Implementation Research to Address the United States Health Disadvantage: Report of a National Heart, Lung, and Blood Institute Workshop.

Four decades ago, U.S. life expectancy was within the same range as other high-income peer countries. However, during the past decades, the United States has fared worse in many key health domains resulting in shorter life expectancy and poorer health-a health disadvantage. The National Heart, Lung, and Blood Institute convened a panel of national and international health experts and stakeholders for a Think Tank meeting to explore the U.S. health disadvantage and to seek specific recommendations for implementation research opportunities for heart, lung, blood, and sleep disorders. Recommendations for National Heart, Lung, and Blood Institute consideration were made in several areas including understanding the drivers of the disadvantage, identifying potential solutions, creating strategic partnerships with common goals, and finally enhancing and fostering a research workforce for implementation research. Key recommendations included exploring why the United States is doing better for health indicators in a few areas compared with peer countries; targeting populations across the entire socioeconomic spectrum with interventions at all levels in order to prevent missing a substantial proportion of the disadvantage; assuring partnership have high-level goals that can create systemic change through collective impact; and finally, increasing opportunities for implementation research training to meet the current needs. Connecting with the research community at large and building on ongoing research efforts will be an important strategy. Broad partnerships and collaboration across the social, political, economic, and private sectors and all civil society will be critical-not only for implementation research but also for implementing the findings to have the desired population impact. Developing the relevant knowledge to tackle the U.S. health disadvantage is the necessary first step to improve U.S. health outcomes.

Country specific associations between social contact and mental health: evidence from civil servant studies across Great Britain, Japan and Finland.

OBJECTIVES: Little is known about which component, such as social contact of social networks is associated with mental health or whether such an association can be observed across countries. This study examined whether the association between frequent social contact and mental health differs by composition (relatives or friends) and whether the associations are similar across three occupational cohorts from Great Britain, Japan, and Finland. STUDY DESIGN: Cross-sectional analysis of data from three prospective cohort studies. METHODS: Participants were civil servants of a prospective cohort study based in London (Men: n = 4519; Women: n = 1756), in the West Coast of Japan (Men: n = 2571; Women: n = 1102), and in Helsinki, Finland (Men: n = 1181; Women: n = 5633); we included the information on study variables which is complete. Mental health function was the study outcome, indicated by the total score from the Mental Health Component on the Short Form Health Survey36. Participants reported frequencies of contacts with their relatives or friends via a questionnaire. Age, marital status, and occupational position were treated as confounders in this study. RESULTS: Findings from multiple regression showed that the associations between social contact and mental health function were different depending on country of origin and gender. Among British or Japanese men, frequent contact with both friends and relatives was positively associated with their mental health function, while only social contact with friends was significantly associated with mental health of Finnish men. In women, the patterns of the associations between social contact and mental health were more distinctive: friends for Great Britain, relatives for Japan, and friends and relatives for Finland. These significant associations were independent of the confounders. CONCLUSIONS: Social contact was related to mental health of working people; however, culture and gender are likely to be tapped into.

Alcohol consumption, drinking patterns, and cognitive function in older Eastern European adults.

OBJECTIVE: To investigate associations of frequency, quantity, binge, and problem drinking with cognitive function in older Eastern European adults. METHODS: The investigation included 14,575 participants, aged 47 to 78 years at cognitive assessment in 2006-2008 from Novosibirsk (Russia), Krakow (Poland), and 6 Czech towns participating in the HAPIEE (Health, Alcohol, and Psychosocial Factors in Eastern Europe) prospective cohort study. Average response rates were 59% at baseline (2002-2005) and 63% in 2006-2008. Alcohol consumption was assessed at baseline and in 2006-2008. Cognitive tests included immediate and delayed word recall, semantic fluency (animal naming), and letter cancellation. Associations between alcohol indices and cognitive scores were analyzed cross-sectionally (all measures from 2006 to 2008) and prospectively (alcohol and covariates from 2002 to 2005 and cognition from 2006 to 2008). RESULTS: In cross-sectional analyses, nondrinkers had lower cognitive scores and female moderate drinkers had better cognitive performance than light drinkers. Heavy, binge, and problem drinking were not consistently associated with cognitive function. Few associations were replicated in prospective analyses. Participants who stopped drinking during follow-up had worse cognition than stable drinkers; in men, regression coefficients (95% confidence interval) ranged from -0.26 (-0.36, -0.16) for immediate recall to -0.14 (-0.24, -0.04) for fluency. CONCLUSION: Regular and episodic heavy drinking were not consistently associated with cognitive function. Worse cognition in participants who stopped drinking during follow-up suggests that inclusion of less healthy ex-drinkers may partly explain poorer cognition in nondrinkers.

Conflicts between work and family life and subsequent sleep problems among employees from Finland, Britain, and Japan.

PURPOSE: Research on the association between family-to-work and work-to-family conflicts and sleep problems is sparse and mostly cross-sectional. We examined these associations prospectively in three occupational cohorts. METHODS: Data were derived from the Finnish Helsinki Health Study (n = 3,881), the British Whitehall II Study (n = 3,998), and the Japanese Civil Servants Study (n = 1,834). Sleep problems were assessed using the Jenkins sleep questionnaire in the Finnish and British cohorts and the Pittsburgh Sleep Quality Index in the Japanese cohort. Family-to-work and work-to-family conflicts measured whether family life interfered with work or vice versa. Age, baseline sleep problems, job strain, and self-rated health were adjusted for in logistic regression analyses. RESULTS: Adjusted for age and baseline sleep, strong family-to-work conflicts were associated with subsequent sleep problems among Finnish women (OR, 1.33 (95 % CI, 1.02-1.73)) and Japanese employees of both sexes (OR, 7.61 (95 % CI, 1.01-57.2) for women; OR, 1.97 (95 % CI, 1.06-3.66) for men). Strong work-to-family conflicts were associated with subsequent sleep problems in British, Finnish, and Japanese women (OR, 2.36 (95 % CI, 1.42-3.93), 1.62 (95 % CI, 1.20-2.18), and 5.35 (95 % CI, 1.00-28.55), respectively) adjusted for age and baseline sleep problems. In men, this association was seen only in the British cohort (OR, 2.02 (95 % CI, 1.42-2.88)). Adjustments for job strain and self-rated health produced no significant attenuation of these associations. CONCLUSION: Family-to-work and work-to-family conflicts predicted subsequent sleep problems among the majority of employees in three occupational cohorts.

Persistent long-standing illness and non-drinking over time, implications for the use of lifetime abstainers as a control group.

BACKGROUND: Non-drinkers are shown to have worse health than moderate drinkers in later life. We examine the preceding health status of non-drinkers in early adulthood, and secondly whether persistent poor health is associated with persistent non-drinking. METHODS: Using two prospective British birth cohort studies established in 1958 (National Child Development Study (NCDS)) and in 1970 (British Cohort Study (BCS)), participants who reported 'never' or 'never had an alcoholic drink' to drinking status questions in successive waves from 23 to 26 years in the NCDS/BCS were derived as 'lifetime abstainers'. Logistic regression on the odds of being a lifetime abstainer was carried out on changes in limiting long-standing illness (LLSI) in the NCDS and long-standing illness (LSI) in the BCS, adjusting for sex, education, poor psychosocial health, marital and parental status. RESULTS: Participants with an LLSI in consecutive waves since 23 years had 4.50 times the odds of someone who did not have an LLSI of being a lifetime abstainer at 33 years (95% CI 1.99 to 10.18) and 7.02 times the odds at 42 years (2.39 to 20.66) after adjusting for all factors. Similarly, in the BCS, having an LSI in consecutive waves resulted in higher odds of being a lifetime abstainer at 30 years (OR 2.80, 1.88 to 4.18) and 34 years (OR 3.33, 2.01 to 5.53). CONCLUSIONS: Persistent LSI was associated with remaining a non-drinker across adulthood. Studies comparing the health outcomes of moderate drinkers to lifetime abstainers that do not account for pre-existing poor health may overestimate the better health outcomes from moderate alcohol consumption.

Childhood socioeconomic position, adult socioeconomic position and social mobility in relation to markers of adiposity in early adulthood: evidence of differential effects by gender in the 1978/79 Ribeirao Preto cohort study.

BACKGROUND: Longitudinal studies drawn from high-income countries demonstrate long-term associations of early childhood socioeconomic deprivation with increased adiposity in adulthood. However, there are very few data from resource-poor countries where there are reasons to anticipate different gradients. Accordingly, we sought to characterise the nature of the socioeconomic status (SES)-adiposity association in Brazil. METHODS: We use data from the Ribeirao Preto Cohort Study in Brazil in which 9067 newborns were recruited via their mothers in 1978/79 and one-in-three followed up in 2002/04 (23-25years). SES, based on family income (salaries, interest on savings, pensions and so on), was assessed at birth and early adulthood, and three different adiposity measures (body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR)) ascertained at follow-up. The association between childhood SES, adult SES and social mobility (defined as four permutations of SES in childhood and adulthood: low-low, low-high, high-low, high-high), and the adiposity measures was examined using linear regression. RESULTS: There was evidence that the association between SES and the three markers of adiposity was modified by gender in both adulthood (P<0.02 for all outcomes) and childhood SES (P<0.02 for WC and WHR). Thus, in an unadjusted model, linear regression analyses showed that higher childhood SES was associated with lower adiposity in women (coefficient (95% confidence intervals) BMI: -1.49 (-2.29,-0.69); WC: -3.85 (-5.73,-1.97); WHR: -0.03 (-0.04,-0.02)). However, in men, higher childhood SES was related to higher adiposity (BMI: 1.03 (0.28,-1.78); WC: 3.15 (1.20, 5.09); WHR: 0.009 (-0.001, 0.019)) although statistical significance was not seen in all analyses. There was a suggestion that adult SES (but not adult health behaviours or birthweight) accounted for these relationships in women only. Upward mobility was associated with protection against greater adiposity in women but not men. CONCLUSION: In the present study, in men there was some evidence that both higher childhood and adulthood SES was related to a higher adiposity risk, while the reverse gradient was apparent in women.

Job strain in relation to body mass index: pooled analysis of 160 000 adults from 13 cohort studies.

BACKGROUND: Evidence of an association between job strain and obesity is inconsistent, mostly limited to small-scale studies, and does not distinguish between categories of underweight or obesity subclasses. OBJECTIVES: To examine the association between job strain and body mass index (BMI) in a large adult population. METHODS: We performed a pooled cross-sectional analysis based on individual-level data from 13 European studies resulting in a total of 161 746 participants (49% men, mean age, 43.7 years). Longitudinal analysis with a median follow-up of 4 years was possible for four cohort studies (n = 42 222). RESULTS: A total of 86 429 participants were of normal weight (BMI 18.5-24.9 kg m(-2) ), 2149 were underweight (BMI < 18.5 kg m(-2) ), 56 572 overweight (BMI 25.0-29.9 kg m(-2) ) and 13 523 class I (BMI 30-34.9 kg m(-2) ) and 3073 classes II/III (BMI ≥ 35 kg m(-2) ) obese. In addition, 27 010 (17%) participants reported job strain. In cross-sectional analyses, we found increased odds of job strain amongst underweight [odds ratio 1.12, 95% confidence interval (CI) 1.00-1.25], obese class I (odds ratio 1.07, 95% CI 1.02-1.12) and obese classes II/III participants (odds ratio 1.14, 95% CI 1.01-1.28) as compared with participants of normal weight. In longitudinal analysis, both weight gain and weight loss were related to the onset of job strain during follow-up. CONCLUSIONS: In an analysis of European data, we found both weight gain and weight loss to be associated with the onset of job strain, consistent with a 'U'-shaped cross-sectional association between job strain and BMI. These associations were relatively modest; therefore, it is unlikely that intervention to reduce job strain would be effective in combating obesity at a population level.

The impact of cash transfers to poor women in Colombia on BMI and obesity: prospective cohort study.

INTRODUCTION: Prevalence of obesity is rising in Latin America and increasingly affecting socially disadvantaged groups, particularly women. Conditional cash transfers are recently established welfare interventions in the region. One, Familias en Accion, transfers ∼20% of average monthly income to women in Colombia's poorest families. Previous work has found that families buy more food as a result. We tested the hypothesis that participation in Familias would be associated with increasing body mass index (BMI) in participating women. METHODS: Women from participating areas and control areas (matched on environmental and socioeconomic criteria) were surveyed in 2002 and 2006. Pregnant, breast-feeding or women aged <18 or with BMI <18.5 kg m(-2) were excluded. The sample comprises 835 women from control and 1238 from treatment areas. Because some treatment areas started Familias shortly before baseline data collection, a dummy variable was created that identified exposure independent of time point or area. Follow-up was 61.5%. BMI was measured by trained personnel using standardized techniques. Overweight was defined as BMI ≥ 25 kg m(-2) and obesity as ≥ 30 kg m(-2). The effect of Familias was estimated using linear regression (or logistic regression for dichotomous outcomes) in a double-difference technique, controlling for several individual, household and area characteristics, including parity and baseline BMI, using robust standard-errors clustered at area-level in an intention-to-treat analysis. RESULTS: At baseline, women's mean age was 33.3 years and mean BMI 25.3 kg m(-2); 12.3% women were obese. After adjustment, exposure to Familias was significantly associated with increased BMI (ß=0.25; 95% confidence interval (CI) 0.03, 0.47; P=0.03). Age (ß=0.09; 95% CI 0.06, 0.13; P<0.001) and household wealth (ß=0.78; 95% CI 0.41, 1.15; P<0.001) were also positively associated with BMI. Familias was also associated with increased odds of obesity (odds ratio (OR)=1.27; 95% CI 1.03, 1.57; P=0.03), as was age (OR=1.04; 95% CI 1.02, 1.06; P=0.001). CONCLUSION: Conditional cash transfers to poor women in Colombia are independently associated with increasing BMI and obesity risk. Although conditional cash transfers are generally regarded as popular and successful schemes, parallel interventions at individual, household and community level are needed to avoid unanticipated adverse outcomes.

Long working hours and symptoms of anxiety and depression: a 5-year follow-up of the Whitehall II study.

BACKGROUND: Although long working hours are common in working populations, little is known about the effect of long working hours on mental health. METHOD: We examined the association between long working hours and the onset of depressive and anxiety symptoms in middle-aged employees. Participants were 2960 full-time employees aged 44 to 66 years (2248 men, 712 women) from the prospective Whitehall II cohort study of British civil servants. Working hours, anxiety and depressive symptoms, and covariates were measured at baseline (1997-1999) followed by two subsequent measurements of depressive and anxiety symptoms (2001 and 2002-2004). RESULTS: In a prospective analysis of participants with no depressive (n=2549) or anxiety symptoms (n=2618) at baseline, Cox proportional hazard analysis adjusted for baseline covariates showed a 1.66-fold [95% confidence interval (CI) 1.06-2.61] risk of depressive symptoms and a 1.74-fold (95% CI 1.15-2.61) risk of anxiety symptoms among employees working more than 55 h/week compared with employees working 35-40 h/week. Sex-stratified analysis showed an excess risk of depression and anxiety associated with long working hours among women [hazard ratios (HRs) 2.67 (95% CI 1.07-6.68) and 2.84 (95% CI 1.27-6.34) respectively] but not men [1.30 (0.77-2.19) and 1.43 (0.89-2.30)]. CONCLUSIONS: Working long hours is a risk factor for the development of depressive and anxiety symptoms in women.

Incidence and prognosis of angina pectoris in South Asians and Whites: 18 years of follow-up over seven phases in the Whitehall-II prospective cohort study.

BACKGROUND: Whether the higher coronary mortality in South Asians compared with White populations is due to a higher incidence of disease is not known. This study assessed cumulative incidence of chest pain in South Asians and Whites, and prognosis of chest pain. METHODS: Over seven phases of 18-year follow-up of the Whitehall-II study (9,775 civil servants: 9,195 White, 580 South Asian), chest pain was assessed using the Rose questionnaire. Coronary death/non-fatal myocardial infarction was examined comparing those with chest pain to those with no chest pain at baseline. RESULTS: South Asians had higher cumulative frequencies of typical angina by Phase 7 (17.0 versus 11.3%, P < 0.001) and exertional chest pain (15.4 versus 8.5%, P < 0.001) compared with Whites. Typical angina and exertional chest pain at baseline were associated with a worse prognosis compared with those with no chest pain in both groups (typical angina, South Asians: HR, 4.67 and 95% CI, 2.12-0.30; Whites: HR, 3.56 95% CI, 2.59-4.88). Baseline non-exertional chest pain did not confer a worse prognosis. Across all types of pain, prognosis was worse in South Asians. CONCLUSION: South Asians had higher cumulative incidence of angina than Whites. In both, typical angina and exertional chest pain were associated with worse prognosis compared with those with no chest pain.

Do the Joint British Society (JBS2) guidelines on prevention of cardiovascular disease with respect to plasma glucose improve risk stratification in the general population? Prospective cohort study.

AIMS: British guidelines on vascular disease prevention recommend adding a random (casual) blood glucose measurement to a lipid profile in those aged > or = 40 years. To assess this recommendation, we compared the predictive value of a risk model based on the Framingham risk score alone to one which additionally included information on fasting blood glucose, with respect to incident coronary heart disease (CHD) over 11 years. METHOD: Men and women aged 40-63 years in Whitehall II were followed up for incident CHD: death/non-fatal myocardial infarction; angina confirmed by doctor diagnosis or electrocardiogram (ECG) and all first events. Fasting blood glucose was specified as a continuous variable or categorized by World Health Organization (WHO) 1999 glycaemic status (normal glucose tolerance, impaired fasting glucose or newly diagnosed diabetes). RESULTS: The hazard ratio for incident CHD was 1.10 (95%CI 1.09; 1.12) in men and 1.13 (1.10; 1.17) in women per percentage point increase in Framingham risk. The excess risk remained unchanged in models which added glycaemic status or continuous fasting glucose. The area under the receiver operating characteristic (ROC) curve for the Framingham score and incident coronary heart disease [0.70 (0.68; 0.73)] did not change when glycaemic status or fasting glucose was added to the prediction model. Reclassification with these modified models improved discrimination based on the Framingham score alone when glycaemic status was added, net reclassification improvement 2.4% (95% CI 0.2%; 4.6%), but not when fasting glucose was added. CONCLUSION: Better detection of unrecognized diabetes is a valuable consequence of including a random blood glucose in a vascular risk profile. Our results suggest that this strategy is unlikely to improve risk stratification for CHD.

Work disability following major organisational change: the Whitehall II study.

BACKGROUND: Privatisation and private sector practices have been increasingly applied to the public sector in many industrialised countries. Over the same period, long-term work disability has risen substantially. We examined whether a major organisational change--the transfer of public sector work to executive agencies run on private sector lines--was associated with an increased risk of work disability. METHODS: The study uses self-reported data from the prospective Whitehall II cohort study. Associations between transfer to an executive agency assessed at baseline (1991-1994) and work disability ascertained over a period of approximately 8 years at three follow-up surveys (1995-1996, 1997-1999 and 2001) were examined using Cox proportional hazard models. RESULTS: In age- and sex-adjusted models, risk of work disability was higher among the 1263 employees who were transferred to an executive agency (HR 1.90, 95% CI 1.46 to 2.48) compared with the 3419 employees whose job was not transferred. These findings were robust to additional adjustment for physical and mental health and health behaviours at baseline. CONCLUSIONS: Increased work disability was observed among employees exposed to the transfer of public sector work to executive agencies run on private sector lines. This may highlight an unintentional cost for employees, employers and society.

Performance of existing risk scores in screening for undiagnosed diabetes: an external validation study.

AIM: To compare the performance of nine published strategies for the selection of individuals prior to screening for undiagnosed diabetes. METHODS: We conducted a validation study, based on a cross-sectional analysis of 6990 participants of the Whitehall II study, an occupational cohort of civil servants in London. We calculated sensitivity, specificity and the area under the receiver operating characteristic (ROC) curve, indicative of the ability of a risk score to correctly identify those with undiagnosed diabetes. RESULTS: The prevalence of unknown diabetes was 2.0%. At a set level of sensitivity (0.70), the specificity of the different scores ranged between 0.41 and 0.57. A reference model, based solely on age and body mass index had an area under the ROC curve of 0.67 [95% confidence interval (CI): 0.62, 0.72]. Four scores had a lower area under the ROC curve (lowest ROC AUC: 0.62; 95% CI: 0.58, 0.67) compared with the reference model, while the other five scores had similar areas (highest ROC AUC: 0.68; 95% CI: 0.63, 0.72). All ROC curve areas were lower than those reported in the original publications and validation studies. CONCLUSIONS: Existing risk scores for the detection of undiagnosed diabetes perform less well in a large validation cohort compared with previous validation studies. Our study indicates that non-invasive risk scores require further refinement and testing before they can be used as the first step in a diabetes screening programme.

Hostility and depressive mood: results from the Whitehall II prospective cohort study.

BACKGROUND: The psychosocial vulnerability model of hostility posits that hostile individuals, given their oppositional attitudes and behaviours, are more likely to have increased interpersonal conflicts, lower social support, more stressful life events (SL-E) and higher likelihood of depression. However, little research has tested this hypothesis using large-scale prospective samples. The present study aims to assess the predictive value of hostility for depressive mood. METHOD: Data are from 3399 participants in the Whitehall II cohort study, aged 35-55 years at baseline (phase 1 1985-1988). Cynical hostility was measured at phase 1. Depressive mood was assessed at phase 7 (2002-2004). Sociodemographic characteristics, health-related behaviours, common mental disorders and antidepressant medication intake were assessed at phase 1. SL-E and confiding/emotional support were measured at phases 1, 2 (1989-1990) and 5 (1997-1999). RESULTS: Compared with participants in the lowest quartile of cynical hostility, those in the highest quartiles were more likely to have depressive mood [second quartile: odds ratio (OR) 1.58, 95% confidence interval (CI) 1.14-2.20; third quartile: OR 2.78, 95% CI 2.03-3.77; fourth quartile: OR 4.66, 95% CI 3.41-6.36] in analysis adjusted for sociodemographic characteristics. This graded association was somewhat attenuated (18%) but remained robust to adjustments for the covariates measured at baseline and follow-up. The association was also evident in participants free of mental health difficulties at baseline. CONCLUSIONS: Cynical hostility is a strong and robust predictor of depressive mood. Consideration of personality characteristics may be crucial to the understanding and management of depression.

The association of cognitive performance with mental health and physical functioning strengthens with age: the Whitehall II cohort study.

BACKGROUND: Cognitive performance has been associated with mental and physical health, but it is unknown whether the strength of these associations changes with ageing and with age-related social transitions, such as retirement. We examined whether cognitive performance predicted mental and physical health from midlife to early old age. METHOD: Participants were 5414 men and 2278 women from the Whitehall II cohort study followed for 15 years between 1991 and 2006. The age range included over the follow-up was from 40 to 75 years. Mental health and physical functioning were measured six times using SF-36 subscales. Cognitive performance was assessed three times using five cognitive tests assessing verbal and numerical reasoning, verbal memory, and phonemic and semantic fluency. Socio-economic status (SES) and retirement were included as covariates. RESULTS: High cognitive performance was associated with better mental health and physical functioning. Mental health differences associated with cognitive performance widened with age from 39 to 76 years of age, whereas physical functioning differences widened only between 39 and 60 years and not after 60 years of age. SES explained part of the widening differences in mental health and physical functioning before age 60. Cognitive performance was more strongly associated with mental health in retired than non-retired participants, which contributed to the widening differences after 60 years of age. CONCLUSIONS: The strength of cognitive performance in predicting mental and physical health may increase from midlife to early old age, and these changes may be related to SES and age-related transitions, such as retirement.

Association between passive jobs and low levels of leisure-time physical activity: the Whitehall II cohort study.

BACKGROUND: There is mixed evidence on the association between psychosocial work exposures (ie, passive jobs) and physical activity, but previous studies did not take into account the effect of cumulative exposures nor did they examine different trajectories in exposure. We investigated whether exposure to passive jobs, measured three times over an average of 5 years, is associated with leisure-time physical activity (LTPA). METHODS: Data were from working men (n = 4291) and women (n = 1794) aged 35-55 years who participated in the first three phases of the Whitehall II prospective cohort. Exposure to passive jobs was measured at each phase and LTPA at phases 1 and 3. Participants were categorised according to whether or not they worked in a passive job at each phase, leading to a scale ranging from 0 (non-passive job at all three phases) to 3 (passive job at all three phases). Poisson regression with robust variance estimates were used to assess the prevalence ratios of low LTPA. RESULTS: An association was found in men between exposure to passive jobs over 5 years and low LTPA at follow-up, independently of other relevant risk factors. The prevalence ratio for low LTPA in men was 1.16 (95% CI 1.01 to 1.33) times greater for employees with three reports of passive job than for those who had never worked in passive jobs. No association was observed in women. CONCLUSION: This study provides evidence that working in passive jobs may encourage a passive lifestyle in men.

Socioeconomic position and cognitive decline using data from two waves: what is the role of the wave 1 cognitive measure?

BACKGROUND: Analysis of change in health status using data from two waves can be examined either adjusted or unadjusted for baseline health status. The effect of socioeconomic position (SEP) on cognitive change was assessed using both these strategies and the implications of the analyses are discussed. METHODS: Data from 1261 men and 483 women of the Whitehall II cohort study, aged 50-55 years at wave 1, were used. Cognition was assessed at both waves using a test of verbal memory, and two tests of verbal fluency. Analysis of variance (ANOVA) was used to estimate the effect of SEP on change score and analysis of covariance (ANCOVA) was used to estimate this effect adjusted for the baseline cognitive score. The ANCOVA estimates were corrected for bias due to measurement error (estimated based on 3-month test-retest). Finally, ANCOVA estimates were examined for increasing levels of measurement error. RESULTS: The results of the ANOVA suggest no effect of SEP on cognitive decline. In contrast, the ANCOVA suggests significantly greater cognitive decline in the lower SEP groups. However, the ANCOVA estimates for the effect of wave 1 cognition show evidence for regression to the mean due to the presence of measurement error. The corrected ANCOVA estimates show no association between SEP and cognitive decline. CONCLUSIONS: Results from analysis of change using two waves of observational data, when adjusted for baseline, should be interpreted with caution.