acase@tulane.edu

Measles is a serious infectious disease, with a global annual burden of approximately 30 million cases and about a million deaths. The vast majority of this burden falls on Africa. An effective vaccine has been available for over fifty years, but the coverage of measles immunization in Africa lags behind the rest of the world. Due to weaknesses in the World Health Organization's cluster sampling surveys of immunization coverage, determinants are best gleaned from national sample surveys, such as the Demographic and Health Surveys (DHS). Previous studies show that covariates of immunization, exist at all levels in the hierarchy of such data; namely at the level of children, mothers, families and communities. There has been little analysis of such data taking into account this hierarchical structure. Multilevel models allow for the design effect of such studies and produce estimates that are less biased. This study used data from nine sub-Saharan African DHS countries to investigate possible determinants of measles immunization. Models using standard logistic regression were compared to multilevel models. The multilevel models identified the same major determinants of children's measles immunization status, as found in the standard models. However, the multilevel models were more conservative in that they often made the fixed-effects standard errors larger. This sometimes led to reductions in the number of parameters in the fitted models. In particular, the multilevel models tended to reduce the number of variables at the higher levels of the hierarchy of the data, such as the variable identifying which region the child lived in. With respect to the variable of the distance to the nearest immunization facility, using the standard logistic regression model this parameter was significant in four of the five countries where it was measured. However, when the multilevel model was applied to this same data, the variable remained statistically significant in just one of the countries, Mali. Multilevel modeling techniques appear to be better suited than standard regression techniques when questions concerning resource availability or allocation for populations are important. In the application of multilevel models, it is common and perhaps best practice, to first examine data using standard logistic regression, which allows for the assessment of the goodness of fit of the parameters. The subsequent use of multilevel models, give more realistic estimates of the model parameters and their standard errors that take into account the structure of the data / acase@tulane.edu

This study examined the hypothesis that immunity to P. falciparum glycosylphophatidylinositol (GPI) could be the basis of the semi-immune state. The hypothesis is based on the fact that GPI, a candidate for the 'malaria toxin' may be responsible for the paroxysmal release of tumor necrosis factor alpha (TNF-alpha), which is associated with severe malaria. We examined the age-specific sero-prevalence of antibodies to GPI among 233 children and 66 adults of the village of Bancoumana, Mali, in March 2001. We also tested in-vitro the ability of these natural antibodies (from 9 subjects) to block the TNF-alpha release by murine macrophages stimulated with GPI. Antibodies to GPI were uncommon among children less than 5 years of age (&sim;20%) who are at risk of severe disease, but frequent among children &ge;6 years of age and adults who are protected from severe disease (80% and 86%, respectively), p < 0.001. Plasma samples with high antibodies titers to GPI inhibited TNF-alpha release by macrophages in vitro, which is consistent with the hypothesis that antibodies to GPI protect against disease by inhibiting TNF-alpha release. To support our hypothesis, we assessed factors associated with anti-parasite immunity including the prevalence, incidence and recovery rates for asymptomatic P. falciparum infection. Data from a village wide-cohort (2000--2500 children 0--14 years old followed from 1996--2001) show a peak of severe malaria in children 2--5 years of age. In contrast, the incidence of infection (estimated from the reversible catalytic method and by person-month of observation) and the prevalence peak among children 6--9 years old, decrease only after the age of 15 years These findings are consistent with the idea that there are two epidemiologically distinct immune responses to the malaria parasite. Anti-disease (antitoxic) immunity, which is based on antibodies that neutralize the parasite GPI, is acquired within the first 5--6 years of life in malaria-endemic areas. In contrast, anti-parasite immunity develops more slowly in older children and becomes effective only after 15 years of age, resulting in lower parasite densities. Vaccine development should not be focused merely on anti-parasite immunity. It should strive to create the semi-immune state in order to prevent disease and death and produce anti-parasite immunity in order to reduce simultaneously the prevalence and the magnitude of infection / acase@tulane.edu

acase@tulane.edu

The fourth Millennium Development Goal (MDG-4) recommends the reduction of mortality in children under five by two-thirds between 1990 and 2015. A substantial number of deaths occur in the neonatal period. Preterm deliveries and low birth weight are predictors of infant mortality This study aimed to improve understanding of Preterm delivery and Low Birth Weight determinants for effective interventions design. The objectives were: to determine the prevalence of low birth weight and preterm delivery, and to identify the determinants of these outcomes Secondary analysis of the First Breath Trial data were conducted using Wilcox-Russel method to analyze the birth weight distribution, calculate the residual distributions, and estimate of small preterm births. Binary logistic regression was run to determine relevant risk factors From 14551 newborns 11.6% were low birth weight and 3.3% were preterm. The predominant distribution mean was 3153 grs +/- 480 grs and the residual distribution represented 4.3% Young primiparous without any formal education or at least primary level, with multiple birth pregnancy, and who did not attend prenatal care were at higher risk of low birth weight. Female infants also were at greater risk Unmarried young mothers with multiple birth pregnancy who did not attend prenatal care and gave birth at their own house were at higher risk of preterm birth. The study assessed the association between maternal characteristics and two important pregnancy outcomes: low birth weight and preterm delivery. They should be considered as public health issues that need urgent strategies to deal with / acase@tulane.edu

acase@tulane.edu

acase@tulane.edu

Objective. The prevalence of four Diabetes Quality Improvement Project (DQIP) measures was explored in Maine diabetics enrolled in the state Medicaid program or the Maine Health Management Coalition (MHMC) Methods. Diabetics were identified in the billing claims database for Medicaid and the MHMC. The four DQIP measures being studied were glycosylated hemoglobin (HbA1c) test, lipid profiles, dilated eye exam, and nephropathy monitoring. These measures were calculated for three separate analyses: (1) using different denominator entry criteria, (2) trends from 1994--1999, and (3) comparing hypertensive with non-hypertensive diabetics. The chi-square test, multiple logistic regression, and the Mantel-Haenszel chi-square test for trend were used for these analyses Results. The DQIP measure rates varied widely among denominators within each cohort and diabetics in the MHMC cohort had significantly higher rates than those in the Medicaid cohort (p < 0.001). All DQIP measure rates had a positive trend over tire (p < 0.05) in both cohorts with the exception of nephropathy monitoring in the Medicaid cohort, which had a negative trend (p < 0.05). These trends were more prominent in the MHMC cohort. Within the Medicaid cohort, 36 percent of the diabetics also had hypertension. These hypertensive diabetics were significantly more likely to have had the DQIP measures than non-hypertensive diabetics (p < 0.001) Conclusions. The DQIP measures reported here confirm the urgency in identifying effective mechanisms for delivering preventive care and ongoing monitoring that are congruent with defined standards of benefit to all diabetics. The rates of these DQIP measures are improving, but remain less than optimal This research was supported by the Maine Diabetes Control Program through a Centers for Disease Control and Prevention/Division of Diabetes Translation grant (Cooperative Agreement U32-CCU100335), the Maine Health Management Coalition, and the Maine Medicaid Program / acase@tulane.edu

The ability of arthropod vectors to support the development and subsequent transmission of two or more species of viruses or protozoa simultaneously is well-documented in the literature, but comparable observations on filarial parasites are lacking. The purpose of this study was to evaluate the ability of a laboratory mosquito, in this case, a filaria-susceptible strain of Aedes aegypti, to support the development of two filarial parasites, Brugia malayi and Dirofilaria immitis, concomitantly, and to assess the effects of these infections on the mosquito vector. Mosquitoes which fed on blood containing 50 microfilariae per 20 cmm of one or the other species ingested an average of 12 microfilariae of B. malayi or 14 microfilarie of D. immitis. Approximately 70% of microfilariae ingested migrated to their appropriate site of development. Dissections of mosquitoes at 9 to 14 days post-feeding yielded an average of 7 to 8 third-stage larvae per mosquito, indicating that most of the microfilariae which migrated from the midgut completed their development to the infective stage. Upon developing to the third stage, the larvae of D. immitis migrated directly from the Malpighian tubules to the head and mouthparts, whereas B. malayi larvae apparently entered the abdominal hemocoel for 1 to 2 days before completing the migration. Loss of third-stage larvae from the mouthparts of mosquitoes feeding on 10% sucrose solution was negligible. Mosquitoes which ingested blood containing 50 microfilariae of each species per 20 cmm yielded an average of 6.5 larvae of B. malayi and of D. immitis. Similar numbers of larvae were recovered from control mosquitoes which ingested blood with 50 microfilariae of one species per 20 cmm. Furthermore, measurements of larvae from mixed and single infections revealed no significant differences in the size of either species of filaria, suggesting that growth processes were not altered. The deleterious effects of mixed infections on the vector were assessed by comparisons of mean daily mortality among mosquitoes harboring either one or both species of filaria. Mosquitoes harboring both filariae experienced mortality (2.32% per day) approximately equal to the cumulated mortality of the 2 groups harboring B. malayi only (.84% per day) and D. immitis only (1.58% per day). The results of the study indicate that there was no interference with development of 2 species of filaria within the same mosquito vector, and that the mortality of mosquitoes resulting from mixed infection was no higher than would be expected given the additive deleterious effects of infection with each species of filaria / acase@tulane.edu

The economic impact of occupational injuries is a significant burden to employees, industry, and society in general. Despite this burden, epidemiologic analyses of occupational injuries is an area that has only recently received an increasing amount of attention, and few studies have applied epidemiologic methods of analysis to the study of occupational injury costs. The following reports the results of an epidemiologic analysis of factors associated with the direct costs to a single company of medically-treated job related injuries in the offshore oil and gas exploration industry. Injury records maintained by the company were examined and 1726 claims for medically-treated injuries occurring between 1979 and 1981 were selected. Doctors' diagnoses were abstracted from accompanying medical reports and injury severity scores assigned according to the Abbreviated Injury Scale (AIS). Information regarding characteristics of the injured workers, the circumstances of the injury event, and the direct costs to the company were available from a database maintained by the company. The relative odds of an injury incurring costs above the 95th percentile for four categories of cost (compensation, medical treatment, legal costs, settlement awards) and total costs were found to be significantly increased for employees hired less than three months, for injuries incurred in falls, for injuries sustained while moving from one area of the rig to another, for strain and impact related injuries, and for injuries involving the back. While not all of these factors were significantly related to costs in all four categories, for those categories in which they were associated with high costs, odds ratios remained significant after controlling for the effects of injury severity, and number of injuries sustained. Injuries to recently hired employees were approximately twice as likely to incur high legal, settlement, and total costs as were injuries to employees with two or more years with the company. This finding is independent of body part affected, injury severity as measured by the AIS, and number of injuries sustained / acase@tulane.edu