With nearly four million women in the United States giving birth each year, maternal health is an important public health concern for the nation. Although maternal mortality has been the traditional indicator used to monitor maternal health, the decline in the risk of death from pregnancy complications1 has prompted researchers, clinicians, and program and policy makers to call for other measures to monitor women’s health during pregnancy.2 This has led to the expansion of maternal health indicators from those that focus solely on mortality to include others that reflect various aspects of morbidity.3,4
Since 1979, the U.S. Public Health Service has periodically published Healthy People objectives for the nation. The maternal health objective used through Healthy People 2000 was a reduction in the rate of antepartum hospitalization.5 However, decreases in antepartum hospitalizations, likely related to the increased use of care provided in outpatient settings, have made this indicator difficult to interpret. Thus, in Healthy People 2010, the maternal health objective to decrease the antepartum hospitalization rate was changed to an objective to reduce the percentage of intrapartum or delivery hospitalizations with a reported maternal morbidity, due to either an obstetric complication or the presence of a preexisting medical condition that can be adversely affected by pregnancy.6 The goal for this objective is 24.0 deliveries with intrapartum complications per 100 deliveries. The benchmark measure for this new objective was based in part on a 2003 analysis of data from the National Hospital Discharge Survey (NHDS) for 1993–1997.7 No detailed analysis of U.S. intrapartum morbidity rates has been published since that analysis.
MATERIALS AND METHODS
We analyzed data from the NHDS for the period 2001–2005. Because some of the conditions examined were relatively rare, we used the most recent 5-year analysis period of NHDS to compare with the original published 5-year analysis period, 1993–1997. By combining 5 years of data, almost all estimates were statistically reliable (ie, with relative standard error less than 30%). The NHDS is conducted by the Centers for Disease Control and Prevention’s National Center for Health Statistics, and its design is described in detail elsewhere.8,9 Briefly, the National Center for Health Statistics collects inpatient utilization data annually from noninstitutional, nonfederal general hospitals, children’s general hospitals, and short-stay hospitals (ie, those with an average length of stay of fewer than 30 days for all patients) in the United States. The NHDS uses a three-stage sampling design, which selects its sample based on region of the country, hospital, and discharge. The NHDS data are then weighted to produce estimates of the number of national inpatient hospitalizations. During 2001–2005, data were collected from an average of 440 hospitals per year (range 426–448), with information on approximately 345,000 discharges per year.9–13 When appropriate sampling weights are applied, the NHDS represents all hospitalizations in the United States. The NHDS, which contains administrative, not clinical, data, includes up to seven International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) diagnosis codes and up to four ICD-9-CM procedure codes for each record sampled. The NHDS samples hospitalizations and not persons. Our analysis focused on delivery hospitalizations. Because we analyzed de-identified unlinked data that were determined not to involve human participants, our study did not require institutional review board review.
To compare intrapartum morbidity rates during the two time periods, we followed the methodology used in the original analysis of the 1993–1997 data.7 For that analysis, maternal morbidity during labor and delivery was defined as “a condition that adversely affects a woman’s physical health during childbirth beyond what would be expected in a normal delivery.” A V27.0–V27.9 code was used to identify hospitalizations during which a delivery occurred; all deliveries associated with hydatidiform mole, ectopic pregnancy, spontaneous or induced abortion, as indicated by ICD-9-CM codes 630–639, were excluded. For this analysis, ICD-9-CM codes indicating morbidity, either due to obstetric complication or to preexisting conditions that can be adversely affected by pregnancy, were identified from the standard pregnancy chapter (630–676) as well as from other chapters of the ICD-9-CM (see Appendix). Cesarean delivery, considered a major operative procedure, was also included as a type of morbidity in the original and in the current analyses. Conditions that affect the fetus but not the physical health of the mother were excluded, as were codes indicating mental health conditions. Cesarean delivery was not included in the morbidities for the Healthy People 2010 objectives; in addition, their definition of delivery morbidity only used codes from the ICD-9-CM pregnancy chapter (630–676).
We divided the selected ICD-9-CM codes indicating morbidity into clinically meaningful categories (eg, hemorrhage, obstetric trauma, chronic hypertension). Clinical categories were combined into obstetric complications (conditions unique to pregnancy or its management), preexisting medical conditions that can be adversely affected by pregnancy, and cesarean delivery. We then calculated the following rates: individual morbidities within clinical categories, obstetric complications, preexisting medical conditions, cesarean delivery, and overall complication (obstetric complications and preexisting medical conditions), including and excluding cesarean delivery. Rates were calculated as the number of delivery hospitalizations with at least one of the conditions in the particular diagnosis group per 100 delivery hospitalizations (percent). Because some women had more than one complication during their delivery hospitalization, the summary rates (obstetric complications, preexisting medical conditions, and overall morbidity including and excluding cesarean delivery) are less than the sum of the individual components. To account for the complex sampling in the NHDS, we used SUDAAN14 software (Research Triangle Institute, Research Triangle Park, NC) for all analyses. We estimated morbidity rates for the two time periods; we then calculated rate ratios comparing the 2001–2005 and 1993–1997 rates. Standard errors were obtained for all rates; 95% confidence intervals for the rate ratios were then calculated using a 2-tailed z test.
During 2001–2005, the NHDS included data on 183,431 sampled delivery hospitalizations, which represented an estimated 19,986,000 deliveries after applying the weighting factors. Combined with the data for 1993–1997, this analysis reflects the reported morbidity among women for almost 39,067,000 deliveries.
Almost 29% of the delivery hospitalizations during 2001–2005 involved an obstetric complication, and 4.9% had a preexisting medical condition that could be aggravated by pregnancy. Almost one-third of delivery hospitalizations involved at least one morbidity, exclusive of method of delivery. With a cesarean delivery rate of 28.3%, the rate of overall morbidity including cesarean delivery was 48.5% (Table 1).
During the period 2001–2005, the most common obstetric complications, each affecting more than 3% of deliveries, or at least 130,000 delivery hospitalizations per year, were preeclampsia/eclampsia, third- and fourth-degree laceration, other obstetric trauma, and gestational diabetes. More than 2% of delivery hospitalizations involved postpartum hemorrhage. The most common preexisting medical condition, chronic hypertension, affected 1.9% of delivery hospitalizations.
For the two time periods, 1993–1997 and 2001–2005, we compared the rates of reported morbidity at delivery hospitalization for each of the categories of morbidity; we found statistically significant changes, both increases and decreases, in the rates for many of the morbidities (Table 1). Most notably, the percentage of delivery hospitalizations with postpartum hemorrhage, severe preeclampsia, transient hypertension of pregnancy, postpartum fever of unknown origin, gestational diabetes, preexisting diabetes mellitus, and asthma each increased significantly. On the other hand, the percentage of delivery hospitalizations with combined third- or fourth-degree lacerations, fourth-degree laceration alone, genitourinary tract infection, amnionitis, major puerperal infection, and other infection decreased. The cesarean delivery rate increased from 21.8% to 28.3% (and to 31.1% in 200615). Although the percentage of delivery hospitalizations with an obstetric complication remained unchanged (28.6%), the percentage with a preexisting medical condition increased by one fifth—from 4.1% to 4.9% of deliveries. The overall morbidity rate excluding cesarean delivery was not statistically different between the two time periods. The overall morbidity rate including cesarean delivery showed a statistically significant increase from 43.0% in 1993–1997 to 48.5% in 2001–2005.
The rate of overall maternal morbidity excluding cesarean delivery at delivery hospitalization, used as the Healthy People 2010 maternal health indicator, did not change appreciably between the time periods 1993–1997 and 2001–2005. Of the two components of overall morbidity, the rate of obstetric complications was identical for the two time periods, whereas the rate of deliveries to women with preexisting medical conditions that can be adversely affected by pregnancy showed a statistically significant 20% increase. The cesarean delivery rate increased by 30%. Embedded within these overarching summary measures are changes in the rates of many specific complications and categories of complications, including increases in chronic hypertension and preeclampsia, gestational and preexisting diabetes, asthma, and postpartum hemorrhage. Between the two time periods decreases were found in the rates of third- and fourth-degree lacerations and various types of infection, including genitourinary infection.
In the 13-year period covered by this analysis, several changes in the U.S. population and in maternity care occurred that may have had an effect on the rates of various complications. One of the most discussed changes in the health status of Americans over the past 30 years has been an increase in the prevalence of obesity, among pregnant women16 as well as in the general population.17 In addition, the percentage of U.S. births to women at older ages has steadily increased. In 1993, 10.4% of U.S. deliveries were to women aged 35 years and older18; in 2005, this rate was 14.4%.19 Between the two time periods of this analysis, the rates of pregnancy-induced hypertension, chronic hypertension, and both preexisting and gestational diabetes, all more common with increasing body mass index (BMI)20 and with increasing maternal age, have increased among pregnant women. We could not explore the relationship between women’s BMI and their risk for these conditions further because information about weight is only present in the NHDS when obesity is coded as a discharge diagnosis. The likely bias in this coding makes discharge data unreliable for population-based inferences about the association between obesity and morbidity. Changes in delivery practices, specifically a decrease in the percentage of women having an episiotomy21 and an increase in the percentage having cesarean delivery,22 may have contributed to the decrease in the percentage of woman having third- and fourth-degree lacerations.
Interestingly, the increase in the rate of postpartum hemorrhage between 1993–1997 and 2001–2005 is similar to that found in both Canadian23 and Australian24 studies. These findings are reinforced by a U.S. report on severe maternal morbidity, which demonstrated that the frequency of blood transfusions during delivery hospitalizations, an indicator of clinically significant hemorrhage, increased from 3 to 5 per 1,000 deliveries between 1991 and 2003.3 The reasons for this increase in the reported rate of postpartum hemorrhage occurring in many countries is unclear. Possible factors include changes in the underlying risk profiles of women (eg, age, parity, BMI, and previous cesarean deliveries or uterine surgery) and changes in clinical practice, such as increased inductions, use of newer agents for labor induction or increased length of labor.
A major issue in obstetrics for the past two decades has been the rate of cesarean delivery. After attempts to decrease the cesarean delivery rate by promoting vaginal birth after cesarean (VBAC) in the early 1990s, the percentage of cesarean births in the United States has steadily increased since 1996, reaching 30.2 % in 2005.19 In addition to obstetric and medical indications for surgical delivery, cesarean delivery at patient request or convenience is an area of significant debate, leading to a National Institutes of Health State of Science Conference on Cesarean Delivery on Maternal Request in 2006.25 When results of the analysis of 1993–1997 NHDS data were published, cesarean delivery was used in calculating a “total morbidity rate.” Although cesarean delivery remains a major operative procedure, in 2005, when over 30% of deliveries are by cesarean, the appropriateness of including cesarean delivery as a morbidity may be in question. Thus, in this analysis, we have reported “overall morbidity including and excluding cesarean delivery.” However, given the question of the presence or absence of a causal relationship between cesarean delivery and morbidity, how the increase in cesarean delivery rates relates to other pregnancy complications is an area for future research.
The use of intrapartum morbidity rates at delivery hospitalization as an indicator of maternal health has advantages and disadvantages. We relied on a list of previously selected ICD-9-CM codes and their reporting in records of delivery hospitalizations to identify intrapartum morbidity. In the United States, ICD-9-CM codes are still used to classify morbidity, and many of the codes related to pregnancy are broadly defined, assigning a single code to several clinical entities. Thus, placing conditions into groups is necessary. It is possible that over time, coding practices have changed, which has affected the rates in NHDS. However, the direction of the changes in the different complications were not consistent, which suggests the absence of a systematic bias in the coding. Also, the fact that many codes were grouped into coarser aggregates should alleviate some of the lack of precision of the codes as well any changes in their use. Finally, codes that were problematic in 1997–1993 probably remained so in 2000–2005.
Using the methodology developed for the publication based on 1993–1997 data as well this analysis, each year since 1998, Healthy People has published “maternal complications during labor and delivery,” (http://ist-cent-wondr/data2010, focus area 16, objective 5A; retrieved February 26, 2009) a measure consistent with our obstetric complication rate. However, rates for the clinical entities that make up the overall morbidity rate have not been published since the original analysis of the 1993–1997 data. Thus, this analysis allows the first more detailed look at progress toward the Healthy People 2010 maternal morbidity objective. Morbidity can be affected by both individual health status and medical practice, and both types of factors may have affected the rate of maternal morbidity. Although the overall morbidity rate is clearly a heterogeneous and relatively crude measure, its usefulness in monitoring women’s well-being at delivery can be improved by breaking it into rates of clinically meaningful categories, as we did in this study. By doing so, we were able to show that although the overall morbidity rate excluding cesarean delivery was essentially unchanged, the rates of some categories of complications had decreased while others had increased. The increase in the percentage of pregnancies with a preexisting medical condition that can be adversely affected by pregnancy, such as those related to obesity and maternal age, represents an emerging challenge for obstetric care. Such information will be meaningful to clinicians as well as to public health practitioners and policy makers.
As is the case with public health surveillance data, we hope that our findings can serve as a springboard for the use of more nuanced data for in-depth investigations of the effect of patient and health care factors on maternal morbidity. Although the NHDS does not contain data that could answer the questions raised by this overview, such as the reasons for the increase in postpartum hemorrhage rates or the relationship between cesarean delivery and morbidity, research institutions could use their large perinatal databases to study such issues.
1. Centers for Disease Control and Prevention. Achievements in public health, 1900–1999: healthier mothers and babies. MMWR Morb Mortal Wkly Rep 1999;48:849–58.
2. Berg CJ, Bruce FC, Callaghan WM. From mortality to morbidity: the challenge of the twenty-first century. J Am Med Womens Assoc 2002;57:173–4.
3. Callaghan WM, MacKay AP, Berg CJ. Identification of severe morbidity during delivery hospitalizations, United States, 1991–2003. Am J Obstet Gynecol 2008;199:133.e1–8.
4. Geller SE, Rosenberg D, Cox SM, Kilpatrick S. Defining a conceptual framework for near-miss maternal morbidity. J Am Med Womens Assoc 2002;57:135–9.
5. Public Health Service. Healthy people 2000: national health promotion and disease prevention objectives. DHHS publication no. (PHS)90-50212. Washington (DC): U.S. Department of Health and Human Services, Public Health Service; 1990.
6. U.S. Department of Health and Human Services. Healthy people 2010: with understanding and improving health and objectives for improving health. 2nd edition. 2 vols. Washington (DC): U.S. Government Printing Office; 2000.
7. Danel I, Berg C, Johnson CH, Atrash H. Magnitude of maternal morbidity during labor and delivery: United States, 1993–1997. Am J Public Health 2003;93:631–4.
8. Dennison C, Pokras R. Design and operation of the National Hospital Discharge Survey: 1988 redesign. Vital Health Stat 1 2000;39:1–42.
9. DeFrances CJ, Cullen KA, Kozak LJ. National Hospital Discharge Survey: 2005 annual summary with detailed diagnosis and procedure data. Vital Health Stat 13 2007;(165):1–209.
10. Kozak LJ, Owings MF, Hall MJ. National Hospital Discharge Survey: 2001 annual summary with detailed diagnosis and procedure data. Vital Health Stat 13 2004;(156):1–198.
11. Kozak LJ, Owings MF, Hall MJ. National Hospital Discharge Survey: 2002 annual summary with detailed diagnosis and procedure data. Vital Health Stat 13 2005;(158):1–199.
12. Kozak LJ, Lees KA, DeFrances CJ. National Hospital Discharge Survey: 2003 annual summary with detailed diagnosis and procedure data. Vital Health Stat 13 2006;(160):1–206.
13. Kozak LJ, DeFrances CJ, Hall MJ. National Hospital Discharge Survey: 2004 annual summary with detailed diagnosis and procedure data. Vital Health Stat 13 2006;(162):1–209.
14. Shah BV, Barnwell BG, Bieler GS. SUDAAN user’s manual, release 7.5. Research Triangle Park (NC): Research Triangle Institute; 1997.
15. Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2006. Natl Vital Stat Rep 2007;56 (7):4–24.
16. Kim SY, Dietz PM, England L, Morrow B, Callaghan WM. Trends in pre-pregnancy obesity in nine states, 1993-2003. Obesity (Silver Spring) 2007;15:986–93.
17. Ogden CL, Fryar CD, Carroll MD, Flegal KM. Mean body weight, height, and body mass index, United States 1960–2002. Adv Data 2004;347:1–17.
18. National Center for Health Statistics. Vital statistics of the United States, 1993, vol 1, natality. Hyattsville (MD): National Center for Health Statistics; 1999.
19. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Kirmeyer S et al. Births: final data for 2005. Natl Vital Stat Rep 2007;56:1–103.
20. Chu SY, Bachman DJ, Callaghan WM, Whitlock EP, Dietz PM, Berg CJ et al. Association between obesity during pregnancy and increased use of health care. N Engl J Med 2008:358;1444–53.
21. Hartmann K, Viswanathan M, Palmieri R, Gartlehner G, Thorp J, Lohr KN. Outcomes of routine episiotomy: a systematic review. JAMA 2005;293:2141–8.
22. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Kirmeyer S. Births: final data for 2004. Natl Vital Stat Rep 2006:55:1–101.
23. Joseph KS, Rouleau J, Kramer MS, Young DC, Liston RM, Baskett TF, et al. Investigation of an increase in postpartum hemorrhage in Canada. BJOG 2007;114:751–9.
24. Ford JB, Roberts CL, Simpson JM, Vaughan J, Cameron CA. Increased postpartum hemorrhage rates in Australia. Int J Gynaecol Obstet 2007;98:237–43.
25. National Institutes of Health state-of-the-science conference statement: cesarean delivery on maternal request March 27–29, 2006. Obstet Gynecol 2006;107:1386–97.