Health care disparities among races in maternal morbidity and mortality rates in the United States have been well documented. The reasons underlying differences in disease incidence and outcomes are likely multifactorial and may include genetic predisposition, comorbid medical conditions, and access to health care. The article by Howell et al in this issue of the American Journal of Obstetrics & Gynecology posits another possible contributor: hospital quality. The authors demonstrate that, after adjustment for comorbidities, the risk of severe maternal morbidity for African American women was doubled compared with white women. Notably, risk-standardized maternal morbidity rates among New York City hospitals varied by >7-fold (from 0.8 to 5.7 per 100 deliveries). Most deliveries among white women occurred in low morbidity hospitals; only one-quarter of black women delivered in such settings. In contrast, twice as many black women delivered in high morbidity hospitals. Furthermore, the authors predicted the probability of morbidity for black patients if they delivered at the same hospitals as white mothers and found that maternal morbidity rates would be predicted to fall by almost 50%, which would save nearly 1000 black women from a severe morbid event annually.

These results provide us with actionable data with which to reduce health care disparities. When research points to non-modifiable risk factors that increase the chance of an adverse outcome, it can be a challenge to determine how to respond. However, this study points to a factor over which we as physicians may exert some control: hospital quality. It is our challenge (1) to determine the reasons that patients choose low-performing hospitals, (2) to assist patients in getting care at hospitals that provide the needed services, and (3) to improve quality so there are no low-performing hospitals.

The reasons that patients choose certain hospitals are not well-defined. Survey data indicate that quality is important to patients, but follow up questions indicate that patients may not know how to judge quality. Despite available information about hospital measures on “consumer report cards” and publically available outcome data, such information may not be easy for patients to access. Instead, patients may choose hospitals based on word of mouth and health plan recommendation. Alternatively, patients may choose a specific health care provider who is affiliated with or refers to only 1 hospital. Furthermore, patients may have a preference for receiving care near where they live. In addition to convenience, patients may have previous experience with a neighborhood hospital. Local hospitals may be better equipped to provide culturally sensitive care and additional support, such as translation services. Patients may desire doctors who are similar to them in gender or ethnicity in the hope that they will get better care if their physician has a personal connection. However, there is no evidence that not adhering to the standard of care out of deference to patients’ wishes improves care. In fact, patients may suffer as a result of doctors making exceptions. Medicare and other organizations publish information on picking a hospital, with information made available on sites such as “hospital compare.” However, general guidelines for determining hospital quality may not predict quality of obstetric care accurately. For example, reducing hospital readmission is a basic marker of quality in surgical patients ; however, in obstetrics, the reasons for readmission may differ and could represent unique pregnancy physiology that is not captured in more general quality indicators. Some well-described measures of obstetric quality include elective delivery at <39 weeks of gestation, singleton vertex cesarean delivery rate, and prophylactic antibiotic use before cesarean delivery ; however, these data are not generally available and may be difficult for patients to interpret. Furthermore, it is not clear that existing Joint Commission indicators of obstetric quality correlate with maternal or neonatal morbidity.

How do we assist patients in getting care at a “quality” hospital? One proposed solution to “poor quality” hospitals is closure. In New York City and around the nation, many hospitals have been closed. However, such actions may be met with resistance from communities that fear loss of access to local care. Furthermore, the decision to close hospitals may reflect financial priorities more than quality of care. Notably, data on hospital closures do not point to increasing mortality rates as a consequence. Although that finding may reassure communities and policy makers that there is little harm in closing these hospitals, if low volume and poor quality hospitals are being closed, we might expect decreases in overall morbidity: this has not been demonstrated. The potential benefits of closure of low-quality hospitals may be offset by delays in care. Such concerns could be mitigated via the opening of local clinics, assistance with transportation to high-quality hospitals, and scheduling appointments at times that take into consideration the difficulties of travel. Another paradigm has been the merging of hospitals under umbrella organizations. In addition to cost savings, such consolidation could improve quality by unifying protocols, educational opportunities, and quality improvement initiatives. When hospitals share infrastructure, it may be easier to provide necessary referrals and enable patient transfers for complicated cases. Both physicians and patients may be more willing to accept a transfer for a “high-risk” patient if a relationship with the hospital or system has already been established.

As an alternative to hospital closure, it may be beneficial to encourage transfer of some patients to regional centers and to work to improve overall hospital quality. The health care professional caring for the patient must have the foresight to transfer patients whose condition requires specialized care to specialty centers. Although high-risk patients may benefit from transfer to specialty centers, many obstetric emergencies cannot be predicted or prevented. Although a patient with known risk factors (eg, placenta accreta) may benefit from delivery at a specialty center, so many patients are at risk for postpartum hemorrhage that transfer of all such patients is not realistic. It is our responsibility to ensure that patients get the best possible care, regardless of where they deliver. There are numerous resources available to improve maternal morbidity and mortality rates. American College of Obstetricians and Gynecologists District 2 has instituted the “safe motherhood initiative,” which is an educational program that includes algorithms for postpartum hemorrhage, venous thromboembolism, and hypertensive emergencies. Such programs are designed not only to provide a plan of care for these emergencies but also to enable staff educational opportunities. Other initiatives that have been demonstrated to improve outcomes in specific medical settings are checklists, team training, simulation drills, and early warning trigger tools. Low volume hospitals or hospitals that care mainly for low-risk patients may benefit from simulation training and early warning systems to both keep emergency skills sharp and to identify which patients may benefit from transfer.

In the current study, hospital-level variables that were associated with lower morbidity rates were teaching status, level 3/4 nursery, private ownership, and very high volume status. Although factors such as very high volume likely are related directly to quality of care, nursery level may be a surrogate for depth and breadth of available services. Improved quality of care at teaching hospitals has been described previously, but the reasons for these differences are not well-delineated. It is possible that quality improvement initiatives may be easier to institute in a teaching-focused setting.

The current study is limited by reliance on administrative data and the limitations of such databases in capturing true risk factors and outcomes. Nonetheless, the authors should be congratulated for focusing attention on an understudied and actionable issue. Even if the role of hospital quality in obstetric outcomes is overestimated, the goals to provide patient-centered care and improve quality of care cannot be overemphasized.