Search tips
Search criteria 


Logo of jgimedspringer.comThis journalToc AlertsSubmit OnlineOpen Choice
J Gen Intern Med. 2011 January; 26(1): 45–50.
Published online 2010 August 31. doi:  10.1007/s11606-010-1488-5
PMCID: PMC3024109

Literacy Skills and Calculated 10-Year Risk of Coronary Heart Disease

Laurie T. Martin, ScD MPH,corresponding author1 Matthias Schonlau, PhD,1,2 Ann Haas, MS,1 Kathryn Pitkin Derose, PhD, MPH,1 Rima Rudd, ScD,3 Eric B. Loucks, PhD,4 Lindsay Rosenfeld, ScD ScM,5 and Stephen L. Buka, ScD4



Coronary heart disease (CHD) is a leading cause of morbidity and mortality. Reducing the disease burden requires an understanding of factors associated with the prevention and management of CHD. Literacy skills may be one such factor.


To examine the independent and interactive effects of four literacy skills: reading, numeracy, oral language (speaking) and aural language (listening) on calculated 10-year risk of CHD and to determine whether the relationships between literacy skills and CHD risk were similar for men and women.


We used multivariable linear regression to assess the individual, combined, and interactive effects of the four literacy skills on risk of CHD, adjusting for education and race.


Four hundred and nine English-speaking adults in Boston, MA and Providence, RI.


Ten-year risk of coronary heart disease was calculated using the Framingham algorithm. Reading, oral language and aural language were measured using the Woodcock Johnson III Tests of Achievement. Numeracy was assessed through a modified version of the numeracy scale by Lipkus and colleagues.

Key Results

When examined individually, reading (p = 0.007), numeracy (p = 0.001) and aural language (p = 0.004) skills were significantly associated with CHD risk among women; no literacy skills were associated with CHD risk in men. When examined together, there was some evidence for an interaction between numeracy and aural language among women suggesting that higher skills in one area (e.g., aural language) may compensate for difficulties in another resulting in an equally low risk of CHD.


Results of this study not only provide important insight into the independent and interactive effects of literacy skills on risk of CHD, they also highlight the need for the development of easy-to use assessments of the oral exchange in the health care setting and the need to better understand which literacy skills are most important for a given health outcome.

KEY WORDS: cardiovascular disease, health literacy, communication skills, vulnerable populations


Coronary heart disease (CHD) remains a leading cause of death and disability worldwide1,2. The high disease burden of CHD has prompted the setting of 15 objectives for Healthy People 2010 under a broader goal “to improve cardiovascular health and quality of life through the prevention, detection, and treatment of risk factors...and prevention of recurrent cardiovascular events”3. However, this requires a thorough understanding of modifiable factors associated with CHD prevention and management.

Health literacy, the “degree to which individuals have the capacity to obtain, process, and understand basic health information and services needed to make appropriate health decisions”4, may be one such factor. Low health literacy remains a formidable barrier to reducing gaps in health care quality and improving outcomes. Individuals with low health literacy find it difficult to understand directions for taking medicine, calculate a dose of an over-the-counter medication or comprehend a consent form5,6. Low health literacy may also contribute to suboptimal care and outcomes through lower participation in screening programs7, reduced ability to act on and understand the advice of a health professional8, and limited ability to navigate the health care system9,10. Reports from the 2003 National Assessment of Adult Literacy (NAAL) suggest that only 12% of the population is considered to have ‘proficient’ health literacy11.

Evidence linking health literacy to CHD outcomes comes from studies of Medicare managed care enrollees12; rates of cardiovascular death were higher for those with inadequate literacy (19.3%) and marginal literacy (16.7%) compared with those with adequate health literacy (7.9%). Individuals with lower health literacy are also more likely to have or engage in CHD risk factors including smoking13,14, hypertension15,16, and diabetes16.

While these studies provide important insight into the relation between literacy and CHD, the majority assessed reading skills as a proxy for health literacy. This has two potential limitations. First, reading skills are only one component of health literacy. Other skills such as numeracy, writing, oral language (speaking) and aural language (listening) are less studied despite our dependence on the oral exchange to communicate symptoms, concerns, and treatment options. Second, it is unclear whether health literacy can, or should, be represented by a single dimension.

The objective of this study was to provide insight into the individual and combined effects of four health literacy skills on the calculated 10-year risk of CHD, using data from 409 individuals from the New England Family Study (NEFS). We hypothesized that in addition to reading skills, other measures of literacy (e.g., numeracy, oral language, aural language) would be associated with CHD risk. Because many people depend on non-written means of communication for health information17,18, we also hypothesized that stronger oral or aural language skills may compensate for weaker reading or numeracy skills.


Study Population

In 2001, the NEFS was established to locate and follow-up adult offspring of pregnant women originally enrolled in the Providence, Rhode Island and Boston, Massachusetts sites of the Collaborative Perinatal Project (CPP)19. The NEFS is a center comprised of numerous research projects, each following-up a different set of individuals20. Data for the current analyses came from a NEFS study designed to assess pathways linking education and health. Eight hundred participants in their mid-40s were selected from the NEFS population through a multi-stage sampling procedure to ensure an adequate mix of educational attainment. Six hundred eighteen (77.3%) individuals were successfully located and interviewed. Study assessments had two components. All participants completed a structured interview in English, and they underwent a clinical assessment, which included a blood draw. Although study participants have been followed since birth, the current analyses are cross-sectional.


Ten-Year Calculated Risk of Coronary Heart Disease

We calculated the 10-year risk of having a hard coronary heart disease outcome (i.e., myocardial infarction or coronary death) as a percentage using the validated Framingham risk algorithm. The algorithm uses gender-specific Cox regression models that incorporate age, diabetes, smoking, total and HDL cholesterol, and systolic and diastolic blood pressure21,22. The Framingham algorithm has good predictive validity for CHD events in the Framingham Heart Study (c-statistic = 0.74 and 0.77 for men and women, respectively)22. External validity tests with other samples have been performed where the risk algorithm was found to perform reasonably well in both black and white men (c-statistic ranged from 0.63 to 0.75) and women (c-statistic ranged from 0.66 to 0.83)23.

Lipids were measured in non-fasting plasma samples and total cholesterol was measured enzymatically as described elsewhere24. HDL cholesterol was determined using a direct enzymatic colorimetric assay25. Presence of diabetes and current smoking status were self-reported. Systolic and diastolic blood pressure were measured in seated, resting participants, using automated blood pressure monitors with good validity and reliability compared with auscultation26. Five blood pressure readings were obtained; systolic and diastolic values were calculated as the mean of the lowest three systolic or diastolic blood pressure readings, excluding the first recorded blood pressure.

Literacy Skills

We measured four literacy skills: reading comprehension, numeracy, oral language (speaking), and aural language (listening). Reading comprehension, oral language, and aural language were assessed using subtests of the Woodcock Johnson III (WJ III) Tests of Achievement, a standardized test normed against a representative U.S. population, ages 24 months to 90 years and older27. The WJ III produces both age- and grade-equivalent scores; grade equivalent scores were used in these analyses. The Tests of Achievement were administered to study participants by one of three interviewers trained on the administration of the WJ III by a psychologist external to the research team. Standard administration procedures were used; subjects answered questions of increasing difficulty until a ceiling was reached. As a result, the number of items for the WJ III subtests varies by study participant. Numeracy was assessed using an eight-item scale adapted from Lipkus et al.28, described in more detail below.

Oral language was assessed by the WJ III test “Story Recall,” in which participants listen to a pre-recorded short story (1–3 paragraphs) and are then asked to repeat the story back to the interviewer27. Scores are based on the correct number of words or phrases spoken, and the reliability of the test is good with a one-year test–retest reliability of 0.7027. This test is similar to commonly used “teach-back” methods that providers use to gauge patient communication, understanding, and recall of health messages and instructions.

Aural language, or listening comprehension, was assessed by the WJ III test “Understanding Directions,” in which participants are given an illustrated drawing and are asked to complete a pre-recorded direction followed by a few-second pause to complete the task, when another direction is given. Difficulty increases as drawings become more complex and the tasks increase in number of components27. Scores are based on the number of correct tasks completed. Reliability of this test is good with a one year test-retest reliability of 0.8827. This test was selected for its similarity to experiences in the health care setting, where providers give patients a set of instructions for taking their medicine or preparing for a medical procedure, for example.

Reading comprehension was assessed using the “Passage Comprehension” test from the WJ III. This is a cloze test, where individuals fill in missing words from a sentence. Reliability for this test is good with a one-year test-retest reliability of 0.9227.

Numeracy was assessed using eight items from the scale developed by Lipkus and colleagues28,29. Three items were dropped from the Lipkus scale (item number 2 from the general numeracy scale, and items 3 and 7 from the expanded numeracy scale) due to time constraints for the full protocol, which took approximately 2 1/2 hours to complete. As such, three items were dropped to fit within the allotted timeframe for the construct. Based on the study by Lipkus et al.28, the items with the highest (90.5%) and lowest (48.6%) correct were deleted from the expanded numeracy scale, and the item with the highest percent correct (59.8%) was dropped from the general numeracy scale to minimize the bias that could arise from dropping the three easiest or hardest items.

Analogously to Lipkus28, we conducted a factor analysis using the 8x8 matrix of tetrachoric correlations, which are used to assess the correlation among dichotomous variables. Factor analysis confirmed a one-factor solution, with all 8 items loading on the first factor (individual loadings each exceeded 0.4). The first factor accounted for 86.25% of the variance. Numeracy scores were based on the total number of correct responses (range: 0-8, mean = 5.47, SD = 1.85).

Demographic Covariates

Educational attainment represents the number of years of formal schooling a person reported, ranging from 7-21 years. Race was self-reported and categorized as non-Hispanic White, non-Hispanic Black, and Hispanic or other. Age of respondent and tobacco use were not included in the multivariable models as they are components of the Framingham risk score.


Correlations, t-tests and chi-square tests were used to describe sample characteristics. We used multivariable regression to assess the individual, combined, and interactive effects of the four literacy skills on risk of CHD, adjusting for education and race. The regressions contained random intercepts to account for within family clustering. Analyses were stratified by gender to parallel the computation of the Framingham risk algorithm. Analyses of the 10-year risk of CHD were conducted on a log scale because the distribution of Framingham risk was right skewed. Because the interpretation of model coefficients on a log scale is not intuitive, we used adjusted predictions to illustrate the association between individual variables and CHD risk31. Specifically, we set each literacy skill in turn to its 10th, 50th, and 90th sample percentile (women and men combined) and computed sample average 10-year CHD risk predictions from the regression model.


Twenty-one individuals with previous heart disease were excluded from the analyses as were an additional 188 participants lacking one or more components used to calculate the Framingham risk. Accordingly, the sample size for the analysis was 409 (66% of the selected sample). No significant differences in gender, race, education, numeracy or aural language were found between those included and not included in the analytic sample (Table 1). Individuals included in the analyses were, however, slightly younger and had higher average grade equivalent scores for reading (12.7 vs. 11.6, p = 0.02), oral language skills (7.9 vs. 6.9, p = 0.02) and aural language skills (8.1 vs. 7.2, p = 0.04) than those excluded from the study. An examination of individual components of the Framingham risk algorithm revealed no differences between the analytic sample and those excluded from analysis.

Table 1
Sample Characteristics*

Correlations among literacy skills ranged from 0.38 (reading and listening) to 0.60 (numeracy and reading). The mean estimated 10-year risk of CHD was significantly higher for men (6.7%) than women (2.6%) (p < 0.001 on the log scale). Several literacy skills showed significant gender differences. Men scored significantly higher on both reading (men = 13.3, women = 11.9, p < 0.001) and numeracy (men = 6.1, women = 5.1, p < 0.001) while women scored higher, on average, on aural language skills (men = 7.4, women = 8.2, p = 0.02). There were no gender differences in oral language skills.

Multivariable Regression

Figure 1 presents model-based predictions, based on the log model of 10-year risk of CHD, separately for men and women at the 10th, 50th and 90th sample percentile of each literacy skill, adjusted for demographic characteristics. For easier interpretation, log transformed CHD risk has been re-transformed to the original scale in Figure 1. In women, higher reading (β = -0.027, p = 0.007), numeracy (β = -0.085, p = 0.001), and aural language (β = -.028, p = 0.004) skills were associated with significantly lower 10-year risk of CHD. In men, however, reading (β = -0.013, p = 0.20), numeracy (β = -0.027, p = 0.36), and aural language (β = 0.003, p = 0.77) skills had smaller log scale coefficients and were not significantly associated with risk of CHD. Oral language skills were not associated with risk for CHD for either gender.

Figure 1
Predicted 10-year CHD risk for the 10th, 50th, and 90th percentile of health literacy skills. Predictions (of CHD risk) illustrate separate regressions (of log CHD risk) for each literacy skill by gender1. Black bar: 10th percentile of literacy skill. ...

For women, an increase in aural language skills from the sample 10th to the 90th percentile was associated with a reduction in predicted 10-year risk of CHD from 2.3% to 1.8%, an estimated reduction of 5 CHD events out of 1,000 women. An analogous increase in reading skills was associated with a reduction in predicted 10-year risk of CHD from 2.4% to 1.7%, an estimated reduction of 7 CHD events out of 1,000 women. An increase in numeracy skills from the 10th to the 90th percentile was also associated with a reduction in predicted 10-year risk of CHD from 2.4% to 1.6%, a projected reduction of 8 CHD events out of 1,000 women.

We then simultaneously examined reading, numeracy, oral and aural language in relation to predicted 10-year CHD risk among women. The final model included numeracy (β = -0.134, SE = 0.047, p = 0.005), aural language (β = -0.071, SE = 0.036, p = 0.046) and the interaction between numeracy and aural language (β = 0.01, SE = 0.006, p = 0.119). Reading did not contribute significantly to the model after controlling for numeracy. Although the interaction between numeracy and aural language was only marginally statistically significant, further data analyses (e.g., regression models stratified by aural language) confirmed that the interaction was present in the data, and not an artifact of multi-variable modeling (Figure 2). Retaining the interaction also avoids potential bias due to model misspecification32. Women with the lowest 10-year risk of CHD had either high numeracy or high aural language skills. Women having difficulty with both numeracy and aural language skills were at highest risk, followed by those with low numeracy but moderate (50th percentile) aural language skills. The final multivariable model explained about 12% of the variation in CHD risk (adjusted R2= 0.119).

Figure 2
Predicted 10-year CHD risk among women based on the final multivariable regression (of log CHD risk) on numeracy, aural language and numeracy by aural language interaction1. Solid line: aural language 10th percentile. Dashed line: aural language 50th ...


Literacy skills are important determinants of health, and affect one’s ability to prevent, manage, and treat disease effectively. This is the first study to examine the individual and combined effects of four literacy skills simultaneously in relation to the calculated 10-year risk of CHD. Findings suggest that when examined individually, higher reading, numeracy, and aural language scores were associated with a lower risk of CHD, although these associations were significant only for women. When examined jointly, numeracy and aural language skills were the strongest predictors. Our results also highlight the importance of listening skills. As hypothesized, higher skills in one area (e.g., aural language) may compensate for difficulties in another (e.g., numeracy) resulting in an equally low risk of CHD. Our study also suggests that reading may not be the strongest literacy measure associated with CHD risk.

The lack of association between literacy skills and 10-year CHD risk among men was surprising, but consistent with other literature illustrating stronger associations of socioeconomic position (correlated with literacy skills) with obesity33 and metabolic syndrome34 (a risk factor for CHD) among women as compared to men. This is unlikely due to lack of power alone because coefficients for men were consistently smaller than those for women. One potential explanation relates to parity. Prospective studies have demonstrated that childbirth is associated with increased long-term central adiposity and decreased HDL cholesterol levels35,36. Given the inverse association between a woman’s educational attainment and number of children born37, and strong correlation between literacy skills and educational attainment, one potential explanation may be the literacy-related differences in parity. We, however, did not ask women in our study about number of pregnancies, and as a result are not able to test this hypothesis empirically.

This study has several limitations. First, the sample was not nationally representative. In particular, the sample contains few Hispanics and did not include immigrants or individuals for whom English is a second language. Second, risk factors for CHD are more prevalent among individuals with lower literacy and while we were able to account for race/ethnicity and educational attainment, there may be other unmeasured factors that may account for the association between literacy skills and risk of CHD. It is not clear, however, whether such factors such as obesity or other measures of socioeconomic status such as income should be treated as potential confounding factors or whether they are better described as mediators on the causal pathway between literacy skills and CHD risk. While education was used as a proxy for socioeconomic status in these analyses, the strong link between education and literacy skills and the uncertainty about their temporal order (i.e., do higher literacy skills result in higher education or vice-versa) suggests that by controlling for education, we may be underestimating the relationship between literacy and CHD risk.

Another limitation is that this study did not include writing which has also been cited as an important component of health literacy(38). It is also not clear whether our results were affected by the use of a modified numeracy scale. Finally, most of our literacy measures were not assessed within a health context and, thus, cannot be considered measures of “health literacy.” While there are no currently available measures of the oral exchange in a health context, the measures used in this study are good proxies to capture such skills as they are readily available, normed, and validated. However, we do acknowledge that context is extremely important when studying health literacy and that the complexity of the health care system, the medical jargon used by many providers, and the exposure to novel health concepts all have the potential to exceed one’s health literacy skills, even among those with adequate literacy. As such, findings from our study may only underestimate the true association between health literacy skills and risk of CHD.

In addition to further examining the association between literacy skills and risk of CHD, more work must be done to understand the mechanisms for the association. It is not known, for example, who among the sample had access to a physician, whether food or cigarette labels or risks were clearly understood, whether antihypertensive mediations were taken as prescribed, or whether conversations with and recommendations by health care providers were effective and useful. Understanding why the association exists has important implications for interventions to lower risk of CHD, particularly among women.

While the results from this study provide important insight into the independent and interactive effects of literacy skills on risk of CHD, they also highlight the need for the development of easy-to use assessments of the oral exchange in the health care setting and the need to better understand which literacy skills specifically are most important for given outcomes.



Conflict of Interest None disclosed.

Funding Sources This work was supported by NHLBI grant 1R21HL094297-01 (Martin)


An erratum to this article can be found at


1. Mensah GA, Brown DW. An overview of cardiovascular disease burden in the United States. Health Aff. 2007;26(1):38–48. doi: 10.1377/hlthaff.26.1.38. [PubMed] [Cross Ref]
2. The World Health Organization. Cardiovascular diseases. Available at: Accessed August, 2010.
3. Healthy people 2010: Understanding and improving health. 2. Washington, DC: US. Government Printing Office; 2000.
4. Ratzan S, Parker R. Introduction. In: Selden C, Zorn M, Ratzan S, Parker R, eds. National library of medicine current bibliographies in medicine: Health literacy. Bethesda, MD: National Institutes of Health, U.S. Department of Health and Human Services; 2000.
5. Kirsch I. The international adult literacy survey (IALS): Understanding what was measured. Princeton, NJ: Statistics and Research Division of the Educational Testing Service; December 2001. RR-01-25.
6. Kirsch I, Jungeblut A, Jenkinds L, Kolstad A. Adult literacy in America: A first look at the results of the national adult literacy survey. Washington, DC: National Center for Education Statistics, US Department of Education; 1993.
7. Dolan N, Ferreira M, Davis TC, et al. Colorectal cancer screening knowledge, attitudes, and beliefs among veterans: Does literacy make a difference? J Clin Oncol. 2004;22(13):2617–2622. doi: 10.1200/JCO.2004.10.149. [PubMed] [Cross Ref]
8. Chew L, Bradley K, Flum D, Cornia P, Koepsell TD. The impact of low health literacy on surgical practice. Am J Surg. 2004;188(3):250–253. doi: 10.1016/j.amjsurg.2004.04.005. [PubMed] [Cross Ref]
9. Davis TC, Wolf MS, Bass PF, et al. Low literacy impairs comprehension of prescription drug warning labels. J Gen Intern Med. 2006;21(8):847–851. doi: 10.1111/j.1525-1497.2006.00529.x. [PMC free article] [PubMed] [Cross Ref]
10. Kripalani S, Henderson L, Chiu E, Robertson R, Kolm P, Jacobson T. Predictors of medication self-management skill in a low-literacy population. J Gen Intern Med. 2006;21(8):852–856. doi: 10.1111/j.1525-1497.2006.00536.x. [PMC free article] [PubMed] [Cross Ref]
11. National Center for Education Statistics. The health literacy of America's adults: Results from the 2003 national assessment of adult literacy. Washington, DC: U.S. Department of Education;2006.
12. Baker DW, Wolf MS, Feinglass J, Thompson JA, Gazmararian JA, Huang J. Health literacy and mortality among elderly persons. Arch. Intern. Med. 2007;167(14):1503. doi: 10.1001/archinte.167.14.1503. [PubMed] [Cross Ref]
13. Fredrickson D, Washington R, Pham N, Jackson T, Wiltshire J, Jecha L. Reading grade levels and health behaviors of parents at child clinics. Kansas Med. 1995;96:127–129. [PubMed]
14. Hawthorne G. Preteenage drug use in Australia: The key predictors and school-based drug education. J Adolesc Health. 1997;20:384–395. doi: 10.1016/S1054-139X(96)00181-4. [PubMed] [Cross Ref]
15. Battersby C, Hartley K, Fletcher A, Markowe H, Brown R, Styles W. Cognitive function in hypertension: A community based study. J Hum Hypertens. 1993;7:117–123. [PubMed]
16. Williams M, Baker D, Parker R, Nurss J. Relationship of functional health literacy to patients' knowledge of their chronic disease: A study of patients with hypertension and diabetes. Arch. Intern. Med. 1998;158(2):166–172. doi: 10.1001/archinte.158.2.166. [PubMed] [Cross Ref]
17. Doak C, Doak L, Root J. Teaching patients with low literacy skills. 2. Philadelphia: J.B. Lippincott; 1996.
18. Weiss BD. Communicating with patients who cannot read. N. Engl. J. Med. 1997;337(4):272–274. doi: 10.1056/NEJM199707243370411. [PubMed] [Cross Ref]
19. Niswander KR, Gordon M. The women and their pregnancies: The collaborative perinatal study of the national institute of neurological diseases and stroke. Washington, D.C.: National Institute of Health; 1972.
20. Gilman S, Martin L, Abrams D, et al. Educational attainment and smoking: A causal association? Int. J. Epidemiol. 2008;37(3):615–624. doi: 10.1093/ije/dym250. [PubMed] [Cross Ref]
21. Sullivan L, Massaro J, D'Agostino R. Presentation of multivariate data for clinincal use: The Framingham Study risk score functions. Stat. Med. 2004;23(10):1631–1660. doi: 10.1002/sim.1742. [PubMed] [Cross Ref]
22. Wilson P, D'Agostino R, Levy D, Belanger A, Sibershatz H, Kannel W. Prediction of coronary heart disease using risk factor categories. Circulation. 1998;97:1837–1847. [PubMed]
23. D'Agostino R, Grundy S, Sullivan L, Wilson P. Validation of the Framingham Coronary Heart Disease prediction scores - results of a multiple ethnic groups investigation. JAMA. 2001;286(2):180–187. doi: 10.1001/jama.286.2.180. [PubMed] [Cross Ref]
24. Allain C, Poon L, Chan C, Richmond W, Fu P. Enzymatic determination of total serum cholesterol. Clin Chem. 1974;20(4):470–475. [PubMed]
25. Rifai N, Cole T, Iannotti E, et al. Assessment of interlaboratory performance in external proficiency testing programs with a direct hdl-cholesterol assay. Clin Chem. 1988;44(7):1452–1458. [PubMed]
26. Mattu G, Heran B, Wright J. Overall accuracy of the BpTRU–an automated electronic blood pressure device. Blood Press Monit. 2004;9(1):47–52. doi: 10.1097/00126097-200402000-00009. [PubMed] [Cross Ref]
27. Woodcock RW, McGrew KS, Mather N. Technical manual: Woodcock-Johnson III. Itasca, IL: Riverside Publishing; 2001.
28. Lipkus IM, Samsa G, Rimer BK. General performance on a numeracy scale among highly educated samples. Med. Decis. Making. 2001;21(1):37–44. doi: 10.1177/0272989X0102100105. [PubMed] [Cross Ref]
29. Schwartz L, Woloshin S, Black W, Welch G. The role of numeracy in understanding the benefit of screening mammography. Ann. Intern. Med. 1997;127:966–971. [PubMed]
30. Graubard B, Korn E. Predictive margins with survey data. Biometrics. 1999;55(2):652–659. doi: 10.1111/j.0006-341X.1999.00652.x. [PubMed] [Cross Ref]
31. Barry W, Feldman S. Multiple regression in practice. Thousand Oaks, CA: Sage Publications; 1985.
32. McLaren L. Socioeconomic status and obesity. Epidemiol Rev. 2007;29:29–48. doi: 10.1093/epirev/mxm001. [PubMed] [Cross Ref]
33. Loucks E, Rehkopf D, Thurston R, Kawachi I. Socioeconomic disparities in metabolic syndrome differ by gender: Evidence from NHANES III. Ann. Epidemiol. 2007;17:19–26. doi: 10.1016/j.annepidem.2006.07.002. [PubMed] [Cross Ref]
34. Gunderson E, Lewis C, Murtaugh M, Quesenberry C, Smith West D, Sidney S. Long-term plasma lipid changes associated with a first birth: The Coronary Artery Risk Development in Young Adults Study. Am J Epidemiol. 2004;159:1028–1039. doi: 10.1093/aje/kwh146. [PubMed] [Cross Ref]
35. Gunderson E, Murtaugh M, Lewis C, Quesenberry C, West D, Sidney S. Excess gains in weight and waist circumference associated with childbearing: The Coronary Artery Risk Development in Young Adults Study (CARDIA) Int. J Obes Relat Metab Disord. 2004;28:525–535. doi: 10.1038/sj.ijo.0802551. [PMC free article] [PubMed] [Cross Ref]
36. Mathews T, Venura S. Birth and fertility rates by educational attainment: United States, 1994. Vital Stat Rep (Suppl) 1997;45:100–107.
37. Health literacy: A prescription to end confusion. Washington, DC: National Academies Press; 2004.

Articles from Journal of General Internal Medicine are provided here courtesy of Society of General Internal Medicine