Zambia University Teaching Hospital (UTH) is located in Lusaka, a capital city of Zambia. At the time of this study it provided the only inpatient unit for children with complicated SAM in Lusaka district. As such children receiving service in the unit come from all corners of Lusaka district. The unit has a 59 bed capacity. However, due to the large number of children needing inpatient treatment, year round, the unit has more children than it can accommodate. This is forcing cot sharing.
According to the inpatient unit audit, close to 2,000 children with SAM receive treatment in the inpatient unit annually. This constitutes close to 30% of the children with SAM that annually receive treatment in Lusaka; outpatient and inpatient combined (personal experience).
The mortality rate of SAM children admitted to the inpatient unit is over 30% (ward audit). This is despite efforts since 2001 to reduce mortality in the unit through training of staff in inpatient management of SAM as per the 1999 WHO guideline[17
All children 6-59 months of age admitted to the inpatient unit were eligible for the study. Children were admitted to the ward based on the presence of bilateral pitting edema and/or weight for height Z-scores (WHZ) < -3 standard deviations (SD). Weight for height Z-scores were calculated using NCHS/WHO normalized charts.
Study design and period
This was a cohort study involving children 6-59 months old with SAM admitted to the UTH inpatient unit. The study was conducted from 1st August to 31st December 2009. Part of the study period (October to December) fall within the malnutrition period; December being the peak month for SAM in Lusaka.
Out of a total of 1041 admission that occurred between August and December 2009, 430 children between 6 and 59 months old were enrolled into the study. Children were enrolled into the study up on consent of their caregivers. Children admitted over the weekend were missed as study protocol required enrolling children within 24 hours of admission.
Trained ward attendants measured the nutritional status of the children. Height was measured using a stadiometer, and weight was measured to the nearest 100 g using a UNISCALE. Social and demographic data were collected using structured questionnaires. HIV serology was done using the Determine® HIV-1/2 test. DNA PCR (for children under 18 months old with a positive HIV serology) was done after parental consent was obtained. MUAC was not measured as it was not part of the inpatient protocol.
On admission, all children were examined by the attending physicians. Clinical evaluation was done to assess co-morbidities. Fever was defined as an admission axilliary temperature of greater than 37.5°C. Diarrhea was diagnosed based on caregiver assessment or three or more loose stools a day.
Clinical and nutritional care
Children were managed by a team of physicians comprising of three rotating resident physicians (average stay in the ward of 4 months) and two junior resident medical officers, supervised by one senior registrar and one consultant pediatrician. In addition, three to five nurses attended to the children in the ward.
Children were managed using WHO standard guidelines for the management of severe malnutrition. Oral vitamin A (200,000 IU if ≥ 12 months old or 100,000IU if < 1 year old) was given on admission; those with clinical signs of vitamin A deficiency received further doses on days two and 14. Children with diarrhea were given ReSoMal. A nasogastric tube was inserted into children who were assessed to be too sick to feed voluntarily or who had persistent vomiting. Children received 10% dextrose upon admission. Intravenous fluids (often 1/2 strength Darrow's solution) were used for management of shock or in children with persistent diarrhea with dehydration.
F75 therapeutic milk was used in the first phase of treatment. F75 prepared in the ward using fermented milk was given to children who continued to have diarrhea after admission. During the second phase of treatment, children were treated either with ready-to-use therapeutic food (RUTF) or F100 therapeutic milk depending on appetite test result.
Children exited from the unit on one of the following criterion; "Stabilized" if they were able to consume RUTF and were referred to one of the 25 outpatient therapeutic programs (OTP) for full recovery; "Absconders" if they were absent from the unit for two consecutive days; "Deaths" if they died while in the unit; "Transfer to AO5" if the child had tuberculosis or measles and was referred to the isolation ward.
Outpatient treatment service
At the time of this study, outpatient service for the management of children with uncomplicated SAM was available in 25 health centre in Lusaka. Children admitted to the inpatient unit were discharged into these centers upon stabilization of their condition and were able to consume RUTF.
Variables in the dataset included binary (sex, HIV, fever, WHZ score < -3SD, diarrhea, and outcome) and categorical data (nutritional status, and admission edema). Weight, height, and age were numeric data but were grouped as categorical data for purposes of analysis. During the analysis, a variable called "nutstat" was created based on a combination of children's admission edema and WHZ. Accordingly, children were classified as "Marasmic" if they had WHZ less than -3 SD but not edema, or "Kwashiorkor" if they had edema but their WHZ was ≥-3 SD, or "Marasmic-Kwashiorkor" if they had both edema and WHZ < -3 SD.
Binary outcome variable (Alive or Dead) was created. Exposure factors used included age, sex, HIV status, nutritional status, diarrhea on admission, and fever. Baseline data were compared between the two groups using mean with Standard Deviation (SD) or percentage. T-test and chi-square test was used to compare difference in mean and percentage, respectively. Variables that had a P-value of < 0.2 were modeled using logistic regression. Univariate and multivariate analysis were done by adjusting for sex, HIV, WHZ score, nutritional status, and age group. Likelihood ratio test and associated P-values were used to test association. Kaplan-Meier curves were used to estimate survival probability. Adjusted and unadjusted odds ratio, 95% confidence interval, and P-values were calculated and reported. Analysis was done using STATA 11.
Permission to conduct the study was provided by UTH, and ethical clearance was granted by University of Zambia Biomedical Research Ethics Committee. As part of the provider initiated counseling and HIV testing service offered by the hospital to all admitted patients, HIV counseling and testing was done by trained health personnel up on consent of the caretakers of children.