. . . . . Hospitals . . . . .

Trends in Rural and Urban Deliveries and VaginalBirths: California 1998-2002Susan Hughes, MS;1 John A. Zweifler, MD, MPH;1 Alvaro Garza, MD, MPH;2 and Matthew A. Stanich, MPH3

ABSTRACT: Context: Pregnant women in rural areasmay give birth in either rural or urban hospitals.Differences in outcomes between rural and urban hospitalsmay influence patient decision making. Purpose: Trendsin rural and urban obstetric deliveries and neonatal andmaternal mortality in California were compared to informpolicy development and patient and provider decisionmaking in rural health care settings. Methods: Deliveriesin California hospitals identified by the CaliforniaDepartment of Health Services, Birth Statistical MasterFiles for years 1998 through 2002 were analyzed. Threegroups of interest were created: rural hospital births to allmothers, urban hospital births to rural mothers, and urbanhospital births to urban mothers. Findings: Of 2,620,096births analyzed, less than 4% were at rural hospitals.Neonatal death rates were significantly higher in babiesborn to rural mothers with no pregnancy complicationswho delivered a normal weight baby vaginally at an urbanhospital compared to urban mothers delivering at anurban hospital (0.2 [CI 0.2-0.4] deaths per 1,000 birthsversus 0.1 [CI 0.1-0.1]). Logistic regression analysisshowed that delivery in a rural hospital was a protectivefactor compared to urban mothers delivering in an urbanhospital, with an odds ratio of 0.8 (CI 0.6-0.9). Maternaldeath rates were not different. Conclusions: Ruralobstetric services in this period showed favorable neonataland maternal safety profiles. This information shouldreassure patients considering a rural hospital delivery,and aid policy makers and health care providers striving toensure access to obstetric services for rural populations.

Pregnant women in rural areas often have theoption to give birth in either rural or urbanhospitals. Adverse pregnancy and deliveryoutcomes among women who reside anddeliver in rural areas have been associated

with lack of access to prenatal and obstetric care.1 Ruralpatients may bypass rural hospitals to receive necessarycare.2 Although in 1995, a greater percentage ofnon-metropolitan hospitals nationwide providedobstetrical services than metropolitan hospitals (68% vs63%), 23% of rural obstetric patients bypassed local

hospitals between the years 1991 and 1996, while earlierstudies found higher rates of obstetric bypassbehavior.3-5 However, other studies have suggestedthat rural obstetric patients who “outflow” to deliver inurban areas may be more likely to experience adverseoutcomes than those who deliver in their ruralcommunity.6,7

Infant mortality rates in rural areas vary regionallyand are influenced by poverty, insurance status,income, and race/ethnicity.8,9 Other studies have foundan association between rural residence and inadequateprenatal care.10,11 When these factors are controlled,investigators have generally found infant mortalityrates are similar in rural and urban areas and adjustedrates of potentially avoidable maternal complicationsmay be lower in rural hospitals.1,12-14 Despite this, itmay be challenging for rural hospitals to maintainobstetric services. Nationally and within California,rural areas have fewer physicians per capita than urbanareas.15-17 Rural communities have long relied onfamily physicians to provide the bulk of primary care.16

Family physicians also perform a greater proportion ofdeliveries in rural than in urban areas.18 However,fewer family physicians are providing obstetric carenow than in the past.19,20

The viability of rural hospitals, in general, alsothreatens rural obstetric services. Between 1999 and2001, total discharges at urban hospitals in California

1University of California, San Francisco, Fresno Family andCommunity Medicine, Fresno, Calif.2University of California, San Francisco, Fresno Latino Center forMedical Education and Research, Fresno, Calif.3University of California, San Francisco, School of Medicine, SanFrancisco, Calif.

We wish to thank the following people for commenting on ourmanuscript: Norman Hearst, MD, MPH; Kathleen Dyer, PhD; andSean Schafer, MD. For further information, contact: Susan Hughes,MS, UCSF Fresno Family and Community Medicine, 155 N FresnoSt., Fresno, CA 93701; e-mail shughes@fresno.ucsf.edu.

C© 2008 National Rural Health Association 416 Fall 2008

. . . . . Hospitals . . . . .

increased by 2.3%, but fell by 17.8% at rural hospitals.21

Seventy-six percent of California rural hospitals lostmoney on operations in 2000.22 In 2001, rural hospitalsaccounted for less than 5% of California’s dischargesand licensed beds, while California’s rural populationwas 6% in 2000.21

In order to inform policy development and patientand provider decision making in rural health caresettings, rural and urban obstetric services in Californiawere compared by delivery method for neonatal andmaternal mortality.

MethodsBirth data were obtained from the California

Department of Health Services, Birth Statistical MasterFiles for years 1998 through 2002. These filesincorporated information obtained from the State ofCalifornia Certificate of Live Birth—VS-10A MedicalData Supplemental Work Sheet. Analysis was limited todeliveries in California hospitals. The variables used foranalyses included maternal demographics, birthspecific measures, and hospital designation as rural orurban. All variables used were taken or calculated fromthe Birth Statistical Master file except for thedesignation of rural or urban hospital status and therural status of a ZIP code. Maternal variables included:age, race/ethnicity, education, delivery insurance(government or private), number of prenatal visits, andZIP code of residence. Birth measures were: date ofbirth, date of newborn death, type of birth (single ormultiple), birth weight, method of delivery (vaginal orcesarean), pregnancy complications, newborn adverseoutcomes, hospital code, and California county of birth.

Rural hospital names were identified according tothe California Office of Statewide Health Planning andDevelopment designation of “small/rural hospital.”This designation includes small acute care hospitals incounties with rural populations and certain small urbanhospitals with a primary emphasis on simpler medicaland surgical care.23 Because 4 hospitals were notconsistently designated, hospitals ever designated asrural in this time frame were considered rural for thisstudy. Hospital codes that did not match the rural listwere considered urban. The Washington, Wyoming,Alaska, Montana, Idaho Rural Health Research CenterZIP code rural-urban commuting area code (RUCA)approximation was used to classify a particular zipcode as rural or urban.24 All values of the RUCA of 4 orgreater were considered rural. The University ofCalifornia, San Francisco Institutional Review Boardapproved this as an exempt study.

Data were analyzed using SAS (Version 9.1; SASInstitute Inc, Cary, NC). Neonatal death was defined as

those newborns living less than 28 days. Low birthweight was defined as less than 2,500 grams, with aseparate analysis of birth weights <1,500, 1,500-1,999,and 2,000-2,499 g. Normal birth weight was defined as2,500 to 4,000 g and large birth weight as greater than4,000 g. Combined selected complications includingmedical conditions (eg, preeclampsia), history (large,small or premature previous pregnancy outcome), andprocedures (eg, tocolysis) were used to create thecomposite variable pregnancy complications.25

Clinically insignificant procedures excluded from thepregnancy complications composite variable includedelectronic fetal monitoring and ultrasound. Informationon maternal death within 72 hours of delivery andtransport of mother from another facility prior todelivery was available and included in our analysis.Poor prenatal care was defined as less than 6 prenatalvisits during pregnancy.

To better compare rural hospital outcomes to urbanhospital outcomes, we created 3 groups by combiningthe rural/urban hospital designation with therural/urban status of the mother’s residence: ruralhospital/all mothers, urban hospital/rural mothers,and urban hospital/urban mothers. To determine if the3 groups differed in demographic characteristics,chi-square tests were used on categorical data, andnonparametric analysis of variance on continuous datausing a 5% significance level. To assess urban and ruralbirth outcomes, 5-year neonatal and maternal deathrates were compared. Rates were stratified by birthweight and presence of pregnancy complications tocontrol for the possible confounding effects of morecomplicated pregnancies preferentially occurring inurban hospitals. Ninety-five percent confidenceintervals (CI) were calculated assuming a Poissondistribution for all rates. Multiple logistic regressionwas used to examine the significance of contributingfactors in neonatal death while controlling for knownconfounding variables. The 10 variables used were:mother’s age, education, race/ethnicity, pregnancycomplications, type of birth (single or multiple),method of delivery (vaginal or cesarean), weight of thebaby, poor prenatal care (less than 6 visits), type ofdelivery insurance (government or private), andhospital/mother’s location. A corresponding multiplelogistic Poisson regression was unsuccessful atconverging on a model for maternal death.

ResultsEpidemiology of Rural and Urban Hospital

Births. During the 5-year study period (1998 to 2002)there were 2,620,096 births at hospitals in California.Mother’s ZIP code was known for 99.4% of births. Less

Hughes, Zweifler, Garza and Stanich 417 Fall 2008

. . . . . Hospitals . . . . .

Table 1. Maternal and Neonatal Characteristics by Hospital/Mother’s Location California 1998-2002

Urban Hospital BirthCharacteristic Rural Hospital BirthMaternal Location All Rural Urban

Total births 99,285 148,600 2,357,552Mother’s age (median years, Q1-Q3∗) 26 (21-31) 26 (22-31) 28 (23-32)Mother’s educational level (%)

Less than high school 32.0 31.3 29.8High school 36.9 33.7 28.3Some college 25.9 29.1 32.1College 5.2 6.0 9.8

Mother’s race/ethnicity (%)White 46.2 41.8 31.3Hispanic 45.6 51.1 49.2Black 2.4 2.2 6.8Asian 1.3 1.3 7.5Other 4.5 4.0 5.3

Pregnancy complications (%) 22.6 13.9 16.0Smoked during pregnancy (%) 5.1 2.0 1.1Vaginal births (%) 76.2 75.7 75.9Singleton births (%) 98.1 97.2 97.2Delivery insurance (%)

Government 55.8 50.3 42.3Private 44.2 49.7 57.7

Poor prenatal care (<6 prenatal visits) (%) 6.2 5.2 4.2Birth weight (%)

<2,500 g 4.3 6.2 6.42,500-4,000 g 85.1 82.2 83.3>4,000 g 10.6 11.6 10.4

∗Q1 is the number at the 25th percentile and Q3 is the number at the 75th percentile.Note: All categories were significantly different for all characteristics. Categorical variables based on chi-square test, Mother’s age based

on median one way analysis, all tested at 5% significance level.

than 4% of these births were delivered at rural hospitals(N = 99,285). More rural mothers delivered in urbanthan rural hospitals (N = 148,600). Close to one-third ofall rural hospital deliveries were to mothers with anurban ZIP code (N = 34,307). Table 1 details variousdemographic and birth characteristics byhospital/mother’s location. All demographic variableswere significantly different for the 3 groups. Motherswho delivered at rural hospitals, tended to be youngerand less educated, were less likely to be a minority, hadmore pregnancy complications, smoked more duringpregnancy, had government insurance, and were morelikely to have had less than 6 prenatal visits.

Proportionally, there were more low birth weightneonates born in urban hospitals to urban mothers(6.4%) and rural mothers (6.2%) than in rural hospitals(4.3%). Fewer normal birth weight neonates were bornin urban hospitals to urban mothers (83.3%) and ruralmothers (82.2%) than in rural hospitals (85.1%). Similarpercentages of large birth weight neonates were born inurban hospitals to urban mothers (10.4%) and in rural

hospitals (10.6%), while urban hospital births to ruralmothers was higher than both (11.6%).

Neonatal Death. Table 2 shows the 1998-2002neonatal death rates per 1,000 births, stratified bypregnancy complications, delivery method, birthweight, and hospital/mother’s location. Rates ofneonatal death at rural hospitals were not statisticallydifferent to respective rates at urban hospitals for eitherrural or urban mothers when stratified by pregnancy orno pregnancy complications, vaginal or cesareandelivery, and birth weight strata that included: large,normal, and low birth weight. The only significantlyhigher death rate occurred in babies born to ruralmothers with no pregnancy complications whodelivered a normal-weight baby vaginally at an urbanhospital compared to babies born to urban mothersdelivering at an urban hospital (0.2 [95% CI 0.2-0.4]deaths per 1,000 births and 0.1 [95% CI 0.1-0.1],respectively).

The Journal of Rural Health 418 Vol. 24, No. 4

. . . . . Hospitals . . . . .

Table 2. Neonatal Deaths by Pregnancy Complication, Delivery Method, Birth Weight, andHospital/Mother’s Location, California, 1998-2002

No Pregnancy Complication Pregnancy Complications

Delivery Birth Weight Hospital/Mother’s Neonatal Deaths per Neonatal Deaths perMethod (grams) Location N 1,000 Births (95% CI) N 1,000 Births (95% CI)

Cesarean >4,000 Rural/all 2,401 0.4 (0.0-2.3) 843 1.1 (0.0-6.6)Urban/rural 4,496 0.4 (0.1-1.6) 829 0.0 (0.0-4.6)Urban/urban 62,651 0.2 (0.1-0.4) 14,936 0.9 (0.5-1.6)

2,500-4,000 Rural/all 13,699 0.4 (0.1-0.9) 5,025 0.4 (0.0-1.4)Urban/rural 22,616 0.5 (0.2-0.9) 4,038 1.2 (0.4-2.9)Urban/urban 343,162 0.4 (0.3-0.5) 80,875 1.1 (0.9-1.3)

<2,500 Rural/all 704 15.6 (7.8-28.0) 980 15.3 (8.6-25.2)Urban/rural 2,255 17.7 (12.7-24.2) 1,943 15.4 (10.4-22.0)Urban/urban 36,464 13.6 (12.4-14.8) 29,364 23.3 (21.5-25.0)

Vaginal >4,000 Rural/all 6,090 0.3 (0.0-1.2) 1,220 0.0 (0.0-3.1)Urban/rural 10,720 0.1 (0.0-0.5) 1,253 0.8 (0.0-4.4)Urban/urban 146,106 0.1 (0.1-0.2) 21,631 0.5 (0.2-0.9)

2,500-4,000 Rural/all 52,546 0.1 (0.0-0.2) 13,212 0.4 (0.1-0.9)Urban/rural 84,718 0.2 (0.2-0.4) 10,716 0.4 (0.1-1.0)Urban/urban 1,339,059 0.1 (0.1-0.1) 200,180 0.4 (0.3-0.4)

<2,500 Rural/all 1,403 17.8 (11.5-26.3) 1,161 61.2 (46.9-75.4)Urban/rural 3,078 25.7 (20.0-31.3) 1,937 64.0 (52.7-75.3)Urban/urban 54,113 27.2 (25.8-28.6) 29,010 75.9 (72.7-79.1)

Note: Bolded values represent statistically significant differences between groups within weight and delivery category at 5%significance level.

95% CI = 95% Poisson confidence interval, N is the total number of births in the strata.

Statewide, low birth weight babies sufferedsignificantly higher death rates. Vaginal, but notcesarean, deliveries with pregnancy complicationstended to have a 2- to 3-fold higher neonatal death ratecompared to those without pregnancy complications.Rates of neonatal death were higher among vaginaldeliveries compared to cesarean deliveries only for lowbirth weight infants. Further breakdown of the lowbirth weight category into <1,500 g, 1,500-1,999 g, and2,000-2,499 g showed a gradient in rates with thesmallest weight group having the highest death rates.No differences in neonatal mortality were seen betweenthe 3 hospital/mother’s groups in these smaller birthweight categories. Newborn adverse outcomes alsoshowed no differences when considered bystratification (data not shown).

Table 3 shows the results of the multiple logisticregression of neonatal death. All of the 10 factors werestatistically significant (C-statistic = 0.936; n =2,483,834). Small birth weight had the largestassociation with mortality with an odds ratio of 101.4(95% CI 91.4-112.5). Babies born to mothers with lessthan 6 prenatal visits had an odds ratio of 6.7 (95% CI6.3-7.1). Being born in a rural hospital was a protective

factor compared to urban mothers delivering in anurban hospital with an odds ratio of 0.8 (95% CI0.6-0.9).

Maternal Death. Five-year rates of maternal deathper 100,000 births are shown in Table 4. Less than 300mothers died within 72 hours of giving birth. Twelvemothers who delivered in a rural hospital, 24 ruralmothers who delivered in an urban hospital, and 255urban mothers who delivered in an urban hospital diedbetween 1998 and 2002. Urban mothers in urbanhospitals had cesarean delivery death rates significantlyhigher than mothers who delivered vaginally, whetherpregnancy complications were present or not. Amongmothers who delivered by cesarean section, maternaldeath rates were highest for rural mothers delivering atan urban hospital with pregnancy complications (117.5,95% CI 50.7-231.5). Rural mothers delivering at anurban hospital had the highest mortality rates for allstrata except cesarean deliveries with no pregnancycomplications. There was no recorded transport ofmothers prior to delivery from rural hospitals or ruralmothers delivered at an urban hospital that ended inmaternal death within 72 hours.

Hughes, Zweifler, Garza and Stanich 419 Fall 2008

. . . . . Hospitals . . . . .

Table 3. Multiple Logistic Regression of Neonatal Death

Characteristic Odds Ratio (95% CI)

Birth weight (grams) >4,000 vs 2,500-4,000 1.2 (0.9-1.6)<2,500 vs 2,500-4,000 101.4 (91.4-112.5)

Pregnancy complication Present vs absent 1.4 (1.4-1.5)Delivery method Cesarean vs vaginal 0.5 (0.4-0.5)Mother’s race/ethnicity Asian vs White 0.9 (0.7-1.0)

Black vs White 1.2 (1.1-1.4)Hispanic vs White 1.3 (1.2-1.4)Other vs White 0.9 (0.8-1.1)

Mother’s education <High School vs HS 0.9 (0.8-0.9)College vs HS 1.0 (0.9-1.1)Advanced degree vs HS 0.9 (0.8-1.0)

Type of birth Multiple vs single 1.4 (1.3-1.5)Mother’s age (y) 1.0 (1.0-1.0)Poor prenatal care <6 prenatal visits vs ≥ 6 6.7 (6.3-7.1)Delivery insurance Government vs private 0.9 (0.8-0.9)Hospital/mother’s location Rural/all vs Urban/urban 0.8 (0.6-0.9)

Urban/rural vs Urban/urban 1.0 (0.8-1.1)

Note: Bolded values are statistically significant at 5% significance level;C-statistic = 0.936; n = 2,483,834.

Table 4. Maternal Deaths by Presence of Pregnancy Complication, Delivery Method, Hospital andMother’s Residence, California 1998-2002

Maternal Deaths per 100,000 Births (95% CI)

No Pregnancy Complications Pregnancy Complications

Vaginal Cesarean Vaginal Cesarean

Rural hospital birth/all maternal residences 8.3 0 25.7 43.8(2.7-19.4) (0-21.5) (7.0-65.7) (9.0-128.0)

Urban hospital birth/rural maternal residence 9.1 10.2 28.8 117.5(4.2-17.3) (2.1-29.9) (7.8-73.6) (50.7-231.5)

Urban hospital birth/urban maternal residence 3.7 17.0 21.9 54.3(2.7-4.7) (13.1-20.8) (16.1-27.7) (41.4-67.2)

95% CI is the 95% Poisson confidence interval.

DiscussionWe found hospitals that fit the state’s definition of a

rural hospital account for only a small proportion ofdeliveries in California. We chose to use the state’sdefinition of a “small/rural” hospital since presumablythe hospitals that were small and urban would providelimited services in a fashion similar to truly ruralhospitals. While 6% of Californians live in rural areas asdefined by the 2000 US Census, less than 4% ofdeliveries occurred in rural hospitals as defined by thestate.26 Of interest, we found that close to one third ofall rural hospital births were to urban mothers. Usingthe RUCA definition, 85% of rural hospitals had a rural

ZIP code. The majority of births in rural hospitals withurban ZIP codes was in outlying high-growth areas orsmaller coastal towns. More than 75% of the urbanmothers who delivered in rural hospitals came fromZIP codes in these same areas. This may represent thedifficulty of applying a static definition to a changingenvironment.

We stratified our comparison of neonatal andmaternal mortality by birth weight and pregnancycomplications because of the possibility that ruralpregnancies involving high-risk conditions identifiedprenatally would be preferentially referred ortransferred to deliver in urban settings better prepared

The Journal of Rural Health 420 Vol. 24, No. 4

. . . . . Hospitals . . . . .

to handle complications. In spite of a larger proportionof pregnancy complications in rural than urban births,rural hospital neonatal death rates were not statisticallydifferent than rates in urban hospitals for all birthweight infants when we stratified by birth weight,delivery method (cesarean or vaginal), and pregnancycomplication (present or absent). We found the odds ofneonatal mortality to be 20% less among all ruralhospital births compared to births involving urbanmothers in urban hospitals. This may reflect successfulefforts at rural hospitals to identify pregnant patients atrisk for the more severe birth complications andtransfer them to urban hospitals. We are not able tocomment on whether the availability of rural hospitalsimproved obstetric outcomes because it is still possiblethat residual confounding from factors such as ruralhospitals being less likely to operate their own neonatalintensive care unit might make them compare morefavorably than they would otherwise.

We found that rural mothers who delivered normalbirth weight infants vaginally without pregnancycomplications at urban hospitals had higher neonatalmortality rates than urban mothers who delivered aturban hospitals. This finding is of interest becausenormal birth weight vaginal deliveries withoutpregnancy complications accounted for more than halfof all deliveries in the study population. We also foundthat close to 60% of rural mothers delivered in urbanhospitals, higher than obstetric bypass rates found inprevious studies.3,4 However, neonatal mortality ratesfor rural women who delivered at urban hospitals werenot significantly different from mothers who deliveredat rural hospitals. Based on these findings, we cannotconclude that rural mothers who outflow to urbanhospitals had worse outcomes than if they haddelivered at rural hospitals. It may be that unmeasuredvariables such as referrals of higher risk rural mothersfor care in urban areas would explain why thesemothers had higher neonatal mortality rates than urbanmothers delivering at urban hospitals.

The only significant finding from our analysis ofmaternal mortality was that urban mothers whodelivered vaginally were less likely to die than urbanmothers delivered by cesarean section when controllingfor pregnancy complications. No rural-urban hospitaldifferences in maternal mortality were observed.

Our conclusions are limited by several factors. Ourresults depend on proper coding of birth certificates,which is subject to misclassification.27 Birth certificatedata include information on maternal mortality within72 hours, which is not consistent with the World HealthOrganization’s definition of maternal death within 42days. The high percentage of rural outflow and urban“inflow” may reflect the inadequacies of ZIP code data.

Our results may not necessarily be generalized topopulations outside of California, since our analysisused State of California births that may be more or lessethnically diverse or rural than other populations. Norcan we generalize our findings to years beyond1998-2002, or control for the bias of pregnant womenself-selecting to deliver in rural or urban settings.

Our study benefited from the availability ofextensive outcome data over a 5-year period. Futurestudies might benefit from controlling for factors suchas access to prenatal care, which emerged as asignificant contributor to neonatal mortality in ourlogistic regression analysis, and potential differences inrural versus urban delivery preferences stratified byrace/ethnicity and age. Further analysis of theproviders of obstetric services (obstetrician, familyphysician, or nurse midwife) would aid in targetingtraining programs to prepare providers for ruralobstetric practices.

In summary, rural hospitals provided a relativelysmall portion of total births in California between 1998and 2002. Neonatal mortality rates were lower at ruralhospitals compared to urban hospitals after controllingfor a number of potentially confounding variables. Nodifferences were seen in rural versus urban maternalmortality rates. The encouraging outcomes of our studyshould reassure patients considering a rural hospitaldelivery, and aid policy makers and health careproviders aiming to ensure access to obstetric servicesfor rural populations.

References1. Peck J, Alexander K. Maternal, infant, and child health in rural

areas: a literature review. In: Gamm LD, Hutchison LL, DabneyBJ, Dorsey AM, eds. Rural Healthy People 2010: A CompanionDocument to Healthy People 2010. Vol 2. College Station, Tex: TheTexas A&M University System Health Science Center, School ofRural Public Health, Southwest Rural Health Research Center;2003.

2. Liu JJ, Bellamy GR, McCormick M. Patient bypass behavior andcritical access hospitals: implications for patient retention. J RuralHealth. 2007;23(1):17-24.

3. Bronstein JM, Morrisey MA. Bypassing rural hospitals forobstetrics care. J Health Polit Policy Law. 1991;16(1):87-118.

4. Radcliff TA, Brasure M, Moscovice IS, Stensland JT.Understanding rural hospital bypass behavior. J Rural Health.2003;19(3):252-259.

5. Ricketts TC. Rural Health in the United States. New York, NY:Oxford University Press; 1999.

6. Nesbitt T, Connell F, Hart LG, Rosenblatt RA. Access toobstetrical care in rural areas: effects on birth outcomes. Am JPublic Health. 1990;80(7):814-818.

7. Nesbitt TS, Larson EH, Rosenblatt RA, Hart LG. Neonataloutcomes and resource utilization in rural areas: effect of accessto maternity care. Am J Public Health. 1997;87(1):85-90.

8. Eberhardt M, Ingram D, Makuc D, et al. Urban and Rural HealthChartbook. Health, United States, 2001. Hyattsville, Md: NationalCenter for Health Statistics; 2001.

Hughes, Zweifler, Garza and Stanich 421 Fall 2008

. . . . . Hospitals . . . . .

9. Larson E, Hart L, Rosenblatt R. Is non-metropolitan residence arisk factor for poor birth outcome in the U.S.? Soc Sci Med.1997;45(2):171-188.

10. Clarke LL, Bono CA, Miller MK, Malone SC. Prenatal careuse in nonmetropolitan and metropolitan America: racial/ethnic differences. J Health Care Poor Underserved. 1995;6(4):410-433.

11. Miller MK, Clarke LL, Albrecht SL, Farmer FL. The interactiveeffects of race and ethnicity and mother’s residence on theadequacy of prenatal care. J Rural Health. 1996;12(1):6-18.

12. Clarke LL, Coward RT. A multivariate assessment of the effectsof residence on infant mortality. J Rural Health. 1991;7(3):246-265.

13. Laditka SB, Laditka JN, Bennett KJ, Probst JC. Deliverycomplications associated with prenatal care access forMedicaid-insured mothers in rural and urban hospitals. J RuralHealth. 2005;21(2):158-166.

14. Laditka SB, Laditka JN, Probst JC. Racial and ethnic disparities inpotentially avoidable delivery complications among pregnantMedicaid beneficiaries in South Carolina. Matern Child Health J.2006;10(4):339-350.

15. Coffman JM, Young JQ, Vranizan K, Blick N, Grumbach K.California Needs Better Medicine: Physician Supply and MedicalEducation in California. San Francisco, Calif: California PrimaryCare Consortium and the Center for the Health ProfessionsUniversity of California; 1997.

16. Council on Graduate Medical Education. Tenth Report: PhysicianDistrubution and Health Care Challenges in Rural and Inner-CityAreas. Washington, DC: US Dept of Health and Human Services,Public Health Service, Health Resources and ServicesAdministration; 1998.

17. Hart LG, Salsberg E, Phillips DM, Lishner DM. Rural health careproviders in the United States. J Rural Health.2002;18(suppl):211-232.

18. Cohen D, Guirguis-Blake J, Jack B, et al. Family phsyicians makea substantial contribution to maternity care: the case of the stateof Maine. Am Fam Physician. 2003;68(3):405.

19. Guirguis-Blake J, Fryer GE, Deutchman M, Green LA, Dobey SM,Phillips RL. Family physicians’ declining contribution to prenatalcare in the United States. Am Fam Physician. 2002;66(12):2192.

20. Topping DB, Hueston WJ, MacGilvray P. Family physiciansdelivering babies: what do obstetricians think? Fam Med.2003;35(10):737-741.

21. California Office of Statewide Health Planning and Development.Hospital Annual Utilization Data. Available at: http://www.oshpd.ca.gov/HID/Products/Hospitals/Utilization/Hospital_Utilization.html. Accessed January 11, 2008.

22. Avery S. Rural Health Care at Risk: California Small & RuralHospitals. Sacramento, Calif: California Healthcare Association;2001. Special Report.

23. Office of Statewide Health Planning & Development. Californiaperspectives in healthcare. Available at: http:// www.oshpd.ca.gov/HID/Perspectives/2004/2004ReadMe.pdf. AccessedJanuary 11, 2008.

24. ZIP code rural-urban commuting area code approximation. May2006. Available at: http://depts.washington.edu/uwruca/index.html. Accessed January, 2007.

25. California Department of Health Services Health and WelfareAgency. Certificate of Live Birth—VS-10A Medical DataSupplemental Work Sheet. Available at: http://www.avss.ucsb.edu/layouts/vs10a.htm. Accessed November 10, 2003.

26. U S Census Bureau. P2. Urban and Rural [6]—Universe: TotalPopulation. Available at: http://factfinder.census.gov/home/saff/main.html?_lang=en. Accessed Nov 10, 2003.

27. Green DC, Moore JM, Adams MM, Berg CJ, Wilcox LS, McCarthyBJ. Are we underestimating rates of vaginal birth after previouscesarean birth? The validity of delivery methods from birthcertificates. Am J Epidemiol. 1998;147(6):581-586.

The Journal of Rural Health 422 Vol. 24, No. 4