prepartum work, job characteristics, and risk of cesarean delivery

Prepartum Work, Job Characteristics, andRisk of Cesarean Delivery

Shirley Hung, MPP, Donna Ruane Morrison, PhD, Leslie A. Whittington, PhD*,and Sara Beck Fein, PhD

ABSTRACT: Background: Reducing the rate of cesarean deliveries in the United States is ahigh priority among public health o�cials and members of the medical community. Manyfactors known to contribute to an individual woman's risk of having a cesarean rather than avaginal delivery are not readily altered by public policy intervention. In this study we exploredthe e�ects on type of delivery of prepartum work practices, a category of factors that hasa potential to a�ect the likelihood of cesarean delivery and to be amenable to change.Methods: Data are from U.S. Food and Drug Administration's Infant Feeding PracticesStudy, using questions on mail surveys administered prenatally and at 1 month postpartum.The sample comprised 1194 women who worked during pregnancy. The outcome measure istype of delivery. Predictor variables are characteristics of prepartum work: how far into theirpregnancy the women work, number of hours worked, and occupation. Results: For mostwomen, maintaining employment through the third trimester, working long hours, andworking in certain occupations are not independently associated with the odds of having acesarean delivery. However, we found marginally signi®cant evidence that those women whoworked more than 40 hours a week in a sales job were more likely to have cesarean deliveriesthan women who worked in other occupations. Conversely, women working part-time in salesjobs were less likely to have a cesarean delivery. Conclusion: This study provides evidencethat prenatal work does not substantially increase the probability of having a cesareandelivery in most occupational categories. (BIRTH 29:1 March 2002)

A major public health objective of the United Statesgovernment is to reduce the proportion of births thatare cesarean deliveries rather than vaginal births (1,objective 16-9). The 1970s saw a 300 percent increasein cesarean section rates from 5.5 percent in 1970 to16.5 percent by the end of the decade (2). That rate

continued to climb during the 1980s before peakingat 24.7 percent in 1988 (3). In 1999, the overall UnitedStates cesarean rate stood at 22 percent, and theprimary cesarean rate was 15.5 percent (4).

The dramatic rise in cesarean delivery rates sincethe 1970s has resulted in a substantial body ofliterature seeking to explain the increase. Largelyabsent from the published research is an examinationof a mother's employment circumstances and thepossibility that her prepartum work practices couldin¯uence her type of delivery. This is a striking gapbecause the marked rise in the cesarean section ratesince the 1970s virtually mirrors the mass entry ofmarried women into the labor force generally (5), andspeci®cally, the trend toward sustaining prepregnan-cy work practices through pregnancy. One study thatexamines this issue found that women who wereemployed outside the home faced an increasedprobability of cesarean section, relative to thosewho did not work outside the home (6).

Shirley Hung is from Cable News Network, Washington, DC. DonnaMorrison and Leslie Whittington are from the Georgetown PublicPolicy Institute, Georgetown University, Washington, DC. Sara Feinis from the Division of Market Studies, Food and Drug Administra-tion, U.S. Department of Health and Human Services, College Park,Maryland.

* See Tribute at the end of the paper.

Address correspondence and requests for reprints to Sara B. Fein,PhD, Room 2C-103, HFS-727, Center for Food Safety and AppliedNutrition, Food and Drug Administration, 5100 Paint BranchParkway, College Park, Maryland 20740-3835.

Ó 2002 Blackwell Science, Inc.

10 BIRTH 29:1 March 2002

Most research about prepartum work has focusedon maternal employment in relation to potentialoutcomes for the infant, in particular, prematurebirth and low birthweight, and much of this literatureis based on data from countries other than the UnitedStates (7±9). The job characteristics related to infantoutcomes provide a list of factors that may also a�ectdelivery method. Those found to a�ect outcomenegatively include working late into the pregnancy(10); working long hours (11,12); working in sales,service, or manufacturing (13); prolonged standing(14); and heavy physical e�ort (15). In addition,evidence suggests that certain combinations of pre-partum work practices may have to be in place forbirth outcomes to be adversely a�ected, such as thecombination of working late into the pregnancy andworking consistently in high-demand but low-controlemployment (16). Given that some characteristics ofprenatal work have been found to a�ect infantoutcomes negatively, the possibility exists that somematernal job characteristics a�ect delivery mode byincreasing the risk of cesarean section.

In this study we investigated the following threehypotheses: ®rst, sustaining maternal employmentlate into the pregnancy increases the likelihood ofcesarean delivery; second, a positive relation existsbetween prepartum hours worked and the likelihoodof cesarean delivery; and third, certain occupationsare associated with an increased likelihood of cesar-ean delivery.

In addition, we explored the possibility thatcombinations of certain work characteristics havesigni®cant e�ects on the method of delivery. Inparticular, we assessed the impact of sustainingmaternal employment late into pregnancy in con-junction with working long hours and the e�ect ofworking long hours in an occupation that requiresprolonged standing.

Methods

Sample

The data for this study are from the Food and DrugAdministration's Infant Feeding Practices Study,which followed mother-infant pairs from late preg-nancy through the infant's ®rst year of life. TheInfant Feeding Practices Study sample was drawnfrom an existing national consumer mail panel of500,000 households. Beginning in February 1993, allwomen in the panel who had reached the thirdtrimester of pregnancy were mailed prenatal ques-tionnaires. Intake continued through July 1993, whenthe desired sample size was reached. Of the 3,155women who were mailed a prenatal questionnaire,

2,615 were considered eligible for this study and 2,358completed the prenatal questionnaire. Respondentswere declared ineligible for administrative reasons,such as a due date over 3 months away, or for healthreasons that arose after delivery, including birth-weight of less than 5 pounds, multiple births, infantstayed in intensive care for more than 3 days, motherwas too ill to feed the infant (in any way) for morethan 1 week, or death of mother or infant.

For the purposes of this study, data from theprenatal questionnaire, the birth screener, and theneonatal questionnaire are used, in addition todemographic data. The prenatal survey includedquestions about prenatal health care and employ-ment, maternity leave, expectations regarding post-natal employment, and certain behavior, such assmoking. The birth screener established the infant'sbirth date and the state of maternal and infant health.The neonatal questionnaire asked about pregnancycomplications and characteristics of the labor anddelivery.

Because this study was intended to look atpatterns of work during pregnancy, the subsampledrawn from the data set has a base of 1,678 womenwho were employed during pregnancy. Cases wereeliminated because of missing data on the dependentvariable, method of delivery (n � 413), almost en-tirely caused by failure to return the neonatalquestionnaire, or missing data on one or moreindependent variables (n � 71). To preserve as manycases as possible, means were imputed on thefollowing independent (continuous) variables: ciga-rettes smoked per day (53 cases), body mass index (26cases), and annual income (4 cases). Descriptivestatistics for the ®nal sample of 1,194 women arepresented in Table 1.

An important limitation of these data is that theyare not nationally representative. Compared with thenational probability sample of the 1988 NationalMaternal and Infant Health Study, respondents wereless likely to be from lower income and educationgroups than the United States population of mothersand were more likely to be white, older, and married.

A comparison of the women in the ®nal analysissample with those 413 women for whom method ofdelivery was unknown was revealing. The twogroups did not di�er markedly with respect to mostof the measured characteristics. A larger percentageof the excluded women came from the South thanthe analysis sample (34.6% compared with 28.5%,respectively), however, and a greater percentage hadno previous children (54.4% compared with 44.6%).Both of these characteristics are believed to increasethe likelihood of cesarean delivery, suggesting thatthe prevalence of cesarean section in the analysis

BIRTH 29:1 March 2002 11

sample may have increased had the 413 women beenincluded in the study.

Variables

The dependent variable was method of delivery, adichotomous variable for which the possible valuesare cesarean delivery or vaginal birth. The keyindependent variables measure characteristics of thewoman's work in the months preceding childbirth.Month stopped work is a series of dummy variablescoded as listed in Table 1. These variables arebased on a question asked on the prenatal ques-tionnaire, typically administered during the seventhmonth of pregnancy. Consequently, if a womanplanned to stop work later in her pregnancy thanthe seventh month, her response actually measuredher expected departure from the workforce. Thisserved as a proxy for her actual departure in ouranalyses.

Hours worked (prepartum) and occupation are alsoseries of dummy variables as listed in Table 1. Hoursworked were captured at one point prenatally andwere assumed to remain constant. Women wereasked to classify their positions in one of 9 standardoccupational categories. We collapsed these into 5categories as listed in Table 1 and also included amarker for occupation unknown.

Table 1. Descriptive Statistics of 1,194 United StatesWomen Who Were Employed During Pregnancy, 1993

VariablePercent

of SampleMean(SD)

Dependent variableMethod of deliveryVaginal birth 77.1Cesarean section 22.9

Background characteristicsAge 28.7 (4.6)Marital statusMarried 89.3Singlea 10.7

Education levelHigh school or less 27.7Some college 36.1Completed college 20.6Graduate school 15.6

Annual household income ($) 41,719.54*(23,816.25)

RegionCentral 32.6Northeast 19.8South 28.5West 19.2

Table 1. Continued

VariablePercent

of SampleMean(SD)

RaceWhite 96.5Nonwhite 3.5Unknown 1.3

ParityNo other children 44.6Other children 55.4

Health statusCigarettes/day 1.4* (4.7)Body mass index 24.5* (5.3)Pregnancy complicationsGestational diabetes 6.3High blood pressure 9.0Toxemia 2.9Vaginal bleeding 2.7

Birthweight of baby (lb) 7.8 (1.1)

Health care servicesPrenatal care timingBegan before wk 13 94.8Began after wk 13 5.2

Prenatal care settingPrivate physician's/

midwife's o�ce78.9

Health maintenanceorganization

7.9

Otherb 13.2

Maternity leavePaid or unpaid leaveAvailable 69.3Not available 30.7

Prepartum work practicesMonth stopped workLess than 3 mo 7.13±5 mo 6.36±8 mo 11.4Over 8 mo 25.2No plans to stop 50.0

Hours worked1±19 hr/wk 13.120±40 hr/wk 60.7Over 40 hr/wk 26.1

OccupationProfessional,

executive,administrative

35.6

Clerical 23.5Sales 8.8Service 10.0Otherc 9.5Unknown 12.6

* The means and standard deviations reported here exclude thosecases for which missing values were imputed.a Widowed, divorced/separated, or never married; b clinic, communityhealth center, public health department, or birth center; c craft/repair,operator/laborer, technician, farming/forestry.


Intensity and duration of work measures the com-bination of working a certain number of hours perweek and the duration for which that level ofintensity is sustained: work more than 40 hours aweek throughout the pregnancy (as proxied by noplans to stop); work more than 40 hours a week butstop before childbirth; work 40 hours a week or lessthroughout the pregnancy; and work 40 hours a weekor less and stop before childbirth. It was hypothesizedthat those with the highest level of work inten-sityÐthat is, those who worked more than 40 hours aweek throughout pregnancyÐwould have the great-est risk of cesarean section, when controlling forother factors.

It is possible that occupations characterized byprolonged standing, such as sales, and those charac-terized by some higher level of physical e�ort, namelythose occupations labeled ``other,'' are associatedwith an increased risk of cesarean birth. Theseinteractions are measured by hours of work andoccupation dummy variables.

The control variables are drawn from the existingliterature on determinants of cesarean section riskand fall essentially into three broad categories:background characteristics, health status of themother and the payment for and delivery of herhealth care. The following background characteristicsare included in the analyses: woman's age, maritalstatus, education level, annual household income,region of residence, race, and parity. Except for ageand household income, which are measured ascontinuous variables, the background variables areincorporated into the models as dummy variables.

The health status variables are as follows: thenumber of cigarettes smoked per day during preg-nancy, prepregnant body mass index, infant birth-weight, and several pregnancy complications, such asgestational diabetes, high blood pressure, toxemia(pre-eclampsia or eclampsia), and vaginal bleeding.These pregnancy complications have been linked tocesarean rates above 30 percent. Two variables relateto the kind of health care women receive: whenwomen began prenatal care visits (before or after13 weeks) and the setting where they receivedprenatal care (categories are listed in Table 1). Settingof prenatal care was included in the model as animperfect proxy for health insurance type, which wasnot measured in this data set.

A ®nal variable, availability of maternity leave,which included both paid and unpaid leave, wasmeasured as a dummy variable. This variable wasincluded under the assumption that the availability ofleave may factor into the decision (by physician andmother) to have a cesarean delivery rather than avaginal birth.

Statistical Analysis

Because the dependent variable, method of delivery, isdichotomous, logistic regression was employed in thisanalysis. We used SAS 6.10 (17) to estimate threeseparate models. Each predicted the adjusted oddsratio of having a cesarean section and incorporatedthe same control variables. Model 1 was the baselinemodel and explored the independent e�ects, if any, ofmonth stopped work, hours worked, and occupation onthe odds of having a cesarean section. In models 2 and3 we explored the joint impact of some of thevariables on the likelihood of cesarean section.

Results

The method of delivery was cesarean section for 23percent of the sample (Table 1). The results, presentedin Table 2, indicate no increased risk of a cesareansection for women who worked up to the date ofdelivery. The distribution of method of delivery byhours of work suggests a positive relation: amongwomen who worked less than 20 hours a week, 16.6percent have cesarean births, compared with 23percent who worked 20 to 40 hours a week, and 26percent more than 40 hours a week (data not shown).

Table 2. Model 1ÐIndependent E�ects of Prepartum WorkPractices on Type of Delivery (n = 1,194)

Work PracticesAdjusted

Odds Ratioa 95% CI

Month stopped workb

3±5 mo 0.98 0.44±2.196±8 mo 0.75 0.37±1.56Over 8 mo 0.88 0.74±1.04No plans to stop 0.71 0.37±1.35

Hours workedc

20±40 hr/wk 1.18 0.71±1.95Over 40 hr/wk 1.18 0.67±2.07

Occupationd

Clerical 0.90 0.58±1.38Sales 0.67 0.37±1.23Service 1.18 0.68±2.03Other 1.05 0.61±1.77Unknown 0.55 0.30±1.04

± 2 log likelihood 1183.62Model chi-square 102.67*Percent correctly classi®ed 68.40

* p < 0.001. p < 0.05.a Analysis controlled for age; marital status; education; householdincome; race; parity; cigarettes smoked per day; body mass index;pregnancy complications: gestational diabetes, high blood pressure,toxemia, and vaginal bleeding; infant's birthweight, prenatal caretiming, prenatal care setting, and availability of maternity leave.Reference categories are as follows: b< 3 mo; c 1±19 hr/wk;d professional, executive, or administrative.


Table 3. Model 2ÐJoint E�ects of Month Stopped Work and Number of Hours Worked on Type of Delivery (n = 1,194)

Independent Variable Adjusted Odds Ratio 95% CI

Background characteristicsAge 1.04à 1.00±1.08Marital statusa

Single 0.96 0.58±1.55

Education levelb

Some college 0.98 0.68±1.42College 0.77 0.47±1.26Graduate school 0.82 0.50±1.35

Annual household income ($) 1.00à 0.99±1.01Regionc

Northeast 0.93 0.60±1.42South 1.51à 1.04±2.19West 0.76 0.49±1.20

Raced

Nonwhite 0.78 0.35±1.74Unknown 0.87 0.22±3.36

Paritye

Other children 0.63 0.46±0.86

Health statusCigarettes/day 1.02 0.98±1.06Body mass index 1.06* 1.04±1.08Pregnancy complicationsf

Gestational diabetes 2.07 1.22±3.52High blood pressure 1.11 0.66±1.84Toxemia 2.51à 1.10±5.72Vaginal bleeding 1.51 0.65±3.50

Birthweight of baby (lb) 1.21 1.05±1.39

Health care servicesPrenatal care timingg

Began after wk 13 1.61 0.86±3.03

Prenatal care settingh

Health maintenance organization 0.75 0.43±1.33Other 0.80 0.50±1.28

Maternity leavePaid or unpaid leavei

Not available 0.88 0.61±1.27

Prepartum work practicesHours worked and month stopped workj

More than 40 hr/wk with plans to work until birth 0.95 0.62±1.45More than 40 hr/wk with plans to stop work 0.62§ 0.04±1.04Less than 40 hr/wk with plans to work until birth 0.64à 0.45±0.91

Occupationk

Clerical 0.91 0.59±1.39Sales 0.66 0.358±1.21Service 1.14 0.66±1.99Other 1.08 0.64±1.84Unknown 0.53à 0.29±0.98

)2 log likelihood 1177.98Model chi-square (improvement) 108.32*Percent correctly classi®ed 68.80

* p < 0.001. p < 0.01.à p < 0.05.§ p < 0.10.Reference categories are as follows: a married; b less than or completed high school; c central; d white; e no other children; f without statedpregnancy complication; g began before wk 13; h private physician's or midwife's o�ce; i leave available; j less than 40 hr/wk with plans to stop;k professional, executive, or administrative.


However, the estimated results of model 1(Table 2), which controlled for the e�ects of otherfactors, show that working more than 40 hours aweek during a woman's pregnancy does not increaseher odds of delivering by cesarean section. There isalso no increased risk of cesarean delivery for womenin any particular occupation, although those withunknown occupation have signi®cantly reduced odds.

The second and third models (Tables 3 and 4)explored the possibility that combinations of the keywork variables might have an impact on a woman'slikelihood of having a cesarean section. The questionasked in the second model was whether a woman whoworked straight through her pregnancy at a pace ofmore than 40 hours a week faced an increased risk ofcesarean delivery. Women are grouped as listed inTable 3. The results of this analysis suggest thatwomen with the most intense work e�ort did nothave an increased risk of cesarean delivery, comparedwith those with the least work e�ort. Surprisingly,both groups of women with a middle level of workintensityÐthat is, women who worked more than40 hours a week but stopped at some point andthose who worked less than 40 hours a week butmaintained employment throughout their preg-nancyÐwere less likely than women with the leastwork e�ort to undergo a cesarean. One likelyexplanation is that those women with the least worke�ortÐthose who worked less than 40 hours a weekand stopped before childbirthÐmight have had ahealth condition, other than those measured, thatprevented them from working more intensely andthat condition also made them more susceptible tocesarean delivery.

The control variables performed largely as expec-ted, and the values of their statistics are similar in allthree models; hence, they are shown only for model 2.The control variables that consistently emerge in theliterature as being signi®cantly related to method ofdelivery are also signi®cant in this analysis. Amongbackground characteristics, age increases the oddsthat a woman will undergo a cesarean section; womenwho live in the southern United States are 50 percentmore likely to undergo a cesarean than women livingin the central United States, and women who alreadyhave at least one other child are 0.63 times as likely, or37 percent less likely, than women who have no otherchildren to deliver the new child by cesarean section.

Health status is the other category of controlvariables that appears to be important in explaining apropensity toward cesarean delivery. We found apositive relationship between a woman's body massindex and the odds of having a cesarean section. Apositive relationship also exists between the birth-weight of a woman's baby, which physicians havesome knowledge of before birth, and method ofdelivery. Of the pregnancy complications examined inthis analysis, both gestational diabetes and toxemiaincreased the odds of having a cesarean birth.

Model 3 was structured to test whether or notworking long hours in certain occupations increaseda woman's odds of having a cesarean section. Inparticular, we examined the impact of working morethan 40 hours a week in a sales occupation. Hours ofwork and occupation comprise a four-level variablethat combines working more or less than 40 hours aweek with working in sales or in an occupation otherthan sales. We found weak evidence of an increased

Table 4. Model 3ÐJoint E�ects of Number of Hours Worked and Occupation on Type of Delivery (n = 1,194)

Work Practices Adjusted Odds Ratioa 95% CI

Month stopped workb

3±5 mo 1.07 0.48±2.406±8 mo 0.86 0.42±1.73Over 8 mo 1.00 0.52±1.91No plans to stop 0.83 0.44±1.55

Hours worked and occupationc

More than 40 hr/wk in sales 2.59à 0.92±7.31More than 40 hr/wk not in sales 0.93 0.67±1.30Less than 40 hr/wk in sales 0.45 0.22±0.92

)2 log likelihood 1182.36Model chi-square (improvement) 103.98*Percent correctly classi®ed 68.00

* p < 0.001. p < 0.05.à p < 0.10.a Analysis controlled for age; marital status; education; household income; race; parity; cigarettes smoked per day; body mass index; pregnancycomplications: gestational diabetes, high blood pressure, toxemia, and vaginal bleeding; infant's birthweight; prenatal care timing; prenatal caresetting, and availability of maternity leave. Reference categories are as follows: b less than 3 mo; c less than 40 hr/wk not in sales.


risk of cesarean section associated with working longhours in sales (Table 4). Women who worked lessthan 40 hours a week in sales, however, were 55percent less likely than their counterparts in otheroccupations to have a cesarean delivery.

Discussion and Conclusions

This study adds to the body of research on method ofdelivery by exploring some characteristics that, unlikemany factors associated with high cesarean sectionrisk, are within the reach of public policymakers,health care professionals, and the women themselvesto modifyÐnamely, prepartum work practices. The®ndings suggest that for most women, working duringpregnancyÐeven at intense levels and throughout thethird trimesterÐdoes not put them at greater risk for acesarean delivery rather than a vaginal birth.

For a small population of women, however,prepartum work of high physical intensity may beinadvisable. Speci®cally, women working in jobs thatrequire prolonged standing, measured in this study asworking in sales, appear to face a slightly greater riskof cesarean delivery. Coupled with previous ®ndingsof increased risk of unfavorable infant outcome forsuch women, this result suggests that prepartum workpractices may need to be adjusted to limit the numberof hours spent standing in the later months ofpregnancy.

When considering the results of this study, it isimportant to recognize several limitations in the datathat may a�ect their robustness. Because the InfantFeeding Practices Study was not designed to answerthe speci®c question investigated by this research,some of the work characteristics are not measured inthe most accurate and meaningful way. The measuresavailable, such as intended duration of work duringpregnancy for duration of work, and work in sales forwork requiring prolonged standing, may result inmisclassi®cation of some of the work characteristics.Also of concern is the lack of information on severalvariables that the literature suggests may beimportant determinants of method of delivery, suchas clinical indications for cesarean delivery andde®nitive measures of health care insurance andother hospital- and physician-related characteristics.Furthermore, the elimination from the study ofwomen who had any adverse outcome of pregnancyalmost certainly decreased the number of observedcesarean deliveries. In addition, it is possible thatwomen may have had health conditions other thanthose measured that simultaneously limited theirability to participate in the labor market andincreased their risk of cesarean delivery. Finally, thesample underrepresents the lower socioeconomic

groups. Therefore, the ®ndings may not apply topregnant women of low socioeconomic status, whomay have a di�erent distribution of jobs within thejob categories used in the study and who probablyhave fewer resources to help counter the stress ofworking while pregnant.

Nevertheless, the ®ndings of this study o�er somelevel of reassurance to working women, as well as totheir employers and public policymakers concernedwith the di�cult work-family balance. It appears thatparticipation in the labor market by pregnant womendoes not expose them to a greater likelihood ofundergoing a cesarean section, and therefore, tothe added costs and risks associated with such aprocedure.

Acknowledgments

No additional funding or support was necessary forthis project. We are grateful for the helpful commentsof two anonymous reviewers.

References

1. U.S. Department of Health and Human Services. Healthy

People 2010. 2nd ed. With Understanding and Improving

Health and Objectives for Improving Health. Objective 16-9.

2 vols. Washington, DC: U.S. Government Printing O�ce,

2000.

2. Goldfarb M. Who Receives Cesareans: Patient and Hospital

Characteristics. Washington, DC: National Center for Health

Services Research, 1984.

3. Clarke S, Ta�el S. Changes in cesarean delivery in the United

States: 1988 and 1993. Birth 1995;22:63±67.

4. Ventura S, Martin J, Curtin S, Menacker F, Hamilton B.

Births: Final data for 1999. Natl Vital Stat Rep 2001;49(1):

1±99.

5. Bianchi S, Spain D. Women, work, and family in America.

Population Bull 1996;51(3):2±47.

6. Tussing AD, Wojtowycz M. The cesarean decision in New

York State, 1986: Economic and noneconomic aspects. Med

Care 1992;30:529±540.

7. Berkowitz G. Employment-related physical activity and

pregnancy outcome. J Am Med Women Ass 1995;50(5):167±

169, 174.

8. Gabbe S, Turner LP. Reproductive hazards of the American

lifestyle: Work during pregnancy. Am J Obstet Gynecol

1997;176:826±832.

9. Henriksen TB, Savitz D, Hedegaard M, Secher NJ.

Employment during pregnancy in relation to risk factors

and pregnancy outcome. Br J Obstet Gynaecol 1994;101:

858±865.

10. Fortier I, Marcoux S, Brisson J. Maternal work during

pregnancy and the risks of delivering a small-for-gestational-

age or preterm infant. Scand J Work Environ Health

1995;21:412±418.

11. Armstrong BG, Nolin AD, McDonald AD. Work in preg-

nancy and birth weight for gestational age. Br J Indust Med

1989;46:196±199.

12. Peoples-Sheps M, Siegel E, Suchindran C, et al. Charac-

teristics of maternal employment during pregnancy: E�ects


on low birthweight. Am J Public Health 1991;81:1007±

1012.

13. Saurel-Cubizolles MJ, Subtil D, Kaminski M. Is preterm

delivery still related to physical working conditions in

pregnancy? J Epidemiol Community Health 1991;45:29±34.

14. Henriksen TB, Hedegaard M, Secher NJ. Standing and

walking at work and birthweight. Acta Obstet Gynecol Scand

1995;74:509±516.

15. Spinillo A, Capuzzo E, Baltaro F, et al. The e�ect of work

activity in pregnancy on the risk of fetal growth retardation.

Acta Obstet Gynecol Scand 1996;75:531±536.

16. Brett K, Strogatz D, Savitz D. Employment, job strain, and

preterm delivery among women in North Carolina. Am J

Public Health 1997;87:199±204.

17. SAS Institute Inc. Windows Version 6.10. Cary, North

Carolina: Author, 1996.

Tribute to Leslie A. Whittington

Leslie A. Whittington was killed in the terrorist attack on the Pentagon onSeptember 11, 2001, while on her way to a sabbatical in Australia with herhusband and two young daughters. Dr. Whittington was an Associate Professorat the Georgetown Public Policy Institute of Georgetown University in Washington,DC. She was a dynamic and gifted teacher and researcher who consistently inspiredthose around her to produce better work than they had imagined was possible.


prepartum work, job characteristics, and risk of cesarean delivery

Documents