Building on a unique exposure assessment project in New York, New York, we examined the relationship of particulate matter with aerodynamic diameter less than 2.5 μm and nitrogen dioxide with birth weight, restricting the population to term births to nonsmokers, along with other restrictions, to isolate the potential impact of air pollution on growth. We included 252,967 births in 2008–2010 identified in vital records, and we assigned exposure at the residential location by using validated models that accounted for spatial and temporal factors. Estimates of association were adjusted for individual and contextual sociodemographic characteristics and season, using linear mixed models to quantify the predicted change in birth weight in grams related to increasing pollution levels. Adjusted estimates for particulate matter with aerodynamic diameter less than 2.5 μm indicated that for each 10-µg/m3 increase in exposure, birth weights declined by 18.4, 10.5, 29.7, and 48.4 g for exposures in the first, second, and third trimesters and for the total pregnancy, respectively. Adjusted estimates for nitrogen dioxide indicated that for each 10-ppb increase in exposure, birth weights declined by 14.2, 15.9, 18.0, and 18.0 g for exposures in the first, second, and third trimesters and for the total pregnancy, respectively. These results strongly support the association of urban air pollution exposure with reduced fetal growth.
air pollution; birth weight; nitrogen dioxide; particulate matter; pregnancy
The question of whether cigarette smoking was associated with lung cancer was central to the expansion of epidemiology into the study of chronic diseases in the 1950s. The culmination of this era was the 1964 report of the Advisory Committee to the Surgeon General, a landmark document that included an objective synthesis of the evidence of the health consequences of smoking according to causal criteria. The report concluded that cigarette smoking was a cause of lung cancer in men and sufficient in scope that “remedial action” was warranted at the societal level. The 2014 Surgeon General's report commemorates the 50th anniversary of the 1964 report. The evidence on the health consequences of smoking has been updated many times in Surgeon General's reports since 1964. These have summarized our increasingly greater understanding of the broad spectrum of the deleterious health effects of exposure to tobacco smoke across most major organ systems. In turn, this evidence has been translated into tobacco control strategies implemented to protect the public's health. The Surgeon General report process is an enduring example of evidence-based public health in practice. Substantial progress has been made, but cigarette smoking remains one of the most pressing global health issues of our time.
causal inference; cigarette smoking; epidemiology; evidence-based public health; history; lung cancer; research methods; secondhand smoke exposure
Limited information on age- and sex-specific estimates of influenza-associated death with different underlying causes is currently available. We regressed weekly age- and sex-specific US mortality outcomes underlying several causes between 1997 and 2007 to incidence proxies for influenza A/H3N2, A/H1N1, and B that combine data on influenzalike illness consultations and respiratory specimen testing, adjusting for seasonal baselines and time trends. Adults older than 75 years of age had the highest average annual rate of influenza-associated mortality, with 141.15 deaths per 100,000 people (95% confidence interval (CI): 118.3, 163.9), whereas children under 18 had the lowest average mortality rate, with 0.41 deaths per 100,000 people (95% CI: 0.23, 0.60). In addition to respiratory and circulatory causes, mortality with underlying cancer, diabetes, renal disease, and Alzheimer disease had a contribution from influenza in adult age groups, whereas mortality with underlying septicemia had a contribution from influenza in children. For adults, within several age groups and for several underlying causes, the rate of influenza-associated mortality was somewhat higher in men than in women. Of note, in men 50–64 years of age, our estimate for the average annual rate of influenza-associated cancer mortality per 100,000 persons (1.90, 95% CI: 1.20, 2.62) is similar to the corresponding rate of influenza-associated respiratory deaths (1.81, 95% CI: 1.42, 2.21). Age, sex, and underlying health conditions should be considered when planning influenza vaccination and treatment strategies.
age; influenza; mortality; sex; underlying cause
The objective of this study was to determine the association between maternal 25-hydroxyvitamin D (25(OH)D) and the risk of spontaneous preterm birth (sPTB) before 35 weeks’ gestation. A random subcohort from the US Collaborative Perinatal Project (1959–1965) was sampled (n = 2,629) and augmented with all remaining cases of sPTB before 35 weeks’ gestation for a total of 767 cases. Banked serum samples collected at 26 weeks’ gestation or earlier were assayed for 25(OH)D. Constructs for vascular histology and inflammatory histology were developed from placental pathology examinations. There was no relationship between 25(OH)D and sPTB among white women. Among nonwhite mothers, serum 25(OH)D levels of 30–<50, 50–<75, and ≥75 nmol/L were associated with reductions of 1.0–1.6 cases of sPTB per 100 live births and 20%–30% reductions in risk of sPTB compared with 25(OH)D levels less than 30 nmol/L after adjustment for prepregnancy body mass index (weight (kg)/height (m)2), season, and other confounders. This association was driven by inflammation-mediated cases of sPTB and sPTB cases without placental lesions. A sensitivity analysis for unmeasured confounding by exercise, fish intake, and skin color suggested some bias away from the null in the conventional results, but conclusions were generally supported. The vitamin D–sPTB relationship should be examined in modern cohorts with detailed data on skin pigmentation and other covariates.
gestational age; 25-hydroxyvitamin D; inflammation; placenta; pregnancy; preterm birth
To maximize statistical power in studies of mammographic density and breast cancer, it is advantageous to combine data from several studies, but standardization of the density assessment is desirable. Using data from 4 case-control studies, we describe the process of reassessment and the resulting correlation between values, identify predictors of differences in density readings, and evaluate the strength of the association between mammographic density and breast cancer risk using different representations of density values. The pooled analysis included 1,699 cases and 2,422 controls from California (1990–1998), Hawaii (1996–2003), Minnesota (1992–2001), and Japan (1999–2003). In 2010, a single reader reassessed all images for mammographic density using Cumulus software (Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada). The mean difference between original and reassessed percent density values was −0.7% (95% confidence interval: −1.1, −0.3), with a correlation of 0.82 that varied by location (r = 0.80–0.89). Case status, weight status, age, parity, density assessment method, mammogram view, and race/ethnicity were significant determinants of the difference between original and reassessed values; in combination, these factors explained 9.2% of the variation. The associations of mammographic density with breast cancer and the model fits were similar using the original values and the reassessed values but were slightly strengthened when a calibrated value based on 100 reassessed radiographs was used.
breast cancer; epidemiologic methods; ethnicity; mammographic density; pooling; risk
Studies have suggested that exposure to ultraviolet (UV) light may increase risk of herpes simplex virus (HSV) recurrence. Between 1993 and 1997, the Herpetic Eye Disease Study (HEDS) randomized 703 participants with ocular HSV to receipt of acyclovir or placebo for prevention of ocular HSV recurrence. Of these, 308 HEDS participants (48% female and 85% white; median age, 49 years) were included in a nested study of exposures thought to cause recurrence and were followed for up to 15 months. We matched weekly UV index values from the National Oceanic and Atmospheric Administration to each participant's study center and used marginal structural Cox models to account for time-varying psychological stress and contact lens use and selection bias from dropout. There were 44 recurrences of ocular HSV, yielding an incidence of 4.3 events per 1,000 person-weeks. Weighted hazard ratios comparing persons with ≥8 hours of time outdoors to those with less exposure were 0.84 (95% confidence interval (CI): 0.27, 2.63) and 3.10 (95% CI: 1.14, 8.48) for weeks with a UV index of <4 and ≥4, respectively (ratio of hazard ratios = 3.68, 95% CI: 0.43, 31.4). Though results were imprecise, when the UV index was higher (i.e., ≥4), spending 8 or more hours per week outdoors was associated with increased risk of ocular HSV recurrence.
cohort studies; herpes simplex virus; recurrence; sunlight; ultraviolet light; UV index
The method of maximum likelihood is widely used in epidemiology, yet many epidemiologists receive little or no education in the conceptual underpinnings of the approach. Here we provide a primer on maximum likelihood and some important extensions which have proven useful in epidemiologic research, and which reveal connections between maximum likelihood and Bayesian methods. For a given data set and probability model, maximum likelihood finds values of the model parameters that give the observed data the highest probability. As with all inferential statistical methods, maximum likelihood is based on an assumed model and cannot account for bias sources that are not controlled by the model or the study design. Maximum likelihood is nonetheless popular, because it is computationally straightforward and intuitive and because maximum likelihood estimators have desirable large-sample properties in the (largely fictitious) case in which the model has been correctly specified. Here, we work through an example to illustrate the mechanics of maximum likelihood estimation and indicate how improvements can be made easily with commercial software. We then describe recent extensions and generalizations which are better suited to observational health research and which should arguably replace standard maximum likelihood as the default method.
epidemiologic methods; maximum likelihood; modeling; penalized estimation; regression; statistics
Previous studies evaluating the association of prenatal exposure to diethylstilbestrol (DES), a potent endocrine disruptor, with incidence of uterine leiomyomata (UL) have had conflicting results. We evaluated the association between prenatal DES exposure and incident UL in women in the Nurses' Health Study II from 1989 to 2009. Women were aged 25–42 years at enrollment and had a prenatal exposure window corresponding to DES use. The analytical sample was larger than previous studies and included 102,164 premenopausal women with intact uteri, no prior history of UL or cancer, and prenatal DES exposure. Multivariable-adjusted Cox proportional hazard models were used to estimate the relationship between DES exposure and UL risk. During 1,273,342 person-years of follow-up, there were 11,831 incident cases of UL. Women with prenatal exposure to DES had a higher incidence of UL compared with unexposed women, with an adjusted hazard ratio of 1.12 (95% confidence interval: 0.98, 1.27). Risk was strongest for women exposed to DES in the first trimester, when exposure corresponds to early stages of fetal Müllerian development (adjusted hazard ratio = 1.21, 95% confidence interval: 1.02, 1.43). These results suggest that first-trimester DES exposure may be associated with an increased risk of UL, but they must be interpreted with concern for detection and recall biases.
diethylstilbestrol; prospective cohort; uterine fibroid
Using data from the National Institutes of Health-AARP Diet and Health Study, we evaluated the influence of adulthood weight history on mortality risk. The National Institutes of Health-AARP Diet and Health Study is an observational cohort study of US men and women who were aged 50–71 years at entry in 1995–1996. This analysis focused on 109,947 subjects who had never smoked and were younger than age 70 years. We estimated hazard ratios of total and cause-specific mortality for recalled body mass index (BMI; weight (kg)/height (m)2) at ages 18, 35, and 50 years; weight change across 3 adult age intervals; and the effect of first attaining an elevated BMI at 4 successive ages. During 12.5 years' follow-up through 2009, 12,017 deaths occurred. BMI at all ages was positively related to mortality. Weight gain was positively related to mortality, with stronger associations for gain between ages 18 and 35 years and ages 35 and 50 years than between ages 50 and 69 years. Mortality risks were higher in persons who attained or exceeded a BMI of 25.0 at a younger age than in persons who reached that threshold later in adulthood, and risks were lowest in persons who maintained a BMI below 25.0. Heavier initial BMI and weight gain in early to middle adulthood strongly predicted mortality risk in persons aged 50–69 years.
body mass index; body weight; mortality; obesity; weight gain; weight loss
We investigated associations between short-term exposure to air pollution and central augmentation index and augmentation pressure, correlates of arterial stiffness, in a cohort of elderly men in the Boston, Massachusetts, metropolitan area. This longitudinal analysis included 370 participants from the Veterans Affairs Normative Aging Study with up to 2 visits between 2007 and 2011 (n = 445). Augmentation index (as %) and augmentation pressure (in mmHg) were measured at each visit by using radial artery applanation tonometry for pulse wave analysis and modeled in a mixed effects regression model as continuous functions of moving averages of air pollution exposures (over 4 hours and 1, 3, 7, and 14 days). The results suggest that short-term changes in air pollution were associated with augmentation index and augmentation pressure at several moving averages. Interquartile range (IQR) increases in 3-day average exposure to particles with aerodynamic diameter less than 2.5 μm (3.6-μg/m3 IQR increase) and sulfate (1.4-μg/m3 IQR increase) and 1-day average exposure to particle number counts (8,741-counts/cm3 IQR increase) were associated with augmentation index values that were 0.8% (95% confidence interval (CI): 0.2, 1.4), 0.6% (95% CI: 0.1, 1.2), and 1.7% (95% CI: 0.4, 2.9) higher, respectively. Overall, the findings were similar for augmentation pressure. The findings support the hypothesis that exposure to air pollution may affect vascular function.
air pollution; particulate matter; pulse wave analysis
Several studies have examined associations between particulate matter with aerodynamic diameter of 2.5 µm or less (PM2.5) and preterm birth, but it is uncertain whether results were affected by individual predispositions (e.g., genetic factors, social conditions) that might vary considerably between women. We tested the hypothesis that a woman is at greater risk of preterm delivery when she has had elevated exposure to ambient PM2.5 during a pregnancy than when she has not by comparing pregnancies in the same woman. From 271,204 births, we selected 29,175 women who had vaginal singleton livebirths at least twice in Connecticut in 2000–2006 (n = 61,688 births). Analyses matched pregnancies to the same woman. Adjusted odds ratios per interquartile range (2.33-µg/m3) increase in PM2.5 in the first trimester, second trimester, third trimester, and whole pregnancy were 1.07 (95% confidence interval (CI): 1.00, 1.15), 0.96 (95% CI: 0.90, 1.03), 1.03 (95% CI: 0.97, 1.08), and 1.13 (95% CI: 1.01, 1.28), respectively. Among Hispanic women, the odds ratio per interquartile range increase in whole-pregnancy exposure was 1.31 (95% CI: 1.00, 1.73). Pregnancies with elevated PM2.5 exposure were more likely to result in preterm birth than were other pregnancies to the same woman at lower exposure. Associations were most pronounced in the first trimester and among Hispanic women.
air pollution; environmental pollution; longitudinal studies; particulate matter; pregnancy; pregnancy outcomes; preterm birth
Data from animal models, historical cohorts, and modern epidemiologic studies have suggested that maternal characteristics can affect reproductive health of offspring; however, distinguishing between prenatal and postnatal contributions is difficult. Anogenital distance (AGD), the distance from the anus to the genitals, is believed to be a biomarker of prenatal androgen exposure in many species, and in humans it has been associated with several adult reproductive health outcomes. We used data from a pregnancy cohort study conducted in 4 US cities from 1999–2005 to examine whether AGD measurements in infants were associated with maternal self-reported age at conception, age at menarche, age at first birth, parity, and gravidity. AGD was measured in 289 infants (140 male, 149 female) born to study participants. After adjustment for relevant covariates, in linear regression models stratified by infant sex, maternal age was positively associated with AGD in male infants (AGD, anus to penis: β = 0.50, P = 0.002; AGD, anus to scrotum: β = 0.29, P = 0.02) but not female infants. Parity was inversely associated with AGD (anus to scrotum; β = −1.68, P = 0.03) in male infants. No other maternal characteristic predicted AGD in either sex. The mechanism underlying the unexpected relationship between maternal characteristics and AGD is unknown; however, we suggest several possibilities for future study.
anogenital distance; maternal age; prenatal hormones; reproduction; sex differences
The majority of adults aged 18–34 years have only cellular phones, making random-digit dialing of landline telephones an obsolete methodology for surveillance of this population. However, 95% of this group has cellular phones. This article reports on the 2011 National Young Adult Health Survey (NYAHS), a pilot study conducted in the 50 US states and Washington, DC, that used random-digit dialing of cellular phones and benchmarked this methodology against that of the 2011 Behavioral Risk Factor Surveillance System (BRFSS). Comparisons of the demographic distributions of subjects in the NYAHS and BRFSS (aged 18–34 years) with US Census data revealed adequate reach for all demographic subgroups. After adjustment for design factors, the mean absolute deviations across demographic groups were 3 percentage points for the NYAHS and 2.8 percentage points for the BRFSS, nationally, and were comparable for each census region. Two-sided z tests comparing cigarette smoking prevalence revealed no significant differences between NYAHS and BRFSS participants overall or by subgroups. The design effects of the sampling weight were 2.09 for the NYAHS and 3.26 for the BRFSS. Response rates for the NYAHS and BRFSS cellular phone sampling frames were comparable. Our assessment of the NYAHS methodology found that random-digit dialing of cellular phones is a feasible methodology for surveillance of young adults.
cellular phone; random-digit dialing; sample quality; sampling; single-frame sampling design; surveillance; survey methodology; young adults
Evidence for the association of type 2 diabetes mellitus (DM) with colorectal neoplasms is contradictory, and African Americans have been underrepresented in the studies published to date. In a nested case-control study (1995–2009), we examined DM and insulin therapy as risk factors for colorectal adenomas in African American women enrolled in the ongoing Black Women's Health Study. From women reporting ever having undergone a gastrointestinal endoscopy, 917 cases of colorectal adenoma were compared with 2,751 controls without a colorectal polyp, matched on age and follow-up time. Cases were verified by medical record review. We used multivariable logistic regression analyses that included DM exposures and selected confounders. There were no overall associations between DM and adenoma risk or between insulin use and adenoma risk. However, DM without insulin use was inversely associated with risk of colon adenomas (odds ratio (OR) = 0.71, 95% confidence interval (CI): 0.52, 0.97) but not rectal adenomas. DM was inversely associated with adenoma risk in women older than 55 years (OR = 0.64, 95% CI: 0.44, 0.91) but not in women 55 years or younger (OR = 1.24, 95% CI: 0.81, 1.89). Future research should attempt to replicate the unexpected inverse association of DM with colon adenoma risk among older African American women.
African American women; colorectal adenoma; colorectal cancer; diabetes; insulin
High body mass index (BMI) (calculated as weight (kg)/height (m)2) is associated with increased asthma risk, but uncertainty persists about the role of physical activity. We estimated the independent and joint associations of hypothetical interventions on BMI and physical activity with the risk of adult-onset asthma in 76,470 asthma-free women from the Nurses’ Health Study who were followed between 1988 and 1998. Information about asthma, BMI, physical activity, and other factors was updated every 2 years. We used the parametric g-formula to estimate the 10-year asthma risk in the following 4 scenarios: no intervention, 5% BMI reduction in a 2-year period for those who were overweight or obese, at least 2.5 hours/week of moderate-to-vigorous physical activity, and both of the previous 2 interventions. At baseline, women had a mean age of 55 (standard deviation, 7) years and a mean BMI of 25.4 (standard deviation, 4.8). Median time spent in physical activity was 0.7 hours/week. During follow-up, 1,146 women developed asthma. The 10-year asthma risk under no intervention was 1.5%. Compared with no intervention, the population risk ratios were 0.96 (95% confidence interval (CI): 0.93, 0.99) under the BMI intervention, 0.96 (95% CI: 0.81, 1.10) under the physical activity intervention, and 0.92 (95% CI: 0.78, 1.06) under the joint intervention. Interventions on BMI and physical activity may have a modest impact on the risk of adult-onset asthma in this population of US women.
asthma; body mass index; g-formula; hypothetical interventions; physical activity
We wanted to compare cancer incidence rates between Parkinson's disease (PD) patients and persons without PD, as well as between siblings of these groups. We conducted a family-based matched cohort study based on nationwide Swedish health registries and the Swedish Multi-Generation Register. We assessed risk of incident cancer in PD patients (n = 11,786) during 1964–2009 versus a matched cohort of PD-free individuals (n = 58,930) and in siblings of PD patients (n = 16,841) versus siblings of PD-free individuals (n = 84,205). Hazard ratios with 95% confidence intervals were estimated using Cox proportional hazards regression. Cancer occurrence was slightly higher in PD patients than in PD-free individuals (hazard ratio (HR) = 1.05, 95% confidence interval (CI): 1.00, 1.10), largely because of cancers arising within 1 year before or after the index date for PD, but risk of smoking-related cancers was lower (HR = 0.87, 95% CI: 0.79, 0.96). PD patients had a higher risk of melanoma both up to 1 year before the PD index date (HR = 1.53, 95% CI: 1.23, 1.91) and from 1 year after the index date onward (HR = 1.46, 95% CI: 1.01, 2.10). In the sibling comparison, cancer occurrence was largely similar. These results indicate that melanoma risk is higher among PD patients and that mechanisms other than familial ones explain the association.
cancer; cohort studies; hazard ratio; Parkinson's disease; siblings
We investigated body size and survival by race/ethnicity in 11,351 breast cancer patients diagnosed from 1993 to 2007 with follow-up through 2009 by using data from questionnaires and the California Cancer Registry. We calculated hazard ratios and 95% confidence intervals from multivariable Cox proportional hazard model–estimated associations of body size (body mass index (BMI) (weight (kg)/height (m)2) and waist-hip ratio (WHR)) with breast cancer–specific and all-cause mortality. Among 2,744 ascertained deaths, 1,445 were related to breast cancer. Being underweight (BMI <18.5) was associated with increased risk of breast cancer mortality compared with being normal weight in non-Latina whites (hazard ratio (HR) = 1.91, 95% confidence interval (CI): 1.14, 3.20), whereas morbid obesity (BMI ≥40) was suggestive of increased risk (HR = 1.43, 95% CI: 0.84, 2.43). In Latinas, only the morbidly obese were at high risk of death (HR = 2.26, 95% CI: 1.23, 4.15). No BMI–mortality associations were apparent in African Americans and Asian Americans. High WHR (quartile 4 vs. quartile 1) was associated with breast cancer mortality in Asian Americans (HR = 2.21, 95% CI: 1.21, 4.03; P for trend = 0.01), whereas no associations were found in African Americans, Latinas, or non-Latina whites. For all-cause mortality, even stronger BMI and WHR associations were observed. The impact of obesity and body fat distribution on breast cancer patients' risk of death may vary across racial/ethnic groups.
adiposity; body mass index; breast cancer; mortality; obesity; race/ethnicity; survival; waist-hip ratio
Risks of noncancer causes of death, particularly cardiovascular disease, associated with exposures to high-dose ionizing radiation, are well known. Recent studies have reported excess risk in workers who are occupationally exposed to low doses at a low dose rate, but the risks of moderately fractionated exposures, such as occur during diagnostic radiation procedures, remain unclear. The Canadian Fluoroscopy Cohort Study includes 63,707 tuberculosis patients exposed to multiple fluoroscopic procedures in 1930–1952 and followed-up for death from noncancer causes in 1950–1987. We used a Poisson regression to estimate excess relative risk (ERR) per Gy of cumulative radiation dose to the lung (mean dose = 0.79 Gy; range, 0–11.60). The risk of death from noncancer causes was significantly lower in these subjects compared with the Canadian general population (P < 0.001). We estimated small, nonsignificant increases in the risk of death from noncancer causes with dose. We estimated an ERR/Gy of 0.176 (95% confidence interval: 0.011, 0.393) (n = 5,818 deaths) for ischemic heart disease (IHD) after adjustment for dose fractionation. A significant (P = 0.022) inverse dose fractionation effect in dose trends of IHD was observed, with the highest estimate of ERR/Gy for those with the fewest fluoroscopic procedures per year. Radiation-related risks of IHD decreased significantly with increasing time since first exposure and age at first exposure (both P < 0.05). This is the largest study of patients exposed to moderately fractionated low-to-moderate doses of radiation, and it provides additional evidence of increased radiation-associated risks of death from IHD, in particular, significantly increased radiation risks from doses similar to those from diagnostic radiation procedures. The novel finding of a significant inverse dose-fractionation association in IHD mortality requires further investigation.
cardiovascular disease; dose fractionation; ionizing radiation; ischemic heart disease; noncancer diseases; radiation dose-response relationship
Elevated alanine aminotransferase (ALT) activity, an important marker of liver injury, has been associated inconsistently with higher mortality. We evaluated whether persons with nonelevated ALT levels are the most appropriate comparison group by examining the relationships of low ALT with mortality and body composition in the US National Health and Nutrition Examination Survey (NHANES). In NHANES 1988–1994, the mortality risk of persons in ALT deciles 1, 2, 3, and 10 was compared with that of persons in deciles 4–9 (mortality was relatively flat across these deciles) over an 18-year period (through 2006) among 14,950 viral-hepatitis-negative adults. In NHANES 1999–2006, low ALT was evaluated in association with dual-energy x-ray absorptiometry body composition measures among 15,028 adults. Multivariate-adjusted mortality was higher for decile 1 (hazard ratio (HR) = 1.42, 95% confidence interval (CI): 1.24, 1.63), decile 2 (HR = 1.27, 95% CI: 1.06, 1.53), and decile 3 (HR = 1.25, 95% CI: 1.04, 1.50) and nonsignificantly higher for decile 10 (HR = 1.21, 95% CI: 0.91, 1.61) than for deciles 4–9. Adjusted appendicular lean mass was decreased among the lowest ALT deciles. In the US population, low ALT was associated with higher mortality risk, possibly attributable to decreased appendicular lean mass. For mortality studies of elevated ALT levels, the most appropriate comparison group is persons in the middle range of ALT rather than all persons with nonelevated ALT.
alanine aminotransferase; body composition; mortality; National Health and Nutrition Examination Survey
Novel diagnostic tests hold promise for improving tuberculosis (TB) control, but their epidemiologic impact remains uncertain. Using data from the World Health Organization (2011–2012), we developed a transmission model to evaluate the deployment of 3 hypothetical TB diagnostic tests in Southeast Asia under idealized scenarios of implementation. We defined diagnostics by their sensitivity for smear-negative TB and proportion of patients testing positive who initiate therapy (“point-of-care amenability”), with tests of increasing point-of-care amenability having lower sensitivity. Implemented in the public sector (35% of care-seeking attempts), each novel test reduced TB incidence by 7%–9% (95% uncertainty range: 4%–13%) and mortality by 20%–22% (95% uncertainty range: 14%–27%) after 10 years. If also deployed in the private sector (65% of attempts), these tests reduced incidence by 13%–16%, whereas a perfect test (100% sensitivity and treatment initiation) reduced incidence by 20%. Annually detecting 20% of prevalent TB cases through targeted screening (70% smear-negative sensitivity, 85% treatment initiation) also reduced incidence by 19%. Sensitivity and point-of-care amenability are equally important considerations when developing novel diagnostic tests for TB. Novel diagnostics can substantially reduce TB incidence and mortality in Southeast Asia but are unlikely to transform TB control unless they are deployed actively and in the private sector.
diagnostic techniques and procedures; epidemiologic methods; Southeast Asia; theoretical models; tuberculosis
In a recent issue of the Journal, Kirkeleit et al. (Am J Epidemiol. 2013;177(11):1218–1224) provided empirical evidence for the potential of the healthy worker effect in a large cohort of Norwegian workers across a range of occupations. In this commentary, we provide some historical context, define the healthy worker effect by using causal diagrams, and use simulated data to illustrate how structural nested models can be used to estimate exposure effects while accounting for the healthy worker survivor effect in 4 simple steps. We provide technical details and annotated SAS software (SAS Institute, Inc., Cary, North Carolina) code corresponding to the example analysis in the Web Appendices, available at http://aje.oxfordjournals.org/.
causal inference; healthy worker effect; marginal structural models; occupational epidemiology; structural nested models
In the summer of 1924, an outbreak of scarlet fever occurred in Flint, Michigan. Unable to trace it to the usual causes, particularly fresh milk, the Michigan Department of Health used a novel approach to disentangle the enigma: The 116 cases of scarlet fever were compared with 117 “controls” selected from neighbors of the quarantined cases and from patients at the City Health Center who had been treated for ailments unrelated to scarlet fever. The extraordinary culprit was ice cream, which had a frequent/occasional/none consumption prevalence of 60%, 34%, and 6% among the cases and 24%, 51%, and 25% among the controls, respectively. The 1925 report reads, “Detailed epidemiological investigation, by means of case histories and control histories on well persons, confirmed early suspicions and established the fact that the epidemic was spread by ice cream” (Am J Hyg. 1925;5(5):669–681). This forgotten epidemiologic study is the oldest study using the case-control design to have been resurrected thus far. The case-control study design may have been conceived simultaneously, but independently and for different purposes, in England (Janet Lane-Claypon's 1926 report on the determinants of breast cancer) and the United States.
case-control studies; disease transmission; epidemiologic methods; history of medicine; ice cream; scarlet fever