Measles

The study definitions reflected national guidelines for management of uncomplicated malaria, which were valid at the time of the study [11 (link)]. In summary, the guidelines recommend that in high malaria risk areas, all children under five years of age with fever or history of fever should be presumptively treated with AL. In low malaria risk areas, presumptive treatment with AL is recommended in all febrile children in the absence of measles, runny nose and other obvious causes of fever. For clinical management purposes, all parts of Kenya were classified as high malaria risk areas, except the highlands of Central and Nairobi provinces. In patients aged five years and older and regardless of the malaria risk, all febrile patients in the absence of another obvious cause of fever should be tested for malaria (microscopy or RDT) and only test positive patients should be treated with AL. In the same age group, presumptive AL treatment is recommended in absence of malaria diagnostics.
Therefore, to reflect criteria for testing and AL treatment, all analyses were restricted to febrile, non-pregnant patients weighing 5 kg and above, presenting for an initial outpatient visit without being referred or admitted for hospitalization. Patients aged five years and older presenting with another obvious cause of fever, as well as the children less than five years of age meeting the same criteria in low risk areas, were excluded from the analysis. Another obvious cause of fever was defined as febrile patients presenting concomitantly with runny nose, sore throat, oral thrush, wounds, urinary problem, skin problem or abscesses. Fever was defined as axillary temperature of ≥37.5°C or a history of fever during the present illness.
To ensure comparable evaluation of anti-malarial treatment practices based on different age-specific recommendations for malaria diagnosis, the focus of the analysis was observations from health facilities where AL was in stock during the survey, stratified by the availability of diagnostics and patients' age (under five and over five years of age). At facilities with malaria diagnostics, anti-malarial treatment practices were further stratified by the use of malaria diagnostic tests and the patient's test result. Combined microscopy and RDT results are presented since the small number of patients with malaria RDTs precluded a meaningful analysis stratified by type of diagnostics.
Data entry and management was undertaken using Access (Microsoft, USA), through customized data entry screens with in-built range and consistency checks. All forms were entered twice by independent data entry clerks and data files were compared for errors using a verification programme and referring to original questionnaires. All analyses were performed using STATA, version 11. The precision of proportions (95% confidence interval [CI]) was determined adjusting for the cluster sampling at the health facility level.

We examined the association between past measles cases and tetanus antibody levels among children participating in the 2013–2014 DRC DHS. Our sample consisted of 711 children 9–59 months old.

A priori confounders, such as age, sex and geographic location, were included in descriptive and regression analyses. Other potential confounding variables were identified in the literature and based largely on variables representing social determinants of health and found by others to capture health disparities, such as chronic malnutrition (stunting), maternal education, wealth index and birth order.^{16 ,23 ,24} Malaria status (via blood smear) was included due to its association with host immunosuppression,^{25 (link)} and the authors previously found that rural versus urban residence impacted measles vaccination coverage.^{26 (link)} Finally, due to the history of conflict in the DRC, we created a dichotomous variable of children living in provinces experiencing high levels of conflict versus all other provinces. The 5 (old) provinces (Katanga, Maniema, Nord-Kivu, Orientale and Sud-Kivu) with the highest levels of human displacement in 2014 were classified as provinces experiencing high levels of conflict, as most of the human displacement (97%) was conflict-related.^{27 ,28} Due to the challenges inherent with directly measuring a household’s wealth, the DHS assumes that the economic state of a household is related to whether the household has access to certain amenities or services. Such variables can easily be included into a questionnaire or directly observed by the interviewer. Within the DHS, households are placed into quintiles according to household ownership of previously selected assets and via principal components analysis. Asset scores are standardized relative to a standard normal distribution and each household receives a summed score based on the standardized score for each asset. Individuals receive a score reflective of their household score, and the resulting quintiles make up the wealth index for the DHS survey.²⁹ In this study, we dichotomized wealth index to compare the most disadvantaged individuals to all others in the sample, resulting in a wealth index that compares the poorest children to all others, or the lowest quintile children to those in the combined group of second lowest, middle, fourth lowest and highest quintile.
Collinearity diagnostic statistics were performed on all explanatory variables included in the final regression models, with acceptable results (tolerance for all variables was ≥ 0.75, variance inflation factor for all variables was ≤ 1.33) for all variables of interest.