David Izrael

Learn More
The SAS® macro RAKING for balancing a weighted sample, presented at SUGI 25 by Izrael et al. [1], has attracted many users by its simplicity and versatility. Raking is a technique that almost all survey researchers use, but they may encounter slow convergence when raking by multiple variables and multiple categories. Therefore, we have enhanced the macro by(More)
Pregnant women and infants are at increased risk for influenza-related complications and hospitalization. Influenza vaccination among pregnant women can reduce their risk for respiratory illness and reduce the risk for influenza in their infants aged <6 months. Since 2004, the Advisory Committee on Immunization Practices and the American College of(More)
UNLABELLED Estimates of geographic variation among states and counties in the prevalence of opioid prescribing are developed using data from a large (135 million) representative national sample of opioid prescriptions dispensed during 2008 by 37,000 retail pharmacies. Statistical analyses are used to estimate the extent to which county variation is(More)
Routine influenza vaccination of health-care personnel (HCP) every influenza season can reduce influenza-related illness and its potentially serious consequences among HCP and their patients. To protect HCP and their patients, the Advisory Committee on Immunization Practices (ACIP) recommends that all HCP be vaccinated against influenza during each(More)
Raking is a widely used technique for developing survey weights. It assigns a weight value to each sampling unit such that the weighted distribution of the sample is in very close agreement with two or more marginal control variables. For example, in household surveys the control variables are typically sample design and socio-demographic variables. Raking(More)
survey opportunities in exchange for nominal cash and rewards.* Among the 2,518 HCP who completed the screening questions and entered the two panel survey sites, 2,348 (93.2%) completed the survey. † Of those, 1,724 (73.4%) were clinical professionals, and 624 (26.6%) were other HCP. Survey categories included demographics, occupation type, work setting,(More)
Item non-response is a challenge faced by virtually all surveys. Item non-response occurs when a respondent skips over a question, refuses to answer a question, or indicates that they do not know the answer to a question. Hot deck imputation is one of the primary item non-response imputation tools used by survey statisticians. Recently, new competitor in(More)
In analyzing data from a survey, researchers often need to compare the effectiveness of several logistic regression models. The receiver operating characteristic curve offers one way to measure effectiveness of prediction, by calculating the area under the curve (AUC). We present a SAS  macro for calculating AUC that takes the survey weights into account.(More)