Measurements, estimates, and inferences in reporting epidemiologic study results.
Explore the source record for details and available documents.
Data Interpretation, Statistical↗
Biomedical subjects
D A Savitz
Publications and source records attributed to D A Savitz.
Is electromagnetic fields and cancer an issue worthy of study?
Explore the source record for details and available documents.
Electromagnetic Phenomena↗
Effects of nondifferential exposure misclassification in ecologic studies.
Although many authors have argued against inferring individual-level exposure-disease relations from ecologic data because of the potential "ecological fallacy." the availability of data from diverse populations promotes the continued use of this rapid and inexpensive study design. In ecologic studies, the exposure status of groups is often defined by the proportion of individuals exposed. In these studies, nondifferential exposure misclassification of individuals is shown to produce overestimation of exposure-disease associations that may be extreme when the ecologically derived rate ratios are applied to individuals. This overestimation contrasts with the bias toward the null resulting from nondifferential misclassification of a binary exposure in epidemiologic studies conducted at the individual level. Given the magnitude of the potential bias from nondifferential exposure misclassification and other sources, quantitative estimates of individual-level rate ratios from ecologic data should be interpreted with extreme caution.
Environmental Health↗
Association of paternal alcohol use with gestational age and birth weight.
Paternal alcohol use has been associated with a number of adverse reproductive outcomes in laboratory animals and there is one epidemiologic report of a detrimental effect on infant birth weight. To expand the epidemiologic evidence, data from the Child Health and Development Studies were analyzed. Data collected from the onset of prenatal care in 10,232 women enrolled in the Kaiser Foundation Health Plan and residing in the San Francisco East Bay area between June 1959 and September 1966 were available, including information on the mother's report of paternal alcohol consumption and a number of potential confounders. Pregnancy outcomes included preterm delivery (< 37 weeks completed gestation), moderately low birth weight (1,501-2,500 g), very low birth weight (< or = 1,500 g), small-for-gestational-age (< 10th percentile of weight for gestational age), and mean birth weight. Paternal alcohol use, analyzed in intervals from 0 to 2.0 or more drinks per day, showed no association with any of the outcomes of interest. Adjusted prevalence odds ratios ranged from 0.7 to 1.5, with no indication of a monotonic dose-response gradient. Mean birth weight was also virtually unrelated to paternal alcohol use. Compared with the earlier report, this population had a very modest level of alcohol consumption. Nonetheless, within the range that was studied there appears to be no association between paternal alcohol use and birth outcome.
Adult↗
Use of a job-exposure matrix to evaluate parental occupation and childhood cancer.
We examined the association between parental occupation and childhood cancer among 252 incident cases of childhood cancer (ages 0-14, diagnosed 1976-83) and 222 controls selected by random digit dialing in Denver, Colorado (USA). A job-exposure matrix was used to assign parental exposures based on job titles, emphasizing chemicals that were implicated in previous studies. All cancers, acute lymphocytic leukemia (ALL), and brain cancer were examined in relation to parental occupation during the year prior to the birth of the child. Elevated odds ratios (OR), all with confidence intervals extending below the null, were found for maternal exposure to benzene (OR = 1.9), petroleum/coke pitch/tar (OR = 2.2), and soot (OR = 3.3) in relation to total cancers. The ORs for total cancer and paternal exposure to all hydrocarbons combined was 1.0. Results for individual hydrocarbons and ALL showed larger odds ratios, including aniline (OR = 2.1), benzene (OR = 1.6), and petroleum/coke pitch/tar (OR = 1.6). Potential exposure to creosote was strongly associated with brain cancer (OR = 3.7) based on five exposed cases (95 percent confidence interval = 0.8-16.6). Control for other potential childhood cancer risk factors did not alter the results substantially. In spite of uncertainties due to small numbers and errors in exposure classification, results tend to corroborate past research that suggests an association between specific parental occupational exposures and childhood cancer.
Adolescent↗
A case-control study of paternal smoking and birth defects.
Although the influence of paternal smoking on birth defects is of great public interest, epidemiological evidence concerning this potential relationship is extremely limited. A stratified random sample of 29 hospitals in the Shanghai Municipality, China, was used to select 1012 birth defects cases and controls. Mothers of the cases and controls were interviewed in the hospitals from October 1986 to September 1987. A modest relationship between paternal smoking and overall birth defects in offspring was identified [odds ratio (OR) = 1.21, 95% confidence interval (CI): 1.01-1.45]. More markedly elevated risks were identified for anencephalus (OR = 2.1), spina bifida (OR = 1.9), pigmentary anomalies of the skin (OR = 3.3) and varus/valgus deformities of the feet (OR = 1.8). Our analysis also shows that paternal smoking is more likely to be associated with multiple rather than isolated malformations. A paternally-mediated effect of smoking on birth defects is suggested and further studies are encouraged.
Adult↗
Preterm birth subtypes among blacks and whites.
The differences in preterm birth between blacks and whites are poorly understood. Our study examined subtypes of moderately preterm delivery (34-36 completed weeks of gestation) and very preterm delivery (20-33 weeks) in blacks and whites using North Carolina birth certificate data for 1988-1989. We divided the causes of preterm birth into three categories: preterm premature rupture of the membranes, indication of pregnancy complication, and idiopathic preterm deliveries. The overall prevalence of preterm birth was 8.0% and 16.7% for whites and blacks, respectively. The entire gestational age distribution of blacks was shifted to earlier ages relative to whites. More highly educated blacks still had higher risks of moderately and very preterm deliveries than less educated whites. Multivariate analysis, controlling for other factors, showed that blacks had 3.3, 2.5, and 3.5 times the risk of whites to have preterm premature rupture of the membranes, complication-related, and idiopathic delivery, respectively, among very preterm births, and 1.6, 1.9, and 2.0 times the risk of whites for moderately preterm births of the same three types.
Black People↗
The effects of nondifferential confounder misclassification in ecologic studies.
In ecologic studies, covariate levels of groups are often quantified as the prevalence of a dichotomous covariate. We show that, under certain conditions, nondifferential misclassification of such a binary covariate does not reduce the ability to control confounding by the covariate in ecologic studies. Thus, any remaining exposure-disease association in an adjusted ecologic analysis cannot be ascribed to incomplete control for confounding due to nondifferential misclassification of the dichotomy under those conditions, although residual confounding by the underlying covariate may still be present. This point is illustrated by ecologic analyses of the association between population density and mortality from lung cancer in women in 30 administrative districts of the Federal Republic of Germany, in which control for cigarette smoking is required.
Confounding Factors, Epidemiologic↗
Prenatal events and the risk of subependymal/intraventricular haemorrhage in very low birthweight neonates.
The effects of prenatal factors on the risk of subependymal and/or intraventricular brain haemorrhage in very low birthweight (VLBW) neonates were studied. Data were collected on 201 consecutively born VLBW neonates without major congenital anomalies, who were born at a regional obstetric referral centre. Brain haemorrhage was identified by cranial ultrasound examinations. The reliability of these examinations (concordance among readers' interpretations) was assessed and found to be moderate (kappa = 0.47 for the finding of subependymal haemorrhage (SEH); kappa = 0.50 for the finding of intraventricular haemorrhage (IVH)). Prenatal factors were more strongly associated with IVH than with SEH. In univariable analyses, maternal pre-eclampsia, multiple gestation and maternal treatment with betamethasone were associated most strongly with a decreased risk of haemorrhage whereas labour and vaginal delivery were associated most strongly with an increased risk. These associations remained in a multivariable analysis which included prenatal events (maternal illnesses, fetal presentation and obstetrical interventions), as well as gestational age, birthweight, gender, treatment with assisted ventilation, and the occurrence of pneumothorax. Further aetiological study of the effects of prenatal factors could provide information useful in preventing SEH/IVH.
Cerebral Hemorrhage↗
Pregnancy-induced hypertension in North Carolina, 1988 and 1989.
INTRODUCTION: Pregnancy-induced hypertension (PIH) is a highly prevalent pregnancy complication with adverse effects on maternal and infant health. Epidemiologic research concerning its etiology is limited. METHODS: Birth records from North Carolina for the period 1988 through 1989 included an indication of the presence of PIH. The risk of PIH was examined in relation to several maternal characteristics and exposures, including reproductive history, demographic characteristics, and tobacco use during pregnancy. Risk ratio estimates, adjusted for confounders, were calculated contrasting PIH among exposed vs unexposed women. RESULTS: The overall risk of PIH was 43.1 per 1000 births, with multiple pregnancies, nulliparity, and advanced maternal age associated with markedly increased risks. Tobacco use was inversely associated with PIH, and Blacks and Whites were at virtually equal risk. CONCLUSIONS: Problems in diagnosis and classification impede research in this area, with birth certificates limited in quality and breadth of information. Nonetheless, several patterns emerged that are worthy of further epidemiologic evaluation using more sophisticated designs.
Adult↗
Parental occupation and risk of prematurity.
Explore the source record for details and available documents.
Female↗
Re: "Associations are not effects".
Explore the source record for details and available documents.
Causality↗
Prenatal exposure to parents' smoking and childhood cancer.
The relation between parents' tobacco smoking prior to birth and cancer in the offspring was investigated with the use of data from a case-control study. Incident cases included all children (aged 0-14 years) diagnosed in Denver, Colorado from 1976 to 1983. Controls were selected through random digit dialing, and matched to cases on age, sex, and geographic area. Information on smoking by parents and other household members was obtained by personal interview for 223 cases and 196 controls. After adjustment for father's education, mother's smoking during the first trimester of pregnancy was associated with an increased risk for all cancers combined (odds ratio (OR) = 1.3, 95% confidence interval (CI) 0.7-2.1), acute lymphocytic leukemia (OR = 1.9, 95% CI 0.9-4.1), and lymphomas (OR = 2.3, 95% CI 0.8-7.1). Adjusting for father's education, associations with father's smoking in the absence of mother's smoking were found for all cancers combined (OR = 1.2, 95% CI 0.8-2.1), acute lymphocytic leukemia (OR = 1.4, 95% CI 0.6-3.1), lymphomas (OR = 1.6, 95% 0.5-5.4), and brain cancer (OR = 1.6, 95% CI 0.7-3.5). In spite of imprecision resulting from small numbers of cases in diagnostic subgroups, these results are suggestive of a possible influence of parents' smoking on childhood cancer.
Adolescent↗
Occupation and leukemia mortality among men in 16 states: 1985-1987.
The relationship between leukemia and occupation was investigated in a case-control study using death certificates of 5,147 men who died of leukemia (ICD-9 codes 204-208) and 51,470 who died of other causes in 16 U.S. states from 1985 to 1987. Of six occupational activities identified previously as potentially increasing the risk of leukemia, only petroleum refining and rubber manufacturing had excess deaths for all leukemias combined (odds ratios (ORs) = 1.3, 95% confidence intervals (CI) = 0.6-2.8 and 0.9-1.8, respectively). Meat workers and wood workers had elevated mortality from acute lymphocytic leukemia (OR = 2.2, 95% CI = 0.7-7.0 and OR = 1.3, 95% CI = 0.8-2.2, respectively). There was no leukemia excess among farmers or auto mechanics. A survey of 43 other occupational groups indicated a widespread excess of leukemia among white collar occupations, primarily managers and professionals, but none among blue collar workers. This pattern was most pronounced for men under 65 years of age, and existed for all leukemia subtypes and among both blacks and whites. Despite the lack of specific exposure information and other limitations of death certificate data, these results encourage further examination of occupational causes of leukemia.
Black or African American↗
Influence of paternal age, smoking, and alcohol consumption on congenital anomalies.
The potential effects of paternal exposures on fetal development are of great public and scientific concern, yet few epidemiologic studies have examined this association. Single live births from 1959 to 1966 among 14,685 Kaiser Foundation Health Plan members who participated in the Child Health and Development Studies were analyzed to assess the impact of paternal age, cigarette smoking, and alcohol consumption on the occurrence of birth defects in the offspring. Prevalence odds ratios for anomalies identified by age 5 were analyzed, contrasting exposed to unexposed fathers with adjustment for maternal age, race, education, smoking, and alcohol use. Advanced paternal age was associated with increased risk of preauricular cyst, nasal aplasia, cleft palate, hydrocephalus, pulmonic stenosis, urethral stenosis, and hemangioma. Father's cigarette smoking was more common among children with cleft lip +/- cleft palate, hydrocephalus, ventricular septal defect, and urethral stenosis. Alcohol use by the father was most positively related to the offspring's risk of ventricular septal defect. For both smoking and alcohol use, inverse associations were more common than positive associations. These data generally do not indicate strong or widespread associations between paternal attributes and birth defects. However, because of this study's imprecision, limited ability to isolate defects most likely to be of paternal origin, and the identification of several suggestive associations with age and smoking, further study of this issue would be of value.
Alcohol Drinking↗
Hospital setting and fetal death during labor among women at low risk.
Intrapartum fetal death in low-risk women at term is a rare obstetric outcome. This is difficult to study because few data sets contain an adequate number of cases for meaningful analysis. This study used data from the 1980 National Natality Survey and National Fetal Mortality Survey, merged with an American Hospital Association annual survey for the same year, to determine whether the frequency of intrapartum fetal death in low-risk women varied by the hospital setting for birth. Stratified analysis was used to assess the relation of level of hospital for delivery with intrapartum fetal death, with control for potential confounding factors. As the level of available perinatal technology decreased, the frequency of intrapartum fetal death increased (odds ratio, 2.0 for Level II and 3.3 for Level I, as compared with Level III hospitals). Even when early neonatal deaths were considered, perinatal mortality remained lowest at Level III facilities (odds ratio, 1.6 for Level II and 2.7 for Level I, as compared with Level III hospitals). The components of intrapartum surveillance that are most effective in the reduction of perinatal mortality have not been identified.
Adolescent↗
Epidemiologic characteristics of preterm delivery: etiologic heterogeneity.
Preterm delivery (less than 37 weeks completed gestation) is known to result from diverse etiologic pathways, which can be grouped into idiopathic preterm labor, preterm premature rupture of the membranes, and medical complications. Data from publications providing sufficient detail to subdivide preterm delivery cases into these groups were tabulated. In spite of inconsistent terminology and incomplete reporting, patterns were identified. Black women have a markedly higher risk of preterm delivery, which is especially pronounced for preterm premature rupture of the membranes. Idiopathic preterm labor is predominant in lower-risk, white populations. These observations encourage consideration of subcategories of preterm delivery in studies of etiology and prevention.
Abortion, Induced↗