Objective
The purpose of this study was to evaluate the use of a childbirth composite morbidity indicator for monitoring childbirth morbidity at hospital and regional levels in California.
Study Design
Study data were obtained from the 2005 linked maternal and neonatal discharge dataset for California hospitals. The study population was limited to laboring women with singleton, term (≥37 weeks’ gestation), inborn, and live births. Women with and without pregnancy complications were stratified into high- and low-risk groups. The composite outcome was defined as any significant morbidity of the mother or newborn infant during the childbirth admission. Submeasures for maternal and neonatal composite morbidity and for severe maternal morbidity were examined with both aggregate and hospital-level analyses.
Results
Of 377,869 eligible deliveries, 120,218 (31.8%) were categorized as high risk and 257,651 (68.2%) were categorized as low risk. High-risk women had higher morbidity rates for all comparisons. The mean childbirth composite morbidity rate was 21% overall: 28% for high-risk women and 18% for low-risk women. For high- and low-risk strata, the rates of maternal complications were 18% and 13%, and the rates of severe maternal morbidity were 1.4% and 0.5%, respectively. There was substantial variation across hospitals for all measures.
Conclusion
The childbirth composite morbidity rate is designed to report childbirth complication rates that combine maternal and neonatal morbidity. This measure and its submeasures met the criteria for quality indicator evaluation as specified by the Agency for Healthcare Research and Quality and can be used for benchmarking or for monitoring childbirth outcomes at regional levels.
Maternal morbidity and mortality rates are on the rise nationally, and specifically in California. In particular, rates of severe maternal morbidity (including renal failure, adult respiratory distress syndrome, shock, and ventilation) have been rising, and racial/ethnic disparities in these clinical outcomes have widened. Given that childbirth is the most common reason for hospitalization in the United States, at 4 million births per year, improved efforts are needed to monitor and address this observed increase in childbirth-related morbidity.
To date, endeavors to improve childbirth safety and health care quality have had variable success and have been relegated largely to learning collaboratives that use benchmarking strategies to rank hospital performance. For example, The Joint Commission has encouraged use of its Perinatal Core Measures dataset ; other organizations, such as the Agency for Healthcare Research and Quality (AHRQ), the National Quality Forum and the Leapfrog Group, have developed and/or promoted individual measures for tracking the quality and safety of perinatal services. Although these measures address specific childbirth concerns, they are not comprehensive, and many require medical chart review to calculate and report indicator rates, which makes hospital adoption more challenging and less feasible. “Feasibility” is a key criterion for the establishment of measures for health care quality and patient safety as specified by the AHRQ.
The measurement of childbirth safety and health care quality requires a unique approach because the childbirth experience differs from other types of hospitalizations, which is evidenced by the complete or partial exclusion of pregnant women from more than one-half of the AHRQ Patient Safety Indicators. First, childbirth involves at least 2 patients, and consequently, data for both patients must be linked. Second, although childbirth is generally perceived to be safe, a wide array of complications (eg, hemorrhage, trauma, or infection) may affect 1 or the other (or both), and tradeoffs in care may occur to optimize outcomes. Approximately 1 in 5 low-risk women experience at least 1 maternal or newborn infant complication; this percentage appears to vary widely across hospitals. Third, preventability of these complications is poorly understood. Many childbirth complications are considered unpredictable, and the line between what may and may not be preventable is often indistinct. The inability to judge preventability should not be a deterrent to monitoring this increasing childbirth-associated morbidity because improvements in multiple components of the health care delivery system, and specifically the hospital’s safety culture, have been shown to yield better childbirth outcomes.
All of these factors argue for the use of a composite measure to estimate childbirth outcome morbidity: to combine the results for mother and fetus, to capture the wide range of potential morbidity (including rare events), and to focus attention, not on a single condition, but on the opportunity to recognize and explore systems-related causes of childbirth morbidity. Such a comprehensive measure should add an important dimension to current childbirth services assessments at both regional and hospital levels.
We build on the efforts of multiple investigators and report the further development and application of a composite measure for tracking both state and hospital-level childbirth outcomes that was first introduced by Gregory et al in 2009, the ideal delivery rate. The ideal delivery rate used administrative data to measure the proportion of women who underwent childbirth and experienced no morbidity to themselves or their newborn infant. The ideal delivery rate initially was conceptualized to be meaningful and interpretable from the patient’s perspective. Here, using the California linked perinatal cohort dataset, we aimed to describe a new measure, the childbirth composite morbidity rate (which is the reverse, or complement, of the ideal delivery rate) and to report the potential for its use in monitoring the childbirth morbidity that occurs among high- and low-risk laboring women at both state and hospital levels. This measure potentially could identify areas for improvement from the facility or health system perspective.
Materials and Methods
Data were obtained from the 2005 linked maternal and neonatal discharge dataset for California hospitals. This dataset combines hospital discharge data with vital statistics data and links maternal and newborn infant data for >97% of California births. The derivation of the study population is based on the original work of Gregory et al and includes women with singleton, term (≥37 weeks’ gestation), inborn, live births in California hospitals; however, for purposes of this study, the study population was limited to women in labor for improved interpretability because hospital rates of elective cesarean deliveries can vary greatly. The presence of labor was defined by a previously validated algorithm using International Classification of Diseases -version 9-Clinical Modification (ICD-9-CM) codes.
High- and low-risk groups were represented by patients who were stratified into groups of women with and without pregnancy complications. Table 1 provides a summary of the conditions that were used to assign patients into the high-risk stratum; the Appendix ( Supplementary Tables 1-3 ) shows the applicable ICD-9-CM codes. These conditions have been augmented since the original publication by Gregory et al. Laboring women with a history of a cesarean delivery and no high-risk condition were classified as “low risk,” and analyses were stratified by route of delivery (ie, vaginal with no previous cesarean delivery, primary cesarean delivery, repeat cesarean delivery, or failed vaginal birth after cesarean [VBAC], and VBAC) to ensure interpretability.
| Condition |
|---|
| Antepartum bleeding |
| Asthma and other pulmonary conditions |
| Chromosome abnormality |
| Diabetes mellitus (includes gestational) |
| Drug withdrawal (neonatal chart) |
| Heart disease |
| Hemolytic disease of newborn infant (neonatal chart) |
| Herpes |
| Hydrops because of isoimmunization (neonatal chart) |
| Idiopathic hydrops (neonatal chart) |
| Intrauterine growth restriction |
| Intrauterine growth restriction, small for gestational age (neonatal chart) |
| Isoimmune disease |
| Kidney disorder |
| Laryngeal stenosis (neonatal chart) |
| Liver disorders |
| Macrosomia |
| Malignancy (neonatal chart) |
| Malpresentation |
| Mental illness |
| Newborn infant affected by vasa previa, velamentous insertion (neonatal chart) |
| Newborn infant affected by previa/abruption (neonatal chart) |
| Oligohydramnios |
| Other congenital anomalies (neonatal chart) |
| Other hypertension: chronic or mild gestational hypertensive diseases of pregnancy |
| Phototherapy of newborn infant (neonatal chart) |
| Polyhydramnios |
| Severe hypertension (ie, severe preeclampsia or eclampsia) |
| Soft tissue conditions: disorders of the organs and soft tissue of the pelvis |
| Substance use |
| Thyroid disorder |
| Unengaged head at term |
a See Supplementary Table 1 for International Classification of Diseases -Version 9-Clinical Modification codes.
The outcome was defined as death or morbidity of the mother or newborn infant and is specified in Table 2 . Details regarding coding are included in Supplementary Tables 2 and 3 ; these conditions have been augmented since the original publication by Gregory et al. The childbirth composite morbidity rate was defined as the prevalence of childbirth hospital admissions with this outcome. For purposes of this investigation, we specified and reported submeasures for maternal and neonatal composite morbidity and for severe maternal morbidity. Maternal morbidity classified as severe was determined by consensus of the investigators after thorough review of the literature regarding the continuum of obstetrical morbidity and was defined as a life-threatening condition or lifesaving procedure at the time of labor and delivery. The definition of neonatal morbidity that was used by Gregory et al was augmented based on the study team’s collaboration in the development of the “healthy term newborn” measure that was approved by the National Quality Forum. A revised ideal delivery rate, the complement, or reverse of the childbirth composite morbidity rate is also reported. The revised ideal delivery rate reported here is modified from the original ideal delivery rate of Gregory et al with respect to the following characteristics: (1) the conditions that were used in categorizing patients with high- and low-risk strata and the conditions that signify morbidity of the mother or newborn infant have been augmented; the women who had a history of cesarean delivery were included in the low-risk stratum and examined separately; and (2) submeasures for severe maternal morbidity and neonatal morbidity have been included. The Figure provides an illustration of the stratification and derivation of the study population for the calculation of the childbirth composite morbidity rate and its submeasures.
| Maternal complications a |
| 3rd-/4th-degree laceration, repair |
| Acute psychiatric condition |
| Anesthesia complications (ie, pulmonary, b cardiac, b central nervous system b ) |
| Arterial embolization/ligation b |
| Cardiac arrest, b conversion of cardiac rhythm b |
| Cerebral edema |
| Cerebral vascular accident b |
| Coma b |
| Disseminated intravascular coagulation b |
| Hemorrhage |
| High vaginal laceration |
| Hysterectomy b |
| Length of stay >5 days or transfer to another acute hospital |
| Liver failure b |
| Maternal death b |
| Maternal distress (metabolic disturbance) |
| Maternal intensive care unit diagnostic codes b |
| Maternal intensive care unit procedures (eg, invasive hemodynamic monitoring b ) |
| Maternal infection septicemia b |
| Obstetric hematoma |
| Obstetric shock, b postoperative shock b |
| Other lacerations |
| Pneumonia |
| Postpartum deep venous thrombosis |
| Pulmonary embolism b |
| Renal failure b |
| Respiratory failure, b pulmonary insufficiency, b mechanical ventilation b |
| Surgical complications (eg, laceration/repair bladder, bowel) |
| Transfusion: whole blood, b packed red blood cells b |
| Uterine dehiscence |
| Uterine inversion |
| Uterine rupture |
| Wound complication |
| Neonatal complications c |
| Birth trauma/injuries |
| Disseminated intravascular coagulation |
| Hypoxia/asphyxia |
| Intensive care unit procedures (eg, arterial or umbilical vein catheterization) |
| Infection |
| Necrotizing enterocolitis |
| Neonatal death b |
| Neonatal length of stay >5 d |
| Neonatal transfer to another acute facility |
| Neurological complications |
| Neurological procedures including electroencephalogram, imaging |
| Renal failure |
| Respiratory conditions |
| Respiratory procedures (eg, mechanical ventilation, respiratory therapy, chest tube) |
| Shock, resuscitations |
| Total peripheral nutrition, gavage feeding, gastrotomy |
a See Appendix Supplementary Table 2 for maternal International Classification of Diseases -Version 9-Clinical Modification codes
b Items are included in the maternal Severe Composite Morbidity Indicator submeasure; severe maternal morbidity is defined as life-threatening diagnoses or lifesaving procedures (includes death)
c See Appendix Supplementary Table 3 for neonatal International Classification of Diseases -Version 9-Clinical Modification codes.
Unadjusted analyses were performed at the aggregate level. Within high- and low-risk groups, patients were stratified by parity (nulliparous vs multiparous) and route of delivery. Hospital-level analyses were also performed and were limited to those hospitals with >200 annual deliveries to ensure that hospital-level data were comparable, to stabilize estimates and error in estimation, and to ensure that hospital-level data were not uniquely identifiable when reported. Direct standardization for race/ethnicity and maternal age based on all denominator deliveries (standard population) was used to adjust the hospital-specific rates. Within the low-risk stratum, subanalyses were performed by parity and route of delivery.
The distribution of the hospital-level childbirth composite morbidity rates was divided into the top quartile, the 2 middle quartiles, and the bottom quartile. Hospitals with a childbirth composite morbidity rate (and a 95% CI around that rate) under the 25th percentile for their rate distribution were considered “high performers” because of their “low” childbirth composite morbidity rate. Similarly, hospitals with a childbirth composite morbidity rate (and a 95% CI) >75th percentile were considered “under-performers” because of their relatively “high” childbirth composite morbidity rate. All other hospitals were defined as having a “mid-range” rate.
The study was approved by all relevant institutional review boards for the Protection of Human Subjects.
Results
Deliveries at 269 hospitals with a total of 377,869 births were evaluated. Of 377,869 eligible gestations, 120,218 (31.8%) were categorized as high risk, and 257,651 (68.2%) were categorized as low risk. High-risk women had higher morbidity rates than low-risk women for all comparisons ( Table 3 ). The mean childbirth composite morbidity rate for all deliveries was 21%: 28% for high-risk women and 18% for low-risk women.
| Composite measure | Risk stratum, n | All (n = 377,869) c | |
|---|---|---|---|
| Low (n = 257,651) | High (n = 120,218) | ||
| Ideal delivery rate b | 212,028 (82.29%) | 87,038 (72.40%) | 299,066 (79.15%) |
| Childbirth composite morbidity rate | 45,623 (17.71%) | 33,180 (27.60%) | 78,803 (20.85%) |
| Maternal complications | 34,583 (13.42%) | 21,637 (18.00%) | 56,220 (14.88%) |
| Severe maternal complications | 1346 (0.52%) | 1713 (1.42%) | 3059 (0.81%) |
| Neonatal complications | 14,277 (5.54%) | 15,791 (13.14%) | 30,068 (7.96%) |
| Risk stratum: low | All (n = 257,612) | ||
| Multiparous (n = 149,345) | Nulliparous (n = 108,267) | ||
| Ideal delivery rate b | 131,334 (87.94%) | 80,662 (74.50%) | 211,996 (82.29%) |
| Childbirth composite morbidity rate | 18,011 (12.06%) | 27,605 (25.50%) | 45,616 (17.71%) |
| Maternal complications | 12,107 (8.11%) | 22,472 (20.76%) | 34,579 (13.42%) |
| Severe maternal complications | 631 (0.42%) | 715 (0.66%) | 1346 (0.52%) |
| Neonatal complications | 6793 (4.55%) | 7481 (6.91%) | 14,274 (5.54%) |
| Risk stratum: high | All (n = 120,192) | ||
| Multiparous (n = 61,691) | Nulliparous (n = 58,501) | ||
| Ideal delivery rate b | 48,379 (78.42%) | 38,643 (66.06%) | 87,022 (72.40%) |
| Childbirth composite morbidity rate | 13,312 (21.58%) | 19,858 (33.94%) | 33,170 (27.60%) |
| Maternal complications | 7232 (11.72%) | 14,401 (24.62%) | 21,633 (18.00%) |
| Severe maternal complications | 825 (1.34%) | 888 (1.52%) | 1713 (1.43%) |
| Neonatal complications | 7479 (12.12%) | 8305 (14.20%) | 15,784 (13.13%) |
a Rates are unadjusted, aggregate outcome rates (across all hospitals) for laboring women with singleton, term, inborn, or liveborn newborn infants delivered in California hospitals in 2005
b The ideal delivery rate and the childbirth composite morbidity rate are mathematic complements
c The reduced number of deliveries for the summaries by parity are due to parity missing values.
Maternal complications contributed the most to the childbirth composite morbidity rate. For high and low-risk women respectively, the aggregate rates of maternal complications were 18% and 13%, and the rates of neonatal complications were 13% and 6%. Severe maternal morbidity rates were 1.4% vs 0.5% among high- and low-risk women, respectively. There were 10 deaths overall (0.003%); 9 of these deaths were among women in the high-risk group. For every death, there were 308 women with severe maternal morbidity; this ratio was 1:1261 for the low-risk group and 1:202 for the high-risk group.
Within both high- and low-risk groups, nulliparous women had higher complication rates compared with multiparous women ( Table 3 ). Regarding route of delivery, women who underwent cesarean delivery during labor had more complications compared with women who underwent vaginal birth, regardless of a history of cesarean delivery ( Table 4 ).
| Composite measure | Risk stratum: low, n | All (n = 257,651) | |||
|---|---|---|---|---|---|
| Failed vaginal birth after cesarean delivery (n = 2930) | Cesarean delivery during labor (n = 23,553) | Vaginal delivery: no previous cesarean delivery (n = 226,838) | Vaginal birth after cesarean delivery (n = 4330) | ||
| Ideal delivery rate b | 2255 (76.96%) | 17,492 (74.27%) | 188,761 (83.21%) | 3520 (81.29%) | 212,028 (82.29%) |
| Childbirth composite morbidity rate | 675 (23.04%) | 6,061 (25.73%) | 38,077 (16.79%) | 810 (18.71%) | 45,623 (17.71%) |
| Maternal complications | 381 (13.00%) | 4412 (18.73%) | 29,152 (12.85%) | 638 (14.73%) | 34,583 (13.42%) |
| Severe maternal complications | 96 (3.28%) | 448 (1.90%) | 787 (0.35%) | 15 (0.35%) | 1346 (0.52%) |
| Neonatal complications | 368 (12.56%) | 2462 (10.45%) | 11,225 (4.95%) | 222 (5.13%) | 14,277 (5.54%) |
| Risk stratum: high | All (n = 120,218) | ||||
| Failed vaginal birth after cesarean delivery (n = 3154) | Cesarean delivery during labor (n = 29,545) | Vaginal delivery: no previous cesarean delivery (n = 85,734) | Vaginal birth after cesarean delivery (n = 1785) | ||
| Ideal delivery rate b | 1977 (62.68%) | 19,750 (66.85%) | 64,023 (74.68%) | 1288 (72.16%) | 87,038 (72.40%) |
| Childbirth composite morbidity rate | 1177 (37.32%) | 9795 (33.15%) | 21,711 (25.32%) | 497 (27.84%) | 33,180 (27.60%) |
| Maternal complications | 614 (19.47%) | 6360 (21.53%) | 14,336 (16.72%) | 327 (18.32%) | 21,637 (18.00%) |
| Severe maternal complications | 169 (5.36%) | 868 (2.94%) | 655 (0.76%) | 21 (1.18%) | 1713 (1.42%) |
| Neonatal complications | 770 (24.41%) | 5188 (17.56%) | 9597 (11.19%) | 236 (13.22%) | 15,791 (13.14%) |
a Rates are unadjusted, aggregate outcome rates (across all hospitals) for laboring women with singleton, term
b The ideal delivery rate and the childbirth composite morbidity rate are mathematic complements.
Hospital-specific rates were calculated for 252 hospitals with >200 annual deliveries and a total of 375,303 deliveries. Adjusted hospital rates are shown in Table 5 ; findings were similar to those noted for aggregate rates. Table 5 also shows the cutoffs and numbers for outlier hospitals. For example, among low-risk women and with adjustment for age and race/ethnicity, the median hospital childbirth composite morbidity rate was 17% (range, 3–58%). When stratified into quartiles, 31 hospitals were high performers (ie, had few patients with complications), with a mean rate below the 25th percentile (12%); 33 hospitals were under-performers (ie, had many patients with complications), with a rate >75th percentile (21%). For all comparisons, high-risk women had a higher childbirth composite morbidity rate compared with low-risk women. The median severe maternal complication rate was 1.16% and 0.42% for high- and low-risk women, respectively. The severe maternal complication rate varied by hospital and risk category; the maximum for any hospital was 19% for high-risk women and 5% for low-risk women.
| Composite measure | Low- vs high-risk strata, % | |||
|---|---|---|---|---|
| Low risk | Quartiles and no. of low-risk outlier hospitals | High risk | Quartiles and no. of high-risk outlier hospitals | |
| Ideal delivery rate b | 82.40 ± 8.40 83.47 (41.85–97.35) | Q1 = 78.64 (Low = 33) Q3 = 87.87 (High = 31) | 74.36 ± 9.47 75.81 (28.94–93.60) | Q1 = 69.93 (Low = 36) Q3 = 80.19 (High = 19) |
| Childbirth composite morbidity rate | 17.60 ± 8.40 16.53 (2.65–58.15) | Q1 = 12.13 (Low = 31) Q3 = 21.36 (High = 33) | 25.64 ± 9.47 24.19 (6.40–71.06) | Q1 = 19.81 (Low = 19) Q3 = 30.07 (High = 36) |
| Maternal complications | 13.17 ± 6.41 12.44 (1.39–46.71) | Q1= 8.90 (Low = 36) Q3 = 17.19 (High = 26) | 16.35 ± 6.81 15.40 (1.85–36.08) | Q1 = 11.38 (Low = 20) Q3 = 20.52 (High = 28) |
| Severe maternal complications | 0.55 ± 0.54 0.42 (0.00–4.90) | Q1 = 0.19 (Low = 27) Q3 = 0.70 (High = 3) | 1.50 ± 1.73 1.16 (0.00–19.38) | Q1 = 0.54 (Low = 24) Q3 = 1.90 (High = 7) |
| Neonatal complications | 5.83 ± 6.04 4.45 (0.56–47.86) | Q1 = 2.73 (Low = 23) Q3 = 6.29 (High = 33) | 12.42 ± 7.82 10.82 (0.00–60.69) | Q1 = 7.58 (Low = 16) Q3 = 14.93 (High = 32) |
| Low-risk stratum stratified by nulliparous vs multiparous groups, % | ||||
| Low-risk nulliparous | Quartiles and no. of low-risk nulliparous outlier hospitals | Low-risk multiparous | Quartiles and no. of low-risk multiparous outlier hospitals | |
| Ideal delivery rate b | 74.46 ± 10.72 75.51 (21.07–97.74) | Q1 = 69.29 (Low = 35) Q3 = 81.72 (High = 28) | 87.72 ± 7.14 89.48 (53.90–97.99) | Q1 = 85.04 (Low = 29) Q3 = 92.35 (High = 22) |
| Childbirth composite morbidity rate | 25.54 ± 10.72 24.49 (2.26–78.93) | Q1 = 18.28 (Low = 28) Q3 = 30.71 (High = 35) | 12.28 ± 7.14 10.52 (2.01–46.10) | Q1 = 7.65 (Low = 22) Q3 = 14.96 (High = 29) |
| Maternal complications | 20.55 ± 9.05 20.52 (1.71–44.49) | Q1= 14.12 (Low = 35) Q3 = 26.40 (High = 26) | 8.14 ± 4.59 7.52 (0.73–33.15) | Q1 = 5.30 (Low = 33) Q3 = 10.40 (High = 25) |
| Severe maternal complications | 0.68 ± 0.80 0.43 (0.00–4.29) | Q1 = 0.14 (Low = 54) Q3 = 0.90 (High = 4) | 0.44 ± 0.53 0.30 (0.00–4.10) | Q1 = 0.07 (Low = 56) Q3 = 0.63 (High = 1) |
| Neonatal complications | 7.47 ± 7.65 5.40 (0.00–67.50) | Q1 = 3.46 (Low = 22) Q3 = 8.80 (High = 31) | 4.82 ± 5.46 3.39 (0.00–44.21) | Q1 = 2.06 (Low = 27) Q3 = 5.57 (High = 26) |
| Low-risk stratum stratified by route of delivery, % | ||||
| Failed vaginal birth after cesarean delivery (n = 241) | Emergent cesarean delivery | Vaginal | Vaginal birth after cesarean delivery (n = 214) | |
| Ideal delivery rate b | 78.37 ± 24.49 86.21 (0.00–100.00) | 73.67 ± 15.47 76.19 (24.22–100.00) | 83.33 ± 7.94 84.71 (51.39–97.25) | 81.87 ± 19.99 85.56 (0.00–100.00) |
| Childbirth composite morbidity rate | 21.63 ± 24.49 13.79 (0.00–100.00) | 26.33 ± 15.47 23.81 (0.00–75.78) | 16.67 ± 7.94 15.29 (2.75–48.61) | 18.13 ± 19.99 14.44 (0.00–100.00) |
| Maternal complications | 12.57 ± 21.00 2.29 (0.00–100.00) | 19.09 ± 13.95 16.43 (0.00–71.94) | 12.61 ± 6.11 12.04 (1.65–41.75) | 13.27 ± 16.53 8.65 (0.00–100.00) |
| Severe maternal complications | 3.07 ± 9.91 0.00(0.00–100.00) | 2.12 ± 3.45 0.83 (0.00–21.58) | 0.36 ± 0.47 0.23 (0.00–4.87) | 0.43 ± 2.67 0.00 (0.00–29.70) |
| Neonatal complications | 11.64 ± 18.82 4.97 (0.00–100.00) | 10.97 ± 10.65 7.55 (0.00–58.77) | 5.19 ± 5.51 3.87 (0.31–44.28) | 6.07 ± 14.02 0.00 (0.00–100.00) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
