Each year it is estimated that more than 600,000 women die during pregnancy and in child birth.1,2 Worldwide, obstetric hemorrhage is the leading cause of maternal death and is associated with nearly one-third of deaths in Africa and Asia.3–5 Even in developed countries, hemorrhage at the time of delivery remains an important cause of morbidity and mortality. In the United States, hemorrhage was the reported etiology of 10% of maternal deaths directly related to obstetric causes.4
Errors in the management of obstetric hemorrhage are common. Delays in diagnosis, inappropriate medical and surgical treatments, and ineffective teamwork all contribute to the morbidity associated with obstetric hemorrhage.6 In a study of hemorrhage-associated morbidity in Scotland, nearly 35% of women received suboptimal care.2 A review of pregnancy-related deaths in North Carolina found that 90% of hemorrhage-related deaths were preventable.7
For women who experience obstetric hemorrhage, peripartum hysterectomy can be a life-saving procedure.8–12 Hysterectomy is usually considered a measure of last resort and is only performed in women in whom treatments have failed or in those who are not candidates for medical treatments of hemorrhage or more conservative surgical alternatives. When peripartum hysterectomy is required, it is often in the setting of heavy bleeding, hemodynamic instability, and in emergent circumstances without the presence of adequate support staff. Given these technical challenges, peripartum hysterectomy is associated with substantial morbidity and mortality. Reported perioperative maternal death rates associated with the procedure range from 1% to 6%.8,10,13
In the United States, initiatives to improve patient safety and quality have been directed toward developing minimum case volumes for complex surgical procedures. These imperatives are based on data that outcomes are improved for patients who are undergoing high-risk surgical procedures when they are treated by surgeons at high-volume centers. In an analysis of 14 high-risk surgical procedures, the absolute difference in adjusted mortality between low- and high-volume hospitals was up to 12%.14 Given the substantial morbidity associated with peripartum hysterectomy and the intensive support services required to care for women with massive obstetric hemorrhage, hospital procedure volumes may impact the outcome of women who require peripartum hysterectomy. The aim of our study was to determine if the volume of peripartum hysterectomies performed by hospitals influences maternal morbidity and mortality.
MATERIALS AND METHODS
This study was approved by the Columbia University Institutional Review board. Data from the Perspective (Premier, Charlotte, NC) database were utilized. The Perspective database is a nationwide, voluntary, fee-supported database developed to measure quality and resource utilization. The database collects inpatient data from more than 500 acute-care hospitals throughout the United States. Each participating institution submits electronic updates quarterly. The data are audited regularly to ensure quality and integrity. In addition to patient demographics and disease and procedure characteristics, the database collects information on all billed services. As such, the database contains data on all drugs and medications, devices, radiologic tests, laboratory tests, and therapeutic services rendered during a patient's hospital stay. Perspective is validated and has been utilized in a number of outcomes studies.15
Our analysis included data for women aged 50 years or less treated between 2002 and 2007. The study cohort consisted of women who underwent a peripartum hysterectomy that we defined as a hysterectomy within 2 days of cesarean delivery. Selection criteria for the study cohort included patients who underwent a cesarean delivery (International Classification of Diseases, 9th edition [ICD-9] 74.x, 669.70, 669.71) in combination with an abdominal hysterectomy (ICD-9 68.3x, 68.4x). The type of hysterectomy performed was further stratified as subtotal (ICD-9 68.3x) or total (ICD-9 68.4x). Patients with a concomitant diagnosis of an invasive malignancy were excluded.
Procedure-associated complications were assessed using ICD-9 codes for known hysterectomy-related morbidity. Complications were classified into the following groups: 1) intraoperative complications (bladder injury, ureteral injury, intestinal injury, vascular injury, other operative injury); 2) perioperative surgical complications (reoperation, postoperative hemorrhage, wound complication, venous thromboembolism); and 3) medical complications (cardiovascular, pulmonary, gastrointestinal, renal, infectious). Length of hospital stay was defined as the number of days from hysterectomy to discharge. Rates of transfusion and intensive care unit use were calculated. Perioperative death was defined as death during the hospitalization during which the patient underwent hysterectomy.
Predictor variables included demographic, clinical, and hospital characteristics. Age (less than 30 and 30 years or older) and year of surgery (2003–2005, 2006–2007) were dichotomized, and race was stratified into white, African American, and other. Each patient's insurance and marital status were recorded. Comorbidity was estimated using the Deyo index, validated index measuring comorbidity based on ICD-9 coding.16 Hospital characteristics including location (urban, rural), type (teaching, nonteaching), size (fewer than 300, 300–450, more than 450 beds), and region (midwest, northeast, south, west) were recorded. The principal diagnoses for women who underwent peripartum hysterectomy were classified into the following: placenta accreta (666.0x, 667.0x), uterine atony (666.1x), uterine rupture (665.0x, 665.1x), hysterotomy extension (665.3x, 665.4x), delayed hemorrhage (666.2x), and leiomyoma (218.x, 219.x). These diagnoses were not mutually exclusive and patients may have had multiple codes.
We hypothesized that hospital volume was a significant predictor of outcome for women who underwent peripartum hysterectomy. Because not all hospitals contributed patient data during the entire study period, we calculated annual hospital volume by dividing the total number of procedures by the number of years in which the individual hospital had at least one peripartum hysterectomy recorded. The distribution of annual surgical volume was generated and reviewed, and cutpoints were selected to divide the hospitals into approximately equal volume tertiles based on procedure volume and complications (low, intermediate, and high). Because results may have improved with time, we adjusted for year of diagnosis in our analysis.
Frequency distributions between categorical variables were compared using χ2 test, whereas continuous variables were compared with one-way analysis of variance. To account for clustering at the hospital level, we used generalized estimating equations.17 Generalized estimating equations logistic regression analysis was performed to determine independent predictors of complications and mortality. Individual generalized estimating equations analyses were performed for the complications of interest and for mortality. Length of stay was estimated with linear regression models using ordinary least squares. We report both unadjusted and adjusted odds ratios (ORs). All analyses were performed with SAS version 9.2 (SAS Institute, Cary, NC).
A total of 2,209 women who underwent peripartum hysterectomy were identified (Table 1). Maternal mortality among the 2,209 women who underwent peripartum hysterectomy was 1.2%. The median age of the cohort was 33 years (age range 14–50 years). Half of the patients were white and 63% were married. The majority of women (61%) had commercial insurance, 32% had Medicaid, and 3% were without insurance. The majority of patients were treated at urban centers, with a similar distribution between teaching and nonteaching hospitals. The most frequent reasons for performance of hysterectomy were placenta accreta (35%) and uterine atony (35%).
The cohort was treated at 320 different hospitals. Two hundred twenty-one (69%) facilities were in the lowest volume tertile, 73 (23%) were in the intermediate volume tertile, and 26 (8%) were in the highest volume tertile. The interquartile peripartum hysterectomy ranges for the three strata were low volume (1.0–2.0 procedures), intermediate volume (2.8–3.7 procedures), and high volume (5.0–7.2 procedures).
The unadjusted rates of operative and perioperative complications are displayed in Table 2. The rate of bladder injury ranged from 7% in the low-volume tertile to 9% in the high-volume group (P=.53). There were no statistically significant differences in the rates of any individual based on hospital volume. Perioperative surgical complications including reoperation (6% compared with 3%) and wound complications (10% compared with 7%) were more frequent in the women treated at low-volume centers (P<.05). Likewise, postoperative pulmonary morbidity was noted in 14% of women cared for at low-volume hospitals compared with 10% in high-volume centers (P=.05).
The transfusion rate was 57% at low-volume hospitals, 47% at intermediate-volume centers, and 45% at high-volume hospitals (P<.001). Intensive care was required in 45% of women at low-volume facilities vs 27% at high-volume hospitals (P<.001). The mean length of stay in the low-volume cohort was 3.5 days compared with 4.1 days for the high-volume group (P=.22). Mortality ranged from 1.8% (13 deaths) for women at low-volume facilities to 0.8% (5 deaths) who underwent operation at high-volume hospitals (P=.02).
Adjusted estimates of morbidity and mortality are shown in Table 3. In the adjusted analyses, volume had no effect on the rates of intraoperative injuries, medical complications, length of stay, or transfusion. In contrast, compared with women treated at low-volume centers, patients who underwent operation at high-volume hospitals had a lower incidence of perioperative surgical complications (OR 0.66, 95% confidence interval 0.47–0.93) and a lower rate of intensive care unit usage (OR 0.53, 95% confidence interval 0.34–0.83). Hospital volume was also associated with maternal mortality. Compared with women who underwent surgery at low-volume hospitals, those who underwent operation at high-volume centers had a 71% (OR 0.29, 95% confidence interval 0.10–0.88) lower risk of death.
A number of studies have demonstrated the association between surgical volume and outcome; patients who underwent operation performed by surgeons at high-volume centers have decreased morbidity and mortality.14,18–21 In a meta-analysis of patients with gastrointestinal cancer, it was estimated that only 16 patients with colon cancer would need to be moved from a low-volume to high-volume hospital to prevent one volume-associated death.18 The association between volume and outcome is strongest for high-risk oncologic and cardiovascular procedures.14,19–21 In contrast, the link between volume and outcome is not as strong for less complex procedures.22,23
The effect of hospital and provider volume on obstetric outcomes is less clear. Whereas several studies have shown decreased neonatal death rates in higher-volume labor units, there does not appear to be an association between volume and outcomes for gestational diabetes, vaginal birth after cesarean delivery, or in the rates of cesarean delivery.24–28 Our analysis of obstetric hemorrhage indicates that maternal morbidity and mortality are lower for women who undergo peripartum hysterectomy at high-volume hospitals. In the adjusted analysis, mortality was reduced by more than 70% for women who underwent operation at high-volume facilities.
Successful management of obstetric hemorrhage requires extensive hospital resources and an interdisciplinary approach. In addition to the collaboration of physicians from a number of specialties, adequate nursing support, blood bank services with the ability to replace large volumes of blood, critical care support, and the ability to perform invasive cardiac monitoring are often required. Because obstetric hysterectomy is only performed as a last resort after the failure of medical interventions and conservative surgical techniques, our study likely includes only the most severe cases of hemorrhage.2 One possible explanation for our findings is that hospitals that perform higher volumes of peripartum hysterectomy have systems in place to deal with obstetric hemorrhage and are more prepared to manage these complex cases. It is important to recognize that these findings are not merely attributable to hospital size or teaching status because the relationship between volume and mortality was independent of both of these variables.
The difference in mortality we noted cannot be explained simply by higher complications rates in low-volume hospitals. We found no difference in the rates of intraoperative and most postoperative medical complications between low-volume and high-volume facilities. Recently, attention has focused on how well hospital systems recognize and manage major complications. So-called failure to rescue patients with complications may be an important mechanism that underlies variation in perioperative mortality.29–31 Timely recognition and effective management of major complications require an extensive infrastructure. Given the rarity of peripartum hysterectomy and the resources required to manage these patients, treatment of complications may partly explain the mortality differences we identified.
Our study had several important limitations. As with any study of administrative data, claims files are meant to provide billing information and may not provide detailed clinical data. Based on available coding, we cannot distinguish preexisting diseases from complications. However, given the young age of the patients included in the study it is unlikely that significant numbers of women had major medical comorbidities. We provided estimates of the effect of hospital volume on outcome. Whereas the data set used provides data on specific physicians, we chose not to perform separate analyses stratified by physician volume because of the rarity of the procedure. Because physician volume has been shown to influence perioperative outcomes for other procedures, we cannot exclude the possibility that physician experience and volume also affect the morbidity of peripartum hysterectomy. Despite use of generalized estimating equations, there may be unmeasured biases and differences among hospitals that influenced our findings. Finally, although we noted differences in mortality based on hospital volume, it must be recognized that there were only 26 maternal deaths in our cohort.
Our data highlight the morbidity and mortality associated with obstetric hemorrhage. A number of patient safety initiatives have been suggested to improve the care of women with obstetric hemorrhage.32–35 After two maternal deaths, the New York Hospital Queens developed a patient safety initiative aimed at improving the care of obstetric hemorrhage. After implementation of the program there was a significant decrease in maternal mortality.32 In addition to standardized protocols, several groups have described the formation of obstetric rapid response teams to better-facilitate and coordinate the care of women with obstetric hemorrhage.32,33
Perhaps the most important question raised by our study is whether obstetric care for women at risk of massive hemorrhage should be regionalized. Whereas uterine atony is frequently unpredictable, placenta accreta is often diagnosed or at least suspected antenatally. Given the resources required to manage patients with placenta accreta and the improved outcomes achieved at high-volume centers, we believe that these women should be referred to tertiary centers with experience in the management of the disease. Although outcomes may be improved, strategies to regionalize care for complex conditions are fraught with difficulty, particularly among ethnic minorities and poorly insured patients.36,37 Given these limitations, all hospitals should have systems in place to deal with parturients at risk of massive obstetric hemorrhage.
In conclusion, our study noted that peripartum hysterectomy is associated with significant morbidity and mortality. Maternal mortality is lower when the procedure is performed in high-volume hospital settings. The risk of maternal mortality is independent of hospital size and does not appear to be attributable to differences in the rates of operative injuries or postoperative medical complications. Further initiatives to reduce maternal mortality from obstetric hemorrhage are clearly needed.
1. Wagaarachchi PT, Graham WJ, Penney GC, McCaw-Binns A, Yeboah Antwi K, Hall MH. Holding up a mirror: changing obstetric practice through criterion-based clinical audit in developing countries. Int J Gynaecol Obstet 2001;74:119–31.
2. Brace V, Kernaghan D, Penney G. Learning from adverse clinical outcomes: major obstetric haemorrhage in Scotland, 2003–05. BJOG 2007;114:1388–96.
3. Weindling AM. The confidential enquiry into maternal and child health (CEMACH). Arch Dis Child 2003;88:1034–7.
4. Heron M, Hoyert DL, Murphy SL, Xu J, Kochanek KD, Tejada-Vera B. Deaths: final data for 2006. Natl Vital Stat Rep 2009;57:1–134.
5. Khan KS, Wojdyla D, Say L, Gulmezoglu AM, Van Look PF. WHO analysis of causes of maternal death: a systematic review. Lancet 2006;367:1066–74.
6. Fuchs KM, Miller RS, Berkowitz RL. Optimizing outcomes through protocols, multidisciplinary drills, and simulation. Semin Perinatol 2009;33:104–8.
7. Berg CJ, Harper MA, Atkinson SM, Bell EA, Brown HL, Hage ML, et al. Preventability of pregnancy-related deaths: results of a state-wide review. Obstet Gynecol 2005;106:1228–34.
8. Yucel O, Ozdemir I, Yucel N, Somunkiran A. Emergency peripartum hysterectomy: a 9-year review. Arch Gynecol Obstet 2006;274:84–7.
9. Stanco LM, Schrimmer DB, Paul RH, Mishell DR Jr. Emergency peripartum hysterectomy and associated risk factors. Am J Obstet Gynecol 1993;168(3 Pt 1):879–83.
10. Kwee A, Bots ML, Visser GH, Bruinse HW. Emergency peripartum hysterectomy: A prospective study in The Netherlands. Eur J Obstet Gynecol Reprod Biol 2006;124:187–92.
11. Zelop CM, Harlow BL, Frigoletto FD Jr, Safon LE, Saltzman DH. Emergency peripartum hysterectomy. Am J Obstet Gynecol 1993;168:1443–8.
12. Clark SL, Yeh SY, Phelan JP, Bruce S, Paul RH. Emergency hysterectomy for obstetric hemorrhage. Obstet Gynecol 1984;64:376–80.
13. Knight M. Peripartum hysterectomy in the UK: management and outcomes of the associated haemorrhage. BJOG 2007;114:1380–7.
14. Birkmeyer JD, Siewers AE, Finlayson EV, Stukel TA, Lucas FL, Batista I, et al. Hospital volume and surgical mortality in the United States. N Engl J Med 2002;346:1128–37.
15. Lindenauer PK, Pekow P, Wang K, Mamidi DK, Gutierrez B, Benjamin EM. Perioperative beta-blocker therapy and mortality after major noncardiac surgery. N Engl J Med 2005;353:349–61.
16. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–9.
17. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121–30.
18. Gruen RL, Pitt V, Green S, Parkhill A, Campbell D, Jolley D. The effect of provider case volume on cancer mortality: systematic review and meta-analysis. CA Cancer J Clin 2009;59:192–211.
19. Schrag D, Cramer LD, Bach PB, Cohen AM, Warren JL, Begg CB. Influence of hospital procedure volume on outcomes following surgery for colon cancer. JAMA 2000;284:3028–35.
20. Begg CB, Riedel ER, Bach PB, Kattan MW, Schrag D, Warren GL, et al. Variations in morbidity after radical prostatectomy. N Engl J Med 2002;346:1138–44.
21. Schrag D, Panageas KS, Riedel E, Hsieh L, Bach PB, Guillem JG, et al. Surgeon volume compared to hospital volume as a predictor of outcome following primary colon cancer resection. J Surg Oncol 2003;83:68–79.
22. Peterson ED, Coombs LP, DeLong ER, Haan CK, Ferguson TB. Procedural volume as a marker of quality for CABG surgery. JAMA 2004;291:195–201.
23. Anger JT, Rodriguez LV, Wang Q, Pashos CL, Litwin MS. The role of provider volume on outcomes after sling surgery for stress urinary incontinence. J Urol 2007;177(4):1457–62.
24. Moster D, Lie RT, Markestad T. Neonatal mortality rates in communities with small maternity units compared with those having larger maternity units. BJOG 2001;108(9):904–9.
25. Moster D, Lie RT, Markestad T. Relation between size of delivery unit and neonatal death in low risk deliveries: population based study. Arch Dis Child Fetal Neonatal Ed 1999;80:F221–5.
26. Nicholson WK, Fox HE, Cooper LA, Strobino D, Witter F, Powe NR. Maternal race, procedures, and infant birth weight in type 2 and gestational diabetes. Obstet Gynecol 2006;108(3 Pt 1):626–34.
27. Chang JJ, Stamilio DM, Macones GA. Effect of hospital volume on maternal outcomes in women with prior cesarean delivery undergoing trial of labor. Am J Epidemiol 2008;167:711–8.
28. Tracy SK, Sullivan E, Dahlen H, Black D, Wang YA, Tracy MB. Does size matter? A population-based study of birth in lower volume maternity hospitals for low risk women. BJOG 2006;113:86–96.
29. Ghaferi AA, Birkmeyer JD, Dimick JB. Variation in hospital mortality associated with inpatient surgery. N Engl J Med 2009;361:1368–75.
30. Silber JH, Williams SV, Krakauer H, Schwartz JS. Hospital and patient characteristics associated with death after surgery. A study of adverse occurrence and failure to rescue. Med Care 1992;30:615–29.
31. Silber JH, Rosenbaum PR, Schwartz JS, Ross RN, Williams SV. Evaluation of the complication rate as a measure of quality of care in coronary artery bypass graft surgery. JAMA 1995;274:317–23.
32. Skupski DW, Lowenwirt IP, Weinbaum FI, Brodsky D, Danek M, Eglinton GS. Improving hospital systems for the care of women with major obstetric hemorrhage. Obstet Gynecol 2006;107:977–83.
33. Gosman GG, Baldisseri MR, Stein KL, Nelson TA, Pedaline SH, Waters JH, et al. Introduction of an obstetric-specific medical emergency team for obstetric crises: implementation and experience. Am J Obstet Gynecol 2008;198:367 e1–7.
34. Homer C, Clements V, McDonnell N, Peek M, Sullivan E. Maternal mortality: What can we learn from stories of postpartum haemorrhage? Women Birth 2009;22:97–104.
35. Cameron CA, Roberts CL, Bell J, Fischer W. Getting an evidence-based post-partum haemorrhage policy into practice. Aust N Z J Obstet Gynaecol 2007;47:169–75.
36. Finlayson SR. Delivering quality to patients. JAMA 2006;296:2026–7.
37. Liu JH, Zingmond DS, McGory ML, SooHoo NF, Ettner SL, Brook RH, et al. Disparities in the utilization of high-volume hospitals for complex surgery. JAMA 2006;296:1973–80.
Figure. No caption available.