Objectives. We derived trajectories of the substantive complexity (SC) of work across mid-adult life in women and determined their association with term birth weight. SC is a concept that encompasses decision latitude, active learning, and ability to use and expand one’s abilities at work.

Methods. Using occupational data from the National Longitudinal Survey of Youth 1979 and O*NET work variables, we used growth mixture modeling (GMM) to construct longitudinal trajectories of work SC from the ages of 18 to 34 years. The association between work trajectories and birth weight of infants born to study participants was modeled using generalized estimating equations, adjusting for education, income, and relevant covariates.

Results. GMM yielded a 5-class solution for work trajectories in women. Higher work trajectories were associated with higher term birth weight and were robust to the inclusion of both education and income. A work trajectory that showed a sharp rise after age 24 years was associated with marked improvement in birth weight.

Conclusions. Longitudinal modeling of work characteristics might improve capacity to integrate occupation into a life-course model that examines antecedents and consequences for maternal and child health.

Low birth weight (LBW) and preterm delivery have long been acknowledged as a source of infant and early childhood morbidity, mortality, and increased health care costs. Additionally, life-course epidemiological evidence supports associations between these birth outcomes and increased risk for adult and chronic disease. Intervention on proximal causes—adverse maternal exposures including smoking, alcohol, poor nutrition, and inadequate prenatal care—have partly but not wholly reduced LBW incidence. Alleviation of more distal material factors has been proposed to relieve the burden of adverse birth outcomes; these factors include socioeconomic status (SES) measures, principally, education, income, and occupational standing associated with LBW.1–6

In general, studies of work (as an SES indicator) and pregnancy indicate an overall protective effect of working,7–10 and change in working status to inadequate employment11 has been considered more deleterious than working conditions. Poerksen and Petitti8 found that employment status explained more of the variance in LBW than did educational mobility in the Alameda County study. Receipt of unemployment benefits was similarly associated with a small increased risk for small-for-gestational age.12

Women’s health studies operationalize occupation in several forms but traditionally use social–hierarchy models of managerial responsibility versus manual work,13 societal perceptions of occupational prestige, or hierarchical classification schemes based on manual or nonmanual differences.14 Investigation into the association between reproductive outcomes and work has begun to focus on contrasts in tasks entailed within a set of jobs.

Job strain, the adverse combination of high working demands with low job control, as described by Karasek,15,16 has demonstrated a variable relationship with poor pregnancy outcome.17–24 The components of low job control, skill discretion, and decision latitude at work, may be better predictors of adverse outcomes than overall job strain.24,25 Health effects associated with the social-class standing of occupation are attenuated by job control, suggesting the mediating influence of this construct.26

This investigation was based on our previous findings of differential work trajectories by race/ethnicity across early and mid-adult life and evidence that these were associated with LBW.25 This initial work was only suggestive, however, because the trajectory patterns we observed were based on cross-sectional, rather than longitudinal, data. Our aim was to derive a better quantitative description of these pathways and their association with birth outcomes. We used data from the National Longitudinal Survey of Youth 1979 (NLSY79), a long-running survey of employment and health, to derive trajectories of the substantive complexity (SC) of work across mid-adult life among women. SC is a concept derived from the work of Kohn and Schooler27–29 that encompasses decision latitude, active learning, and ability to use and expand one’s abilities at work. It represents a similar construct to control over work but was chosen in preference because of a more robust portrayal of occupational characteristics in a service-based economy.30–33 Preparatory to this study, we validated the association of a measure of SC with self-rated health and hypertension and examined the role of work trajectories with incident hypertension.32,33 Having established these, we hypothesized that work trajectories in women were associated with differentials in birth weight, net of other SES influences.

The NLSY79 is a nationally representative sample of 12 686 participants (6283 females) that was initiated and is maintained by the US Bureau of Labor Statistics. Its primary objective is to assess the labor market activities and significant life events of a cohort of adolescents and young adults across time.

Initiated in 1979, the NLSY79 enrolled participants residing in US households who represented a birth cohort from 1957 to 1964. Sample characteristics were extensively described.33 Between 1979 and 1994, participants were interviewed annually, after which interviews were biennial. Retention rates were greater than 80% across the 25 years (1979–2004) of the survey used here.

Occupational Information Resource Center

The successor to the Dictionary of Occupational Titles (DOT), the Occupational Information Resource Center (O*NET) publishes new occupational descriptors based on worker survey data on skills, generalized work activities, work context, and knowledge. A set of 225 variables that describe these work domains have been developed and are published online.30,34–36

The O*NET supplants the DOT model based upon expert opinion and aggregates similar jobs by type of work tasks and their requisite education, skills, or training. More than 1000 occupations are described in detail in the version (version 13.0, June 2008) of the O*NET used here and can be used to classify nearly every job outside of those in military service.

Measures
Independent variables.

The primary independent variable of interest was individual yearly participant scores for the SC of work. To create a metric for work SC from O*NET variables, a factor analysis was performed as outlined previously.30,32 This resulted in a 3-factor solution describing SC, people versus things, and physical demands. These factors were scaled using regression methods to derive z-scores (mean 0, standard deviation ±1) and retaining variables with loadings greater than 0.7. O*NET variables with the highest loadings on the SC factor were deductive reasoning, inductive reasoning, critical thinking, analyzing information, and complex problem solving. SC scores were linked by crosswalks, via O*NET job codes, to census occupational codes for NLSY79 participants’ job data.35

Annual or biennial job data (up to 5 jobs per year) were downloaded from the NLSY79 data set and recoded by age so that longitudinal patterns were uniform by age beginning at age 18 years. SC scores, linked by 1980 or 2000 census occupational codes, were imputed to the job histories extracted from the NLSY79 data set, and the mean work SC score was calculated at each age point for each participant.

Educational attainment was measured at each survey in the NLSY79. We used the level of education recorded closest to the time of each delivery in the analysis. Educational level was defined and recoded as an ordinal variable with 4 categories that corresponded to discrete metrics of educational attainment: non–high school graduate, high school graduate, some (1–3 years) college, and college graduate or beyond.

Outcome variables.

The principal outcome of interest was birth weight for each live singleton term delivery in the NLSY database from age 18 years through the 2004 survey. Beginning in 1983, a fertility section in the NLSY included additional questions about each pregnancy experienced by the participants (including gestational age at birth and child’s birth weight and size) and participants' health and health care during pregnancy (alcohol and tobacco use in pregnancy, time, number, and visits for prenatal care, and additional details about the birth). Maternal recall for birth data in other data sets was demonstrated to be reliable after much longer intervals.37,38 Multiple births in the same pregnancy (e.g., twins) were excluded from the analyses. Birth weight was recorded as a continuous variable, and preterm births, defined as delivery before 37 weeks' gestation, were excluded.

Other variables.

Codes from the NLYS79 were used for self-identified race/ethnicity. Foreign-born status was obtained from the initial (1979) NLSY interview. Individual income (wages and other income) was obtained annually or biennially from the NLSY data set and inflation-adjusted to 1980 dollar values for a consistent metric across years.39 Income was matched to year of each delivery and natural-logarithm (Ln) transformed to approximate a normal distribution. A smoking variable was constructed from the fertility survey; participants who indicated smoking tobacco during the year of the pregnancy or delivery were coded as current smokers.

Analyses
Longitudinal modeling of occupational trajectories.

Growth mixture modeling (GMM) was used to construct longitudinal trajectories of work characteristics across the time period of the multiple waves of the NLSY79, using the mean SC score for jobs held every 2 years from ages 18 to 34 years. GMM allows for the estimation of growth trajectories based upon repeated measures across time. This modeling assumes an overall survey population composed of a finite set of unobserved subpopulations that differ in their initial status and trajectory (slope).40,41 These subpopulations are hereafter referred to as classes in keeping with standard terminology on the subject. Growth curve modeling estimates latent trajectories that describe intraindividual patterns of change in work SC over time. The modeling then reduces interindividual heterogeneity in growth patterns to describe a probable set of group trajectory classifications in the absence of a priori assumptions as to how the data may be split. The purpose of GMM analysis is to (1) estimate the number and size of trajectory classes and (2) assign latent class membership to individuals in the population based on the posterior probability of each individual’s growth trajectory fitting with 1 of the set of estimated classes.

A GMM using mean work SC in NLSY79 female participants biennially from ages 18 to 34 years was constructed using the mixture modeling function in Mplus Version 6.11 (Muthen and Muthen, Los Angeles, CA). Our modeling of NLSY79 data disclosed markedly differing trajectory patterns for men and women.33 Models from 1 to 7 classes were constructed to test and contrast the model fit. A quadratic growth model was specified for both substantive and statistical reasons because trajectory slopes were steeper in the participants’ early to mid-20s, after which a leveling off was observed, consistent with early advancement and later career equilibrium. Little substantive information was gained by constructing trajectories past participants’ mid-30s. Determination of the best-fitting trajectory class solution was based on 3 factors: adjusted Bayesian Information Criterion (aBIC) scores for each model; the Vuong-Lo–Mendell–Rubin likelihood ratio test (VLMR); and the entropy score, a measure of classification certainty. We used aBIC with a smaller-is-better indication of model fit. The VLMR statistic compares an estimated model with a model with 1 less class; smaller P values indicate that the estimated model is an improvement from the preceding (fewer class) model. Entropy estimates classification precision from 0 (completely random) to 1.0 (perfect classification).

The association of occupational SC class with birth weight was modeled using generalized estimating equations (GEEs) that accounted for the nonindependence of births nested within mothers. The lowest-intercept, lowest-slope class was used as a referent category for contrasts in the model. Models incorporated maternal age, education, income, race/ethnicity, smoking, foreign-born status, and prenatal care as covariates. GEE analyses were performed in SPSS version 20 (SPSS/IBM Analytics 2012, Armonk, NY).

Construction of occupational histories and imputation of work SC scores in participants aged 18 to 34 years across the surveys from 1979 to 2004 yielded 3292 women (52.4% of female NLSY79 participants) with at least 2 data points on occupation during that period. Of these 3292 women participants, 2602 (79%) delivered a total of 5731 liveborn infants from age 18 years to the 2004 survey, at which point the age range for NLSY79 participants was between 40 and 47 years.

GMM of work SC yielded a 5-class solution as the best-fitting model. Although the maximum-likelihood estimation approach of GMM could model trajectories from all participants with at least 1 data point,40,41 the gains in model fit using 2 or more data points led us to exclude participants with only 1 year of recorded occupation. Analysis of aBIC and entropy across the range of class specifications, from 1 to 7, disclosed an inflection point at 5 classes, beyond which there was no distinct reduction in aBIC. Entropy, the degree to which class assignment was stable, showed an initial expected decline with increasing class specification, with stability (0.632) at a 5-class solution. The P value for the VLMR likelihood test was of borderline significance (P = .10) for a 5-class solution compared with 4 classes, whereas solutions with classes beyond 5 were uniformly nonsignificant (P = .3 and higher).

The resultant 5-class trajectories are shown in Figure 1, denoted by their relative position at age 34 years, with relevant demographic variables for each class and the entire group of participants in Table 1. The model partitioned participants into a low-intercept, downward-slope class (class 1, 12.8% of participants); a more normative flat-trajectory class comprising nearly one half (44.2%) of the participants (class 2); 2 classes with higher intercepts (class 5, 24.9%; and class 3, 11.7%); and finally one class, class 4, which had an initial flat trajectory followed by a sharp upward slope beginning at age 24 years. Black participants were underrepresented in the high-trajectory class 5. Educational attainment at delivery was unequally distributed by class. College-educated participants were found disproportionately in class 5, whereas classes 1 and 2 were composed of participants in whom the majority of individuals achieved at most a high school education. Education at delivery was generally stable: only 13% increased by 1 category or more from ages 20 to 34 years, the majority being individuals who completed a college degree. Nonsmokers were disproportionately represented in classes 3 and 5.

Table

TABLE 1— Demographic Characteristics of Participants and Birth Outcomes, by Occupational Trajectory Class: National Longitudinal Survey of Youth 1979, United States, 1979–2004

TABLE 1— Demographic Characteristics of Participants and Birth Outcomes, by Occupational Trajectory Class: National Longitudinal Survey of Youth 1979, United States, 1979–2004

VariableClass 1Class 2Class 3Class 4Class 5Total
No. of participants (% of total)333 (12.8)1150 (44.2)304 (11.7)166 (6.4)649 (24.9)2602
At first delivery after age 17 y
 Age, y, mean232325252624
 Education, y, median121213131412
 Income, $, median 1980878610 93319 93315 41725 40715 623
Race/ethnicity, no. (%)
 White182 (54.7)660 (57.4)175 (57.6)94 (56.6)443 (68.3)1554 (59.7)
 Black95 (28.5)266 (23.1)70 (23.0)51 (30.7)97 (14.9)579 (22.3)
 Hispanic56 (16.8)224 (19.5)59 (19.4)21 (12.7)109 (16.8)469 (18.0)
Foreign-born, no. (%)
 No307 (92.2)1073 (93.3)292 (96.1)160 (96.4)597 (92.0)2429 (93.4)
 Yes26 (7.8)77 (6.7)12 (3.9)6 (3.6)52 (8.0)173 (6.6)
Education, no. (%)
 < HS88 (26.4)248 (21.6)33 (10.9)31 (18.7)44 (6.8)444 (17.1)
 Completed HS185 (55.6)586 (51.0)136 (44.7)68 (41.0)170 (26.2)1145 (44.0)
 Some college48 (14.4)250 (21.7)87 (28.6)47 (28.3)174 (26.8)606 (23.3)
 Bachelor degree12 (3.6)66 (5.7)48 (15.8)20 (12.0)261 (40.2)407 (15.6)
Smoking in year of first pregnancy, no. (%)
 No197 (59.2)773 (67.2)229 (75.3)111 (66.9)501 (77.2)1811 (69.6)
 Yes136 (40.8)377 (32.8)75 (24.7)55 (33.1)148 (22.8)791 (30.4)
Alcohol in year of first pregnancy, no. (%)
 Never71 (21.3)261 (22.7)79 (26.0)49 (29.5)186 (28.7)646 (24.8)
 ≤ Once/mo220 (66.1)768 (66.8)195 (64.1)94 (56.6)409 (63.0)1686 (64.8)
 ≥ Once/wk42 (12.6)121 (10.5)30 (9.9)23 (13.9)54 (8.3)270 (10.4)
Outcomes
All births: mean BW, g332634093479341934753421
 Normal BW (%)765 (93.1)2518 (95.7)607 (96.7)349 (95.9)1245 (96.9)5484 (95.7)
 Low BW (%)57 (6.9)114 (4.3)21 (3.3)15 (4.1)40 (3.1)247 (4.3)
Term births: mean BW, g340734743545348735353489
 Normal BW (%)716 (95.7)2400 (97.6)579 (98.8)329 (97.3)1180 (98.3)5204 (97.6)
 Low BW (%)32 (4.3)58 (2.4)7 (1.2)9 (2.7)20 (1.7)126 (2.4)

Note. BW = birth weight; HS = high school. Demographic data are given for participants at the time of first delivery at or after age 18 years. Birth outcome data include maternal liveborn deliveries at or after age 18 years through the 2004 survey and are adjusted for births nested within maternal participants. Term births are liveborn deliveries ≥ 37 weeks' gestation. Low BW are liveborn deliveries weighing < 2500 grams.

Results of regression models using GEE are shown in Table 2. Birth weight estimates by occupational trajectory classes generally tracked the hierarchy of trajectory classes seen by age 34 years, and with the exception of class 4, were all significantly increased from the referent class 1. Birth weight estimates for SC were additionally robust to the inclusion of education and income. Black and Hispanic ethnicity remained independent risk factors for lower birth weight compared with White participants.

Table

TABLE 2— Parameter Estimates for the Association of Birth Weight With Maternal SC Trajectory and Related Covariates: National Longitudinal Survey of Youth 1979, United States, 1979–2004

TABLE 2— Parameter Estimates for the Association of Birth Weight With Maternal SC Trajectory and Related Covariates: National Longitudinal Survey of Youth 1979, United States, 1979–2004

VariableModel 1, b (95% CI)
Model 2, b (95% CI)
Model 3, b (95% CI)
Work SC trajectory class
 1 (lowest; Ref)
 283.7 (32.7, 134.7)84.1 (33.2, 135.0)72.7 (22.4, 123.0)
 3149.1 (81.7, 216.4)143.2 (73.2, 213.1)124.5 (55.0, 193.9)
 476.3 (−5.8, 158.4)86.1 (2.4, 169.8)80.2 (−2.4, 162.8)
 5124.7 (67.2, 182.4)101.5 (38.0, 165.0)86.4 (23.4, 149.3)
Maternal age at delivery52.8 (26.7, 78.9)50.0 (22.8, 77.3)51.4 (24.6, 78.3)
Income at delivery (natural logarithm [Ln])−5.1 (−17.0, 6.7)−7.0 (−18.6, 4.7)
Education at delivery2.5 (−7.6, 12.5)−4.6 (−14.8, 5.5)
Race/ethnicity
 White (Ref)
 Black−179.2 (−221.0, −137.4)−193.8 (−235.7, −151.9)
 Hispanic−34.7 (−79.3, 9.9)−61.5 (−108.0, −14.9)
Foreign-born−12.8 (−90.8, 65.2)
Smoker in year of delivery−149.5 (−187.1, −111.9)
Prenatal care begun in trimester 1−8.9 (−126.3, 108.4)

Note. CI = confidence interval; SC = substantive complexity. Parameter estimates represent regression of birth weight on predictors using generalized estimating equations modeling to account for the nesting of deliveries within mothers. Trajectory classes correspond to those noted in Figure 1.

The nonlinear form of the trajectories seen in Figure 1 suggests that a change in trajectory, even within the same class, might be associated with changes in outcome. Table 3 shows contrasts in mean predicted birth weight by SC trajectory class using the same methods as Table 2 but dichotomized by maternal age at delivery (up to age 24 years vs age 25 years and higher). Of particular note is the change in mean birth weight for class 4; up to age 24 years both the work trajectory and infant birth weight tracked the lower 2 trajectory classes,1,2 whereas afterward, the class 4 birth weight estimate was between that of the 2 high-trajectory classes 3 and 5. This group remained more similar to classes 1 and 2 in education at delivery after age 24 years, with a median educational attainment of a high school diploma, in contrast with the median attainment of classes 3 (some college) and 5 (college graduate). This change therefore likely represented a group for whom rapid occupational advancement occurred in young adulthood despite lower educational attainment, and suggests that this work advancement represented an important factor in birth outcomes.

Table

TABLE 3— Parameters for Birth Weight Associated With Trajectory Class, Stratified by Maternal Age at Delivery: National Longitudinal Survey of Youth 1979, United States, 1979–2004

TABLE 3— Parameters for Birth Weight Associated With Trajectory Class, Stratified by Maternal Age at Delivery: National Longitudinal Survey of Youth 1979, United States, 1979–2004

Trajectory ClassMaternal Age < 25 Years, b (95% CI)Maternal Age ≥ 25 Years, b (95% CI)
1 (Ref)
270.8 (30.8, 140.1)57.2 (19.3, 135.4)
3127.3 (42.5, 202.4)147.1 (51.1, 208.7)
454.7 (-36.7, 147.7)103.5 (16.3, 190.3)
5132.2 (50.0, 196.4)85.5 (27.6, 160.0)

Note. CI = confidence interval. Parameter estimates represent regression of birth weight on trajectory class, with Class 1 as referent. All results were adjusted for maternal age, race/ethnicity, educational attainment, and maternal smoking status.

Consistent with our hypothesis and previous work that suggested differential birth outcomes associated with differing patterns of occupational progression, we found that maternal work trajectories could be described by a set of distinct or discrete classes, which were in turn associated with differences in birth weight at term. Additionally, we showed that a sharp increase in occupational trajectory, as seen in participants in class 4, was associated with an increase in birth weight consistent with movement into a higher occupational class. Our results were robust to the inclusion of other SES indicators, specifically educational attainment and income level, which often covary with most measures of occupational standing. These findings suggested an ongoing dynamic effect of work, when contrasted with educational attainment, whose effect remained more static with increasing age.

We believed that examination of longitudinal patterns of occupational standing might have some advantages in the determination of work’s effect on health and birth outcomes. Latent classes and latent growth curves provided methods of modeling a hypothesized set of underlying changes and social mobility within a population across time that might be more difficult to discern with discrete individual data.41 The “noise” introduced into exposure assessment by participants’ frequent changes in work could be filtered out, leaving a smaller set of typical patterns of occupational progress or stasis. Virtually all research into occupational psychosocial exposures and pregnancy outcomes has used maternal occupation held during the pregnancy as the exposure metric.9,17–19,21–25,42–44 However, investigation of cumulative effects of work on other health outcomes, including cardiovascular health,45 and the demonstrably erosive effect of racially based stressors on birth outcomes, or “weathering”46,47 suggested the possibility that prolonged work that lacks such occupational attributes as skill discretion or decision authority might accumulate to produced poor health outcomes. Although we did not contrast associations of birth weight with work SC scores during pregnancy with our trajectory-based classes, we previously found that work trajectories of SC in male participants in the NLSY79 had a more robust association with hypertension in men than did the SC of current work.

The latent class trajectory approach additionally suggested that further analyses are warranted to determine the antecedents or predictors of class membership and the determinants of rises or falls in work SC (or change in class membership), given the evident association with birth weight outcome. This was particularly relevant to the findings for class 4, particularly for Black participants who exhibited a rapid rise in work SC in their 20s. Although class 4 Whites had similar educational attainment and a greater mean income at delivery ($28 269 vs $19 203; P = .06), mean age-adjusted birth weight did not differ between Whites and Blacks (3519 grams vs 3434 grams; P = .51). Such findings indicated that the GMM approach might be useful in a search for factors that could positively change the arc of a working life and contribute to more healthy outcomes.

Limitations

Some limitations of the present work should be acknowledged. The NLSY79 preferentially enrolled low-income participants and over-recruited minority participants; our findings might not represent the career trajectories of those from more privileged backgrounds. Additionally, imputed job characteristics from databases remained proxy average measures of exposure and might not reflect individual work circumstances, differential job tasks for minorities or women, or changes in occupational stressors that occurred as jobs changed. However, imputation might help to avoid problems associated with common-instrument bias.15,18 These participants’ job experiences also became more distant in time and might not reflect current labor conditions as service and information jobs supplanted manufacturing. The O*NET was designed in the past 15 years to supplant the DOT; thus O*NET factor scores might not reflect work in the participants’ youth in the 1980s. The DOT system, though, which has more than 12 000 entries, became too cumbersome and unwieldy for similar statistical applications.30 To check on the validity of using more recent job ratings, however, we analyzed a set of SC scores developed from the DOT in the 1970s48 with those we derived from the O*NET and found a correlation coefficient of 0.71 (P < .001) between the 2 data sets, suggesting reasonable interchangeability between DOT and O*NET scored jobs across the 1980s to the 2000s.

We did not incorporate a measure of unemployment or account for periods when the participants were not working. It was unclear as to what the most suitable metric for unemployment (on an equivalent scale to work SC or job control) might be, although the deleterious effects on health are well-described.10,49,50 Prolonged unemployment would, however, almost certainly flatten participants’ occupational trajectory. Further investigation would require detailed work histories accounting for the length of employed and nonemployed periods.

Finally, we did not address questions of precedence and reverse causation in this study. Career trajectories in women have been long known to be affected by childbearing, with the reverse also being possible. We modeled work and pregnancy outcomes from age 18 years in the data set, which might differentially alter participants’ representation in job trajectories if they, for example, became pregnant sooner and left school. Such questions require a more detailed investigation. As an initial sensitivity analysis, we modeled the risk of a change in class trajectory as a consequence of having a child between ages 15 and 19 years. This analysis yielded an odds ratio (adjusted for education) of 1.03 (95% confidence interval = 0.94, 1.13) for change in class with earlier pregnancy, indicating, at least at first pass, little substantive change in class trajectory from childbearing in adolescence in the NLSY79 participants.

Conclusions

The construction of latent class trajectories for occupation and their association with perinatal health outcomes (in this case, with birth weight) suggested useful avenues for future research. We found that work trajectory was strongly associated with birth weight and was robust to the inclusion of other SES indicators, namely, education and income. Antecedents and predictors of trajectory classes could be examined for both their predictive value in latent class assignment, in addition to the determination of their capacity to modify individual trajectories or shift participants between trajectory classes. By a similar process, participants who shift occupational trajectory during the course of working life might be sensitive indicators for the effect of influential factors, such as alcohol, drugs, education, or training opportunities on work and health. The ability to follow work and its relevant characteristics longitudinally might improve our capacity to integrate occupation into a life-course model that examines both its antecedents and consequences for health.

Acknowledgments

We were supported for a portion of this work by the National Institute of Occupational Safety and Health (grant OH966).

Portions of the work were derived and adapted from M. Mutambudzi's PhD dissertation at the University of Connecticut, awarded in May 2012.

We thank Nicholas Warren, ScD, Susan Reisine, PhD, and Vicki Magley, PhD, for suggestions and support during the research process.

Human Participant Protection

An exemption approval for this study was received from the institutional review board of the University of Connecticut Health Center.

References

1. Braveman P, Cubbin C, Marchi K, Egerter S, Chavez G. Measuring socioeconomic status/position in studies of racial/ethnic disparities: maternal and infant health. Public Health Rep. 2001;116(5):449463. Crossref, MedlineGoogle Scholar
2. Nicolaidis C, Ko CW, Saha S, Koepsell TD. Racial discrepancies in the association between paternal vs. maternal educational level and risk of low birthweight in Washington State. BMC Pregnancy Childbirth. 2004;4(1):10. Crossref, MedlineGoogle Scholar
3. Savitz DA, Kaufman JS, Dole N, Siega-Riz AM, Thorp JM Jr, Kaczor DT. Poverty, education, race, and pregnancy outcome. Ethn Dis. 2004;14(3):322329. MedlineGoogle Scholar
4. Astone NM, Misra D, Lynch C. The effect of maternal socio-economic status throughout the lifespan on infant birthweight. Paediatr Perinat Epidemiol. 2007;21(4):310318. Crossref, MedlineGoogle Scholar
5. Villalbi JR, Salvador J, Cano-Serral G, Rodriguez-Sanz MC, Borrell C. Maternal smoking, social class and outcomes of pregnancy. Paediatr Perinat Epidemiol. 2007;21(5):441447. Crossref, MedlineGoogle Scholar
6. Blumenshine P, Egerter S, Barclay CJ, Cubbin C, Braveman PA. Socioeconomic disparities in adverse birth outcomes: a systematic review. Am J Prev Med. 2010;39(3):263272. Crossref, MedlineGoogle Scholar
7. Murphy JF, Dauncey M, Newcombe R, Garcia J, Elbourne D. Employment in pregnancy: prevalence, maternal characteristics, perinatal outcome. Lancet. 1984;323(8387):11631166. CrossrefGoogle Scholar
8. Poerksen A, Petitti DB. Employment and low birth weight in black women. Soc Sci Med. 1991;33(11):12811286. Crossref, MedlineGoogle Scholar
9. Savitz DA, Brett KM, Dole N, Tse CK. Male and female occupation in relation to miscarriage and preterm delivery in central North Carolina. Ann Epidemiol. 1997;7(7):509516. Crossref, MedlineGoogle Scholar
10. Burdorf A, Brand T, Jaddoe VW, Hofman A, Mackenbach JP, Steegers EA. The effects of work-related maternal risk factors on time to pregnancy, preterm birth and birth weight: the Generation R Study. Occup Environ Med. 2011;68(3):197204. Crossref, MedlineGoogle Scholar
11. Dooley D, Prause J. Birth weight and mothers’ adverse employment change. J Health Soc Behav. 2005;46(2):141155. Crossref, MedlineGoogle Scholar
12. Morales-Suárez-Varela M, Kaerlev L, Zhu JL, et al. Unemployment and pregnancy outcomes: a study within the Danish National Birth Cohort. Scand J Public Health. 2011;39(5):449456. Crossref, MedlineGoogle Scholar
13. Martikainen P, Blomgren J, Valkonen T. Change in the total and independent effects of education and occupational social class on mortality: analyses of all Finnish men and women in the period 1971-2000. J Epidemiol Community Health. 2007;61(6):499505. Crossref, MedlineGoogle Scholar
14. Krieger N, Williams DR, Moss NE. Measuring social class in US public health research: concepts, methodologies, and guidelines. Annu Rev Public Health. 1997;18:341378. Crossref, MedlineGoogle Scholar
15. Schwartz JE, Pieper CF, Karasek RA. A procedure for linking psychosocial job characteristics data to health surveys. Am J Public Health. 1988;78(8):904909. LinkGoogle Scholar
16. Karasek R, Brisson C, Kawakami N, Houtman I, Bongers P, Amick B. The Job Content Questionnaire (JCQ): an instrument for internationally comparative assessments of psychosocial job characteristics. J Occup Health Psychol. 1998;3(4):322355. Crossref, MedlineGoogle Scholar
17. Homer CJ, James SA, Siegel E. Work-related psychosocial stress and risk of preterm, low birthweight delivery. Am J Public Health. 1990;80(2):173177. LinkGoogle Scholar
18. Brandt LP, Nielsen CV. Job stress and adverse outcome of pregnancy: a causal link or recall bias? Am J Epidemiol. 1992;135(3):302311. Crossref, MedlineGoogle Scholar
19. Henriksen TB, Hedegaard M, Secher NJ. The relation between psychosocial job strain, and preterm delivery and low birthweight for gestational age. Int J Epidemiol. 1994;23(4):764774. Crossref, MedlineGoogle Scholar
20. Cerón-Mireles P, Harlow SD, Sanchez-Carrillo CI. The risk of prematurity and small-for-gestational-age birth in Mexico City: the effects of working conditions and antenatal leave. Am J Public Health. 1996;86(6):825831. LinkGoogle Scholar
21. Brett KM, Strogatz DS, Savitz DA. Employment, job strain, and preterm delivery among women in North Carolina. Am J Public Health. 1997;87(2):199204. LinkGoogle Scholar
22. Escribá-Agüir V, Perez-Hoyos S, Saurel-Cubizolles MJ. Physical load and psychological demand at work during pregnancy and preterm birth. Int Arch Occup Environ Health. 2001;74(8):583588. Crossref, MedlineGoogle Scholar
23. Bell JF, Zimmerman FJ, Diehr PK. Maternal work and birth outcome disparities. Matern Child Health J. 2008;12(4):415426. Crossref, MedlineGoogle Scholar
24. Meyer JD, Warren N, Reisine S. Job control, substantive complexity, and risk for low birth weight and preterm delivery: An analysis from a state birth registry. Am J Ind Med. 2007;50(9):664675. Crossref, MedlineGoogle Scholar
25. Meyer JD, Warren N, Reisine S. Racial and ethnic disparities in low birth weight delivery associated with maternal occupational characteristics. Am J Ind Med. 2010;53(2):153162. Crossref, MedlineGoogle Scholar
26. Rahkonen O, Laaksonen M, Martikainen P, Roos E, Lahelma E. Job control, job demands, or social class? The impact of working conditions on the relation between social class and health. J Epidemiol Community Health. 2006;60(1):5054. Crossref, MedlineGoogle Scholar
27. Schooler C, Mulatu MS, Oates G. The continuing effects of substantively complex work on the intellectual functioning of older workers. Psychol Aging. 1999;14(3):483506. Crossref, MedlineGoogle Scholar
28. Kohn ML, Schooler C. Work and Personality: An Inquiry Into the Impact of Social Stratification. Norwood, NJ: Ablex Pub. Corp.; 1983. Google Scholar
29. Kohn ML, Schooler C. The reciprocal effects of the substantive complexity of work and intellectual flexibility: a longitudinal assessment. Am J Sociol. 1978;84(1):2452. CrossrefGoogle Scholar
30. Hadden W, Kravetz N, Muntaner C. Descriptive dimensions of US occupations with data from the O*NET. Soc Sci Res. 2004;33(1):6478. CrossrefGoogle Scholar
31. Cifuentes M, Boyer J, Gore R, et al. Inter-method agreement between O*NET and survey measures of psychosocial exposure among healthcare industry employees. Am J Ind Med. 2007;50(7):545553. Crossref, MedlineGoogle Scholar
32. Meyer JD, Cifuentes M, Warren N. Association of self-rated physical health and incident hypertension with O*NET factors: validation using a representative national survey. J Occup Environ Med. 2011;53(2):139145. Crossref, MedlineGoogle Scholar
33. Meyer JD, Mutambudzi M. Construction of life-course occupational trajectories: evidence for work as a mediator of racial disparities in hypertension. J Occup Environ Med. 2012;54(10):12011207. Crossref, MedlineGoogle Scholar
34. O*NET Resource Center. O*NET Database v12.0. Available at: http://www.onetcenter.org/database.html. Accessed February 11, 2008. Google Scholar
35. LaPolice CC, Carter GW, Johnson JW. Linking O*NET descriptors to occupational literacy requirements using job component validation. Pers Psychol. 2008;61(2):405441. CrossrefGoogle Scholar
36. Peterson N, Mumford M, Borman W, et al. Understanding work using the Occupational Information Network (O*NET): implications for practice and research. Pers Psychol. 2001;54(2):451492. CrossrefGoogle Scholar
37. McCormick MC, Brooks-Gunn J. Concurrent child health status and maternal recall of events in infancy. Pediatrics. 1999;104(5 Pt 2):11761181. MedlineGoogle Scholar
38. Tomeo CA, Rich-Edwards JW, Michels KB, et al. Reproducibility and validity of maternal recall of pregnancy-related events. Epidemiology. 1999;10(6):774777. Crossref, MedlineGoogle Scholar
39. Sahr, RC. Conversion factor tables for CPI-U-RS inflation adjustment factors. Corvallis, OR: Oregon State University Department of Political Science; 2012. Google Scholar
40. Muthén B, Muthén LK. Integrating person-centered and variable-centered analyses: growth mixture modeling with latent trajectory classes. Alcohol Clin Exp Res. 2000;24(6):882891. Crossref, MedlineGoogle Scholar
41. Sturgis P, Sullivan L. Exploring social mobility with latent trajectory groups. J R Stat Soc Ser A Stat Soc. 2008;171(1):6588. Google Scholar
42. Oths KS, Dunn LL, Palmer NS. A prospective study of psychosocial job strain and birth outcomes. Epidemiology. 2001;12(6):744746. Crossref, MedlineGoogle Scholar
43. Stein ZA, Susser MW, Hatch MC. Working during pregnancy: physical and psychosocial strain. Occup Med. 1986;1(3):405409. MedlineGoogle Scholar
44. Landsbergis PA, Hatch MC. Psychosocial work stress and pregnancy-induced hypertension. Epidemiology. 1996;7(4):346351. Crossref, MedlineGoogle Scholar
45. Singh-Manoux A, Ferrie JE, Chandola T, Marmot M. Socioeconomic trajectories across the life course and health outcomes in midlife: evidence for the accumulation hypothesis? Int J Epidemiol. 2004;33(5):10721079. Crossref, MedlineGoogle Scholar
46. Geronimus AT. Understanding and eliminating racial inequalities in women’s health in the United States: the role of the weathering conceptual framework. J Am Med Womens Assoc. 2001;56(4):133136, 149–150. MedlineGoogle Scholar
47. Colen CG, Geronimus AT, Bound J, James SA. Maternal upward socioeconomic mobility and black-white disparities in infant birthweight. Am J Public Health. 2006;96(11):20322039. LinkGoogle Scholar
48. Miller A, Treiman D, Cain P, Roos P. Work, Jobs, and Occupations: A Critical Review of the Dictionary of Occupational Titles. Washington, DC: National Academies Press; 1980. Google Scholar
49. Breslin FC, Mustard C. Factors influencing the impact of unemployment on mental health among young and older adults in a longitudinal, population-based survey. Scand J Work Environ Health. 2003;29(1):514. Crossref, MedlineGoogle Scholar
50. Steenland K. Epidemiology of occupation and coronary heart disease: research agenda. Am J Ind Med. 1996;30(4):495499. Crossref, MedlineGoogle Scholar

Related

No related items

TOOLS

SHARE

ARTICLE CITATION

Miriam Mutambudzi, PhD, MPH, and John D. Meyer, MD, MPHMiriam Mutambudzi is with the Johns Hopkins Lupus Center, Johns Hopkins School of Medicine, Baltimore, MD. John D. Meyer is with the Department of Environmental and Occupational Health Sciences, SUNY-Downstate School of Public Health, Brooklyn, NY, and the Department of Preventive Medicine, Icahn School of Medicine at Mount Sinai, New York, NY. “Construction of Early and Midlife Work Trajectories in Women and Their Association With Birth Weight”, American Journal of Public Health 104, no. S1 (February 1, 2014): pp. S58-S64.

https://doi.org/10.2105/AJPH.2013.301401

PMID: 24354827