Easy SAS Calculations for Risk or Prevalence Ratios and Differences

Spiegelman, Donna

doi:10.1093/aje/kwi188

Cited by 1,604 publications

(1,286 citation statements)

References 4 publications

Supporting

Mentioning

1,235

Contrasting

Unclassified

Order By: Relevance

“…We tested the statistical significance of the associations by estimating prevalence ratios (PRs) with 95% confidence intervals. 25,26 Analyses were conducted using SAS statistical software, version 9.1.2. 27…”

Section: Methodsmentioning

confidence: 99%

Are Household Firearms Stored Less Safely in Homes With Adolescents?

Johnson

Miller

Vriniotis

et al. 2006

Arch Pediatr Adolesc Med

View full text Add to dashboard Cite

Objective-To examine whether firearms are more frequently stored loaded, unlocked, or both in households with adolescents only (aged 13-17 years) compared with households with younger children only (aged 0-12 years).Design-Random-digit-dial survey on firearms (n=2770). We computed bivariate associations between the presence of adolescents and firearm storage practices. Statistical significance was assessed using prevalence ratios with 95% confidence intervals. Setting-United States.Participants-Survey respondents with children (aged <18 years) who reported the presence of a household firearm.Main Outcome Measures-Prevalence of firearms in the home stored loaded and/or unlocked.Results-Of the 392 respondents, 22% had a loaded firearm, 32% had an unlocked firearm, and 8% had a firearm stored loaded and unlocked. Compared with households with younger children, households with adolescents only were somewhat more likely to store a firearm unlocked (42% vs 29%; prevalence ratio, 1.48; 95% confidence interval, 1.04-2.02), loaded (26% vs 20%; prevalence ratio, 1.25; 95% confidence interval, 0.82-1.91), or both (10% vs 8%; prevalence ratio, 1.43; 95% confidence interval, 0.64-3.19).Conclusions-Parents of adolescents appear to be more likely to keep household firearms stored unsafely, especially with regard to keeping firearms unlocked. This is of concern because most youth firearm injuries happen to adolescents. Firearm injury prevention programs should directly target parents of adolescents to promote safe firearm storage.Firearms are present in about one third of US households with children and youth. [1][2][3][4] Therefore, safe firearm storage practices-including keeping firearms stored unloaded and locked up-represent important safety behaviors. confer a degree of protection in the risks for suicide and unintentional injury. [5][6][7][8] In keeping with these findings, professional organizations, including the American Academy of Pediatrics 9 and the Society for Adolescent Medicine, 10 recommend that parents keep firearms stored unloaded and locked up.In the United States, 14% to 30% of households with guns and young or adolescent children (aged <18 years) have at least 1 loaded firearm, and about 43% contain an unlocked firearm. 1,3,4,11,12 A recent study 13 estimated that nearly 2 million US children live in homes with loaded, unlocked firearms. The home is the primary place from which young people obtain firearms that are used in unintentional and violent injuries, as well as a primary setting in which pediatric firearm injuries, especially suicides, take place. 7,8,[14][15][16] Keeping children and youth safe represents a strong motivation for parents to not have firearms in the home or to store them locked up and unloaded. [2][3][4]17,18 However, parents' perceptions of how likely their children are to get hurt with a firearm may vary based on characteristics of the child. Qualitative research shows that parents' expectations that their children will not play with or touch guns are based on beliefs abou...

show abstract

Section: Methodsmentioning

confidence: 99%

Are Household Firearms Stored Less Safely in Homes With Adolescents?

Johnson

Miller

Vriniotis

et al. 2006

Arch Pediatr Adolesc Med

View full text Add to dashboard Cite

show abstract

“…Differences between groups were assessed using chi-square test for categorical variables or one-way analysis of variance for continuous variables when appropriate, with alpha less than 0.05 defined a priori. Next, we used generalized estimating equations with a log link and Poisson distribution to estimate the risk of having an indeterminate compared to a readable result (18). Models were adjusted for sex, age, race, concomitant diagnoses, and resource utilization.…”

Section: Significance and Innovationsmentioning

confidence: 99%

“…We also assessed GC use in patients with asthma and/or COPD, because these are 2 comorbidities associated with frequent exposure to GCs. All study analyses were performed using SAS, version 9.3 (18).…”

Section: Significance and Innovationsmentioning

confidence: 99%

Evaluating Indeterminate Interferon‐γ–Release Assay Results in Patients With Chronic Inflammatory Diseases Receiving Immunosuppressive Therapy

Calabrese

Overman

Dusetzina

et al. 2015

Arthritis Care & Research

View full text Add to dashboard Cite

Objective. To analyze the rate of indeterminate interferon-g-release assay (specifically the QuantiFeron-TB Gold InTube [QFT-GIT]) testing for Mycobacterium tuberculosis in patients with chronic inflammatory disease (CID) compared with the general hospital population (GH) and a healthy reference group of hospital employees (HR), and to analyze factors associated with an indeterminate test result. Methods. Adults with a QFT-GIT result within a large regional US health system were included. We compared the likelihood of having an indeterminate test across each patient group. Among patients with CID, we estimated the effect of glucocorticoids, biologic agents, and disease-modifying antirheumatic drug (DMARD) use on the likelihood of having an indeterminate test, controlling for age, sex, comorbidities, and prior health services use. Results. Of the 55,108 patients who met the inclusion criteria, CID made up 5.2% (n 5 2,864), GH 48.2%, and HR 46.6% of the study population. Indeterminate results were present in 5.3% of the CID group, 1.9% of the GH group, and 1.5% of the HR group.

show abstract

“…"), misleadingly suggesting that we used statistical models that did not converge. The whole sentence reads: "Since log-binomial models failed to converge, we used a Poisson model (Greenland 2004;Spiegelman and Hertzmark 2005) with a maximum likelihood estimator." There was no problem of convergence with the Poisson model that we used.…”

mentioning

confidence: 99%

“…There was no problem of convergence with the Poisson model that we used. In the context of cohort or cross-sectional data, the Poisson model yields a more relevant risk estimate than odds-ratios estimated by logistic regression (Greenland 2004;Spiegelman and Hertzmark 2005), making it possible to report results more explicitly.…”

mentioning

confidence: 99%