Testing logistic regression coefficients with clustered data and few positive outcomes

Analysis of Familial Aggregation Using Recurrence Risk for Complex Survey Data

Familial or family aggregation of a disease is important for studying possible genetic etiology of a disease. A popular and useful measure of family aggregation is recurrence risk. Household health surveys with (family) network sampling, which surveyed individuals report about disease status of themselves and specified relatives, has been shown to be useful for estimating prevalence of diseases and more recently for estimating recurrence risk of disease using nonparametric classical survey methods. Because these surveys have complex sample designs with sample weighting for differential sample selection rates, this paper extends the composite-likelihood estimation and hypothesis of parameters of the quadratic exponential model (QEM) for simple random samples to data from these complex sample designs. In addition, the QEM is extended to simultaneously estimate and test parameters and recurrence risk for multiple family relationships, for comparing recurrence risk across family-level covariates (e.g., race) and utilizing propensity score weighting to adjust for confounding by individual-level covariates (e.g., age). Simulations are used to study the finite sample properties of the parameter estimation, variance estimation and level and power of hypothesis testing based on derived Wald and Quasi-Score tests for these extended QEMs. Finally, our methods are illustrated using the 1976 National Health Interview Survey diabetes data set.

Prevalence of oral human papillomavirus infection: Impact of sex, race/ethnicity and vaccination status

BACKGROUND

Prior studies have demonstrated differences by sex and race/ethnicity in oral HPV prevalence. This study examined the impact of vaccination on these disparities.

METHODS

We examined participants 18-59 years old in the National Health and Nutrition Examination Survey from 2011-16 who reported their HPV vaccination status and submitted an adequate oral sample (N=9,437). Oral prevalence of HPV, grouped by any, low risk, high risk, 4vHPV, 9vHPV, and non-vaccine types, was examined by sex, race/ethnicity and vaccination status. Binary logistic regression was used to estimate prevalence ratios by vaccination status. Multivariable logistic regression models controlled for age, sex, and race/ethnicity.

RESULTS

The prevalence of any type, non-vaccine type, low-risk, high-risk, 4vHPV, and 9vHPV types was higher among males than females, even among vaccinated participants. Examination of racial/ethnic differences demonstrated differences in all HPV groups among unvaccinated males and among low-risk types in females. In all but the 2 vaccine-type groups, the prevalence of oral HPV was notably higher among black males compared to other groups. Significant differences were not observed by race/ethnicity among vaccinated males or females.

CONCLUSION

Males tested positive for oral HPV more frequently than females, even among those vaccinated. This may have resulted from a lower frequency of males being vaccinated before initiating oral sex than females. Vaccination of males at the recommended age, therefore, may decrease differences in oral HPV by sex. Racial/ethnic differences were observed only in unvaccinated individuals, suggesting these disparities will decrease as more individuals are vaccinated.

Effect of Prophylactic Human Papillomavirus (HPV) Vaccination on Oral HPV Infections Among Young Adults in the United States

Purpose The incidence of human papilloma virus (HPV)-positive oropharyngeal cancers has risen rapidly in recent decades among men in the United States. We investigated the US population-level effect of prophylactic HPV vaccination on the burden of oral HPV infection, the principal cause of HPV-positive oropharyngeal cancers. Methods We conducted a cross-sectional study of men and women 18 to 33 years of age (N = 2,627) within the National Health and Nutrition Examination Survey 2011 to 2014, a representative sample of the US population. Oral HPV infection with vaccine types 16, 18, 6, or 11 was compared by HPV vaccination status, as measured by self-reported receipt of at least one dose of the HPV vaccine. Analyses accounted for the complex sampling design and were adjusted for age, sex, and race. Statistical significance was assessed using a quasi-score test. Results Between 2011 and 2014, 18.3% of the US population 18 to 33 years of age reported receipt of at least one dose of the HPV vaccine before the age of 26 years (29.2% in women and 6.9% in men; P < .001). The prevalence of oral HPV16/18/6/11 infections was significantly reduced in vaccinated versus unvaccinated individuals (0.11% v 1.61%; P_adj = .008), corresponding to an estimated 88.2% (95% CI, 5.7% to 98.5%) reduction in prevalence after model adjustment for age, sex, and race. Notably, the prevalence of oral HPV16/18/6/11 infections was significantly reduced in vaccinated versus unvaccinated men (0.0% v 2.13%; P_adj = .007). Accounting for vaccine uptake, the population-level effect of HPV vaccination on the burden of oral HPV16/18/6/11 infections was 17.0% overall, 25.0% in women, and 6.9% in men. Conclusion HPV vaccination was associated with reduction in vaccine-type oral HPV prevalence among young US adults. However, because of low vaccine uptake, the population-level effect was modest overall and particularly low in men.

A statistical method for studying correlated rare events and their risk factors

Longitudinal studies of rare events such as cervical high-grade lesions or colorectal polyps that can recur often involve correlated binary data. Risk factor for these events cannot be reliably examined using conventional statistical methods. For example, logistic regression models that incorporate generalized estimating equations often fail to converge or provide inaccurate results when analyzing data of this type. Although exact methods have been reported, they are complex and computationally difficult. The current paper proposes a mathematically straightforward and easy-to-use two-step approach involving (i) an additive model to measure associations between a rare or uncommon correlated binary event and potential risk factors and (ii) a permutation test to estimate the statistical significance of these associations. Simulation studies showed that the proposed method reliably tests and accurately estimates the associations of exposure with correlated binary rare events. This method was then applied to a longitudinal study of human leukocyte antigen (HLA) genotype and risk of cervical high grade squamous intraepithelial lesions (HSIL) among HIV-infected and HIV-uninfected women. Results showed statistically significant associations of two HLA alleles among HIV-negative but not HIV-positive women, suggesting that immune status may modify the HLA and cervical HSIL association. Overall, the proposed method avoids model nonconvergence problems and provides a computationally simple, accurate, and powerful approach for the analysis of risk factor associations with rare/uncommon correlated binary events.