Faculty Bibliography

Suicide is a leading cause of premature mortality. Aspects of the social environment such as incidents of violence in the community may induce psychological distress and affect suicidality, but these determinants are not well understood. We conducted an ecological study using California vital statistics records, geocoded to address of the decedent, to examine whether proximity to homicide was associated with the occurrence of suicide in urban census tracts. For each urban tract (N = 7194) and each month in 2012, we assessed homicides in the tract or within buffer zones around the tract with a 1-month lag. We estimated two risk difference parameters that capture how suicide risk is related to differences in homicide exposure. Proximity to homicides was negatively associated with suicide occurrence after controlling for demographic factors, seasonality, and other confounders. Estimates suggest that the absence of homicides would be associated with a 4.2 % higher number of tract-months with one or more suicides (95 % confidence interval 2.2-6.0). This relationship was stronger in tracts that were wealthier, older, and less civically engaged. Results were robust to a wide variety of sensitivity tests. Contrary to expectations, we identified a consistent negative association of proximity to homicide with suicide occurrence. It may be that a homicide deters or distracts from suicidality or that aggression or hopelessness may be expressed as inward or outward directed violence in different settings. Further investigation is needed to identify the drivers of this association.

BACKGROUND:Studies have shown that children with cancer globally lack access to palliative care. Little is known regarding physicians' perceptions of palliative care, treatment access, and self-reported competence in providing palliative care. PROCEDURE/METHODS:Members of the Global Neuroblastoma Network (online tumor board) were surveyed. Eighty-three respondents met inclusion criteria; 53 (64%) completed the survey. RESULTS:Most respondents trained in high-income countries (HIC) but practice in low- and middle-income countries (LMIC), and care for more than five patients with neuroblastoma annually. WHO Essential Medicines in palliative care varied in availability, with incomplete access across LMIC centers. Nonpharmacologic therapies were inconsistently available. Contrary to international definitions, 17% of respondents inappropriately considered palliative care as that initiated only after curative therapy is stopped. Mean physician competence composite score (Likert scale 1-5, 5 = very competent) in providing symptomatic relief and palliative care across phases of care was 2.93 (95% CI 2.71-3.22). Physicians reported significantly greater competence in symptom management during cure-directed therapy than during end-of-life (P = 0.02) or when patients are actively dying (P = 0.007). Practicing in HIC, prior palliative care training, having access to radiotherapy, and not having to turn patients away due to bed shortages were significantly predictive of perceived competence in providing palliative care at end of life. CONCLUSIONS:An international sample identified gaps in treatment and palliative care service availability, in understanding the definition of palliative care, and in self-reported competence in providing palliative care. Increased perceived competence was associated with training, which supports the need for increased palliative care education and advocacy, especially in LMIC.

Matching methods are common in studies across many disciplines. However, there is limited evidence on how to optimally combine matching with subsequent analysis approaches to minimize bias and maximize efficiency for the quantity of interest. We conducted simulations to compare the performance of a wide variety of matching methods and analysis approaches in terms of bias, variance, and mean squared error (MSE). We then compared these approaches in an applied example of an employment training program. The results indicate that combining full matching with double robust analysis performed best in both the simulations and the applied example, particularly when combined with machine learning estimation methods. To reduce bias, current guidelines advise researchers to select the technique with the best post-matching covariate balance, but this work finds that such an approach does not always minimize mean squared error (MSE). These findings have important implications for future research utilizing matching. To minimize MSE, investigators should consider additional diagnostics, and use of simulations tailored to the study of interest to identify the optimal matching and analysis combination.

Seroepidemiological monitoring of population immunity to vaccine-preventable diseases is critical to prevent future outbreaks. Dried blood spots (DBS), drops of capillary blood dried on filter paper, are an affordable, minimally invasive alternative to venipuncture for collecting blood in field settings. However, few proven methods exist to analyze DBS for the presence of protective antibodies. This study validates a novel technique for measuring measles-specific immunoglobulin G (IgG) in capillary DBS using a commercial ELISA. The predictive performance of a new method for analyzing DBS was tested by comparing matched serum and DBS samples from 50 children. The accuracy, precision, and reliability of the procedure were evaluated, and the optimal cut points to classify positive and negative samples were determined. The method was then applied to 1,588 DBS collected during a large survey of children in Mexico and Nicaragua. Measles-specific IgG in serum samples were 62% negative, 10% equivocal, and 28% positive. In comparisons with matched serum, DBS results were 100% sensitive and 96 · 8% specific, and agreed in 46 of 50 (92%) cases. The inter-assay and intra-assay coefficients of variation from kit-provided controls were greater than desired (24.8% and 8.4%, respectively). However, in predictive simulations the average misclassification was only 3.9%. Procedures were found to be acceptable to surveyors and participants. Analyzing DBS collected in low-resources settings is a feasible and accurate means of measuring population immunity to measles and should be used to generate objective measures of health status and health system performance.

BACKGROUND:Individual income and poverty are associated with poor health outcomes. The poor face unique challenges related to access, education, financial capacity, environmental effects, and other factors that threaten their health outcomes. METHODS:We examined the variation in the health outcomes and health behaviors among the poorest quintile in eight countries of Mesoamerica using data from the Salud Mesomérica 2015 baseline household surveys. We used multivariable logistic regression to measure the association between delivering a child in a health facility and select household and maternal characteristics, including education and measures of wealth. RESULTS:Health indicators varied greatly between geographic segments. Controlling for other demographic characteristics, women with at least secondary education were more likely to have an in-facility delivery compared to women who had not attended school (OR: 3.20, 95 % confidence interval [CI]: 2.56-3.99, respectively). Similarly, women from households with the highest expenditure were more likely to deliver in a health facility compared to those from the lowest expenditure households (OR 3.06, 95 % CI: 2.43-3.85). Household assets did not impact these associations. Moreover, we found that commonly-used definitions of poverty do not align with the disparities in health outcomes observed in these communities. CONCLUSIONS:Although poverty measured by expenditure or wealth is associated with health disparities or health outcomes, a composite indicator of health poverty based on coverage is more likely to focus attention on health problems and solutions. Our findings call for the public health community to define poverty by health coverage measures rather than income or wealth. Such a health-poverty metric is more likely to generate attention and mobilize targeted action by the health communities than our current definition of poverty.

BACKGROUND:Achieving universal health coverage and reducing health inequalities are primary goals for an increasing number of health systems worldwide. Timely and accurate measurements of levels and trends in key health indicators at local levels are crucial to assess progress and identify drivers of success and areas that may be lagging behind. METHODS:We generated estimates of 17 key maternal and child health indicators for Zambia's 72 districts from 1990 to 2010 using surveys, censuses, and administrative data. We used a three-step statistical model involving spatial-temporal smoothing and Gaussian process regression. We generated estimates at the national level for each indicator by calculating the population-weighted mean of the district values and calculated composite coverage as the average of 10 priority interventions. RESULTS:National estimates masked substantial variation across districts in the levels and trends of all indicators. Overall, composite coverage increased from 46% in 1990 to 73% in 2010, and most of this gain was attributable to the scale-up of malaria control interventions, pentavalent immunization, and exclusive breastfeeding. The scale-up of these interventions was relatively equitable across districts. In contrast, progress in routine services, including polio immunization, antenatal care, and skilled birth attendance, stagnated or declined and exhibited large disparities across districts. The absolute difference in composite coverage between the highest-performing and lowest-performing districts declined from 37 to 26 percentage points between 1990 and 2010, although considerable variation in composite coverage across districts persisted. CONCLUSIONS:Zambia has made marked progress in delivering maternal and child health interventions between 1990 and 2010; nevertheless, substantial variations across districts and interventions remained. Subnational benchmarking is important to identify these disparities, allowing policymakers to prioritize areas of greatest need. Analyses such as this one should be conducted regularly and feed directly into policy decisions in order to increase accountability at the local, regional, and national levels.

BACKGROUND:Recent outbreaks of measles in the Americas have received news and popular attention, noting the importance of vaccination to population health. To estimate the potential increase in immunization coverage and reduction in days at risk if every opportunity to vaccinate a child was used, we analyzed vaccination histories of children 11-59 months of age from large household surveys in Mesoamerica. METHODS:Our study included 22,234 children aged less than 59 months in El Salvador, Guatemala, Honduras, Mexico, Nicaragua, and Panama. Child vaccination cards were used to calculate coverage of measles, mumps, and rubella (MMR) and to compute the number of days lived at risk. A child had a missed opportunity for vaccination if their card indicated a visit for vaccinations at which the child was not caught up to schedule for MMR. A Cox proportional hazards model was used to compute the hazard ratio associated with the reduction in days at risk, accounting for missed opportunities. RESULTS:El Salvador had the highest proportion of children with a vaccine card (91.2%) while Nicaragua had the lowest (76.5%). Card MMR coverage ranged from 44.6% in Mexico to 79.6% in Honduras while potential coverage accounting for missed opportunities ranged from 70.8% in Nicaragua to 96.4% in El Salvador. Younger children were less likely to have a missed opportunity. In Panama, children from households with higher expenditure were more likely to have a missed opportunity for MMR vaccination compared to the poorest (OR 1.62, 95% CI: 1.06-2.47). In Nicaragua, compared to children of mothers with no education, children of mothers with primary education and secondary education were less likely to have a missed opportunity (OR 0.46, 95% CI: 0.24-0.88 and OR 0.25, 95% CI: 0.096-0.65, respectively). Mean days at risk for MMR ranged from 158 in Panama to 483 in Mexico while potential days at risk ranged from 92 in Panama to 239 in El Salvador. CONCLUSIONS:Our study found high levels of missed opportunities for immunizing children in Mesoamerica. Our findings cause great concern, as they indicate that families are bringing their children to health facilities, but these children are not receiving all appropriate vaccinations during visits. This points to serious problems in current immunization practices and protocols in poor areas in Mesoamerica. Our study calls for programs to ensure that vaccines are available and that health professionals use every opportunity to vaccinate a child.

Timely and accurate measurement of population protection against measles is critical for decision-making and prevention of outbreaks. However, little is known about how survey-based estimates of immunization (crude coverage) compare to the seroprevalence of antibodies (effective coverage), particularly in low-resource settings. In poor areas of Mexico and Nicaragua, we used household surveys to gather information on measles immunization from child health cards and caregiver recall. We also collected dried blood spots (DBS) from children aged 12 to 23 months to compare crude and effective coverage of measles immunization. We used survey-weighted logistic regression to identify individual, maternal, household, community, and health facility characteristics that predict gaps between crude coverage and effective coverage. We found that crude coverage was significantly higher than effective coverage (83% versus 68% in Mexico; 85% versus 50% in Nicaragua). A large proportion of children (19% in Mexico; 43% in Nicaragua) had health card documentation of measles immunization but lacked antibodies. These discrepancies varied from 0% to 100% across municipalities in each country. In multivariate analyses, card-positive children in Mexico were more likely to lack antibodies if they resided in urban areas or the jurisdiction of De Los Llanos. In contrast, card-positive children in Nicaragua were more likely to lack antibodies if they resided in rural areas or the North Atlantic region, had low weight-for-age, or attended health facilities with a greater number of refrigerators. Findings highlight that reliance on child health cards to measure population protection against measles is unwise. We call for the evaluation of immunization programs using serological methods, especially in poor areas where the cold chain is likely to be compromised. Identification of within-country variation in effective coverage of measles immunization will allow researchers and public health professionals to address challenges in current immunization programs.

BACKGROUND:Health has improved markedly in Mesoamerica, the region consisting of southern Mexico and Central America, over the past decade. Despite this progress, there remain substantial inequalities in health outcomes, access, and quality of medical care between and within countries. Poor, indigenous, and rural populations have considerably worse health indicators than national or regional averages. In an effort to address these health inequalities, the Salud Mesoamérica 2015 Initiative (SM2015), a results-based financing initiative, was established. METHODS:For each of the eight participating countries, health targets were set to measure the progress of improvements in maternal and child health produced by the Initiative. To establish a baseline, we conducted censuses of 90,000 households, completed 20,225 household interviews, and surveyed 479 health facilities in the poorest areas of Mesoamerica. Pairing health facility and household surveys allows us to link barriers to care and health outcomes with health system infrastructure components and quality of health services. RESULTS:Indicators varied significantly within and between countries. Anemia was most prevalent in Panama and least prevalent in Honduras. Anemia varied by age, with the highest levels observed among children aged 0 to 11 months in all settings. Belize had the highest proportion of institutional deliveries (99%), while Guatemala had the lowest (24%). The proportion of women with four antenatal care visits with a skilled attendant was highest in El Salvador (90%) and the lowest in Guatemala (20%). Availability of contraceptives also varied. The availability of condoms ranged from 83% in Nicaragua to 97% in Honduras. Oral contraceptive pills and injectable contraceptives were available in just 75% of facilities in Panama. IUDs were observed in only 21.5% of facilities surveyed in El Salvador. CONCLUSIONS:These data provide a baseline of much-needed information for evidence-based action on health throughout Mesoamerica. Our baseline estimates reflect large disparities in health indicators within and between countries and will facilitate the evaluation of interventions and investments deployed in the region over the next three to five years. SM2015's innovative monitoring and evaluation framework will allow health officials with limited resources to identify and target areas of greatest need.