2024-03-29T06:00:56Zhttp://open-archive.highwire.org/handler
oai:open-archive.highwire.org:amjepid:148/9/8212015-05-11HighWireOUPamjepid:148:9
Early and Late Weight Gain following Smoking Cessation in the Lung Health Study
O'Hara, Peggy
Connett, John E.
Lee, Wondra W.
Nides, Mitchell
Murray, Robert
Wise, Robert
ORIGINAL CONTRIBUTIONS
The authors examine weight gains associated with smoking cessation in the Lung Health Study (1986–1994) over a 5-year follow-up period. A cohort of 5,887 male and female smokers in the United States and Canada, aged 35–60 years, were randomized to either smoking intervention or usual care. Among participants who achieved sustained quitting for 5 years, women gained a mean of 5.2 (standard error, 5.0) kg in year 1 and a mean of 3.4 (standard error, 5.5) kg in years 1–5. Men gained a mean of 4.9 (standard error, 4.9) kg in year 1 and a mean of 2.6 (standard error, 5.8) kg in years 1–5. In regression analyses, smoking-change variables were the most potent predictors of weight change. Participants going from smoking to quit-smoking in a given year had mean weight gains of 2.95 kg/year (3.61%) in men and 3.09 kg/year (4.69%) in women. Over 5 years, 33% of sustained quitters gained ≥10 kg compared with 6% of continuing smokers. Also among sustained quitters, 7.6% of men and 19.1% of women gained ≥20% of baseline weight; 60% of the gain occurred in year 1, although significant weight gains continued through year 5. The average gains and the high proportions of sustained and intermittent quitters who gained excessive weight suggest the need for more effective early interventions that address both smoking cessation and weight control. <it>Am J Epidemiol</it> 1998; 148:821–30.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/821
http://dx.doi.org/10.1093/oxfordjournals.aje.a009706
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/8312015-05-11HighWireOUPamjepid:148:9
Invited Commentary on "Early and Late Weight Gain following Smoking Cessation in the Lung Health Study"
Thun, Michael J.
Colditz, Graham A.
ORIGINAL CONTRIBUTIONS
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/831
http://dx.doi.org/10.1093/oxfordjournals.aje.a009707
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/8332015-05-11HighWireOUPamjepid:148:9
Relation of Cigarette Smoking to Non-Hodgkin's Lymphoma among Middle-aged Men
Freedman, David S.
Tolbert, Paige E.
Coates, Ralph
Brann, Edward A.
Kjeldsberg, Carl R.
ORIGINAL CONTRIBUTIONS
Because of previous inconsistencies in the observed relation of cigarette smoking to non-Hodgkin's lymphoma, this association was investigated in the Selected Cancers Study, a population-based case-control study of 1, 193 non-Hodgkin's lymphoma cases and 1, 903 controls, conducted between 1984 and 1988. Study subjects were men, and the median age of non-Hodgkin's lymphoma cases was 50 years (range, 32–60 years). As compared with the risk among men who had never smoked cigarettes, the risk among ever smokers was not increased (odds ratio (OR) = 1.05, p {small tilde} 0.50), but the risk was significantly elevated among men who reported smoking ≥21/2 packs per day and among men who had smoked for 30–39 years (OR = 1.45 in each group, p < 0.05). The estimated odds ratio among the 350 heavy smokers (>50 pack-years) was 1.41 (95% confidence interval 1.08–1.85) after controlling for educational achievement, various occupational and medical exposures, and other potential confounders. The observed associations, however, tended to vary by age, with the odds ratio among heavy smokers decreasing from 2.8 among 32- to 44-year-olds to 1.1 among men over 55 years of age. These age-related differences, which may account for some of the inconsistencies seen in previous studies of cigarette smoking and non-Hodgkin's lymphoma, should be considered in future investigations. <it>Am J Epidemiol</it> 1998; 148:833–41.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/833
http://dx.doi.org/10.1093/oxfordjournals.aje.a009708
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/8422015-05-11HighWireOUPamjepid:148:9
A Prospective Cohort Study on Vegetable and Fruit Consumption and Stomach Cancer Risk in the Netherlands
Botterweck, Anita A. M.
van den Brandt, Piet A.
Goldbohm, R. Alexandra
ORIGINAL CONTRIBUTIONS
The association between vegetable and fruit consumption and stomach cancer risk was investigated in the Netherlands Cohort Study among 120,852 men and women aged 55–69 years at the start in September 1986. Analyses were based on 282 incident stomach cancer cases after 6.3 years of follow-up. Age- and sex-adjusted rate ratios of stomach cancer in increasing quintiles of combined vegetable and fruit consumption were 1.00, 0.70, 0.65, 0.76, and 0.64 (p trend = 0.04). Multivariate analysis resulted in rate ratios that were somewhat closer to one (p trend = 0.14). Furthermore, inverse associations for total vegetables, pulses, raw leafy vegetables, total fruit, citrus fruit, and apples and pears that were observed in crude analyses became weaker or disappeared in multivariate analyses. Total vegetable, but not fruit, consumption was significantly lower in cases diagnosed in the first follow-up year. In analyses limited to first year cases (resembling a case-control study), rate ratios for increasing tertiles of total vegetable consumption were 1.00, 0.17, and 0.18 (p trend = 0.0001), which may indicate the presence of information bias in case-control studies. This prospective study suggests that vegetable and fruit consumption was not clearly associated with stomach cancer risk in the Dutch population. The findings of this study are comparable with findings of other cohort studies, but they do not support the findings of case-control studies. Am J Epidemiol 1998; 148: 842–53.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/842
http://dx.doi.org/10.1093/oxfordjournals.aje.a009709
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/8542015-05-11HighWireOUPamjepid:148:9
Serum Fatty Acids and the Risk of Fatal Cancer
Simon, Joel A.
Fong, Josephine
Bemert,, John T.
Browner, Warren S.
ORIGINAL CONTRIBUTIONS
To examine the relation between serum fatty acids and cancer, the authors conducted a nested case-control study of 108 middle-aged men who died of cancer and 215 men without cancer, who were enrolled in the Multiple Risk Factor Intervention Trial (MRFIT) between 1973 and 1976. Control subjects were matched to case subjects on age, smoking status, treatment assignment, date of randomization, and clinical center. After confirming the stability of the stored serum samples, the authors measured serum fatty acid levels by gas-liquid chromatography and analyzed their association with cancer. In stepwise logistic regression analyses that controlled for the MRFIT selection criteria variables and for alcohol consumption, no fatty acid was significantly associated with overall risk of cancer death (all p > 0.05). Serum levels of phospholipid dihomo-gammalinolenic acid (20: 3), an essential fatty acid, were inversely associated with the risk of dying from lung cancer; a standard deviation increase was associated with a 32% decrease in risk (p = 0.05). Dietary cholesterol intake was associated with the risk of nonlung, non-digestive tract cancers; a standard deviation increase (331 mg/day) was associated with a 75% increase in risk (p = 0.02). The authors found no evidence to suggest that increased dietary intake or serum levels of polyunsaturated fatty acids were associated with an increased risk of fatal cancer among middle-aged men at high risk for coronary heart disease. The clinical significance of the inverse association between dihomogammalinolenic acid and lung cancer death is uncertain and requires confirmation. <it>Am J Epidemiol</it> 1998; 148:854–8.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/854
http://dx.doi.org/10.1093/oxfordjournals.aje.a009710
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/8592015-05-11HighWireOUPamjepid:148:9
Association of Apolipoprotein E Phenotype with Plasma Lipoproteins in African-American and White Young Adults: The CARDIA Study
Howard, Barbara V.
Gidding, Samuel S.
Liu, Kiang
ORIGINAL CONTRIBUTIONS
Apolipoprotein E phenotype (APOE phenotype) has been demonstrated to be a genetic determinant of cardiovascular disease. This atherogenicity may be a reflection of the association of APOE phenotype and plasma lipoprotein concentrations. The Coronary Artery Risk Development in Young Adults (CARDIA) Study affords the opportunity to assess the frequency of apolipoprotein E alleles in population-based samples of African Americans and whites in the United States and to compare the associations of APOE phenotype with lipoprotein and apoprotein concentrations. Data from 3, 485 African-American and white men and women between the ages of 25 and 37 years who attended the fourth CARDIA Study examination in 1992–1993 were used in this analysis. African-American men and women had significantly higher frequencies of E2 and E4 phenotype and thus higher frequencies of *e2 and *e4 alleles (<it>p</it> > 0.005). Men and women of both races with APOE4 phenotype generally had higher low density lipoprotein cholesterol, apolipoprotein B, and lipopro-tein(a) concentrations and lower high density lipoprotein cholesterol concentration, and individuals with APOE3 phenotype had the lowest triglyceride concentration. Major differences between African Americans and whites were observed in the distribution of APOE phenotypes and *e alleles, but APOE phenotype was associated with similar differences in lipoprotein and apoprotein concentrations in both races. The data suggest that APOE phenotype may be a risk factor for cardiovascular disease in both African Americans and whites because it is associated similarly with an adverse lipoprotein profile. <it>Am J Epidemiol</it> 1998; 148: 859–68.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/859
http://dx.doi.org/10.1093/oxfordjournals.aje.a009711
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/8692015-05-11HighWireOUPamjepid:148:9
Risk Factor Clustering in the Insulin Resistance Syndrome: The Strong Heart Study
Gray, R. Stuart
Fabsitz, Richard R.
Cowan, Linda D.
Lee, Elisa T.
Howard, Barbara V.
Savage, Peter J.
ORIGINAL CONTRIBUTIONS
The objective of this study was to examine how the major components of the insulin resistance syndrome relate to each other and to macrovascular disease in American Indians in the Strong Heart Study. The study cohort (4, 228 resident tribal members 45–74 years old) underwent a personal interview and a physical examination between July 1989 and January 1992 at three centers: Arizona, Oklahoma, and North and South Dakota; blood samples were drawn and a 75-g oral glucose tolerance test was performed. Factor analysis was used to assess the clustering and interdependence of groups of insulin resistance syndrome variables. Within both diabetic and nondiabetic groups, three factors emerged. In nondiabetic participants, a cluster of glucose, body mass index, and insulin accounted for 35% (male) and 32% (female) of the total variance in all variables considered, and a cluster of systolic blood pressure and diastolic blood pressure accounted for 25% and 22% in men and women, respectively. Both clusters were positively associated with coronary heart disease but not peripheral vascular disease. In diabetic participants, the combination of systolic and diastolic blood pressures was the most important factor, but the cluster was not associated with coronary heart disease or peripheral vascular disease. A component containing high density lipoprotein cholesterol, triglycerides, and glucose had a positive association with coronary heart disease in diabetic women and with peripheral vascular disease in both sexes. The association of clusters of risk factors and their relations with coronary heart disease provide important clues that may be used in understanding the metabolic disorders associated with insulin resistance and diabetes. <it>Am J Epidemiol</it> 1998; 148: 869–78.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/869
http://dx.doi.org/10.1093/oxfordjournals.aje.a009712
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/8792015-05-11HighWireOUPamjepid:148:9
Prevalence of Hearing Loss in Older Adults in Beaver Dam, Wisconsin: The Epidemiology of Hearing Loss Study
Cruickshanks, Karen J.
Wiley, Terry L.
Tweed, Theodore S.
Klein, Barbara E.K.
Klein, Ronald
Mares-Perlman, Julie A.
Nondahl, David M.
ORIGINAL CONTRIBUTIONS
There are no recent population-based data on the prevalence of hearing loss in older adults using standard audiometric testing. The population-based Epidemiology of Hearing Loss Study was designed to measure the prevalence of hearing loss in adults aged 48–92 years, residing in Beaver Dam, Wisconsin. Hearing thresholds were measured with standardized protocols using pure-tone air-and bone-conduction audiometry in sound-treated booths. The examination also included an otoscopic evaluation, screening tympanogram, and a questionnaire on hearing-related medical history, noise exposure, other potential risk factors, and self-perceived hearing handicap. Of the 4,541 eligible people, 3,753 (82.6%) participated in the hearing study (1993–1995). The average age of participants was 65.8 years, and 57.7% were women. The prevalence of hearing loss was 45.9%. The odds of hearing loss increased with age (odds ratio (OR) = 1.88 for 5 years, 95% confidence Interval (CI) 1.80–1.97) and were greater for men than women (OR = 4.42, 95% CI 3.73–5.24). The male excess of hearing loss remained statistically significant after adjusting for age, education, noise exposure, and occupation (OR = 3.65). These results demonstrate that hearing loss is a very common problem affecting older adults. Epidemiologic studies are needed to understand the genetic, environmental, and sex-related determinants of age-related hearing loss and to identify potential intervention strategies. <it>Am J Epidemiol 1998</it>: 148: 879–86.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/879
http://dx.doi.org/10.1093/oxfordjournals.aje.a009713
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/8872015-05-11HighWireOUPamjepid:148:9
Cognitive Impairment, Drug Use, and the Risk of Hip Fracture in Persons over 75 Years Old: A Community-based Prospective Study
Guo, Zhenchao
Wills, Phillippa
Viitanen, Matti
Fastbom, Johan
Winblad, Bengt
ORIGINAL CONTRIBUTIONS
The authors examined the effects of cognitive function, as assessed by the Mini-Mental State Examination, and drug use on the incidence of hip fracture in a community-based Swedish population of 1,608 subjects who were aged ≥75 years on October 1,1987, and who had not had a hip fracture. During the 7,123.8 person-year follow-up, 134 first hip fractures were identified. The Cox proportional hazards model was used to estimate the relative risk of developing hip fracture, taking into account several potential confounders. Compared with those without cognitive impairment, subjects with mild impairment (Mini-Mental State Examination scores 18–23) had a relative risk of 2.04 (95% confidence interval (Cl) 1.29–3.24), and subjects with moderate-severe impairment (Mini-Mental State Examination score <18) had a relative risk of 2.09 (95% Cl 1.17–3.72). Subjects using opioid analgesics (97% took propoxyphene) had a relative risk of 2.01 (95% Cl 1.19–3.40). Taking potassium supplements (99% took potassium chloride) was related to a reduced risk of hip fracture (relative risk = 0.55, 95% Cl 0.31–0.98), while diuretics did not have an independent impact. In summary, the results show that cognitive impairment and use of propoxyphene are associated with increased risk of hip fracture. The observed protection of potassium chloride merits further attention. The limitation of the study was that the assessment of drug use was made only at baseline. <it>Am J Epidemiol</it> 1998; 148:887–92.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/887
http://dx.doi.org/10.1093/oxfordjournals.aje.a009714
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/8932015-05-11HighWireOUPamjepid:148:9
Distinguishing the Effects of Maternal and Offspring Genes through Studies of "Case-Parent Triads"
Wilcox, Allen J.
Weinberg, Clarice R.
Lie, Rolv Terje
ORIGINAL CONTRIBUTIONS
A gene variant that increases disease risk will be overrepresented among diseased persons, even compared with their own biologic parents. This insight has led to tests based solely on the asymmetric distribution of a variant allele among cases and their parents (e.g., the transmission/disequilibrium test). Existing methods focus on effects of alleles that operate through the offspring genotype. Alleles can also operate through the mother's genotype, particularly for conditions such as birth defects that have their origins in fetal life. An allele working through the mother would have higher frequency in case-mothers than in case-fathers. The authors develop a log-linear method for estimating relative risks for alleles in the context of case-parent triads. This method is able to detect the effects of genes working through the offspring, the mother, or both. The authors assume Mendelian inheritance, but Hardy-Weinberg equilibrium is unnecessary. Their approach uses standard software, and simulations demonstrate satisfactory power and confidence interval coverage. This method is valid with a self-selected or hospital-based series of cases and helps to protect against misleading inference that can result when cases and controls are randomly sampled from a population not in Hardy-Weinberg equilibrium. <it>Am J Epidemiol</it> 1998; 148:893–901.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/893
http://dx.doi.org/10.1093/oxfordjournals.aje.a009715
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/9022015-05-11HighWireOUPamjepid:148:9
A New Method for Estimating the Risk Ration in Studies Using Case-Partental Control Design
Sun, Fengzhu
Flanders, W. Dana
Yang, Quanhe
Khoury, Muin J.
ORIGINAL CONTRIBUTIONS
The authors describe a new simple noniterative, yet efficient method to estimate the risk ratio in studies using case-parental control design. The new method is compared with two other noniterative methods, Khoury's method and Flanders and Khoury's method, and with a maximum likelihood-based method of Schaid and Sommer. The authors found that the variance of the new estimation method is usually smaller than that of Khoury's method or Flanders and Khoury's method and that it is slightly larger than that of the maximum likelihood-based method of Schaid and Sommer. Despite the slightly large variance of the new estimator compared with that of the maximum likelihood-based method, the simplicity of the new estimator and its variance makes the new method appealing. When genotypic information for only one parent is available, the authors also describe a method to estimate the risk ratio without assuming Hardy-Weinberg equilibrium or random mating. A simple formula for the variance of the estimator is given. <it>Am J Epidemiol</it> 1998; 148:902–9.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/902
http://dx.doi.org/10.1093/oxfordjournals.aje.a009716
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/9102015-05-11HighWireOUPamjepid:148:9
An Industry-wide Pulmonary Study of Men and Women Manufacturing Refractory Ceramic Fibers
Lemasters, Grace Kawas
Lockey, James E.
Levin, Linda S.
McKay, Roy T.
Rice, Carol H.
Horvath, Edward P.
Papes, Diane M.
Lu, John W.
Feldman, Daniel J.
ORIGINAL CONTRIBUTIONS
An industry-wide pulmonary morbidity study was undertaken to evaluate the respiratory health of employees manufacturing refractory ceramic fibers at five US sites between 1987 and 1989. Refractory ceramic fibers are man-made vitreous fibers used for high temperature insulation. Of the 753 eligible current employees, 742 provided occupational histories and also completed the American Thoracic Society respiratory symptom questionnaire; 736 also performed pulmonary function tests. Exposure to refractory ceramic fibers was characterized by classifying workers as production or nonproduction employees and calculating the duration of time spent in production employment. The risk of working in the production of refractory ceramic fibers and having one or more respiratory symptoms was estimated by adjusted odds ratios and found to be 2.9 (95 percent confidence interval 1.4–6.2) for men and 2.4 (95 percent confidence interval 1.1–5.3) for women. The effect of exposure to refractory ceramic fibers on forced vital capacity (FVC), forced expiratory volume in 1 second (FEV<inf>1</inf>), the ratio of the two (FEV<inf>1</inf>/FVC), and forced expiratory flow (liters/second) between 25 percent and 75 percent of the FVC curve (FEF<inf>25-75</inf>) was evaluated by multiple regression analysis using transformed values adjusted for height, by dividing by the square of each individual's height. For men, there was a significant decline in FVC for current and past smokers of 165.4 ml (p < 0.01) and 155.5 ml (p = 0.04), respectively, per 10 years of work in the production of refractory ceramic fibers. For FEV<inf>17</inf> the decline was significant (p < 0.01) only for current smokers at 134.9 ml. For women, the decline was greater and significant for FVC among nonsmokers, who showed a decrease of 350.3 ml (p = 0.05) per 10 years of employment in the production of refractory ceramic fibers. These findings indicate that there may be important sex differences in response to occupational and/or environmental exposure. <it>Am J Epidemiol</it> 1998;148:910–19. occupational exposure; occupations; respiratory function tests; sex; spirometry
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/910
http://dx.doi.org/10.1093/oxfordjournals.aje.a009717
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/9202015-05-11HighWireOUPamjepid:148:9
Reliability of Reported Occupational History Information for US Coal Miners, 1969-1977
Brower, Patricia S.
Attfield, Michael D.
ORIGINAL CONTRIBUTIONS
For estimating reliable exposure-response relations it is necessary that random variation in both the response and the exposure variables be sufficiently small. Variability in cumulative exposures can arise from uncertainties in self-reported work histories from interviews. In most epidemiologic surveys, the information gathered from questionnaires is used without knowing the validity or reproducibility of these data. This paper investigates the reliability of occupational histories reported by the same individuals on two occasions separated by 9 years in the US National Study of Coal Workers' Pneumoconiosis and its implications on the exposure-response relation for simple coal workers' pneumoconiosis. For 480 coal miners, from whom occupational histories were obtained twice (in 1969–1971 and 1977–1981), the reliability (intraclass correlation coefficient) of the cumulative exposures generated from each work history was 87%. Logistic model fitting of simple coal workers' pneumoconiosis prevalence to the cumulative coal dust exposure produced almost identical results. After accounting for intersurvey variability in the occupational histories, the authors found that the exposure-response coefficients estimated from information reported at the surveys were attenuated by 12%. In epidemiologic studies, knowledge of the reproducibility of self-reported occupational history information is important to ascertain whether the true exposure effect is underestimated. <it>Am J Epidemiol</it> 1998; 148: 920–6.
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/920
http://dx.doi.org/10.1093/oxfordjournals.aje.a009718
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/927-a2015-05-11HighWireOUPamjepid:148:9
THREE AUTHORS REPLY
Bender, Ralf
Spraul, Maximillan
Trautner, Christoph
LETTERS TO THE EDITOR
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/927-a
http://dx.doi.org/10.1093/oxfordjournals.aje.a009720
en
Copyright (C) 1998, Oxford University Press
oai:open-archive.highwire.org:amjepid:148/9/9272015-05-11HighWireOUPamjepid:148:9
RE: "ASSESSMENT OF EXCESS MORTALITY IN OBESITY"
Marshall, Julie A.
LETTERS TO THE EDITOR
Oxford University Press
1998-11-01 00:00:00.0
TEXT
text/html
http://aje.oxfordjournals.org/cgi/content/short/148/9/927
http://dx.doi.org/10.1093/oxfordjournals.aje.a009719
en
Copyright (C) 1998, Oxford University Press