Sensitivity analysis for clinical trials with missing continuous outcome data using controlled multiple imputation: A practical guide

Missing data due to loss to follow-up or intercurrent events are unintended, but unfortunately inevitable in clinical trials. Since the true values of missing data are never known, it is necessary to assess the impact of untestable and unavoidable assumptions about any unobserved data in sensitivity analysis. This tutorial provides an overview of controlled multiple imputation (MI) techniques and a practical guide to their use for sensitivity analysis of trials with missing continuous outcome data. These include δ- and reference-based MI procedures. In δ-based imputation, an offset term, δ, is typically added to the expected value of the missing data to assess the impact of unobserved participants having a worse or better response than those observed. Reference-based imputation draws imputed values with some reference to observed data in other groups of the trial, typically in other treatment arms. We illustrate the accessibility of these methods using data from a pediatric eczema trial and a chronic headache trial and provide Stata code to facilitate adoption. We discuss issues surrounding the choice of δ in δ-based sensitivity analysis. We also review the debate on variance estimation within reference-based analysis and justify the use of Rubin's variance estimator in this setting, since as we further elaborate on within, it provides information anchored inference.     

Investing in antibiotics to alleviate future catastrophic outcomes: What is the value of having an effective antibiotic to mitigate pandemic influenza?

Over 95% of post‐mortem samples from the 1918 pandemic, which caused 50 to 100 million deaths, showed bacterial infection complications. The introduction of antibiotics in the 1940s has since reduced the risk of bacterial infections, but growing resistance to antibiotics could increase the toll from future influenza pandemics if secondary bacterial infections are as serious as in 1918, or even if they are less severe. We develop a valuation model of the option to withhold wide use of an antibiotic until significant outbreaks such as pandemic influenza or foodborne diseases are identified. Using real options theory, we derive conditions under which withholding wide use is beneficial, and calculate the option value for influenza pandemic scenarios that lead to secondary infections with a resistant Staphylococcus aureus strain. We find that the value of withholding an effective novel oral antibiotic can be positive and significant unless the pandemic is mild and causes few secondary infections with the resistant strain or if most patients can be treated intravenously. Although the option value is sensitive to parameter uncertainty, our results suggest that further analysis on a case‐by‐case basis could guide investment in novel agents as well as strategies on how to use them.     

School Policies and Practices That Improve Indoor Air Quality*

BACKGROUND: To determine whether schools with a formal indoor air quality management program were more likely than schools without a formal program to have policies and practices that promote superior indoor air quality. METHODS: This study analyzed school-level data from the 2006 School Health Policies and Programs Study, a national study of school health programs and policies at the state, district, and school levels. Using chi-square analyses, the rates of policies and practices that promote indoor air quality were compared between schools with and schools without a formal indoor air quality program. RESULTS: The findings of this study show that 51.4% of schools had a formal indoor air quality management program, and that those schools were significantly more likely than were schools without a program to have policies and use strategies to promote superior indoor air quality. CONCLUSIONS: These findings suggest that schools with a formal indoor air quality program are more likely support policies and engage in practices that promote superior indoor air quality.     

Health Policy,Ethics, and the Kansas Legislative Health Academy

We describe a unique program, the Kansas Legislative Health Academy, that brings together state legislators from across the political spectrum to build their capacity in advancing policies that can improve the health of Kansans.To that end, the academy helps legislators develop new skills to deliberate the ethics of health policy, use systems thinking to understand the long- and short-term effects of policy action and inaction, and engage in acts of civic leadership. The academy also seeks to foster an environment of respectful open dialogue and to build new cross-chamber and cross-party relationships.Among the most important outcomes cited by program participants is the value of sustained, personal interaction and problem solving with individuals holding differing political views.Health policy often elicits controversy. Recent examples include the uproar over recommendations for mammography screening for women 40 to 49 years old and human papillomavirus vaccination for adolescent girls and boys. Perhaps the most dramatic controversy relates to the passage of the Patient Protection and Affordable Care Act (Pub L No. 111-148); despite being signed into law in 2010 and found substantially constitutional by the US Supreme Court in 2012 (Medicaid expansion was made optional for states), this legislation remains subject to vigorous dissent. Such controversies, although no doubt a function of interest-driven politics, also reflect deep differences in ethical values.Ethical values and premises underpin all public policy.1,2 Ideas about individual liberty, personal responsibility, solidarity, justice, and the role of the government are just a few of the moral constructs that often clash in the making of policy. Policy analysis often ignores these dimensions of policy-making, although that is beginning to change.3–5Here we describe a project based in part on the premise that training policymakers to recognize and talk openly about the ethical values entailed in health policy might improve its content and process. This project, the Kansas Legislative Health Academy (hereafter Health Academy), brought together state legislators from across the political spectrum to build their capacity to respond to complex health policy challenges in Kansas. To that end, the curriculum sought to help legislators develop new skills in 3 areas: health policy ethics, systems thinking, and civic leadership. The Health Academy also sought to foster an environment of open, respectful dialogue and to build new cross-chamber and cross-party relationships.To our knowledge, the Health Academy is a unique program. Many educational programs exist for legislators to focus on leadership development or specific health policy issues, but none we are aware of are specifically designed to cover a broad range of health policy issues while also addressing underlying barriers to effective policy-making within legislative bodies. In what follows, we describe the Health Academy’s origins, structure, substance, and lessons learned.     

Association between Lifetime Exposure to Inorganic Arsenic in Drinking Water and Coronary Heart Disease in Colorado Residents

Background: Chronic diseases, including coronary heart disease (CHD), have been associated with ingestion of drinking water with high levels of inorganic arsenic (> 1,000 μg/L). However, associations have been inconclusive in populations with lower levels (< 100 μg/L) of inorganic arsenic exposure.Objectives: We conducted a case-cohort study based on individual estimates of lifetime arsenic exposure to examine the relationship between chronic low-level arsenic exposure and risk of CHD.Methods: This study included 555 participants with 96 CHD events diagnosed between 1984 and 1998 for which individual lifetime arsenic exposure estimates were determined using data from structured interviews and secondary data sources to determine lifetime residence, which was linked to a geospatial model of arsenic concentrations in drinking water. These lifetime arsenic exposure estimates were correlated with historically collected urinary arsenic concentrations. A Cox proportional-hazards model with time-dependent CHD risk factors was used to assess the association between time-weighted average (TWA) lifetime exposure to low-level inorganic arsenic in drinking water and incident CHD.Results: We estimated a positive association between low-level inorganic arsenic exposure and CHD risk [hazard ratio (HR): = 1.38, 95% CI: 1.09, 1.78] per 15 μg/L while adjusting for age, sex, first-degree family history of CHD, and serum low-density lipoprotein levels. The risk of CHD increased monotonically with increasing TWAs for inorganic arsenic exposure in water relative to < 20 μg/L (HR = 1.2, 95% CI: 0.6, 2.2 for 20–30 μg/L; HR = 2.2; 95% CI: 1.2, 4.0 for 30–45 μg/L; and HR = 3, 95% CI: 1.1, 9.1 for 45–88 μg/L).Conclusions: Lifetime exposure to low-level inorganic arsenic in drinking water was associated with increased risk for CHD in this population.Citation: James KA, Byers T, Hokanson JE, Meliker JR, Zerbe GO, Marshall JA. 2015. Association between lifetime exposure to inorganic arsenic in drinking water and coronary heart disease in Colorado residents. Environ Health Perspect 123:128–134; http://dx.doi.org/10.1289/ehp.1307839     

Test–Retest Reliability of Capability Measurement in the UK General Population

Although philosophically attractive, it may be difficult, in practice, to measure individuals' capabilities (what they are able to do in their lives) as opposed to their functionings (what they actually do). To examine whether capability information could be reliably self‐reported, we administered a measure of self‐reported capability (the Investigating Choice Experiments Capability Measure for Adults, ICECAP‐A) on two occasions, 2 weeks apart, alongside a self‐reported health measure (the EuroQol Five Dimensional Questionnaire with 3 levels, EQ‐5D‐3L). We found that respondents were able to report capabilities with a moderate level of consistency, although somewhat less reliably than their health status. The more socially orientated nature of some of the capability questions may account for the difference. © 2014 The Authors Health Economics Published by John Wiley & Sons Ltd.     

Lessons learned from engaging men in sexual and reproductive health as clients,partners and advocates of change in the Hoima district of Uganda

This study examined the impact of a three-year intervention project conducted in the Hoima district of Uganda, which sought to engage men in sexual and reproductive health as clients, equal partners and advocates of change. Structured surveys with 164 self-reported heterosexual men aged 18–54 years were used to assess knowledge and attitudes towards sexual and reproductive health. Data from these were analysed using Stata and SPSS. Additionally, five focus groups were conducted with the female partners and male beneficiaries of the project and with project peer educators. Four interviews were conducted with project staff and male beneficiaries. Data from these and the focus groups were analysed using a thematic approach. Following the intervention, a significantly greater number of men accessed, and supported their partners in accessing sexual health services services, had gained sexual and reproductive health awareness, reported sharing domestic duties and contraceptive decision-making, and displayed a decreased tolerance for domestic violence. It was more difficult to assess men's involvement and behaviours as advocates of change, which sheds light on the complexities of a gender transformative project and the importance of evaluating such projects from both men's and their partners' perspectives and at different levels of the male involvement model in sexual and reproductive health.     

Model averaging for robust extrapolation in evidence synthesis

Extrapolation from a source to a target, eg, from adults to children, is a promising approach to utilize external information when data are sparse. In the context of meta-analyses, one is commonly faced with a small number of studies, whereas potentially relevant additional information may also be available. Here, we describe a simple extrapolation strategy using heavy-tailed mixture priors for effect estimation in meta-analysis, which effectively results in a model-averaging technique. The described method is robust in the sense that a potential prior-data conflict, ie, a discrepancy between source and target data, is explicitly anticipated. The aim of this paper is to develop a solution for this particular application to showcase the ease of implementation by providing R  code, and to demonstrate the robustness of the general approach in simulations.     

Sample size re-estimation for clinical trials with longitudinal negative binomial counts including time trends

In some diseases, such as multiple sclerosis, lesion counts obtained from magnetic resonance imaging (MRI) are used as markers of disease progression. This leads to longitudinal, and typically overdispersed, count data outcomes in clinical trials. Models for such data invariably include a number of nuisance parameters, which can be difficult to specify at the planning stage, leading to considerable uncertainty in sample size specification. Consequently, blinded sample size re-estimation procedures are used, allowing for an adjustment of the sample size within an ongoing trial by estimating relevant nuisance parameters at an interim point, without compromising trial integrity. To date, the methods available for re-estimation have required an assumption that the mean count is time-constant within patients. We propose a new modeling approach that maintains the advantages of established procedures but allows for general underlying and treatment-specific time trends in the mean response. A simulation study is conducted to assess the effectiveness of blinded sample size re-estimation methods over fixed designs. Sample sizes attained through blinded sample size re-estimation procedures are shown to maintain the desired study power without inflating the Type I error rate and the procedure is demonstrated on MRI data from a recent study in multiple sclerosis.