Ann Marie Navar

Effectiveness of measles vaccination and vitamin A treatment

Background The current strategy utilized by WHO/United Nations Childrens Fund (UNICEF) to reach the Global Immunization Vision and Strategy 2010 measles reduction goal includes increasing coverage of measles vaccine, vitamin A treatment and supplementation in addition to offering two doses of vaccine to all children. Methods We conducted a systematic review of published randomized controlled trials (RCTs) and quasi-experimental (QE) studies in order to determine effect estimates of measles vaccine and vitamin A treatment for the Lives Saved Tool (LiST). We utilized a standardized abstraction and grading format in order to determine effect estimates for measles mortality employing the standard Child Health Epidemiology Research Group Rules for Evidence Review. Results We identified three measles vaccine RCTs and two QE studies with data on prevention of measles disease. A meta-analysis of these studies found that vaccination was 85% [95% confidence interval (CI) 83–87] effective in preventing measles disease, which will be used as a proxy for measles mortality in LiST for countries vaccinating before one year of age. The literature also suggests that a conservative 95% effect estimate is reasonable to employ when vaccinating at 1 year or later and 98% for two doses of vaccine based on serology reviews. We included six high-quality RCTs in the meta-analysis of vitamin A treatment of measles which found no significant reduction in measles morality. However, when stratifying by vitamin A treatment dose, at least two doses were found to reduce measles mortality by 62% (95% CI 19–82). Conclusion Measles vaccine and vitamin A treatment are effective interventions to prevent measles mortality in children.

Opportunities and challenges in developing risk prediction models with electronic health records data: a systematic review

Objective: Electronic health records (EHRs) are an increasingly common data source for clinical risk prediction, presenting both unique analytic opportunities and challenges. We sought to evaluate the current state of EHR based risk prediction modeling through a systematic review of clinical prediction studies using EHR data. Methods: We searched PubMed for articles that reported on the use of an EHR to develop a risk prediction model from 2009 to 2014. Articles were extracted by two reviewers, and we abstracted information on study design, use of EHR data, model building, and performance from each publication and supplementary documentation. Results: We identified 107 articles from 15 different countries. Studies were generally very large (median sample size = 26 100) and utilized a diverse array of predictors. Most used validation techniques (n = 94 of 107) and reported model coefficients for reproducibility (n = 83). However, studies did not fully leverage the breadth of EHR data, as they uncommonly used longitudinal information (n = 37) and employed relatively few predictor variables (median = 27 variables). Less than half of the studies were multicenter (n = 50) and only 26 performed validation across sites. Many studies did not fully address biases of EHR data such as missing data or loss to follow-up. Average c-statistics for different outcomes were: mortality (0.84), clinical prediction (0.83), hospitalization (0.71), and service utilization (0.71). Conclusions: EHR data present both opportunities and challenges for clinical risk prediction. There is room for improvement in designing such studies.

Vaccine knowledge and practices of primary care providers of exempt vs. vaccinated children

Objectives: Compare vaccine knowledge, attitudes, and practices of primary care providers for fully vaccinated children and children who are exempt from school immunization requirements. Methods: We conducted a mailed survey of parent-identified primary care providers from four states to measure perceived risks and benefits of vaccination and other key immunization beliefs. Frequencies of responses were stratified by type of provider, identified by exempt versus vaccinated children. Logistic regression was used to calculate odds ratios for responses by provider type. Results: 551 surveys were completed (84.3% response rate). Providers for exempt children had similar attitudes to providers for non-exempt children. However, there were statistically significant increased concerns among providers for exempt children regarding vaccine safety and lack of perceived individual and community benefits for vaccines compared to other providers. Conclusions: The great majority of providers for exempt children had similar attitudes about vaccine safety, effectiveness, and benefits as providers of non-exempt children. Although providers for exempt children were more likely to believe that multiple vaccines weaken a child’s immune system and were concerned about vaccine safety and less likely to consider vaccines were beneficial, a substantial proportion of providers of both exempt and vaccinated children have concerns about vaccine safety and believe that CDC underestimates the frequency of vaccine side effects. Effective continuing education of providers about the risks and benefits of immunization and including in vaccine recommendations more information on pre and post licensing vaccine safety evaluations may overcome some of these perceptions.

Use of Open Access Platforms for Clinical Trial Data

Concerns over bias in clinical trial reporting have stimulated calls for more open data sharing.1 In response, multiple pharmaceutical companies have created mechanisms for investigators to access patient-level clinical trials data. However, if and how these shared clinical trial data are being used is unknown.

Using Seroprevalence and Immunisation Coverage Data to Estimate the Global Burden of Congenital Rubella Syndrome, 1996-2010: A Systematic Review.

Background The burden of Congenital Rubella Syndrome (CRS) is typically underestimated in routine surveillance. Updated estimates are needed following the recent WHO position paper on rubella and recent GAVI initiatives, funding rubella vaccination in eligible countries. Previous estimates considered the year 1996 and only 78 (developing) countries. Methods We reviewed the literature to identify rubella seroprevalence studies conducted before countries introduced rubella-containing vaccination (RCV). These data and the estimated vaccination coverage in the routine schedule and mass campaigns were incorporated in mathematical models to estimate the CRS incidence in 1996 and 2000–2010 for each country, region and globally. Results The estimated CRS decreased in the three regions (Americas, Europe and Eastern Mediterranean) which had introduced widespread RCV by 2010, reaching <2 per 100,000 live births (the Americas and Europe) and 25 (95% CI 4–61) per 100,000 live births (the Eastern Mediterranean). The estimated incidence in 2010 ranged from 90 (95% CI: 46–195) in the Western Pacific, excluding China, to 116 (95% CI: 56–235) and 121 (95% CI: 31–238) per 100,000 live births in Africa and SE Asia respectively. Highest numbers of cases were predicted in Africa (39,000, 95% CI: 18,000–80,000) and SE Asia (49,000, 95% CI: 11,000–97,000). In 2010, 105,000 (95% CI: 54,000–158,000) CRS cases were estimated globally, compared to 119,000 (95% CI: 72,000–169,000) in 1996. Conclusions Whilst falling dramatically in the Americas, Europe and the Eastern Mediterranean after vaccination, the estimated CRS incidence remains high elsewhere. Well-conducted seroprevalence studies can help to improve the reliability of these estimates and monitor the impact of rubella vaccination.

Moving beyond regression techniques in cardiovascular risk prediction: applying machine learning to address analytic challenges

Abstract Risk prediction plays an important role in clinical cardiology research. Traditionally, most risk models have been based on regression models. While useful and robust, these statistical methods are limited to using a small number of predictors which operate in the same way on everyone, and uniformly throughout their range. The purpose of this review is to illustrate the use of machine-learning methods for development of risk prediction models. Typically presented as black box approaches, most machine-learning methods are aimed at solving particular challenges that arise in data analysis that are not well addressed by typical regression approaches. To illustrate these challenges, as well as how different methods can address them, we consider trying to predicting mortality after diagnosis of acute myocardial infarction. We use data derived from our institutions electronic health record and abstract data on 13 regularly measured laboratory markers. We walk through different challenges that arise in modelling these data and then introduce different machine-learning approaches. Finally, we discuss general issues in the application of machine-learning methods including tuning parameters, loss functions, variable importance, and missing data. Overall, this review serves as an introduction for those working on risk modelling to approach the diffuse field of machine learning.

Comparison of Recommended Eligibility for Primary Prevention Statin Therapy Based on the US Preventive Services Task Force Recommendations vs the ACC/AHA Guidelines

Importance There are important differences among guideline recommendations for using statin therapy in primary prevention. New recommendations from the US Preventive Services Task Force (USPSTF) emphasize therapy based on the presence of 1 or more cardiovascular disease (CVD) risk factors and a 10-year global CVD risk of 10% or greater. Objective To determine the difference in eligibility for primary prevention statin treatment among US adults, assuming full application of USPSTF recommendations compared with the American College of Cardiology/American Heart Association (ACC/AHA) guidelines. Design, Setting, and Participants National Health and Nutrition Examination Survey (NHANES) data (2009-2014) were used to assess statin eligibility under the 2016 USPSTF recommendations vs the 2013 ACC/AHA cholesterol guidelines among a nationally representative sample of 3416 US adults aged 40 to 75 years with fasting lipid data and triglyceride levels of 400 mg/dL or less, without prior CVD. Exposures The 2016 USPSTF recommendations vs 2013 ACC/AHA guidelines. Main Outcomes and Measures Eligibility for primary prevention statin therapy. Results Among the US primary prevention population represented by 3416 individuals in NHANES, the median weighted age was 53 years (interquartile range, 46-61), and 53% (95% CI, 52%-55%) were women. Along with the 21.5% (95% CI, 19.3%-23.7%) of patients who reported currently taking lipid-lowering medication, full implementation of the USPSTF recommendations would be associated with initiation of statin therapy in an additional 15.8% (95% CI, 14.0%-17.5%) of patients, compared with an additional 24.3% (95% CI, 22.3%-26.3%) of patients who would be recommended for statin initiation under full implementation of the 2013 ACC/AHA guidelines. Among the 8.9% of individuals in the primary prevention population who would be recommended for statins by ACC/AHA guidelines but not by USPSTF recommendations, 55% would be adults aged 40 to 59 years with a mean 30-year cardiovascular risk greater than 30%, and 28% would have diabetes. Conclusions And Relevance In this sample of US adults from 2009-2014, adherence to the 2016 USPSTF recommendations for statin therapy, compared with the 2013 ACC/AHA guidelines, could lead to a lower number of individuals recommended for primary prevention statin therapy, including many younger adults with high mean long-term CVD risk.

Research needs to improve hypertension treatment and control in African Americans

This report presents findings of an ad hoc working group assembled by the National Heart, Lung, and Blood Institute (NHLBI) to assess research needs to improve prevention, treatment, and control of hypertension among African Americans. Non-Hispanic Blacks (African American and Black will be used for US and international studies, respectively) tend to have an earlier onset, higher prevalence, and disproportionately high risk of complications for hypertension compared with non-Hispanic Whites and Mexican Americans.1 Surveys identify substantial variation in mean blood pressure (BP) among populations of African origin.2 In high-income countries, including the United States, mean BP and prevalence of hypertension are higher in adults self-described,3–6 observer reported,7,8 or otherwise identified9,10 as being black or having darker skin color.11 However, the relationship between African origin and BP is absent or only minimally apparent in reports from middle-income countries.12–14 Research to clarify reasons for this variability may contribute to understanding of hypertension-related racial disparities in the United States. In US National Health and Nutrition Examination Survey (NHANES) reports, crude and age-adjusted prevalence of hypertension (systolic BP [SBP] ≥140 mm Hg, diastolic BP ≥90 mm Hg, or taking antihypertensive medication) in adults has remained fairly constant at ≈30% since 1999 to 2000.3,4 The corresponding prevalence estimate for African Americans is ≈40% and has also remained reasonably stable. In African Americans, hypertension awareness and treatment rates are higher but control rates lower compared with non-Hispanic Whites (85.7% versus 82.7% for awareness, 77.4% versus 76.7% for treatment, and 49.5% versus 53.9% for control in NHANES 2011–2012).4 The lower prevalence of BP control is present despite use of more BP-lowering medications, including thiazide diuretics.15 This contrasts with clinical trial experience, where differences in BP control rates by race/ethnicity …

Statin Utilization and Recommendations Among HIV and HCV-Infected Veterans: A Cohort Study

BACKGROUND Human immunodeficiency virus (HIV) and hepatitis C virus (HCV) infections are associated with increased risk of cardiovascular disease (CVD). The potential impact of recently updated cholesterol guidelines on treatment of HIV- and HCV-infected veterans is unknown. METHODS We performed a retrospective cohort study to assess statin use and recommendations among 13 579 HIV-infected, 169 767 HCV-infected, and 6628 HIV/HCV-coinfected male veterans aged 40-75 years. Prior 2004 Adult Treatment Panel (ATP-III) guidelines were compared with current 2013 American College of Cardiology/American Heart Association (ACC/AHA) cholesterol guidelines and 2014 US Department of Veterans Affairs (VA)/US Department of Defense (DoD) joint clinical practice guidelines using laboratory, medication, and comorbidity data from the VA Clinical Case Registry from 2008 through 2010. RESULTS Using risk criteria delineated by the ATP-III guidelines, 50.6% of HIV-infected, 45.9% of HCV-infected, and 33.8% of HIV/HCV-coinfected veterans had an indication for statin therapy. However, among those eligible, 22.7%, 30.5%, and 31.5%, respectively, were not receiving ATP-III recommended statin therapy. When current cholesterol guidelines were applied by VA/DoD and ACC/AHA criteria, increases in recommendations for statins were found in all groups (57.3% and 66.1% of HIV-infected, 64.4% and 73.7% of HCV-infected, 49.1% and 58.5% of HIV/HCV-coinfected veterans recommended). CONCLUSIONS Statins were underutilized among veterans infected with HIV, HCV, and HIV/HCV according to previous ATP-III guidelines. Current VA/DoD and ACC/AHA guidelines substantially expand statin recommendations and widen the gap of statin underutilization in all groups. These gaps in care present an opportunity to improve CVD prevention efforts in these at-risk populations.

Assessing Cardiovascular Risk to Guide Hypertension Diagnosis and Treatment

Importance The Systolic Blood Pressure Intervention Trial (SPRINT) demonstrated the benefit of lowering systolic blood pressure (SBP) to 120 mm Hg, yet other trials, such as Heart Outcomes Prevention Evaluation-3 (HOPE-3), did not find consistent benefit. How to incorporate these results into the treatment of those with elevated SBP in the general population is not clear. Objectives To assess the representativeness of SPRINT and HOPE-3 relative to patients in the United States and to explore the cardiovascular disease (CVD) risk profiles of various populations with elevated SBP. Design, Setting, and Participants The study examined data from nonpregnant adults aged 20 to 79 years participating in the 2007-2012 National Health and Nutrition Examination Survey (NHANES) who had complete data available (n = 14 142), representing 206.9 million US adults. The study was performed from October 1, 2015, to August 2, 2016. Main Outcomes and Measures The study estimated the number and characteristics of adults with SBP of 120 mm Hg or higher, including SPRINT and HOPE-3 eligibility, and estimated who may have newly required treatment initiation or intensification if various trial or risk-based criteria were applied. Results NHANES included completed clinical evaluations from mobile examination centers on 15 974 adults aged 20 to 79 years (mean [SD] age, 45.9 [15.5] years). The study excluded 182 pregnant women and 1650 adults in whom CVD risk data were unavailable, leaving a final study population of 14 142 (50.5% women [95% CI, 49.6%-51.3%] and 49.5% men [95% CI, 48.6%-50.4%]). An estimated 53.3 million untreated and 19.8 million treated US adults have an SBP in the diagnostic and treatment gray zone (120-139 mm Hg), a small proportion of whom would have been eligible for SPRINT (5.4% untreated, 13.9% treated) or HOPE-3 (13.9% treated, 1.7% untreated). Even among those with prior CVD or high risk of CVD and elevated SBP (120-139 mm Hg), only a few would have qualified for SPRINT (27.0% and 21.9% of untreated and treated patients, respectively) or HOPE-3 (10.6% and 2.1% of untreated and treated, respectively). If blood pressure treatment recommendations were extended to adults with an SBP between 120 and 139 mm Hg, as well as prior CVD or CVD risk of 15% or higher, then 5.8 million untreated adults would be reclassified as treatment eligible; furthermore, 8.5 million treated patients would require medication intensification. Conclusions and Relevance Millions of US adults have elevated SBP and high CVD risk, most of whom would not have been eligible for SPRINT. Until more definitive evidence becomes available, clinicians should consider a management paradigm based on CVD risk in addition to blood pressure measurements.