Materials and Methods
We used 2000-2011 data from the Nationwide Inpatient Sample, which is the largest all-payer hospital inpatient care database in the United States and is maintained by the Agency for Healthcare Research and Quality as part of its Healthcare Cost and Utilization Project. The Nationwide Inpatient Sample is a stratified sample of approximately 20% of all community hospitals in the United States and includes all discharges from an average of 1000 hospitals each year. Thus, when weights are applied during analysis to account for the complex survey design, nationwide estimates can be derived. Each discharge record has the potential to list up to 15 International Classification of Disease , 9th Revision, Clinical Modification (ICD-9-CM) diagnoses and 15 ICD-9-CM procedures as well as diagnosis-related group codes. Because this analysis used publicly available data lacking personal identifiers, the Centers for Disease Control and Prevention’s Institutional Review Board determined it to be research that does not involve human subjects.
We identified all delivery hospitalization records using the algorithm developed by Kuklina et al. This algorithm uses ICD-9-CM diagnosis codes for delivery outcomes, ICD-9-CM procedure codes for selected delivery-related procedures, and diagnosis-related group delivery codes. Cesarean deliveries were identified by the presence of an ICD-9-CM procedure code for cesarean delivery (74.0, 74.1, 74.2, 74.4, 74.99). Among cesarean deliveries, those with an ICD-9-CM diagnosis code 654.2 were classified as repeat cesarean deliveries and those without such code were considered primary cesarean deliveries. We used the ICD-9-CM diagnosis or procedure codes shown in Table 1 to identify records with the following 11 complications or conditions that potentially are associated with cesarean delivery: (1) urinary bladder operations, (2) cystotomy, (3) anesthesia complications, (4) obstetric wound complications, (5) blood transfusion, (6) ventilation support, (7) renal failure, (8) sepsis, (9) shock, (10) prolonged hospital stay defined as stay longer than five days (ie, 90th percentile of the length of hospital stay distributions for all cesarean deliveries in the sample), and (11) death during the delivery hospitalization.
| Condition | Diagnosis or procedure codes |
|---|---|
| Cesarean delivery complications | |
| Urinary bladder operations | 57.x |
| Cystotomy | 57.1 |
| Anesthesia complications | 668.0x, 668.1x, 668.2x |
| Blood transfusion | 99.0x |
| Ventilation support | 93.90, 96.01-96.05, 96.7x |
| Renal failure | 669.3, 584.x |
| Sepsis | 038.0-038.4, 038.8, 038.9, 995.91, 995.92 |
| Shock | 669.1, 998.0, 995.0, 995.4, 785.5x |
| Obstetric wound complications | 674.10, 674.12, 674.14, 674.30, 674.32, 674.34 |
| Chronic conditions | |
| Preexisting diabetes mellitus | 249.x, 250.x, 648.0x |
| Chronic hypertensive disease | 401.x-405.x, 642.0x, 642.1x, 642.2x, 642.7x |
| Chronic heart disease | 412.x, 413.x, 414.x, 394.x, 395.x, 396.x, 397.x, 424.x, 428.22, 428.23, 428.32, 428.33, 428.42, 428.43 |
| Chronic respiratory disease | 491.x-496.x |
| Chronic renal disease | 581.x-583.x, 585.x, 587.x, 588.x, 646.2x |
| Chronic liver disease | 571.x, 572.x |
| HIV/AIDS | 042.x, V08.x |
In addition, of special interest for this analysis was the identification of records from patients with placenta accreta, since this condition poses serious maternal risks in pregnancies after a cesarean delivery and contemporary national-level data on this condition are lacking. Because there is no specific ICD-9-CM code for placenta accreta, we developed and validated an ICD-9-CM code–based algorithm to identify diagnoses and procedures that, when concurrently coded and in the absence of uterine rupture (ICD-9-CM codes: 665.0x, 665.10, 665.11), identify placenta accreta with a high positive predictive value among women with a previous cesarean delivery. This algorithm identified patients with a diagnosis of placenta previa (ICD-9-CM codes: 641.0, 641.1) or retained placenta (ICD-9-CM codes: 666.0, 667.x), and a hysterectomy procedure code (ICD-9-CM codes: 68.3, 68.4, 68.5, 68.6, 68.7, 68.8, 68.9). Specifically for this analysis, this algorithm was validated at the Massachusetts General Hospital and the Stanford University School of Medicine by a comparison of recorded ICD-9-CM codes in administrative hospital data against medical record data for the periods between Jan. 1, 2001, to Dec. 31, 2011, and Jan. 1, 2000, and Dec. 31, 2013, respectively. Validation exercises were approved by Institutional Review Boards at Partner’s Healthcare and Stanford University, respectively. The medical record review was conducted by 2 obstetrics anesthesia residents (A.M. and L.R.), 1 at each validation site. Of the 15 probable accreta cases at Massachusetts General Hospital and 54 probable accreta cases at Stanford University that were identified based on our ICD-9-CM code algorithm, 14 and 47 cases, respectively, were confirmed by 1 of the following items: (1) radiologic evidence, (2) a presumptive clinical diagnosis made surgically during cesarean delivery or laparotomy after vaginal delivery, (3) pathology report, or (4) a description in the patient discharge summary that was written by the obstetrician. The 2-site positive predictive value of our ICD-9-CM–based accreta algorithm in women with a previous cesarean delivery was 88.4% (95% confidence interval [CI], 78.4–94.6). For consistency among all analyses that were performed, we used the same algorithm to identify accreta cases among women with a primary cesarean delivery.
We first examined trends in primary and repeat cesarean delivery between 2000 and 2011. Next, we used pooled 2000-2011 data to estimate rates for each of the 11 complications or conditions noted earlier, placenta accreta that was identified with our validated ICD-9-CM–based algorithm, and a composite cesarean delivery morbidity measure that included at least 1 of these 12 conditions by type of cesarean delivery: primary vs repeat. Also by type of cesarean delivery, we examined rate differences and percentage rate changes between 2000 and 2011 for each of the 12 conditions and for the composite cesarean delivery morbidity measure. On the pooled dataset, we fitted year-adjusted Poisson regression models to assess the relative incidence of each of the 12 conditions and of the composite cesarean delivery morbidity measure among repeat vs primary cesarean deliveries; incidence rate ratios (IRRs) and corresponding 95% CIs were estimated.
Characteristics of women without cesarean delivery complications, those with placenta accreta, and those with at least 1 of the 12 conditions examined in each of the 2 groups (primary vs repeat cesarean deliveries) were examined with chi square tests; similarly, we assessed characteristics of the hospitals where cesarean deliveries were obtained during the study period. The maternal characteristics examined were age (15-19, 20-24, 25-29, 30-34, ≥35 years); race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, Asian, American Indian or Alaska native, other); insurance coverage for delivery (Medicaid, private, self-pay, other); household income quartile for patient by ZIP code; and presence of at least 1 of the following chronic medical conditions: preexisting diabetes mellitus, chronic hypertensive disease, chronic heart disease, chronic respiratory disease, chronic renal disease, chronic liver disease, and HIV/AIDS (the corresponding ICD-9-CM codes are shown in Table 1 ). The hospital characteristics examined were location (urban/rural), teaching status (yes/no), ownership (public, private for-profit, private not-for-profit), bed size (small, medium, large), and total hospital charges as submitted by each hospital and reported in 2011 US dollars after adjustment for inflation with the use of the Consumer Price Indices that are released annually by the US Bureau of Labor Statistics. In separate models, we investigated factors that are associated with placenta accreta and composite cesarean delivery morbidity (defined by the presence of at least 1 of the 12 complications) by fitting multivariable Poisson regression models that are adjusted for all the patient and hospital characteristics mentioned earlier, except for hospital charges. All analyses were conducted in Stata software (version 13; StataCorp, College Station, TX) and were adjusted for the complex survey design of the Nationwide Inpatient Sample.
Results
The percentage of all hospital deliveries in the United States that were cesarean deliveries increased significantly from 22.8% in 2000 to 33.3% in 2011, with a peak at 33.6% in 2009 ( P < .05; Figure 1 ). The increase was steeper for repeat (8.7% in 2000 vs 15.3% in 2011; P < .05) than for primary (14.2% in 2000 vs 18.0% in 2011; P < .05) cesarean deliveries. Moreover, the increase in repeat cesarean deliveries was monotonic from 2000-2011; the percentage of all deliveries that were primary cesarean deliveries peaked at 18.6% in 2009.
From 2000-2011, approximately 76 in 1000 cesarean deliveries (97 in 1000 primary cesarean deliveries and 48 in 1000 repeat cesarean deliveries) were accompanied by at least 1 of the 12 conditions of interest; of these, approximately three-quarters of primary cesarean deliveries and approximately one-half of repeat cesarean deliveries were associated with a hospital stay of >5 days ( Table 2 ). The prevalence of 5 specific medical complications (renal failure, need for ventilation support, urinary bladder operations, obstetric wound complications, and need for blood transfusion) was greater than 1 per 1000 primary cesarean deliveries; notably, approximately 18 in every 1000 women with a primary cesarean delivery received a blood transfusion. Among repeat cesarean deliveries, placenta accreta, urinary bladder operations, obstetric wound complications, and blood transfusion had a prevalence of >1 per 1000 such deliveries; placenta accreta was associated with approximately 1.8 of 1000 repeat cesarean deliveries, and blood transfusion was part of clinical treatment for 13 of 1000 women with a repeat cesarean delivery during the study period.
| Type of morbidity | Pooled 2000-2011 cases, n | Pooled 2000-2011 rate (standard error) a | 2011 vs 2000 rate difference, (95% confidence interval) b | Adjusted-incidence rate ratio (95% confidence interval) c | |||
|---|---|---|---|---|---|---|---|
| Primary cesarean deliveries (n = 8,560,005) | Repeat cesarean deliveries (n = 6,317,911) | Primary cesarean deliveries | Repeat cesarean deliveries | Primary cesarean deliveries | Repeat cesarean deliveries | ||
| Cystotomy | 292 | 729 | 0.17 (0.007) | 0.56 (0.02) | –0.04 (–0.12 to 0.05) | –0.24 (–0.47 to –0.01) d | 3.40 (3.20–3.62) |
| Death | 448 | 93 | 0.25 (0.01) | 0.07 (0.007) | –0.07 (–0.20 to 0.05) | –0.06 (–0.14 to 0.03) | 0.29 (0.26–0.32) |
| Anesthesia complication | 963 | 382 | 0.55 (0.02) | 0.29 (0.01) | –0.54 (–0.75 to –0.35) d | –0.51 (–0.71 to –0.31) d | 0.55 (0.52–0.58) |
| Placenta accreta | 1178 | 2281 | 0.67 (0.02) | 1.76 (0.04) | –0.10 (–0.31 to 0.11) | 0.45 (0.07–0.84) d | 2.61 (2.53–2.70) |
| Shock | 1228 | 559 | 0.70 (0.02) | 0.43 (0.02) | 0.74 (0.53–0.95) d | 0.37 (0.19–0.55) d | 0.60 (0.58–0.63) |
| Sepsis | 1634 | 267 | 0.93 (0.02) | 0.21 (0.01) | –0.07 (–0.29 to 0.14) | –0.17 (–0.30 to –0.03) d | 0.22 (0.21–0.24) |
| Renal failure | 2451 | 496 | 1.39 (0.03) | 0.38 (0.02) | 1.07 (0.79–1.35) d | 0.37 (0.19–0.54) d | 0.27 (0.26–0.28) |
| Ventilation support | 3627 | 1022 | 2.06 (0.03) | 0.79 (0.02) | 0.27 (–0.08 to 0.61) | 0.25 (–0.02 to 0.51) | 0.38 (0.37–0.39) |
| Urinary bladder operation | 5671 | 5014 | 3.22 (0.04) | 3.86 (0.05) | –0.12 (–0.57 to 0.33) | –0.93 (–1.57 to –0.30) d | 1.19 (1.17–1.21) |
| Obstetric wound complication | 17,126 | 14,622 | 9.73 (0.07) | 11.26 (0.09) | –0.73 (–1.47 to –0.02) d | –4.01 (–5.01 to –3.00) d | 1.17 (1.15–1.18) |
| Blood transfusion | 32,048 | 16,403 | 18.22 (0.10) | 12.63 (0.10) | 16.42 (15.44–17.44) d | 12.04 (11.05–13.03) d | 0.68 (0.67–0.68) |
| Prolonged hospital stay | 129,207 | 31,615 | 73.44 (0.20) | 24.33 (0.14) | –8.54 (–10.57 to –6.51) d | –7.09 (–8.55 to –5.62) d | 0.33 (0.33–0.34) |
| ≥1 complication listed | 171,023 | 62,688 | 97.21 (0.22) | 48.25 (0.19) | 3.99 (1.69–6.28) d | –1.63 (–3.66 to 0.39) | 0.50 (0.49–0.50) |
a Rates reported per 1000 deliveries
b Rate differences reported per 1000 deliveries
c Poisson regression models adjusted for type of cesarean delivery (primary cesarean delivery group used as reference) and year
The rate of the composite cesarean delivery morbidity increased by 3.6% ( P < .001; Figures 2 and 3 ) among women with a primary cesarean delivery but did not change significantly among those with repeat cesarean delivery from 2000-2011. The rate of placenta accreta increased significantly only among women with a repeat cesarean delivery, from 1.6 to 2.0 per 1000 repeat cesarean deliveries, a 30.8% increase during the same time period ( P = .025; Figures 2 and 3 ). Several significant complication rate changes stand out among women with both primary and repeat cesarean deliveries ( Figure 2 ). Between 2000 and 2011, rates of anesthesia complications, obstetric wound complications, and especially prolonged hospital stay decreased among women with both primary and repeat cesarean deliveries; rates of sepsis, cystotomy, and urinary bladder operations declined significantly only among women with a repeat cesarean delivery (all P < .05). Conversely, during the same period, the rates of shock, renal failure, and blood transfusion increased by 190.0%, 110.9%, and 154.3%, respectively, among women with primary cesarean deliveries and by 150.0%, 168.2%, and 164.6%, respectively, among women with repeat cesarean deliveries (all P < .05).
After controlling for year-to-year variations, we found that, when compared with women having a first cesarean delivery, those who obtained a repeat cesarean delivery were approximately one-half as likely (IRR, 0.50; 95% CI, 0.49–0.50) to have experienced at least 1 of the conditions of interest, but more likely to have had a cystotomy (IRR, 3.40; 95% CI, 3.20–3.62), placenta accreta (IRR, 2.61; 95% CI, 2.53–2.70), a urinary bladder operation (IRR, 1.19; 95% CI, 1.17–1.21), or wound complications (IRR, 1.17; 95% CI, 1.15–1.18); all other specific complications were significantly less prevalent in women with repeat rather than primary cesarean deliveries ( Table 2 ).
Overall, women with at least 1 cesarean delivery complication, and especially those with a repeat cesarean delivery, were older, in higher proportion black, insured by Medicaid, and with a chronic medical condition than were women without a cesarean delivery complication ( Table 3 ); they also had ≥2 additional diagnoses at discharge and more of their cesarean deliveries took place in urban, teaching, public facilities than in other types of facilities when compared with women without a cesarean delivery complication. Notably, median total hospital charges were overall lower for women with repeat ($10,866.50) than for women with primary ($12,547.70) cesarean deliveries who had no complications. For women with at least 1 cesarean delivery complication, median total hospital charges were more than double ($23,672.30 if primary and $19,033.00 if repeat cesarean delivery). Moreover, for women with placenta accreta, median total hospital charges were $34,035.50 if primary cesarean delivery and $33,944.1 if repeat cesarean delivery.
| Characteristic | Primary cesarean deliveries | Repeat cesarean deliveries | ||||
|---|---|---|---|---|---|---|
| No complication (n = 7,722,632) | ≥1 complication (n = 837,373) | Accreta (n = 5734) | No complication (n = 5,996,357) | ≥1 complication (n = 321,554) | Accreta (n = 11,092) | |
| Maternal characteristics, % | ||||||
| Age, y | ||||||
| 15-19 | 10.6 | 10.9 | 0.7 | 2.3 | 2.3 | 0.6 |
| 20-24 | 22.9 | 21.6 | 6.5 | 17.8 | 16.0 | 7.0 |
| 25-29 | 26.3 | 23.8 | 17.0 | 26.9 | 24.6 | 17.3 |
| 30-34 | 23.8 | 23.4 | 28.6 | 30.2 | 29.2 | 32.9 |
| >35 | 16.4 | 20.2 | 47.1 | 22.8 | 28.0 | 42.2 |
| Race/ethnicity | ||||||
| Non-Hispanic white | 42.6 | 37.9 | 39.4 | 40.7 | 35.4 | 35.7 |
| Non-Hispanic black | 10.8 | 16.8 | 12.3 | 10.6 | 17.7 | 12.4 |
| Hispanic | 16.6 | 15.6 | 19.0 | 21.0 | 19.6 | 27.3 |
| Asian | 4.0 | 3.6 | 6.0 | 3.3 | 3.0 | 3.4 |
| American Indian or Alaska Native | 0.5 | 0.7 | 0.5 | 0.6 | 0.7 | 0.6 |
| Other | 3.8 | 4.0 | 3.3 | 3.6 | 3.7 | 3.6 |
| Missing | 21.7 | 21.6 | 19.6 | 20.3 | 19.9 | 17.1 |
| Insurance coverage | ||||||
| Medicaid | 36.3 | 40.2 | 33.2 | 40.0 | 45.1 | 43.0 |
| Private | 57.7 | 53.3 | 60.3 | 53.5 | 47.4 | 50.4 |
| Self-pay | 2.6 | 2.6 | 3.5 | 3.1 | 3.3 | 2.7 |
| Other | 3.3 | 3.7 | 3.0 | 3.3 | 3.9 | 3.8 |
| Missing | 0.2 | 0.2 | 0.0 | 0.2 | 0.2 | 0.2 |
| Income quartile for patient ZIP code | ||||||
| 1st (lowest) | 24.9 | 27.1 | 23.7 | 26.7 | 30.2 | 27.2 |
| 2nd | 24.1 | 23.7 | 22.3 | 24.2 | 23.9 | 22.5 |
| 3rd | 24.5 | 23.4 | 25.0 | 23.3 | 22.0 | 22.4 |
| 4th (highest) | 24.8 | 23.7 | 27.0 | 24.0 | 21.7 | 25.4 |
| Missing | 1.7 | 2.1 | 1.9 | 1.7 | 2.2 | 2.4 |
| Chronic conditions a | ||||||
| Yes | 7.3 | 16.6 | 12.5 | 7.3 | 17.6 | 10.3 |
| No | 92.7 | 83.4 | 87.5 | 92.7 | 82.4 | 89.7 |
| Hospital/hospitalization characteristics | ||||||
| Discharge diagnoses, mean ± standard deviation | 5.1 ± 2.4 | 7.6 ± 3.2 | 8.9 ± 4.1 | 4.6 ± 2.3 | 7.8 ± 3.3 | 9.3 ± 3.8 |
| Hospital procedures, mean ± standard deviation | 1.9 ± 1.1 | 2.5 ± 1.6 | 4.5 ± 2.5 | 1.7 ± 0.9 | 2.6 ± 1.6 | 4.8 ± 2.4 |
| Location, % | ||||||
| Urban | 88.4 | 92.7 | 95.2 | 87.7 | 90.8 | 94.9 |
| Rural | 11.6 | 7.3 | 4.8 | 12.3 | 9.2 | 5.1 |
| Teaching status, % | ||||||
| Teaching | 45.5 | 62.7 | 68.2 | 43.9 | 59.2 | 66.5 |
| Nonteaching | 54.5 | 37.3 | 31.8 | 56.1 | 40.8 | 33.5 |
| Ownership, % | ||||||
| Public | 63.8 | 74.7 | 77.9 | 62.5 | 72.2 | 77.4 |
| Private not-for-profit | 21.5 | 16.1 | 13.2 | 21.8 | 16.5 | 13.8 |
| Private for profit | 14.7 | 9.2 | 8.9 | 15.7 | 11.3 | 8.8 |
| Bed size, % | ||||||
| Small | 10.5 | 7.9 | 6.5 | 10.7 | 8.2 | 5.8 |
| Medium | 25.9 | 22.7 | 22.2 | 26.0 | 23.7 | 23.6 |
| Large | 63.7 | 69.4 | 71.3 | 63.2 | 68.1 | 70.6 |
| Hospital costs, median (interquartile range) b | 12,547.9 (8833.0–17,767.4) | 23,672.3 (15,002.2–38,609.3) | 34,035.5 (18,624.6–67,803.6) | 10,866.5 (7569.3–15,391.0) | 19,033.0 (11,193.1–32,663.5) | 33,944.1 (17,832.5–64,232.4) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
