Approaches to estimate bidirectional causal effects using Mendelian randomization with application to body mass index and fasting glucose

Mendelian randomization (MR) is an epidemiological framework using genetic variants as instrumental variables (IVs) to examine the causal effect of exposures on outcomes. Statistical methods based on unidirectional MR (UMR) are widely used to estimate the causal effects of exposures on outcomes in observational studies. To estimate the bidirectional causal effects between two phenotypes, investigators have naively applied UMR methods separately in each direction. However, bidirectional causal effects between two phenotypes create a feedback loop that biases the estimation when UMR methods are naively applied. To overcome this limitation, we proposed two novel approaches to estimate bidirectional causal effects using MR: BiRatio and BiLIML, which are extensions of the standard ratio, and limited information maximum likelihood (LIML) methods, respectively. We compared the performance of the two proposed methods with the naive application of UMR methods through extensive simulations of several scenarios involving varying numbers of strong and weak IVs. Our simulation results showed that when multiple strong IVs are used, the proposed methods provided accurate bidirectional causal effect estimation in terms of median absolute bias and relative median absolute bias. Furthermore, compared to the BiRatio method, the BiLIML method provided a more accurate estimation of causal effects when weak IVs were used. Therefore, based on our simulations, we concluded that the BiLIML should be used for bidirectional causal effect estimation. We applied the proposed methods to investigate the potential bidirectional relationship between obesity and diabetes using the data from the Multi-Ethnic Study of Atherosclerosis cohort. We used body mass index (BMI) and fasting glucose (FG) as measures of obesity and type 2 diabetes, respectively. Our results from the BiLIML method revealed the bidirectional causal relationship between BMI and FG in across all racial populations. Specifically, in the White/Caucasian population, a 1 kg/m2 increase in BMI increased FG by 0.70 mg/dL (95% confidence interval [CI]: 0.3517-1.0489; p = 8.43×10-5), and 1 mg/dL increase in FG increased BMI by 0.10 kg/m2 (95% CI: 0.0441-0.1640; p = 6.79×10-4). Our study provides novel findings and quantifies the effect sizes of the bidirectional causal relationship between BMI and FG. However, further studies are needed to understand the biological and functional mechanisms underlying the bidirectional pathway.

Awareness and interest in cannabis use for cancer management among cancer survivors

BACKGROUND

We examined the awareness, interest, and information sources relating to cannabis use for cancer management (including management of cancer symptoms and treatment-related side effects) and determined factors associated with cancer survivors' awareness and interest in learning about cannabis use for cancer management.

METHODS

This was a cross-sectional study of adult cancer survivors (N = 1886) receiving treatment at a comprehensive cancer center. Weighted prevalence and multivariable logistic regression analyses were conducted.

RESULTS

Among cancer survivors, 88% were aware and 60% were interested in learning about cannabis use for cancer management. Common sources of information to learn about cannabis use for cancer management were cancer doctors/nurses (82%), other patients with cancer (27%), websites/blogs (26%), marijuana stores (20%), and family/friends (18%). The odds of being aware of cannabis use for cancer management was lower among male compared to female survivors (adjusted odds ratio [AOR]: 0.61; 95% confidence interval [CI]: 0.41-0.90), non-Hispanic Blacks compared to non-Hispanic Whites (AOR: 0.36; 95% CI: 0.21-0.62), and survivors who do not support the legalization of cannabis for medical use compared to those who do (AOR: 0.10; 95% CI: 0.04-0.23). On the other hand, the odds of being interested in cannabis use for cancer management was higher among non-Hispanic Blacks compared to non-Hispanic Whites (AOR: 1.65; 95% CI: 1.04-2.62), and among cancer survivors actively undergoing cancer treatment compared to patients on non-active treatment (AOR: 2.25; 95% CI: 1.74-2.91).

CONCLUSION

Awareness of cannabis use for cancer management is high within the cancer survivor population. Results indicated health care providers are leading information source and should receive continued medical education on cannabis-specific guidelines. Similarly, tailored educational interventions are needed to guide survivors on the benefits and risks of cannabis use for cancer management.

Prevalence, reasons for use, perceived benefits, and awareness of health risks of cannabis use among cancer survivors - implications for policy and interventions

BACKGROUND

There is increasing discourse on the use of cannabis as a palliative for cancer/cancer-treatment-related symptoms. We described the prevalent reasons for use, perceived benefits, and awareness of health risks from cannabis use for cancer management among cancer survivors.

METHODS

Cross-sectional survey of adult (≥ 18 years) cancer survivors from 41 US states receiving treatment at a comprehensive cancer center.

RESULTS

Of 1,886 cancer survivors included, 17.4% were current users, 30.5% were former users, and 52.2% were never users of cannabis. Among survivors who currently or formerly used cannabis after their cancer diagnosis (n = 510), the reasons for cannabis use in cancer management were; sleep disturbance (60%), pain (51%), stress (44%), nausea (34%), and mood disorder/depression (32%). Also, about a fifth (91/510) of survivors used cannabis to treat their cancer. Across the different symptoms assessed, over half of the survivors who reported a reason for using cannabis currently or after their cancer diagnosis perceived that cannabis was helpful to a great extent in improving their symptoms. However, of the 167 survivors who reported awareness of potential health risks from cannabis use, the awareness of adverse health risks associated with cannabis use was low: suicidal thoughts (5%), intense nausea and vomiting (6%), depression (11%), anxiety (14%), breathing problems (31%), and interaction with cancer drugs (35%).

CONCLUSION

Prevalence of cannabis use among survivors was notable, with most reporting a great degree of symptomatic improvement for the specified reason for use. However, only a few were aware of the health risks of cannabis use during cancer management.

IMPLICATIONS FOR CANCER SURVIVORS

With more cancer survivors using cannabis as a palliative in managing their cancer-related symptoms, future guidelines and policies on cannabis use in cancer management should incorporate cannabis-based interventions to minimize the inadvertent harm from cannabis use during cancer treatment among survivors.

Cannabis Use among Cancer Survivors: Use Pattern, Product Type, and Timing of Use

Despite growing interest in the use of cannabis for the treatment of cancer-related symptoms, there are limited studies that have assessed the use pattern, type, and mode of delivery of cannabis products used by cancer survivors. This study describes the current state of the use pattern, product type, and mode of delivery of cannabis used by cancer survivors. This was a cross-sectional study of cancer survivors from 41 U.S. states who received treatment at the largest NCI-designated comprehensive cancer center. The weighted prevalence of the use patterns, product types, and modes of delivery of cannabis used by cancer survivors was estimated. A total of 1886 cancer survivors were included in the study, with 915 (48% [95% CI: 45-51]) reporting ever using cannabis. Of survivors who had ever used cannabis, 36% (95% CI: 33-40) were current users. Among survivors who reported cannabis use after diagnosis, 40% used cannabis during and after cancer treatment, 35% used cannabis during treatment, and 25% used cannabis after completing their cancer treatment. Additionally, 48% of survivors reported an increase in cannabis use since cancer diagnosis. The commonest types of cannabis products used by cancer survivors were dry leaf cannabis (71%), cannabidiol (CBD) oil (46%), and cannabis candy (40%). Moreover, cancer survivors frequently used baked goods (32%), creams and gels (21%), and tinctures (18%). Furthermore, among ever users, the predominant mode of use was cannabis inhalation/smoking (69%) compared to eating/drinking (59%). More so, the common mode of inhalation/smoking of cannabis products were rolled cannabis cigarettes (79%), pipes (36%), water pipes (34%), vaporizers or vapes (14%), and e-cigarette devices (14%). A substantial number of cancer survivors use cannabis during cancer treatment, with increased use following cancer diagnosis. The forms and modes of delivery of cannabis varied among survivors, with most survivors inhaling or smoking cannabis. There is a need to educate healthcare providers (HCPs) and survivors on current evidence of cannabis use and strengthen cannabis regulatory frameworks to optimize benefits and minimize adverse events from cannabis use during cancer treatment.

Barriers to human papillomavirus vaccine uptake: role of state religiosity and healthcare professionals' participation in a state vaccine program

BACKGROUND

Despite the known benefits of preventing human papillomavirus (HPV)-related cancers, HPV vaccine coverage is low in the United States. Grounded in Social Ecological theory, we assessed the macro-level (state) and meso-level (organization) factors associated with HPV vaccine initiation and up-to-date.

METHODS

Data from 2020 National Immunization Survey-Teen were used to study a sample of 20 163 US adolescents (aged 13-17 years). The data were collected from each teen's parents or guardians and health-care professionals. Weighted prevalence estimates were calculated, and multivariable regression analyses were conducted.

RESULTS

The prevalence of HPV vaccine initiation was 75.1% and of remaining up-to-date was 58.6%. At the macro level, teens living in states with high and moderate religiosity had lower odds of HPV vaccine initiation (high religiosity adjusted odds ratio [AOR] = 0.63, 95% confidence interval [CI] = 0.50 to 0.78; moderate religiosity AOR = 0.68, 95% CI = 0.55 to 0.85) and up-to-date (high religiosity AOR = 0.69, 95% CI = 0.56 to 0.85; moderate religiosity AOR = 0.74, 95% CI = 0.61 to 0.91) than states with low religiosity. At the meso level, when none of their healthcare professionals ordered vaccine from the state, teens had lower odds of initiation (AOR = 0.68, 95% CI = 0.53 to 0.87) and up-to-date (AOR = 0.76, 95% CI = 0.60 to 0.95) than teens whose healthcare professionals ordered vaccine from the state. In addition, race and ethnicity, age, mother's education level, household income, well-child examination status, and doctor's recommendation were significantly associated with HPV vaccine uptake.

CONCLUSION

A multiprong approach is needed to address religious and systemic barriers to HPV vaccination and expand healthcare professionals' access and enrollment in state vaccine initiatives, such as the Vaccine for Children program.

Individual, systemic and state factors associated with provider recommendation of HPV vaccination: Findings from NIS-Teen, 2020

The most important determinant of HPV vaccination uptake is healthcare provider recommendation, yet not all eligible patients receive HPV vaccination recommendations. We used data from the 2020 National Immunization Survey-Teen to determine individual, systemic, and state factors associated with provider recommendation of HPV vaccination. A total of 18,534 teens were included, with 81.4% receiving provider recommendations for HPV vaccination. HPV vaccination recommendation was higher among females compared to males (AOR: 1.57; 95% CI: 1.27-1.93), teens who received a well-child exam at 11 or 12 years compared to those who did not (AOR: 2.10; 95% CI: 1.61-2.74), and teens whose mothers had college, some college or at least a high school education compared to those with less than high school education. In addition to individual factors, provider recommendation of HPV vaccination was higher in hospitals (AOR: 1.51; 95% CI: 1.00-2.29) and STD/school/teen clinics (AOR: 2.47; 95% CI: 1.05-5.78) compared to public facilities. However, the odds of provider recommendation were lower when none of the teen's providers ordered vaccines from the state compared to when all teen providers ordered vaccines from the state (AOR: 0.69; 95% CI: 0.52-0.93). Moreover, the state's mean prevalence of provider recommendations of HPV vaccination was 7.2% lower (Coefficient: -0.072; 95% CI: -0.107 - -0.036) in states with high religious ideology compared to those with low religious ideology. Interventions to increase provider recommendation of HPV vaccination should take a multiprong and comprehensive approach that addresses religious and systemic barriers to HPV vaccination recommendation.

Impact of U.S. Preventive Services Task Force lung cancer screening update on drivers of disparities in screening eligibility

BACKGROUND

In 2021, the U.S. Preventive Services Task Force (USPSTF) updated its recommendation to expand lung cancer screening (LCS) eligibility and mitigate disparities. Although this increased the number of non-White individuals who are eligible for LCS, the update's impact on drivers of disparities is less clear. This analysis focuses on racial disparities among Black individuals because members of this group disproportionately share late-stage lung cancer diagnoses, despite typically having a lower intensity smoking history compared to non-Hispanic White individuals.

METHODS

We used data from the National Health Interview Survey to examine the impact of the 2021 eligibility criteria on racial disparities by factors such as education, poverty, employment history, and insurance status. We also examined preventive care use and reasons for delaying medical care.

RESULTS

When comparing Black individuals and non-Hispanic White individuals, our analyses show significant differences in who would be eligible for LCS: Those who do not have a high school diploma (28.7% vs. 17.0%, p = 0.002), are in poverty (26.2% vs. 14.9%, p < 0.001), and have not worked in the past 12 months (66.5% vs. 53.9%, p = 0.009). Further, our analyses also show that more Black individuals delayed medical care due to not having transportation (11.1% vs. 3.6%, p < 0.001) compared to non-Hispanic White individuals.

CONCLUSIONS

Our results suggest that despite increasing the number of Black individuals who are eligible for LCS, the 2021 USPSTF recommendation highlights ongoing socioeconomic disparities that need to be addressed to ensure equitable access.

Sociodemographic Differences in Menthol Cigarette Use in the United States

BACKGROUND

Menthol cigarette use is associated with higher smoking initiation and reduced smoking cessation. We investigated sociodemographic differences in menthol and nonmenthol cigarette use in the United States.

METHODS

We used the most recent available data from the May 2019 wave of the nationally-representative Tobacco Use Supplement to the Current Population Survey. The survey weights were used to estimate the national prevalence of individuals who currently smoke among individuals who use menthol and nonmenthol cigarettes. Survey-weighted logistic regression methods were used to estimate the association of menthol cigarette use with formerly smoking cigarettes and attempting to quit within the past 12 months while adjusting for several sociodemographic factors associated with smoking.

RESULTS

The prevalence of current smoking in individuals who ever smoked menthol cigarettes was higher 45.6% (44.5%-46.6%) compared to 35.8% (35.2%-36.4%) in individuals who ever smoked nonmenthol cigarettes. Non-Hispanics Blacks who used menthol cigarettes had a higher likelihood of being an individual who currently smokes (OR 1.8, 95% CI: [1.6-2.0], P-value <.001) compared to Non-Hispanic Whites who used nonmenthol cigarettes. However, Non-Hispanics Blacks who used menthol cigarettes were more likely to make a quit attempt (OR 1.4, 95%CI: [1.3-1.6], P-value <.001) compared to non-Hispanic Whites who used nonmenthol cigarettes.

CONCLUSIONS

Individuals who currently use menthol cigarettes are more likely to make attempts to quit smoking. However, this did not translate to successfully quitting smoking, as evidenced by the proportion of individuals who formerly smoked within the population which used menthol cigarettes.

Clinical determinants of intestinal failure and death in preterm infants with surgical necrotizing enterocolitis

OBJECTIVE

We sought to investigate the clinical determinants of intestinal failure and death in preterm infants with surgical NEC.

METHODS

Retrospective comparison of clinical information between Group A = intestinal failure (Parenteral nutrition (PN) >90 days) and death and Group B = survivors and with PN dependence < 90 days in preterm infants with surgical NEC.

RESULTS

Group A (n = 99/143) had a lower mean gestational age (26.4 weeks [SD3.5] vs. 29.4 [SD 3.5]; p = 0.013), lower birth weight (873 gm [SD 427g] vs. 1425 gm [894g]; p = <0.001), later age of NEC onset (22 days [SD20] vs. 16 days [SD 17]; p = 0.128), received surgery later (276 hours [SD 544] vs. 117 hours [SD 267]; p = 0.032), had cholestasis, received dopamine (80.6% vs. 58.5%; p = 0.010) more frequently and had longer postoperative ileus time (19.8 days [SD 15.4] vs. 11.8 days [SD 6.5]; p = <0.001) and reached full feeds later (93 days [SD 45] vs. 44 [SD 22]; p = <0.001) than Group B.On multivariate logistic regression, higher birth weight was associated with lower risk (OR 0.35, 95% CI 0.15-0.82; p = 0.016) of TPN > 90 days or death. Longer length of bowel resected (OR 1.76, 95% CI 1.02-3.02; p = 0.039) and longer postoperative ileus (OR 2.87, 95% CI 1.26-6.53; p = 0.011) were also independently associated with TPN >90days or death adjusted for gestational age and antenatal steroid treatment.

CONCLUSION

In preterm infants with surgical NEC, clinical factors such as lower birth weight, longer bowel loss, and postoperative ileus days were significantly and independently associated with TPN >90 days or death.

Clinical determinants and impact of hemorrhagic lesions on intestinal pathology in preterm infants with surgical necrotizing enterocolitis

OBJECTIVE

We sought to determine the clinical and histopathological factors associated with intestinal hemorrhage and its correlation with clinical outcomes in neonates with surgical necrotizing enterocolitis (NEC).

METHODS

A retrospective study compared clinical and histopathology information in neonates following surgical NEC with severe hemorrhage and those with mild/moderate hemorrhagic lesions seen on resected intestine pathology.

RESULTS

The infants with severe hemorrhage (Grade 3-4, 81/148, 54.7%) had significantly lower exposure to antenatal steroids (52.5 % vs 76.9 % ; p = 0.004), had higher gestational age (28.5 weeks [7.14] vs. 26.58 [2.90]; p = 0.034), lost more bowel length (p = 0.045), had higher CRP levels at 2 weeks (p = 0.035), and had less intestinal failure ([30.3 % vs 52.5 %]; p = 0.014) than mild/moderate (Grade 0-2, 67/148, 45.2%) hemorrhage group. Those with severe hemorrhage had significantly higher mean inflammation score (2.67 [0.94] vs. 1.63 [0.92]; p = <0.001), higher necrosis scores (1.95 [1.28] vs. 1.49 [1.35]; p = 0.037), higher neovascularization (p = 0.01), higher fibroblasts (p = 0.023) and higher lymphocyte percentages up to 48 hours (p <  0.05) following NEC than mild/ moderate hemorrhage group.On multivariable regression, less exposure to antenatal steroids (OR 0.18 [95% CI 0.05-0.58]; p = 0.005), higher inflammation (OR 3.7 [95% CI 2.09-7.32]; p = 0.001), and lymphocyte count on the day of onset/24 hours following NEC (OR 1.06 [95% CI 1.02-1.11]; p = 0.005) were independently associated with a higher odd of severe intestinal hemorrhage.

CONCLUSION

The surgical NEC infants with intestinal hemorrhage were less likely to have antenatal steroid exposure but had higher inflammation grade and lymphocyte counts following NEC onset on multivariable regression modeling.

The influence of parent-child gender on intentions to refuse HPV vaccination due to safety concerns/side effects, National Immunization Survey - Teen, 2010-2019

Amid subpar uptake of HPV vaccination in the United States, gender-generated disparities in HPV vaccination uptake have the potential to perpetuate existing disparities in HPV-associated cancers. Yet few studies have investigated the influence of parent-child gender on intentions to refuse HPV vaccination due to safety concerns/side effects. This study used nationally representative data, spanning 2010-2019, from the National Immunization Survey-Teen (NIS-Teen). NIS-Teen respondents are parents/guardians or primary caregivers of adolescents 13-17 years old living in the United States. Over the study period, intentions to refuse HPV vaccination due to safety concerns rose among all parent-child gender pairings but were highest among respondent mothers regarding their unvaccinated daughters. The results revealed a statistically significant increased likelihood of having intentions to refuse HPV vaccination due to safety concerns among all parent-child combinations compared with father-son pairs. These odds were consistently highest among mother-daughter pairs. In 2019, compared with father-son pairs, fathers were 1.94 (95% CI: 1.21-3.12) times more likely to report the intention to not vaccinate against HPV for their daughters, while mothers were 2.23 (95% CI: 1.57-3.17) and 2.87 (95% CI: 2.02-4.09) times more likely to report intentions to refuse HPV vaccination for their sons and daughters, respectively. These findings were persistent and constantly increased over the 10-year study period. Interventions aimed at correcting gender-based misperceptions and countering misinformation about the safety of the HPV vaccine are warranted.

Post‐kidney transplant cancers: Racial and ethnic differences in sun‐exposed skin versus non‐sun‐exposed anogenital skin

BACKGROUND

Transplant recipients have a 2- to 4-fold increased risk of developing malignancies over the general population. Cancer is the second most common cause of death for recipients. The magnitude of the risk depends on the cancer type and increases in viral-related malignancies. Skin cancer is the most common. However, data in most cancer registries is limited to cutaneous melanomas, thereby limiting the epidemiologic examination of cancer risk in non-melanoma skin cancer. Our goal was to evaluate post-kidney transplant cancer cases and sites in our population to guide screening recommendations.

METHODS

Between 2009 and 2015, a retrospective study of adult kidney recipients transplanted at East Carolina University was conducted. The first cancer diagnosis after transplant through February 18, 2020, was captured and analyzed. Patient demographics, cancer sites, and histological diagnoses were analyzed and compared. p16 immunohistochemistry was used as a surrogate marker for high-risk human papillomavirus (HPV) infection.

RESULTS

Retrospectively, kidney transplant recipients were analyzed (N = 439), the majority were non-Hispanic Black (NHB) individuals, 312 (71.1%), and 127 (28.9%) were non-Hispanic White (NHW) individuals. Of these, 59 (13.4%) developed a posttransplant malignancy, with the majority on sun-exposed skin found in NHW. NHB had all anogenital/mucosa skin cancers on non-sun-exposed skin. Of these detected in NHB, all were squamous cell carcinomas, with five out of six (83.3%) being positive for p16.

CONCLUSIONS

Posttransplant malignancy differed significantly by race, site, and potential source of etiology. The majority of malignancies are likely explained by acceleration of precursor lesions from prior exposure to ultraviolet rays or HPV.

Financial Toxicities Persist for Cancer Survivors Irrespective of Current Cancer Status: An Analysis of Medical Expenditure Panel Survey

This study estimates the out-of-pocket (OOP) expenditures for different cancer types among survivors with current vs no current cancer condition and across sex, which is understudied in the literature. This is a cross-sectional study of Medical Expenditure Panel Survey data for 2009-2018 where the primary outcome was the average per year OOP expenditure incurred by cancer survivors. Of 189 285 respondents, 15 010 (7.93%) were cancer survivors; among them, 46.28% had a current cancer condition. Average per year OOP expenditure for female survivors with a current condition of breast cancer ($1730), lung cancer ($1679), colon cancer ($1595), melanoma ($1783), non-Hodgkin lymphoma ($1656), nonmelanoma/other skin cancer (NMSC, $2118) and two or more cancers ($2310) were significantly higher than that of women with no history of cancer ($853, all P < .05). Similarly, average per year OOP expenditure for male survivors with a current condition of prostate cancer ($1457), lung cancer ($1131), colon cancer ($1471), melanoma ($1474), non-Hodgkin's lymphoma ($1653), NMSC ($1789), and bladder cancer ($2157) were significantly higher compared with the men with no history of cancer ($621, all P < .05). These differences persisted in survivors with no current cancer condition for breast cancer among women; prostate, lung, colon, and bladder cancer among men; and melanoma, NMSC, and two or more cancers among both sexes. OOP expenditure varied across cancer types and by sex for survivors with and without a current cancer condition. These findings highlight the need for targeted interventions for cancer survivors.

Comparing data-driven and hypothesis-driven MRI-based predictors of cognitive impairment in individuals from the Atherosclerosis Risk in Communities (ARIC) study

INTRODUCTION

A data-driven index of dementia risk based on magnetic resonance imaging (MRI), the Alzheimer's Disease Pattern Similarity (AD-PS) score, was estimated for participants in the Atherosclerosis Risk in Communities (ARIC) study.

METHODS

AD-PS scores were generated for 839 cognitively non-impaired individuals with a mean follow-up of 4.86 years. The scores and a hypothesis-driven volumetric measure based on several brain regions susceptible to AD were compared as predictors of incident cognitive impairment in different settings.

RESULTS

Logistic regression analyses suggest the data-driven AD-PS scores to be more predictive of incident cognitive impairment than its counterpart. Both biomarkers were more predictive of incident cognitive impairment in participants who were White, female, and apolipoprotein E gene (APOE) ε4 carriers. Random forest analyses including predictors from different domains ranked the AD-PS scores as the most relevant MRI predictor of cognitive impairment.

CONCLUSIONS

Overall, the AD-PS scores were the stronger MRI-derived predictors of incident cognitive impairment in cognitively non-impaired individuals.

Human Immunodeficiency Virus/Sexually Transmitted Infection Testing Preferences for Young Black Men Who Have Sex With Men in the Southeastern United States: Implications for a Post-COVID-19 Era

BACKGROUND

Little is known regarding human immunodeficiency virus (HIV)/sexually transmitted infection (STI) testing preferences for location, staffing, and hours of operation among Young Black men who have sex with men (YBMSM) in the Southeastern United States, a population at high risk for HIV/STIs. We used a discrete choice experiment to explore these preferences.

METHODS

Young Black men who have sex with men ages 16 to 35 years in Birmingham, AL and Jackson, MS completed online surveys evaluating their preferences (best/worst) for HIV/STI testing locations, staffing, hours, method of results notification, and cost. They also selected preferred combinations of these variables through choice tasks. Results were analyzed using joint best/worst and discrete choice experiment models.

RESULTS

Between June 2018 and December 2019, participants in Alabama (n = 54) and Mississippi (n = 159) completed online surveys. Both groups preferred stationary testing locations over mobile testing vans, with the most significant difference favoring STI testing-only clinics in Mississippi and local health departments in Alabama (P < 0.001). Technician-performed tests or self-testing were significantly less preferred compared with clinician-performed testing for both groups (P < 0.0001 and P < 0.0001, respectively). Free testing and phone results notification (versus text) were preferred by both groups. The most desirable combination among all participants was weekday clinician-performed testing at the health department for $5.

CONCLUSIONS

Young Black men who have sex with men in the Southeastern United States prefer traditional testing locations staffed by experienced personnel. Combination choices are influenced by services that are low or no cost. More research is needed to inform the best way(s) to provide affordable, high-quality HIV/STI testing services for YBMSM, particularly in the post-COVID-19 era when sexual health care delivery models are evolving toward home-based and remote health-focused strategies.

Associations of Pulmonary Function with MRI Brain Volumes: A Coordinated Multi-Study Analysis

BACKGROUND

Previous studies suggest poor pulmonary function is associated with increased burden of cerebral white matter hyperintensities and brain atrophy among elderly individuals, but the results are inconsistent.

OBJECTIVE

To study the cross-sectional associations of pulmonary function with structural brain variables.

METHODS

Data from six large community-based samples (N = 11,091) were analyzed. Spirometric measurements were standardized with respect to age, sex, height, and ethnicity using reference equations of the Global Lung Function Initiative. Associations of forced expiratory volume in 1 second (FEV1), forced vital capacity (FVC), and their ratio FEV1/FVC with brain volume, gray matter volume, hippocampal volume, and volume of white matter hyperintensities were investigated using multivariable linear regressions for each study separately and then combined using random-effect meta-analyses.

RESULTS

FEV1 and FVC were positively associated with brain volume, gray matter volume, and hippocampal volume, and negatively associated with white matter hyperintensities volume after multiple testing correction, with little heterogeneity present between the studies. For instance, an increase of FVC by one unit was associated with 3.5 ml higher brain volume (95% CI: [2.2, 4.9]). In contrast, results for FEV1/FVC were more heterogeneous across studies, with significant positive associations with brain volume, gray matter volume, and hippocampal volume, but not white matter hyperintensities volume. Associations of brain variables with both FEV1 and FVC were consistently stronger than with FEV1/FVC, specifically with brain volume and white matter hyperintensities volume.

CONCLUSION

In cross-sectional analyses, worse pulmonary function is associated with smaller brain volumes and higher white matter hyperintensities burden.

Awareness of heated tobacco products among US Adults - Health information national trends survey, 2020

Introduction: Heated Tobacco Products (HTP) have the potential to attain high uptake in the US. The current study estimated the prevalence of awareness, prevalence of use and the factors associated with awareness of HTP among US adults. Methods: This study included 3201 respondents from the Health Information National Trends Survey (HINTS) 5 Cycle 4 (2020). The prevalence of awareness of HTP and ever use of HTP among US adults were estimated. Multivariable logistic regression was conducted to identify the factors associated with awareness of HTP. Results: About 15% of the adult US population were aware of HTP, while 2.2% had ever used HTP. Age between 35 and 49 years (aOR, 1.9; 95% CI, 1.3-2.9; p-value = 0.003), male sex (aOR, 1.7; 95% CI, 1.0-2.7; p-value = 0.04), lower income ($0-$9,999) (aOR, 3.0; 95% CI, 1.3-6.9; p-value = 0.01), smoking on some days (aOR, 3.4; 95% CI, 1.2-9.4; p-value = 0.02) and moderately or extremely worrying about getting cancer (aOR, 1.7; 95% CI, 1.1-2.7; p-value = 0.03) were associated with higher odds of being aware of HTP; whereas, belief that there are so many cancer prevention recommendations, it's hard to know which ones to follow (aOR, 0.5; 95% CI, 0.3-0.8; p-value = 0.009) was associated with lower odds of being aware of HTP. Conclusions: A significant proportion of US adult population were aware of HTP in 2020. Given the recent high proliferation of e-cigarettes, potential health effects of the HTP products should be monitored by the regulators closely. Adequate surveillance and policy interventions are warranted in this regard.

Safety Concerns or Adverse Effects as the Main Reason for Human Papillomavirus Vaccine Refusal: National Immunization Survey-Teen, 2008 to 2019

This secondary analysis assesses safety concerns and adverse effects as reported reasons for HPV vaccine refusal.

Reflection on modern methods: shared-parameter models for longitudinal studies with missing data

A primary goal of longitudinal studies is to examine trends over time. Reported results from these studies often depend on strong, unverifiable assumptions about the missing data. Whereas the risk of substantial bias from missing data is widely known, analyses exploring missing-data influences are commonly done either ad hoc or not at all. This article outlines one of the three primary recognized approaches for examining missing-data effects that could be more widely used, i.e. the shared-parameter model (SPM), and explains its purpose, use, limitations and extensions. We additionally provide synthetic data and reproducible research code for running SPMs in SAS, Stata and R programming languages to facilitate their use in practice and for teaching purposes in epidemiology, biostatistics, data science and related fields. Our goals are to increase understanding and use of these methods by providing introductions to the concepts and access to helpful tools.

Trends in HPV Vaccination Initiation and Completion Within Ages 9-12 Years: 2008-2018

BACKGROUND AND OBJECTIVES

Routine human papillomavirus (HPV) vaccination is recommended at ages 11 to 12 years and may be initiated as early as 9 years of age.

METHODS

Data were derived from the National Immunization Survey-Teen, spanning 2008-2018. Using health care provider-verified vaccination histories, we examined trends in human papillomavirus vaccination up-to-date (HPV-UTD) rates within ages 9 to 12 years. Furthermore, we assessed vaccination status by sociodemographic factors and US state of residence.

RESULTS

Overall, amid evidence of recent stagnation, HPV vaccination between ages 9 to 12 increased over the years. Initiation rates rose from 17.3% in 2008 to 62.8% in 2018, and HPV-UTD rates rose from 13.5% in 2011 to 32.8% in 2018. After the inception of gender-neutral HPV vaccination, HPV-UTD rates between 2011 and 2018 rose by 31.9% among boys and only 6.6% among girls. For most of the study period, non-Hispanic Black and Hispanic individuals had higher rates of initiation and HPV-UTD than non-Hispanic white individuals. In 2018, vaccination initiation rates exceeded 70% in several states; however, HPV-UTD rates in most US states were <50%, excluding Rhode Island (61.6%), Colorado (58.7%), Hawaii (53.5%), District of Columbia (53.2%), and Ohio (50%).

CONCLUSIONS

HPV vaccination within ages 9 to 12 years is suboptimal. To leverage the substantial benefits of HPV vaccination within this age range, it is imperative that conscious efforts are taken at the national and state levels to promote HPV vaccination for this age group.

Disparities in Secondhand Smoke Exposure in the United States: National Health and Nutrition Examination Survey 2011-2018

This cross-sectional study examines National Health and Nutrition Examination Survey data to assess the prevalence of secondhand smoke exposure in nonsmokers in the US population 3 years and older.

Assessment of Trends in Cigarette Smoking Cessation After Cancer Diagnosis Among US Adults, 2000 to 2017

IMPORTANCE

Continued cigarette smoking after cancer diagnosis has been shown to adversely alter the cancer survivor's health outcomes and quality of life. Although considerable progress has been made in reducing cigarette smoking in the United States over the last decades, trends in cigarette smoking cessation among cancer survivors have not yet been fully investigated.

OBJECTIVE

To evaluate temporal trends in cigarette smoking cessation and their sociodemographic and behavioral correlates in the US population of adult cancer survivors.

DESIGN, SETTING, AND PARTICIPANTS

This investigation was a serial cross-sectional study of the National Health Interview Survey (NHIS) from 2006 to 2018, a household survey of civilian US residents who were 18 years or older. Data analysis was performed from June to October 2019. The NHIS is population based and representative of the US population. Included in this study were 381 989 respondents to the NHIS in 2006 to 2018.

MAIN OUTCOMES AND MEASURES

The primary outcome was the probability of quitting cigarette smoking after first cancer diagnosis. Secondary outcomes were factors associated with quitting cigarette smoking after cancer diagnosis.

RESULTS

Data on 381 989 adults (weighted N = 239 114 051; mean [SD] age, 48.96 [18.28] years; 211 508 [55.37%] female; 61.90% non-Hispanic White, 13.97% non-Hispanic Black, and 16.22% Hispanic individuals) 18 years or older were analyzed from the 2006 to 2018 NHIS, of whom 8.80% (n = 35 524; weighted n = 21 016 720) were diagnosed with cancer. Among cancer survivors diagnosed between 2000 and 2017, the age-adjusted prevalence of current cigarette smoking at the time of first cancer diagnosis was 24.45% (n = 4054; weighted n = 2 395 173). The probability of reporting a cigarette smoking cessation event after first cancer diagnosis statistically significantly increased with each year of cancer diagnosis (hazard ratio [HR], 1.05; 95% CI, 1.02-1.08), indicating upward trends in the prevalence of quitting cigarette smoking over time. Older individuals (HR, 1.02; 95% CI, 1.01-1.03), individuals diagnosed as having a smoking-related cancer (HR, 1.28; 95% CI, 1.06-1.54), individuals with an undergraduate degree (HR, 1.39; 95% CI, 1.08-1.79) or a postgraduate degree (HR, 1.61; 95% CI, 1.18-2.20), and individuals with obesity (HR, 1.32; 95% CI, 1.06-1.63) had a higher probability of reporting a cigarette smoking cessation event after cancer diagnosis, whereas individuals living below the poverty level (HR, 0.62; 95% CI, 0.48-0.81) had a lower probability of reporting a cigarette smoking cessation event after cancer diagnosis.

CONCLUSIONS AND RELEVANCE

In this nationally representative survey of the US adult population, the likelihood of cigarette smoking cessation among cancer survivors increased with the year of cancer diagnosis from 2000 to 2017; however, the improvement is incremental, and the prevalence of smoking remained high among this population. Considering the projected increase in the population of cancer survivors in the United States, urgent action is needed to increase cigarette smoking cessation rates in this high-risk population.

Smoking Behaviors in Survivors of Smoking-Related and Non-Smoking-Related Cancers

IMPORTANCE

The population of cancer survivors is rapidly growing in the US. Tobacco smoking is associated with many cancers; however, whether cigarette smoking behaviors among cancer survivors vary according to cancer type-that is, smoking-related cancers (SRCs) vs non-smoking-related cancers (NSRCs)-remains unclear.

OBJECTIVES

To examine cigarette smoking prevalence and behaviors (ie, continuing or quitting smoking) among cancer survivors and to compare them between survivors of SRCs and NSRCs.

DESIGN, SETTING, AND PARTICIPANTS

This study was a cross-sectional analysis of the 2017 National Health Interview Survey, a household survey of civilian US residents who were aged 18 years or older. The National Health Interview Survey is population based and is representative of the US population. Data analysis was performed from June to October 2019.

MAIN OUTCOMES AND MEASURES

The primary outcomes were prevalence of current cigarette smoking among cancer survivors and prevalence of continuing smoking and quitting smoking after a cancer diagnosis. Secondary outcomes included factors associated with continued smoking vs quitting smoking after a cancer diagnosis.

RESULTS

A total of 26 742 respondents (mean [SD] age, 50.97 [18.61] years; 14 646 women [51.76%]) to the 2017 National Health Interview Survey were included in this study. Of the 3068 individuals (9.42%) in the study population who had cancer, 589 (19.96%) were SRC survivors, 2297 (74.50%) were NSRC survivors, 168 (4.96%) were survivors of both SRC and NSRC, and the remaining 14 (0.58%) had missing information about the type of cancer. Four hundred forty-nine SRC survivors (54.08%) were women, compared with 1412 NSRC survivors (54.30%). Ninety-six SRC survivors (15.69%) and 151 NSRC survivors (7.99%) were younger than 45 years. Overall, 372 cancer survivors (13.16%) were current smokers. Current smoking prevalence was higher among survivors of SRCs (145 survivors [19.78%]) compared with NSRC survivors (251 survivors [10.63%]). Among cancer survivors, 309 current smokers at cancer diagnosis (43.96%) reported having successfully quit smoking and 372 (56.04%) reported continuing smoking. Among the continuing smokers, 176 (56.49%) reported an unsuccessful quit attempt in the last 12 months. After cancer diagnosis, SRC survivors had higher odds of continued smoking compared with NSRC survivors (odds ratio [OR], 2.10; 95% CI, 1.12-3.93; P = .02). Men (OR, 1.93; 95% CI, 1.05-3.57; P = .04), those with angina pectoris (OR, 5.40; 95% CI, 1.33-21.91; P = .02), and those with chronic bronchitis (OR, 2.55; 95% CI, 1.05-6.19; P = .04) had higher odds of continued smoking, whereas Hispanic participants (compared with non-Hispanic white participants: OR, 0.18; 95% CI, 0.05-0.68; P = .01) and married participants (compared with never married participants: OR, 0.33; 95% CI, 0.12-0.96; P = .04) had lower odds of continued smoking.

CONCLUSIONS AND RELEVANCE

These findings suggest that compared with NSRC survivors, SRC survivors may be at higher risk of being cigarette smokers at cancer diagnosis and of continuing smoking afterward. Although smoking cessation interventions are critically important for all cancer survivors, special efforts should target survivors of SRCs.

Physician-office vs home uptake of colorectal cancer screening using FOBT/FIT among screening-eligible US adults

Background

Guidelines of the American Cancer Society and US Preventive Services Task Force specify that colorectal cancer (CRC) screening using guaiac‐based fecal occult blood test (FOBT)/fecal immunochemical test (FIT) should be done at home. We therefore examined the prevalence and correlates of CRC screening using FOBT/FIT in physicians' office vs at home.

Methods

Analysis of 9493 respondents 50‐75 years old from the Cancer Control Supplement of the 2015 National Health Interview Survey was conducted. Weighted multivariable logistic regression was used to identify the determinants of in‐office vs home use of FOBT/FIT for CRC screening.

Results

Of the overall sample of screening‐eligible adults (n = 9403), only 937 (10.4%) respondents underwent CRC screening using FOBT/FIT within the past year; among this screening population, 279 (28.3%) respondents were screened in‐office. We found that sociodemographic factors alone, not CRC risk factors, determined whether FOBT/FIT would be used in‐office or at home. Hispanics had greater odds of being screened in‐office using FOBT/FIT (aOR: 2.04; 95% CI: 1.05‐3.99). Compared with those 50‐59 years old, respondents 70‐75 years old were less likely to be screened in‐office using FOBT/FIT (aOR: 0.44, 95% CI: 0.25‐0.79). Similarly, individuals residing in the Western region of the country had lower odds of in‐office FOBT/FIT (aOR: 0.26; 95% CI: 0.11‐0.58).

Conclusion

Amid low overall uptake rates of FOBT/FIT in the United States, in‐physician office testing is high, indicative of a missed opportunity for effective screening and poor adherence of physicians to national guidelines. Sociodemographic factors are determinants of uptake of FOBT/FIT at home or in‐office and should be considered in designing interventions aimed at providers and the general population.

Amid low overall uptake rates of FOBT/FIT in the United States, in‐physician office testing is high, indicative of a missed opportunity for effective screening and poor adherence of physicians to national guidelines.

Pre-medication before dental procedures: A randomized controlled study comparing intranasal dexmedetomidine with oral midazolam

BACKGROUND

This study aimed to determine if intranasal dexmedetomidine is a superior pre-medication to oral midazolam in older, difficult children.

METHODS

This was conducted as a prospective, single-blind randomized control trial in a tertiary care center. Seventy-five children, age >5 years and weight >20 kg, who needed general anesthesia for dental procedures were randomly assigned to be pre-medicated with either oral midazolam at a dose of 0.5 mg/kg (max 15 mg) or intranasal dexmedetomidine at a dose of 2 mcg/kg (max 100 mcg). The primary outcome studied was the patients' level of sedation when separated from their parents, which was assessed using a 5-point University of Michigan Sedation Scale. Secondary outcome studied was the level of anxiolysis assessed by the acceptance of mask induction using a 4-point scale. All assessments were made by one research person blinded to the study drug.

RESULTS

The two groups were similar in age, sex, weight, pre-anesthetic behavior, time from pre-medication to anesthesia induction, and surgical time. A significantly higher proportion of patients who received dexmedetomidine had satisfactory sedation at separation from parents (69.4% vs 40.5%, P = .03) compared to those who received midazolam. There were no significant differences in the rate of acceptance of mask induction (80.6% vs 78.4%, P = 1.00). Intranasal dexmedetomidine was tolerated well when administered using a mucosal atomizer and without any clinically significant effect on heart rate or systolic blood pressure.

CONCLUSIONS

Intranasal dexmedetomidine provides higher success rate in sedation and parental separation compared to oral midazolam, in older, difficult children.

Identification of Key Bacteria Involved in the Induction of Incident Bacterial Vaginosis: A Prospective Study

Background

The sequence of events preceding incident bacterial vaginosis (iBV) is unclear.

Methods

African American women who have sex with women, who had no Amsel criteria and Nugent scores of 0-3, were followed for 90 days to detect iBV (defined as a Nugent score of 7-10 on at least 2-3 consecutive days), using self-collected vaginal swab specimens. For women with iBV (cases) and women maintaining normal vaginal flora (healthy women), 16S ribosomal RNA gene sequencing targeting V4 was performed. Longitudinal vaginal microbiome data were analyzed.

Results

Of 204 women screened, 42 enrolled; of these, 45% developed iBV. Sequencing was performed on 448 specimens from 14 cases and 8 healthy women. Among healthy women, Lactobacillus crispatus dominated the vaginal microbiota in 75%. In contrast, prior to iBV, the vaginal microbiota in 79% of cases was dominated by Lactobacillus iners and/or Lactobacillus jensenii/Lactobacillus gasseri. The mean relative abundance of Prevotella bivia, Gardnerella vaginalis, Atopobium vaginae, and Megasphaera type I became significantly higher in cases 4 days before (P. bivia), 3 days before (G. vaginalis), and on the day of (A. vaginae and Megasphaera type I) iBV onset. The mean relative abundance of Sneathia sanguinegens, Finegoldia magna, BV-associated bacteria 1-3, and L. iners was not significantly different between groups before onset of iBV.

Conclusion

G. vaginalis, P. bivia, A. vaginae, and Megasphaera type I may play significant roles in iBV.

An approach to estimate bidirectional mediation effects with application to body mass index and fasting glucose

Obesity and type 2 diabetes are major public health issues with known interdependence. Genetic variants have been associated with obesity, type 2 diabetes, or both; thus, we hypothesize that some single nucleotide polymorphisms (SNPs) associated with both conditions may be mediated through obesity to affect type 2 diabetes or vice versa. We propose a framework for bidirectional mediation analyses. Simulations show that this approach accurately estimates the parameters, whether the mediation is unidirectional or bidirectional. In many scenarios, when the mediator is regressed on the initial variable and the outcome is regressed on the mediator and the initial variable, the resulting residuals are correlated because of other unmeasured covariates not in the model. We show that the proposed model provides accurate estimates in this scenario, too. We applied the proposed approach to investigate the mediating effects of SNPs associated with type 2 diabetes and obesity using genetic data from the Multi-Ethnic Study of Atherosclerosis cohort. Specifically, we used body mass index (BMI) as a measure for obesity and fasting glucose as a measure for type 2 diabetes. We evaluated the top 6 SNPs associated with both BMI and fasting glucose. Two SNPs (rs3752355 and rs6087982) had indirect effects on BMI mediated through fasting glucose [0.2677; 95% confidence interval (CI) (0.0007, 0.6548) and 0.3301; 95% CI (0.0881, 0.8544), respectively]. The remaining four SNPs (rs7969190, rs4869710, rs10201400, and rs12421620) directly affect BMI and fasting glucose without mediating effects.

Selection of X-chromosome Inactivation Model

To address the complexity of the X-chromosome inactivation (XCI) process, we previously developed a unified approach for the association test for X-chromosomal single-nucleotide polymorphisms (SNPs) and the disease of interest, accounting for different biological possibilities of XCI: random, skewed, and escaping XCI. In the original study, we focused on the SNP-disease association test but did not provide knowledge regarding the underlying XCI models. One can use the highest likelihood ratio (LLR) to select XCI models (max-LLR approach). However, that approach does not formally compare the LLRs corresponding to different XCI models to assess whether the models are distinguishable. Therefore, we propose an LLR comparison procedure (comp-LLR approach), inspired by the Cox test, to formally compare the LLRs of different XCI models to select the most likely XCI model that describes the underlying XCI process. We conduct simulation studies to investigate the max-LLR and comp-LLR approaches. The simulation results show that compared with the max-LLR, the comp-LLR approach has higher probability of identifying the correct underlying XCI model for the scenarios when the underlying XCI process is random XCI, escaping XCI, or skewed XCI to the deleterious allele. We applied both approaches to a head and neck cancer genetic study to investigate the underlying XCI processes for the X-chromosomal genetic variants.

Using the weighted area under the net benefit curve for decision curve analysis

BACKGROUND

Risk prediction models have been proposed for various diseases and are being improved as new predictors are identified. A major challenge is to determine whether the newly discovered predictors improve risk prediction. Decision curve analysis has been proposed as an alternative to the area under the curve and net reclassification index to evaluate the performance of prediction models in clinical scenarios. The decision curve computed using the net benefit can evaluate the predictive performance of risk models at a given or range of threshold probabilities. However, when the decision curves for 2 competing models cross in the range of interest, it is difficult to identify the best model as there is no readily available summary measure for evaluating the predictive performance. The key deterrent for using simple measures such as the area under the net benefit curve is the assumption that the threshold probabilities are uniformly distributed among patients.

METHODS

We propose a novel measure for performing decision curve analysis. The approach estimates the distribution of threshold probabilities without the need of additional data. Using the estimated distribution of threshold probabilities, the weighted area under the net benefit curve serves as the summary measure to compare risk prediction models in a range of interest.

RESULTS

We compared 3 different approaches, the standard method, the area under the net benefit curve, and the weighted area under the net benefit curve. Type 1 error and power comparisons demonstrate that the weighted area under the net benefit curve has higher power compared to the other methods. Several simulation studies are presented to demonstrate the improvement in model comparison using the weighted area under the net benefit curve compared to the standard method.

CONCLUSIONS

The proposed measure improves decision curve analysis by using the weighted area under the curve and thereby improves the power of the decision curve analysis to compare risk prediction models in a clinical scenario.

Evaluating methods for modeling epistasis networks with application to head and neck cancer

Epistasis helps to explain how multiple single-nucleotide polymorphisms (SNPs) interact to cause disease. A variety of tools have been developed to detect epistasis. In this article, we explore the strengths and weaknesses of an information theory approach for detecting epistasis and compare it to the logistic regression approach through simulations. We consider several scenarios to simulate the involvement of SNPs in an epistasis network with respect to linkage disequilibrium patterns among them and the presence or absence of main and interaction effects. We conclude that the information theory approach more efficiently detects interaction effects when main effects are absent, whereas, in general, the logistic regression approach is appropriate in all scenarios but results in higher false positives. We compute epistasis networks for SNPs in the FSD1L gene using a two-phase head and neck cancer genome-wide association study involving 2,185 cases and 4,507 controls to demonstrate the practical application of the methods.

BAYESIAN SPARSE GRAPHICAL MODELS FOR CLASSIFICATION WITH APPLICATION TO PROTEIN EXPRESSION DATA

Reverse-phase protein array (RPPA) analysis is a powerful, relatively new platform that allows for high-throughput, quantitative analysis of protein networks. One of the challenges that currently limit the potential of this technology is the lack of methods that allow for accurate data modeling and identification of related networks and samples. Such models may improve the accuracy of biological sample classification based on patterns of protein network activation and provide insight into the distinct biological relationships underlying different types of cancer. Motivated by RPPA data, we propose a Bayesian sparse graphical modeling approach that uses selection priors on the conditional relationships in the presence of class information. The novelty of our Bayesian model lies in the ability to draw information from the network data as well as from the associated categorical outcome in a unified hierarchical model for classification. In addition, our method allows for intuitive integration of a priori network information directly in the model and allows for posterior inference on the network topologies both within and between classes. Applying our methodology to an RPPA data set generated from panels of human breast cancer and ovarian cancer cell lines, we demonstrate that the model is able to distinguish the different cancer cell types more accurately than several existing models and to identify differential regulation of components of a critical signaling network (the PI3K-AKT pathway) between these two types of cancer. This approach represents a powerful new tool that can be used to improve our understanding of protein networks in cancer.

A risk prediction model for smoking experimentation in Mexican American youth

BACKGROUND

Smoking experimentation in Mexican American youth is problematic. In light of the research showing that preventing smoking experimentation is a valid strategy for smoking prevention, there is a need to identify Mexican American youth at high risk for experimentation.

METHODS

A prospective population-based cohort of 1,179 adolescents of Mexican descent was followed for 5 years starting in 2005-06. Participants completed a baseline interview at a home visit followed by three telephone interviews at intervals of approximately 6 months and additional interviews at two home visits in 2008-09 and 2010-11. The primary endpoint of interest in this study was smoking experimentation. Information about social, cultural, and behavioral factors (e.g., acculturation, susceptibility to experimentation, home characteristics, and household influences) was collected at baseline using validated questionnaires.

RESULTS

Age, sex, cognitive susceptibility, household smoking behavior, peer influence, neighborhood influence, acculturation, work characteristics, positive outcome expectations, family cohesion, degree of tension, ability to concentrate, and school discipline were found to be associated with smoking experimentation. In a validation dataset, the proposed risk prediction model had an area under the receiver operating characteristic curve (AUC) of 0.719 (95% confidence interval, 0.637-0.801) for predicting absolute risk for smoking experimentation within 1 year.

CONCLUSIONS

The proposed risk prediction model is able to quantify the risk of smoking experimentation in Mexican American adolescents.

IMPACT

Accurately identifying Mexican American adolescents who are at higher risk for smoking experimentation who can be intervened will substantially reduce the incidence of smoking and thereby subsequent health risks.

Calculation of exact p-values when SNPs are tested using multiple genetic models

Background

Several methods have been proposed to account for multiple comparisons in genetic association studies. However, investigators typically test each of the SNPs using multiple genetic models. Association testing using the Cochran-Armitage test for trend assuming an additive, dominant, or recessive genetic model, is commonly performed. Thus, each SNP is tested three times. Some investigators report the smallest p-value obtained from the three tests corresponding to the three genetic models, but such an approach inherently leads to inflated type 1 errors. Because of the small number of tests (three) and high correlation (functional dependence) among these tests, the procedures available for accounting for multiple tests are either too conservative or fail to meet the underlying assumptions (e.g., asymptotic multivariate normality or independence among the tests).

Results

We propose a method to calculate the exact p-value for each SNP using different genetic models. We performed simulations, which demonstrated the control of type 1 error and power gains using the proposed approach. We applied the proposed method to compute p-value for a polymorphism eNOS -786T>C which was shown to be associated with breast cancer risk.

Conclusions

Our findings indicate that the proposed method should be used to maximize power and control type 1 errors when analyzing genetic data using additive, dominant, and recessive models.

Gaussian graphical models for phenotypes using pedigree data and exploratory analysis using networks with genetic and nongenetic factors based on Genetic Analysis Workshop 18 data

Graphical models are increasingly used in genetic analyses to take into account the complex relationships between genetic and nongenetic factors influencing the phenotypes. We propose a model for determining the network structure of quantitative traits while accounting for the correlated nature of the family-based samples using the kinship coefficient. The Gaussian graphical model of age, systolic blood pressure, diastolic blood pressure, hypertension, blood pressure medication use, and smoking status was derived for three time points using real data. We also explored binary sparse graphical models of single-nucleotide polymorphisms (SNPs), covariates, and quantitative traits for exploratory analysis of the data. We validated the applicability of this method by producing a network graph using 20 causal variants, 21 noncausal variants, and 6 binary and quantitative phenotypes using the simulated data. To improve the model's ability to identify associations between the causal variants and the phenotypes, we intend to conduct follow-up studies investigating how to use the relationships between SNPs and between SNPs and phenotypes when analyzing genome wide association data with multiple phenotypes.

Bayesian sparse graphical models and their mixtures

We propose Bayesian methods for Gaussian graphical models that lead to sparse and adaptively shrunk estimators of the precision (inverse covariance) matrix. Our methods are based on lasso-type regularization priors leading to parsimonious parameterization of the precision matrix, which is essential in several applications involving learning relationships among the variables. In this context, we introduce a novel type of selection prior that develops a sparse structure on the precision matrix by making most of the elements exactly zero, in addition to ensuring positive definiteness - thus conducting model selection and estimation simultaneously. More importantly, we extend these methods to analyze clustered data using finite mixtures of Gaussian graphical model and infinite mixtures of Gaussian graphical models. We discuss appropriate posterior simulation schemes to implement posterior inference in the proposed models, including the evaluation of normalizing constants that are functions of parameters of interest, which result from the restriction of positive definiteness on the correlation matrix. We evaluate the operating characteristics of our method via several simulations and demonstrate the application to real data examples in genomics.

A linkage disequilibrium-based approach to selecting disease-associated rare variants

Rare variants have increasingly been cited as major contributors in the disease etiology of several complex disorders. Recently, several approaches have been proposed for analyzing the association of rare variants with disease. These approaches include collapsing rare variants, summing rare variant test statistics within a particular locus to improve power, and selecting a subset of rare variants for association testing, e.g., the step-up approach. We found that (a) if the variants being pooled are in linkage disequilibrium, the standard step-up method of selecting the best subset of variants results in loss of power compared to a model that pools all rare variants and (b) if the variants are in linkage equilibrium, performing a subset selection using step-based selection methods results in a gain of power of association compared to a model that pools all rare variants. Therefore, we propose an approach to selecting the best subset of variants to include in the model that is based on the linkage disequilibrium pattern among the rare variants. The proposed linkage disequilibrium-based variant selection model is flexible and borrows strength from the model that pools all rare variants when the rare variants are in linkage disequilibrium and from step-based selection methods when the variants are in linkage equilibrium. We performed simulations under three different realistic scenarios based on: (1) the HapMap3 dataset of the DRD2 gene, and CHRNA3/A5/B4 gene cluster (2) the block structure of linkage disequilibrium, and (3) linkage equilibrium. We proposed a permutation-based approach to control the type 1 error rate. The power comparisons after controlling the type 1 error show that the proposed linkage disequilibrium-based subset selection approach is an attractive alternative method for subset selection of rare variants.

Bayesian Random Segmentation Models to Identify Shared Copy Number Aberrations for Array CGH Data

Array-based comparative genomic hybridization (aCGH) is a high-resolution high-throughput technique for studying the genetic basis of cancer. The resulting data consists of log fluorescence ratios as a function of the genomic DNA location and provides a cytogenetic representation of the relative DNA copy number variation. Analysis of such data typically involves estimation of the underlying copy number state at each location and segmenting regions of DNA with similar copy number states. Most current methods proceed by modeling a single sample/array at a time, and thus fail to borrow strength across multiple samples to infer shared regions of copy number aberrations. We propose a hierarchical Bayesian random segmentation approach for modeling aCGH data that utilizes information across arrays from a common population to yield segments of shared copy number changes. These changes characterize the underlying population and allow us to compare different population aCGH profiles to assess which regions of the genome have differential alterations. Our method, referred to as BDSAcgh (Bayesian Detection of Shared Aberrations in aCGH), is based on a unified Bayesian hierarchical model that allows us to obtain probabilities of alteration states as well as probabilities of differential alteration that correspond to local false discovery rates. We evaluate the operating characteristics of our method via simulations and an application using a lung cancer aCGH data set.