Dataset Information

Standardized binomial models for risk or prevalence ratios and differences.

ABSTRACT: BACKGROUND:Epidemiologists often analyse binary outcomes in cohort and cross-sectional studies using multivariable logistic regression models, yielding estimates of adjusted odds ratios. It is widely known that the odds ratio closely approximates the risk or prevalence ratio when the outcome is rare, and it does not do so when the outcome is common. Consequently, investigators may decide to directly estimate the risk or prevalence ratio using a log binomial regression model. METHODS:We describe the use of a marginal structural binomial regression model to estimate standardized risk or prevalence ratios and differences. We illustrate the proposed approach using data from a cohort study of coronary heart disease status in Evans County, Georgia, USA. RESULTS:The approach reduces problems with model convergence typical of log binomial regression by shifting all explanatory variables except the exposures of primary interest from the linear predictor of the outcome regression model to a model for the standardization weights. The approach also facilitates evaluation of departures from additivity in the joint effects of two exposures. CONCLUSIONS:Epidemiologists should consider reporting standardized risk or prevalence ratios and differences in cohort and cross-sectional studies. These are readily-obtained using the SAS, Stata and R statistical software packages. The proposed approach estimates the exposure effect in the total population.

SUBMITTER: Richardson DB

PROVIDER: S-EPMC6372130 | biostudies-literature | 2015 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Standardized binomial models for risk or prevalence ratios and differences.

Richardson David B DB Kinlaw Alan C AC MacLehose Richard F RF Cole Stephen R SR

International journal of epidemiology 20150730 5

<h4>Background</h4>Epidemiologists often analyse binary outcomes in cohort and cross-sectional studies using multivariable logistic regression models, yielding estimates of adjusted odds ratios. It is widely known that the odds ratio closely approximates the risk or prevalence ratio when the outcome is rare, and it does not do so when the outcome is common. Consequently, investigators may decide to directly estimate the risk or prevalence ratio using a log binomial regression model.<h4>Methods</ ...[more]

PMID: 26228585

Similar Datasets

Project description:BackgroundThe odds ratio (OR) is used as an important metric of comparison of two or more groups in many biomedical applications when the data measure the presence or absence of an event or represent the frequency of its occurrence. In the latter case, researchers often dichotomize the count data into binary form and apply the well-known logistic regression technique to estimate the OR. In the process of dichotomizing the data, however, information is lost about the underlying counts which can reduce the precision of inferences on the OR.MethodsWe propose analyzing the count data directly using regression models with the log odds link function. With this approach, the parameter estimates in the model have the exact same interpretation as in a logistic regression of the dichotomized data, yielding comparable estimates of the OR. We prove analytically, using the Fisher information matrix, that our approach produces more precise estimates of the OR than logistic regression of the dichotomized data. We also show the gains in precision using simulation studies and real-world datasets. We focus on three related distributions for count data: geometric, Poisson, and negative binomial.ResultsIn simulation studies, confidence intervals for the OR were 56-65% as wide (geometric model), 75-79% as wide (Poisson model), and 61-69% as wide (negative binomial model) as the corresponding interval from a logistic regression produced by dichotomizing the data. When we analyzed existing datasets using our approach, we found that confidence intervals for the OR could be up to 64% shorter (36% as wide) compared to if the data had been dichotomized and analyzed using logistic regression.ConclusionsMore precise estimates of the OR can be obtained directly from the count data by using the log odds link function. This analytic approach is easy to implement in software packages that are capable of fitting generalized linear models or of maximizing user-defined likelihood functions.

Project description:BackgroundStandardized Mortality Ratios (SMRs) are case-mix adjusted mortality rates per hospital and are used to evaluate quality of care. However, acute care is increasingly organized on a regional level, with more severe patients admitted to specialized hospitals. We hypothesize that the current case-mix adjustment insufficiently captures differences in case-mix between non-specialized and specialized hospitals. We aim to improve the SMR by adding proxies of disease severity to the model and by calculating a regional SMR (RSMR) for acute cerebrovascular disease (CVD) and myocardial infarction (MI).MethodsWe used data from the Dutch National Basic Registration of Hospital Care. We selected all admissions from 2016 to 2018. SMRs and RSMRs were calculated by dividing the observed in-hospital mortality by the expected in-hospital mortality. The expected in-hospital mortality was calculated using logistic regression with adjustment for age, sex, socioeconomic status, severity of main diagnosis, urgency of admission, Charlson comorbidity index, place of residence before admission, month/year of admission, and in-hospital mortality as outcome.ResultsThe IQR of hospital SMRs of CVD was 0.85-1.10, median 0.94, with higher SMRs for specialized hospitals (median 1.12, IQR 1.00-1.28, 71%-SMR > 1) than for non-specialized hospitals (median 0.92, IQR 0.82-1.07, 32%-SMR > 1). The IQR of RSMRs was 0.92-1.09, median 1.00. The IQR of hospital SMRs of MI was 0.76-1.14, median 0.98, with higher SMRs for specialized hospitals (median 1.00, IQR 0.89-1.25, 50%-SMR > 1 versus median 0.94, IQR 0.74-1.11, 44%-SMR > 1). The IQR of RSMRs was 0.90-1.08, median 1.00. Adjustment for proxies of disease severity mostly led to lower SMRs of specialized hospitals.ConclusionSMRs of acute regionally organized diseases do not only measure differences in quality of care between hospitals, but merely measure differences in case-mix between hospitals. Although the addition of proxies of disease severity improves the model to calculate SMRs, real disease severity scores would be preferred. However, such scores are not available in administrative data. As a consequence, the usefulness of the current SMR as quality indicator is very limited. RSMRs are potentially more useful, since they fit regional organization and might be a more valid representation of quality of care.

Dataset Information

Standardized binomial models for risk or prevalence ratios and differences.

Publications

Standardized binomial models for risk or prevalence ratios and differences.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets