Rayburn, William F. MD, MBA; Richards, Michael E. MD; Elwell, Erika C. BA
A key to health care reform preparedness is ensuring patient access to care. An uneven distribution of obstetrician–gynecologists (ob-gyns) exists throughout the United States, which may worsen as resident graduates cluster in metropolitan areas.1 Determinations of a sufficient physician workforce are dependent not only on supply–demand analyses but also on the ease of access to ob-gyns by patients. Most ob-gyns' offices are beside or near hospitals with maternity services. Those women seeking the care of the nearest ob-gyn would need to determine the time to drive to the nearest hospital offering maternity and nursery (perinatal) services.
The optimal distribution of hospitals with perinatal centers has long been debated, yet timely access remains essential. Quality care around the time of birth is a resource-intensive endeavor requiring hospitals to have different levels of specialized equipment and experienced personnel. These resources are typically available at perinatal centers dedicated with a well-equipped nursery.
With a potential shortage of obstetricians and the growing complexity of medical and surgical aspects of pregnancy, there is a need to determine how best to ensure that all pregnant women have access to some hospital with perinatal services. As a first and necessary step in this process, we undertook a broad approach by evaluating the driving time by our patients to the various hospitals offering perinatal services in the United States.
MATERIALS AND METHODS
This observational study was conducted after approval from the University of New Mexico Human Research Review Committee (HRRC 11-328). We performed a cross-sectional analysis of access to hospitals offering maternity and nursery care. The 2007 American Hospital Association annual survey includes a self-reported list of hospitals offering any of the following levels of maternity and nursery care.2 Those levels of perinatal care have been described by the American Academy of Pediatrics and the American College of Obstetricians and Gynecologists (the College) and are used by the American Hospital Association: level I, uncomplicated obstetric and nursery care; level II, limited complicated care; and level III, full complement of care.3 Free-standing birth centers were excluded because they are not part of the American Hospital Association database and represent only a very small proportion of deliveries. Point locations for the perinatal care hospitals were geocoded by the hospital latitude and longitude from the American Hospital Association survey using the Arcgeographic information system 10.0 geographic information system.
Demographic data were obtained from the ESRI Census 2000 TIGER/Line Shapefile for census block groups with 2010 demographic estimates. Being subdivisions of census tracts, census block groups are the smallest geographic area for which the U.S. Census Bureau publishes sample data.4 Census block groups generally contain between 600 and 3,000 people. Reproductive age was defined as age 18–39 years using the following census block group age categories of 18–21, 27–29, and 30–39 years. Population data of each census block group were assigned to the block group's geographic centroid (the location representing the “mean center” of the block group). Census block group data were aggregated at state, regional, and national levels for separate analyses. We analyzed regions according to the 11 districts of the College. The only district excluded was District X (Armed Forces).
The U.S. Department of Agricultural rural–urban commuting area codes were used to classify the geographic areas using measures of population density, urbanization, and daily commuting patterns.5 The most recent rural–urban commuting area codes are based on data from the 2000 census and delineate four core areas by population: metropolitan (more than 50,000), micropolitan (10,000–49,999), rural town (2,500–9,999), and rural commuting areas (less than 2,500). Neighboring geographic areas also are assigned to the core area when the primary commuting flow to a core area is more than 10%.
Arcgeographic information system network analyst tool was used to identify census block groups with 30-minute and 60-minute driving times to maternity care hospitals. The two time periods were chosen in studies involving emergent conditions (trauma, stroke, burn) and were considered by these authors as being a reasonable period (30 minutes) and a maximal safe period (60 minutes) for maternity patients.6–9 For each hospital, 30-minute and 60-minute drive times to the hospital buffer were generated using 2009 ESRI Street Map North America road network. Census block group centroids within each drive time buffer zone were identified and included in analysis.
The U.S. census data are considered to be an enumeration of the U.S. population, with negligibility in small sampling variability. For such data, and for the intention of this study, only descriptive statistics of population-level data are presented.10
In 2007, there were 4,141 general medical surgical hospitals in the United States, of which 2,606 (62.9%) hospitals offered some level of perinatal care. Of those hospitals, 1,169 (44.9%) were designated as being level I, 928 (35.6%) were level II, and 509 (19.5%) were level III. A total of 49.8 million women in the United States were of reproductive age. Access was not the same for women of all racial backgrounds. Hispanic and African American women lived the closest to targeted hospitals. Native-American women had the lowest percentage living within a 30-minute drive (68.2%) or a 60-minute drive (83.2%) of the nearest perinatal center. Furthermore, most of those did not offer level III care, because Native-American women were more inclined to live in isolated rural areas.
Access to any maternity center was within a 30-minute drive for 87.5% of the population (43.6 million) and within a 60-minute drive for 97.3% of the population (48.4 million). Any level of hospitals was most accessible in metropolitan settings. Hospitals with capabilities to provide care for more complicated cases (level II or III) were less accessible, within a 30-minute drive (78.6% of the population; 39.1 million) and a 60-minute drive (93.1% of the population; 46.4 million). Centers providing the most comprehensive maternal and newborn care (level III) were least accessible and located only in metropolitan areas (60.8% or 30.3 million of the population within a 30-minute drive; 80.1% or 39.9 million of the population within a 60-minute drive).
Figure 1 displays the percentages of all reproductive-aged women who lived within a 30-minute drive and a 60-minute drive of their nearest hospital offering perinatal care. Commute times became more prolonged as the population area became more rural or less dense. This finding was explained by fewer hospitals being found in more rural commuting areas.
Shown in Figure 2 are the census block groups of women residing within 30-minute and 60-minute drives to those hospitals. Clearly, drive times were shorter in the eastern half of the United States, with nearly all regions having some targeted hospital being within a 60-minute drive. The lowest density of hospitals was in the western Great Plains, noncoastal West, South Central, and Alaska.
The time to drive to perinatal centers varied by College district and state. Shown in Table 1 are the numbers of reproductive-aged women and the percent who live within a 30-minute drive and within a 60-minute drive to their nearest center. Accessibility was best in the Northeast United States (Districts I, II, and III) and California (District IX), where urban density was highest. We found that 92.8–95.2% of the targeted population lived within a 30-minute drive and that 97.7–99.7% lived with a 60-minute drive of any center. In contrast, accessibility was lowest in Districts VII and VIII, where 75.5–81.7% of the targeted population lived within a 30-minute drive and 91.3–95.4% lived with a 60-minute drive of any perinatal center.
Table 1 also displays geographic access to perinatal centers by state in each district. Most of the targeted population lived in states within a 60-minute drive of any center. Approximately 95% or more of reproductive-aged women in Connecticut, Massachusetts, Maryland, New Jersey, Ohio, Rhode Island, and the District of Columbia resided within a 30-minute drive. The lowest proportion of the population in Alaska, Mississippi, Montana, North Dakota, Wyoming, and New Mexico lived within either a 30-minute or a 60-minute drive to the nearest hospital providing perinatal care.
Access to care will be a central issue with health care reform. Maps generated for this report were based on 2007 hospital data and on 2010 U.S. Census Bureau data, the most recent available at that time. Maps were based on driving times from home to the hospital rather than on merely ground distance. We ascertained 30-minute and 60-minute driving times for a targeted population of women who could become pregnant and need some level of obstetric care.
Region-to-region variation in access to the nearest hospital offering perinatal care was considerable. Geographic information service mapping helped identify College districts and states with the greatest need. Access to a perinatal center within a 60-minute drive ranged from as low as 82% of the population in the noncoastal Western states and in the South Central states, to as high as 99% of those living in Northeast states with large metropolitan areas. Overall, 30-minute driving time access was common but followed patterns according to population density. Not surprisingly, access to hospitals with level I, II, or III perinatal centers was highest in the Northeast.
We concluded that hospitals offering perinatal care, especially at II and III levels, are unevenly distributed in the United States but can be meaningfully improved on if additional centers are added strategically. Areas where the population was small and widely distributed, however, require solutions other than simply constructing another hospital. To achieve very high dissemination of perinatal care throughout those regions, alternate strategies would be more practical, such as use of telemedicine to augment late prenatal care at nonmaternity hospitals, scheduling inductions of labor in advance for women with a favorable cervix or who are multiparous, and improving stabilization of the mother or newborn before transfer from an outlying hospital in a rural setting.
Targeting general medical surgical hospitals that do not offer perinatal care and are located outside the 60-minute drive time is worthwhile to consider. We located 136 such hospitals in the western United States. Training at those hospitals could be offered to improve access for emergent perinatal care. It is noteworthy that driving times to hospitals with perinatal care are unlikely to be significantly altered using ambulance services. Furthermore, there are no national data that report times in transporting pregnant women using fixed-wing aircraft.
This investigation demonstrated that use of geographic information system-generated maps has implications for policy-making and planning, especially in adding any new perinatal centers to optimize overall regional access.11 This software package has been utilized by other medical specialties for reporting on regional development of burn, stroke, cardiac, and trauma centers.12 A feature of geographic information system software is its use in a “hot-spot” analysis, ie, if one new hospital with a perinatal center was added to a state, then where should it be placed to have the greatest effect on improving drive times? Factoring in the traffic in metropolitan areas is important and was not included in this report. It is possible using geographic information system software to simulate conditions in which traffic density can vary from time of day and day of the week.
A feature of this investigation was that the data are easily accessible and updated often. Thus, there were no delays in performing computations that apply currently or for predicting trends. No data identified any persons, so Institutional Review Board approval was unnecessary. The hospital and census data sets contained population-level data, so sampling variability was negligible with respect to the presented descriptive statistical analysis.10 Results from these findings should aid in identifying those sites for future hospitals with perinatal care on a needs basis and to support decision-making during the implementation of the recommendations for facility requirements and physician recruitment.
This study addressed access related to driving times to providers but not availability for the next appointment or quality of care. Although accessibility to the nearest of hospitals is an important consideration, it is also necessary to factor in the capabilities of each hospital (and nearby clinics) to provide quality, comprehensive, and safe care.12 We assume that level I hospitals are more likely to serve rural populations where more of the obstetric care may be provided by family physicians and nonphysician clinicians rather than by ob-gyns alone. In general, hospitals treating a higher volume of patients with complicated cases (levels II and III hospitals) make fewer errors and achieve better perinatal outcomes. Verification of quality care criteria should include determining the minimum number of complicated pregnancy admissions, dedicated staff of nurses, prehospital triage and maternal and neonatal transfer guidelines, processes for quality of care review, and directors who have sufficient training or experience in providing the appropriate level of obstetric and newborn care. It is possible that these reported numbers do not accurately reflect currently available staff and capacity to care for pregnant women and their newborns if all beds were filled.
The optimal period in which pregnant women require access to perinatal centers remains to be defined. We chose a 60-minute travel time to be a reasonable upper limit for a pregnant woman to be driven to a hospital after symptom onset (eg, regular contractions, presumed spontaneous rupture of the amniotic membranes, vaginal bleeding). Based on the findings of this study, more than a 1-hour drive was required for patients in isolated rural settings to reach a perinatal center to suit her needs. Several centers provide coverage for large geographic regions, particularly in the western United States. A future investigation would be to determine whether a prolonged delay in accessing a perinatal center would contribute to less safe and to costlier care resulting from more maternal infection, cesarean deliveries, induction of labor, prolonged hospital stays, readmissions, and perhaps maternal or neonatal mortality. Another investigation could also forecast the potential benefit of coordinated networks of transportation between and among those regional centers serving the vast reach of rural America.
In conclusion, this observational study demonstrates that nearly 97.3% of the U.S. population of reproductive-aged women live within a 60-minute drive of a hospital offering some level of perinatal care. However, there exist substantial state and regional variations in geographic access to various levels of perinatal centers. Although the optimal distribution of targeted hospitals remains undetermined, these findings provide important information about population access that may be used to guide resource allocation for present and future inpatient maternity and nursery care.
1. Rayburn WF. Characteristics and distribution of obstetricians-gynecologists in the United States. The obstetrician and gynecologist workforce in the United States. Washington (DC): American Congress of Obstetricians and Gynecologists; 2011 p. 43–66.
2. American Hospital Association, Annual Survey Database Fiscal Year 2007. Chicago (IL): American Hospital Association Resource Center. 2008.
3. Guidelines for perinatal care. 6th ed. Elk Grove Village (IL): American Academy of Pediatrics, American College of Obstetricians and Gynecologists; 2007. p 11–12.
4. United States Census Bureau, Geographic Division. Cartographic products management branch. 2001 Census block groups cartographic boundary files descriptions and metadata. Available at: http://www.census.gov/geo/www/cob/bg_metadata.html
. Retrieved December 8, 2010.
5. Hall SA, Kaufman JS, Richetts TC. Defining urban and rural areas in U.S. epidemiological studies. J Urban Health. March 2006; 83: 162–75.
6. Young JS, Bassam D, Cephas GA, Brady WJ, Butler K, Pomphrey M. Interhospital versus direct scene transfer of major trauma patients in a rural trauma system. Am Surg 1998; 64: 988–91.
7. Rogers FB, Osler TM, Shackford SR, Cohen M, Camp L, Lesage M. Study of the outcome of patients transferred to a level I hospital after stabilization at an outlying hospital in a rural setting. J Trauma 1999; 46: 328–33.
8. Saposnik G, Fang J, O'Donnell M, Hachinski V, Kapral MK, Hill MD, et al.. Escalating levels of access to in-hospital care and stroke mortality. Stroke 2008; 39: 2522–30.
9. Klein MB, Kramer CB, Nelson J, Rivara FP, Gibran NS, Concannon T. Geographic access to burn center hospitals. JAMA 2009; 302: 1774–81.
10. Desbiens NA. The reporting of statistics in medical educational studies: an observational study. BMC Med Res Methodol 2007; 7: .
11. McLafferty SL. Geographic information system and health care. Annu Rev Public Health 2003; 24: 25–42.
12. Graves BA. Integrative literature review: a review of literature related to geographical information systems, healthcare access, and health outcomes. Perspect Health Inf Manag 2008; 5: 11.