the literature state that there is insufficient evidence to evaluate the effectiveness of tests of fetal well being as a mechanism to reduce perinatal mortality,10–12 although a recent review supports antenatal testing for older women.13 Continued monitoring of these trends will be useful for the public health.
Hospital and payor data trends showed little difference in general between 1994 and 2001 for elective primary cesarean deliveries and those procedures associated with an abnormal fetal heart rate. A similar shift from nonteaching to teaching hospitals is seen for all NIS hospital discharges between 1997 and 2001(data not shown). Finally, one could have hypothesized that older, more financially stable mothers were undergoing elective primary cesarean delivery based on reports in the literature.14,15 Although we found a shift toward women aged older than 40 years, our data did not show a change in payor status during the study period.
This study has several limitations. We did not examine indications for cesarean deliveries other than those of the elective primary type, and our work did not attempt to recreate the validated hierarchical model of cesarean delivery indications. Thus, although coding changes did occur, there was an easily identifiable concentration of change in the overwhelming use of a new single code. It is also important to note that there are limitations inherent to hospital discharge data. In this study, the ICD-9 algorithms were developed using information that compared chart review by obstetricians to ICD-9 code assignments. Note, however, that there has not been a chart review of the cases assigned to fetal heart rate abnormalities and that the results are based on a sampling framework.
Terminology to clearly describe elective compared with nonelective procedures is not clear, and public health measures would benefit from standardization. The conventional use of the term elective in surgery can refer to either ease in scheduling or procedures that are medically not necessary. The use of the term elective when referring to certain indications for cesarean delivery, such as breech delivery that could be accomplished vaginally in some cases, has ambiguous connotations. We have used classification by labor status to define whether the procedure is elective because extensive work to develop algorithms had been done warranting national examination, and this convention is supported by many obstetricians. However, eliminating the use of the term elective and then classifying procedures into prelabor or labor cesarean delivery would be an improvement. Maternal-choice cesarean deliveries would then be a subset of prelabor cesarean delivery and should be identified by a unique code.
In conclusion, we found the use of ICD-9 algorithms to be valuable for describing trends in indications for and use of elective primary cesarean delivery. By means of the hierarchical method described here for the 2001 HCUP data, the use of elective primary cesarean delivery has increased markedly from 1994 to 2001, and indications for its use are shifting. Further investigation into the perinatal environment leading to the increase of elective primary cesarean delivery and the outcomes of these procedures is warranted.
1. Martin JA, Hamilton BE, Ventura SJ, Menacker F, Park MM, Sutton PD. Births: final data for 2001. National vital statistics reports; vol 51 no. 2. Hyattsville, MD: National Center for Health Statistics; 2002.
2. Minkoff H, Powderly KR, Chervenak F, et al. Ethical dimensions of elective primary cesarean delivery. Obstet Gynecol 2004;103:387–92.
3. Gregory KD, Korst LM, Gornbein JA, Platt LD. Using administrative data to identify indications for elective primary cesarean delivery. Health Serv Res 2002;37:1387–1401.
4. Henry OA, Gregory KD, Hobel CJ, et al. Using ICD-9 codes to identify indications for primary and repeat cesarean deliveries: agreement with clinical records. Am J Public Health 1995;85:1143–6.
5. International Classification of Diseases, Ninth Revision, Clinical modification. Hyattsville, MD: National Center for Health Statistics;
6. American Hospital Association Hospital Statistics. Chicago (IL): Health Forum LLC, an affiliate of the American Hospital Association; 1995/6-2001.
7. NIS Technical Documentation. June 2003. Agency for Healthcare Research and Quality, Rockville, MD. Available at: http://www.hcup-us.ahrq.gov/db/nation/nis/nisdbdocumentation.jsp
. Retrieved February 3, 2005.
8. Anderson GM, Lomas J. Determinants of the increasing cesarean delivery birth rate: Ontario data 1979-1982. N Engl J Med 1984;311:887–92.
9. Allen VM, O'Connell CM, Liston RM, Baskett TF. Maternal morbidity associated with cesarean delivery without labor compared with spontaneous onset of labor at term. Obstet Gynecol 2003;102:477–82.
10. Pattison P, McCowan L. Cardiotocography for antepartum fetal assessment (Cochrane Review). In: The Cochrane Library, Issue 1, 2004. Chichester, UK: John Wiley & Sons, Ltd.
11. Alfirevic Z, Neilson JP. Biophysical profile for fetal assessment in high risk pregnancies (Cochrane Review). In: The Cochrane Library, Issue 1, 2004. Chichester, UK: John Wiley & Sons, Ltd.
12. Crowley P. Interventions for preventing or improving the outcome of delivery at or beyond term (Cochrane Review). In: The Cochrane Library, Issue 1, 2004. Chichester, UK: John Wiley & Sons, Ltd.
13. Fretts RC, Elkin EB, Myers ER, Heffner LJ. Should older women have antepartum testing to prevent unexplained stillbirth? Obstet Gynecol 2004;104:56–64.
14. Gould JB, Davey B, Stafford RS. Socioeconomic differences in rates of cesarean delivery. N Engl J Med 1989;321:233–9.
15. Mathews TJ, Hamilton BE. Mean age of mother, 1970-2000. National vital statistics reports; Vol 51, No 1. Hyattsville, MD: National Center for Health Statistics; 2002.
The following is a list of the statewide data organizations that participate in the Healthcare Cost and Utilization Project Nationwide Inpatient Sample in 2001: Arizona Department of Health Services; California Office of Statewide Health Planning & Development; Colorado Health & Hospital Association; Connecticut—CHIME, Inc.; Florida Agency for Health Care Administration; Georgia—GHA: An Association of Hospitals & Health Systems; Hawaii Health Information Corporation; Illinois Health Care Cost Containment Council; Iowa Hospital Association; Kansas Hospital Association; Kentucky Department for Public Health; Maine Health Data Organization; Maryland Health Services Cost Review Commission; Massachusetts Division of Health Care Finance and Policy; Michigan Health and Hospital Association; Minnesota Hospital Association; Missouri Hospital Industry Data Institute; Nebraska Hospital Association; New Jersey Department of Health & Senior Services; New York State Department of Health; North Carolina Department of Health and Human Services; Oregon Association of Hospitals & Health Systems; Pennsylvania Health Care Cost Containment Council; Rhode Island Department of Health; South Carolina State Budget & Control Board; Tennessee Hospital Association; Texas Health Care Information Council; Utah Department of Health; Vermont Association of Hospitals and Health Systems; Virginia Health Information; Washington State Department of Health; West Virginia Health Care Authority; and Wisconsin Department of Health & Family Services