Martin R. Petersen

Approaches for estimating prevalence ratios.

Recently there has been much interest in estimating the prevalence (risk, proportion or probability) ratio instead of the odds ratio, especially in occupational health studies involving common outcomes (for example, with prevalence rates above 10%). For example, if 80 out of 100 exposed subjects have a particular disease and 50 out of 100 non-exposed subjects have the disease, then the odds ratio (OR) is (80/20)/(50/50) = 4. However, the prevalence ratio (PR) is (80/100)/(50/100) = 1.6. The latter indicates that the exposed subjects are only 1.6 times as likely to have the disease as the non-exposed subjects, and this is the number in which most people would be interested. There is considerable literature on the advantages and disadvantages of OR versus PR (see Greenland,1 Stromberg,2 Axelson et al 3 and others). In this article we will review the existing methods and give examples and recommendations on how to estimate the PR. The most common method of modelling binomial (no/yes or 0/1) health outcomes today is logistic regression. In logistic regression one models the probability of the binomial outcome (Y = 1) of interest as: P(Y = 1| X1, X2, …, Xk) = eXβ/(1+eXβ) where Xβ = β+β1X1+β2X2+…+βkXk. Then exp(β1) = OR for a 1 unit increase in X1 adjusted for all other variables in the model. Logistic regression yields maximum likelihood estimates (MLEs) of the OR (adjusted for other covariates). If the adjusted OR is the parameter of interest, then these MLEs are generally considered the best estimators available. The adjusted OR can also be used to estimate the adjusted PR, but this should only be done for a rare disease (eg, one with a prevalence of 10% or less). This, together with the fact that …

A comparison of two methods for estimating prevalence ratios

BackgroundIt is usually preferable to model and estimate prevalence ratios instead of odds ratios in cross-sectional studies when diseases or injuries are not rare. Problems with existing methods of modeling prevalence ratios include lack of convergence, overestimated standard errors, and extrapolation of simple univariate formulas to multivariable models. We compare two of the newer methods using simulated data and real data from SAS online examples.MethodsThe Robust Poisson method, which uses the Poisson distribution and a sandwich variance estimator, is compared to the log-binomial method, which uses the binomial distribution to obtain maximum likelihood estimates, using computer simulations and real data.ResultsFor very high prevalences and moderate sample size, the Robust Poisson method yields less biased estimates of the prevalence ratios than the log-binomial method. However, for moderate prevalences and moderate sample size, the log-binomial method yields slightly less biased estimates than the Robust Poisson method. In nearly all cases, the log-binomial method yielded slightly higher power and smaller standard errors than the Robust Poisson method.ConclusionAlthough the Robust Poisson often gives reasonable estimates of the prevalence ratio and is very easy to use, the log-binomial method results in less bias in most common situations, and because it fits the correct model and obtains maximum likelihood estimates, it generally results in slightly higher power, smaller standard errors, and, unlike the Robust Poisson, it always yields estimated prevalences between zero and one.

Worksite Health Promotion Programs in the U.S.: Factors Associated with Availability and Participation

Purpose. To examine how the availability of and participation in worksite health promotion programs varies as a function of individual (e.g., age), organizational (e.g., occupation), and health (e.g., high blood pressure) characteristics. Availability of worksite programs was also compared to that reported in two previous national surveys of private companies. Design. Data analyzed were from the 1994 National Health Interview Survey (NHIS), a national cross-sectional probability sample of the U.S. civilian population. Subjects. Five thousand two hundred nineteen NHIS respondents met the inclusion criteria of (1) being currently employed in a company of at least 50 employees, and (2) completing the NHIS section on worksite health promotion. Measures. Employees indicated the availability of, and their participation in, 33 different types of worksite programs. National Health Interview Survey data were also available regarding general health, blood pressure, body mass index, and medical conditions. Results. Smoking cessation programs had the highest mean availability (43 %), followed by health education programs (31 %) and screening tests (31 %). Overall, availability of worksite programs appeared comparable to that reported in a recent national survey. Participation ranged from 32% for health education programs to 5% for smoking cessation programs. Compared to availability, participation depended less on individual and organizational characteristics. Healthy employees were not consistently more likely to participate in worksite health promotion programs than nonhealthy employees. Conclusions. Although availability of worksite health promotion programs remains high, participation by employees in specific types of programs can vary widely. Attempts to increase participation should look beyond individual, health, and organizational variables, to specific features of the work environment that encourage involvement in health promotion activities.

Update of the NIOSH life table analysis system: A person-years analysis program for the windows computing environment†

BACKGROUND Person-years analysis is a fundamental tool of occupational epidemiology. A life table analysis system (LTAS), previously developed by the National Institute for Occupational Safety and Health, was limited by its platform and analysis and reporting capabilities. We describe the updating of LTAS for the Windows operating system (LTAS.NET) with improved properties. SOFTWARE DEVELOPMENT PROCESS A group of epidemiologists, programmers, and statisticians developed software, platform, and computing requirements. Statistical methods include the use of (indirectly) standardized mortality ratios, (directly) standardized rate ratios, confidence intervals, and P values based on the normal approximation and exact Poisson methods, and a trend estimator for linear exposure-response associations. SOFTWARE FEATURES We show examples using LTAS.NET to stratify and analyze multiple fixed and time-dependent variables. Data import, stratification, and reporting options are highly flexible. Users may export stratified data for Poisson regression modeling. CONCLUSIONS LTAS.NET incorporates improvements that will facilitate more complex person-years analysis of occupational cohort data.

THE NATIONAL INSTITUTE FOR OCCUPATIONAL SAFETY AND HEALTH INDOOR ENVIRONMENTAL EVALUATION EXPERIENCE. PART THREE : ASSOCIATIONS BETWEEN ENVIRONMENTAL FACTORS AND SELF-REPORTED HEALTH CONDITIONS

Abstract Associations between environmental factors and work-related health conditions were assessed using regression techniques with environmental and health data for 2435 respondents in 80 office buildings included in the National Institute for Occupational Safety and Health Health Hazard Evaluation program. The health conditions analyzed included two symptom groupings—multiple lower respiratory symptoms and multiple atopic symptoms—and the presence of asthma diagnosed after beginning work in the building. Four categories of environmental variables were included: heating, ventilation, and air conditioning (HVAC) system design; HVAC maintenance; building design; and building maintenance. Female gender and age over 40 years showed increased relative risks (RRs) for each health condition. In regression models adjusted for age and gender, RRs of multiple lower respiratory symptoms were increased for variables in the HVAC design and maintenance categories, with the highest RR for presence of debris inside th...

Mortality patterns among electrical workers employed in the U.S. Construction Industry, 1982–1987†

BACKGROUND Studies of electrical workers in the utility and manufacturing industries have reported excess site-specific cancer. No previous studies of electrical workers in the construction industry have been conducted. METHODS Our study evaluated the mortality patterns of 31,068 U.S. members of the International Brotherhood of Electrical Workers who primarily worked in the construction industry and died 1982-1987. RESULTS Comparison to the U.S. population by using the NIOSH life table showed significantly elevated proportionate mortality for many causes. Excess mortality for leukemia (proportionate mortality ratio (PMR)=115) and brain tumors (PMR=136) is similar to reports of electrical workers with occupational exposure to electric and magnetic fields in the electric utility or manufacturing industry. Excess deaths due to melanoma skin cancer (PMR=123) are consistent with findings of other PCB-exposed workers. A significantly elevated PMR was observed for the diseases caused by asbestos: lung cancer (PMR=117), asbestosis (PMR=247), and malignant mesothelioma (PMR=356) and from fatal injuries, particularly electrocutions (PMR=1180). The findings of statistically significant excess deaths for prostate cancer (PMR=107), musculoskeletal disease (PMR=130), suicide (PMR=113), and disorders of the blood-forming organs (PMR=141) were unexpected. CONCLUSIONS Results suggest that more detailed investigations of occupational risk factors and evaluation of preventive practices are needed to prevent excess mortality in this hazardous occupation. Am. J. Ind. Med. 36:630-637, 1999. Published 1999 Wiley-Liss, Inc.

Prevalence of respiratory symptoms among female flight attendants and teachers

Background: Potential health effects of the indoor environment in office buildings and aircraft have generated considerable concern in recent years. Aims: To analyse the prevalence of self reported respiratory symptoms and illnesses in flight attendants (FAs) and schoolteachers. Methods: Data were collected as part of a study of reproductive health among female FAs. The prevalences of work related eye, nose, and throat symptoms, wheezing, physician diagnosed asthma, chest illness, and cold or flu were calculated and stratified by smoking status in 1824 FAs and 331 schoolteachers. Results: FAs and teachers were significantly more likely to report work related eye (12.4% and 7.4 %, respectively), nose (15.7% and 8.1%), and throat symptoms (7.5% and 5.7%) than were other working women (2.9% eye, 2.7% nose, and 1.3% throat symptoms). FAs were significantly more likely than teachers and referent working women to report chest illness during the prior three years (32.9%, 19.3%, 7.2%, respectively). Both study groups were more likely to report five or more episodes of cold or flu in the past year than were other working women (10.2% of FAs, 8.2% of teachers, 2.3% of referents), and both groups were more likely to report wheezing than other working women (22.8% of FAs, 28.4% of teachers, 16.4% of referents). FAs were significantly less likely than teachers and other working women to report ever having been diagnosed with asthma (8.2%, 13.3%, 11.8%, respectively). Conclusions: Overall, FAs and schoolteachers report a higher prevalence of work related upper respiratory symptoms, chest illness, and cold or flu than the general working population.

Increased frequency of chromosome translocations in airline pilots with long-term flying experience

Background: Chromosome translocations are an established biomarker of cumulative exposure to external ionising radiation. Airline pilots are exposed to cosmic ionising radiation, but few flight crew studies have examined translocations in relation to flight experience. Methods: We determined the frequency of translocations in the peripheral blood lymphocytes of 83 airline pilots and 50 comparison subjects (mean age 47 and 46 years, respectively). Translocations were scored in an average of 1039 cell equivalents (CE) per subject using fluorescence in situ hybridisation (FISH) whole chromosome painting and expressed per 100 CE. Negative binomial regression models were used to assess the relationship between translocation frequency and exposure status and flight years, adjusting for age, diagnostic x ray procedures, and military flying. Results: There was no significant difference in the adjusted mean translocation frequency of pilots and comparison subjects (0.37 (SE 0.04) vs 0.38 (SE 0.06) translocations/100 CE, respectively). However, among pilots, the adjusted translocation frequency was significantly associated with flight years (p = 0.01) with rate ratios of 1.06 (95% CI 1.01 to 1.11) and 1.81 (95% CI 1.16 to 2.82) for a 1- and 10-year incremental increase in flight years, respectively. The adjusted rate ratio for pilots in the highest compared to the lowest quartile of flight years was 2.59 (95% CI 1.26 to 5.33). Conclusions: Our data suggests that pilots with long-term flying experience may be exposed to biologically significant doses of ionising radiation. Epidemiological studies with longer follow-up of larger cohorts of pilots with a wide range of radiation exposure levels are needed to clarify the relationship between cosmic radiation exposure and cancer risk.

Natural rubber latex: glove use, sensitization, and airborne and latent dust concentrations at a Denver hospital.

Exposure to natural rubber latex may cause immediate hypersensitivity reactions. Published latex sensitization prevalence rates range from 2.9% to 22% among health care workers, and from 0.12% to about 20% of occupationally unexposed populations. In this study, self-administered questionnaires addressed job and personal characteristics, glove use, and symptoms in two groups of hospital workers: those who regularly used latex gloves and those who did not. Serum was tested for latex-specific immunoglobulin E. Air, surface, and air-filter dust samples for natural rubber latex were collected. The prevalence of latex sensitization was 6.3% in the non-users and 6.1% in the latex glove users (P = 0.9); 81.3% of sensitized workers were atopic compared with 59.5% of non-sensitized workers (P < 0.05). Reporting of work-related hand dermatitis was more common in the latex glove users (23.4%) than in the non-users (4.9%), as were rhino-conjunctivitis (16.3% and 7.9%, respectively, [P < 0.01]), and hand urticaria (9.9% and 2.1%, respectively, [P < 0.01]). There was no significant difference in work-related symptoms between the sensitized and non-sensitized workers. Environmental concentrations of latex were higher in the work areas of the non-sensitized workers, but higher in the clinical than in the non-clinical areas. Occupational latex glove use was not a risk factor for sensitization.

Cadmium and lung cancer mortality accounting for simultaneous arsenic exposure

Objectives Prior investigations identified an association between airborne cadmium and lung cancer but questions remain regarding confounding by arsenic, a well-established lung carcinogen. Methods A cadmium smelter population exhibiting excess lung cancer was re-analysed using a retrospective exposure assessment for arsenic (As), updated mortality (1940–2002), a revised cadmium (Cd) exposure matrix and improved work history information. Results Cumulative exposure metrics for both cadmium and arsenic were strongly associated making estimation of their independent effects difficult. Standardised mortality ratios (SMRs) were modelled with Poisson regression with the contribution of arsenic to lung cancer risk constrained by exposure–response estimates previously reported. The results demonstrate (1) a statistically significant effect of Cd independent of As (SMR=3.2 for 10 mg-year/m3 Cd, p=0.012), (2) a substantial healthy worker effect for lung cancer (for unexposed workers, SMR=0.69) and (3) a large deficit in lung cancer mortality among Hispanic workers (SMR=0.27, p=0.009), known to have low lung cancer rates. A supralinear dose-rate effect was observed (contribution to risk with increasing exposure intensity has declining positive slope). Lung cancer mortality was somewhat better predicted using a cadmium burden metric with a half-life of about 20–25 years. Conclusions These findings support an independent effect for cadmium in risk of lung cancer mortality. 1/1000 excess lifetime risk of lung cancer death is predicted from an airborne exposure of about 2.4 μg/m3 Cd.