Implementing tuberculosis patient cost surveys in resource-constrained settings: lessons from Tanzania.

Tuberculosis (TB) disproportionally affects persons and families who are economically and socially disadvantaged. Therefore, a patient cost survey was conducted in Tanzania to evaluate the costs incurred by patients and their households before and after the diagnosis of TB. It was the first survey in Tanzania to ascertain baseline information and experience for subsequent surveys. This paper aims to share the experience encountered during the survey to ensure a standardized approach and elimination of potential barriers for the implementation of future surveys. A total of 777 TB patients from 30 clusters selected based on probability proportional to the size were interviewed during the study period. As the sample size was calculated based on notification data from the previous year, some health facilities experienced an inadequate number of TB patients during the study to meet the allocated cluster size for the survey. Most facilities had poor recording and recordkeeping in TB registers where deaths were not registered, and some patients had not been assigned district identification numbers. Fixed days for TB drug refills in health facilities affected the routine implementation of the survey as the interviews were conducted when patients visited the facility to pick up the drugs. Tablets used to collect data failed to capture the geographic location in some areas. The households of TB patients lost to follow-up and those who had died during TB treatment were not included in the survey. When planning and preparing for patient costs surveys, it is important to consider unforeseen factors which may affect planned activities and findings. During the survey in Tanzania, the identified challenges included survey logistics, communications, patient enrollment, and data management issues. To improve the quality of the findings of future surveys, it may be reasonable to revise survey procedures to include households of TB patients who were lost to follow-up and those who died during TB treatment; the households of such patients may have incurred higher direct and indirect costs than households whose patient was cured as a result of receiving TB treatment.

Economic burden of tuberculosis in Tanzania: a national survey of costs faced by tuberculosis-affected households.

BACKGROUND: Although tuberculosis (TB) care is free in Tanzania, TB-associated costs may compromise access to services and treatment adherence resulting in poor outcomes and increased risk of transmission in the community. TB can impact economically patients and their households. We assessed the economic burden of TB on patients and their households in Tanzania and identified cost drivers to inform policies and programs for potential interventions to mitigate costs. METHODS: We conducted a nationally representative cross-sectional survey using a standard methodology recommended by World Health Organization. TB patients of all ages and with all types of TB from 30 clusters across Tanzania were interviewed during July - September 2019. We used the human capital approach to assess the indirect costs and a threshold of 20% of the household annual expenditure to determine the proportion of TB-affected households experiencing catastrophic cost. We descriptively analyzed the cost data and fitted multivariable logistic regression models to identify potential predictors of catastrophic costs. RESULTS: Of the 777 TB-affected households, 44.9% faced catastrophic costs due to TB. This proportion was higher (80.0%) among households of patients with multi-drug resistant TB (MDR-TB). Overall, cost was driven by income loss while accessing TB services (33.7%), nutritional supplements (32.6%), and medical costs (15.1%). Most income loss was associated with hospitalization and time for picking up TB drugs. Most TB patients (85.9%) reported worsening financial situations due to TB, and over fifty percent (53.0%) borrowed money or sold assets to finance TB treatment. In multivariable analysis, the factors associated with catastrophic costs included hospitalization (adjusted odds ratio [aOR] = 34.9; 95% confidence interval (CI):12.5-146.17), living in semi-urban (aOR = 1.6; 95% CI:1.0-2.5) or rural areas (aOR = 2.6; 95% CI:1.8-3.7), having MDR-TB (aOR = 3.4; 95% CI:1.2-10.9), and facility-based directly-observed treatment (DOT) (aOR = 7.2; 95% CI:2.4-26.6). CONCLUSION: We found that the cost of TB care is catastrophic for almost half of the TB-affected households in Tanzania; our findings support the results from other surveys recently conducted in sub-Saharan Africa. Collaborative efforts across health, employment and social welfare sectors are imperative to minimize household costs due to TB disease and improve access to care, patient adherence and outcomes.

Adolescents and age of consent to HIV testing: an updated review of national policies in sub-Saharan Africa.

OBJECTIVES: In sub-Saharan Africa (SSA) where HIV burden is highest, access to testing, a key entry point for prevention and treatment, remains low for adolescents (aged 10-19). Access may be hampered by policies requiring parental consent for adolescents to receive HIV testing services (HTS). In 2013, the WHO recommended countries to review HTS age of consent policies. Here, we investigate country progress and policies on age of consent for HIV testing. DESIGN: Comprehensive policy review. DATA SOURCES: Policies addressing HTS were obtained through searching WHO repositories and governmental and non-governmental websites and consulting country and regional experts. ELIGIBILITY CRITERIA: HTS policies published by SSA governments before 2019 that included age of consent. DATA EXTRACTION AND SYNTHESIS: Data were extracted on HTS age of consent including exceptions based on risk and maturity. Descriptive analyses of included policies were disaggregated by Eastern and Southern Africa (ESA) and Western and Central Africa (WCA) subregions. RESULTS: Thirty-nine policies were reviewed, 38 were eligible; 19/38 (50%) permitted HTS for adolescents ≤16 years old without parental consent. Of these, six allowed HTS at ≥12 years old, two at ≥13, two at ≥14, five at ≥15 and four at ≥16. In ESA, 71% (n=15/21) allowed those of ≤16 years old to access HTS, while only 24% (n=6/25) of WCA countries allowed the same. Maturity exceptions including marriage, sexual activity, pregnancy or key population were identified in 18 policies. In 2019, 63% (n=19/30) of policies with clear age-based criteria allowed adolescents of 12-16 years old to access HIV testing without parental consent, an increase from 37% (n=14/38) in 2013. CONCLUSIONS: While many countries in SSA have revised their HTS policies, many do not specify age of consent. Revision of SSA consent to HTS policies, particularly in WCA, remains a priority to achieve the 2025 goal of 95% of people with HIV knowing their status.

Intussusception among infants in Tanzania: findings from prospective hospital-based surveillance, 2013-2016.

INTRODUCTION: intussusception surveillance was initiated in Tanzania in 2013 after monovalent rotavirus vaccine was introduced, as part of the 7-country African evaluation to assess whether the vaccine was associated with an increased risk of intussusception. An increased risk from vaccine was not identified. Published data on intussusception in Tanzanian infants are limited. METHODS: prospective intussusception surveillance was conducted at 7 referral hospitals during 2013-2016 to identify all infants with intussusception meeting Brighton Level 1 criteria. Demographic, household and clinical data were collected by hospital clinicians and analyzed. RESULTS: a total of 207 intussusception cases were identified. The median age of cases was 5.8 months and nearly three-quarters were aged 4-7 months. Median number of days from symptom onset to admission at treatment hospital was 3 (IQR 2-5). Seventy-eight percent (152/195) of cases had been admitted at another hospital before transfer to the treating hospital. Enema reduction was not available; all infants were treated surgically and 55% (114/207) had intestinal resection. The overall case-fatality rate was 30% (62/206). Compared with infants who survived, those who died had longer duration of symptoms before admission to treatment hospital (median 4 vs 3 days; p < 0.01), higher rate of intestinal resection (81% [60/82] vs 44% [64/144], p < 0.001), and from families with lower incomes (i.e., less likely to own a television [p < 0.01] and refrigerator [p < 0.05). CONCLUSION: Tanzanian infants who develop intussusception have a high case-fatality rate. Raising the index of suspicion among healthcare providers, allocating resources to allow wider availability of abdominal ultrasound for earlier diagnosis, and training teams in ultrasound-guided enema reduction techniques used in other African countries could reduce the fatality rate.

Risk Factors Associated With Increased Mortality From Intussusception in African Infants.

OBJECTIVES: Morbidity and mortality from intussusception, the leading cause of bowel obstruction in infants, is higher in Africa than in other regions of the world, but the reasons have not been well examined. We sought to identify risk and protective factors associated with death or intestinal resection following intussusception. METHODS: Infants with intussusception from 7 sub-Saharan African countries (Ethiopia, Ghana, Kenya, Malawi, Tanzania, Zambia, and Zimbabwe) were enrolled through active, hospital-based surveillance from February 2012 to December 2016. We examined demographic, clinical, and socioeconomic factors associated with death or intestinal resection following intussusception, using multivariable logistic regression. RESULTS: A total of 1017 infants <1 year of age with intussusception were enrolled. Overall, 13% of children (133/1017) died during the hospitalization, and 48% (467/966) required intestinal resection. In multivariable analyses, female sex [odds ratio (OR) 1.8, 95% confidence interval (CI) 1.2-3.3], longer duration of symptoms before presentation (OR 1.1; 95% CI 1.0-1.2), and undergoing intestinal resection (OR 3.4; 95% CI 1.9-6.1) were associated with death after intussusception. Diagnosis by ultrasound or enema (OR 0.4; 95% CI 0.3-0.7), and employment of a household member (OR 0.7; 95% CI 0.4-1.0) were protective against intestinal resection. CONCLUSIONS: Delays in hospital presentation and female sex were significantly associated with death, whereas higher socioeconomic status and availability of radiologic diagnosis reduced likelihood of undergoing resection. Efforts should be intensified to improve the awareness, diagnosis, and management of intussusception in sub-Saharan African countries to reduce morbidity and mortality from intussusception in these resource-limited settings.

Evaluation of Intussusception after Monovalent Rotavirus Vaccination in Africa.

BACKGROUND: Postlicensure evaluations have identified an association between rotavirus vaccination and intussusception in several high- and middle-income countries. We assessed the association between monovalent human rotavirus vaccine and intussusception in lower-income sub-Saharan African countries. METHODS: Using active surveillance, we enrolled patients from seven countries (Ethiopia, Ghana, Kenya, Malawi, Tanzania, Zambia, and Zimbabwe) who had intussusception that met international (Brighton Collaboration level 1) criteria. Rotavirus vaccination status was confirmed by review of the vaccine card or clinic records. The risk of intussusception within 1 to 7 days and 8 to 21 days after vaccination among infants 28 to 245 days of age was assessed by means of the self-controlled case-series method. RESULTS: Data on 717 infants who had intussusception and confirmed vaccination status were analyzed. One case occurred in the 1 to 7 days after dose 1, and 6 cases occurred in the 8 to 21 days after dose 1. Five cases and 16 cases occurred in the 1 to 7 days and 8 to 21 days, respectively, after dose 2. The risk of intussusception in the 1 to 7 days after dose 1 was not higher than the background risk of intussusception (relative incidence [i.e., the incidence during the risk window vs. all other times], 0.25; 95% confidence interval [CI], <0.001 to 1.16); findings were similar for the 1 to 7 days after dose 2 (relative incidence, 0.76; 95% CI, 0.16 to 1.87). In addition, the risk of intussusception in the 8 to 21 days or 1 to 21 days after either dose was not found to be higher than the background risk. CONCLUSIONS: The risk of intussusception after administration of monovalent human rotavirus vaccine was not higher than the background risk of intussusception in seven lower-income sub-Saharan African countries. (Funded by the GAVI Alliance through the CDC Foundation.).

Trends in diarrhea hospitalizations among infants at three hospitals in Tanzania before and after rotavirus vaccine introduction.

BACKGROUND: The Tanzania Ministry of Health introduced monovalent human rotavirus vaccine in January 2013, to be administered at ages 6 and 10 weeks. Data suggest there was high vaccine uptake. We used hospital ward registers from 3 hospitals to examine trends in diarrhea hospitalizations among infants before and after vaccine introduction. METHODS: Ward registers from Dodoma Regional Referral Hospital (Central Tanzania), and two hospitals in Mbeya (Southwest area), Mbeya Zonal Referral Hospital and Mbalizi Hospital, were used to tally admissions for diarrhea among children by age group, month and year. Rotavirus surveillance had started at these hospitals in early 2013; the proportion of infants enrolled and rotavirus-EIA positive were examined by month to determine peak periods of rotavirus disease post-vaccine introduction. RESULTS: Registers were available for 2-4 prevaccine years and 2-3 post introduction years. At Dodoma Regional Referral Hospital, compared with the mean of 2011 and 2012, diarrhea hospitalizations among infants were 26% lower in 2015 and 58% lower in 2016. The diarrhea peak shifted later in the year first by 1 and then by 2-3 months from prevaccine. At the Mbeya hospitals, the number of diarrhea admissions in prevaccine period varied substantially by year. At Mbeya Referral Hospital, diarrhea hospitalizations among infants were lower by 25-37% in 2014 and 11-26% in 2015, while at Mbalizi Hospital, these hospitalizations were 4% lower in 2014 and 14% higher in 2015. Rotavirus testing data demonstrated a lowering of the prevaccine peak, a shift in timing of the peak months and indicated that other diarrheal peaks in post-introduction years were not due to rotavirus. CONCLUSIONS: In this ecological evaluation, total diarrhea hospitalizations among infants were lower (≥25% lower in ≥1 year) following introduction in 2 of 3 hospitals. There are challenges in using ward registers to ascertain possible impact of rotavirus vaccine introduction on trends in hospitalizations for treatment of all diarrheal illness.

Detection of rotavirus before and after monovalent rotavirus vaccine introduction and vaccine effectiveness among children in mainland Tanzania.

BACKGROUND: Monovalent rotavirus vaccine (RV1) was introduced in Tanzania in January 2013 under the Reach Every Child initiative, to be given at ages 6 and 10 weeks. We used the sentinel hospital rotavirus surveillance system to examine the rotavirus detection rate before and after vaccine introduction and estimate vaccine effectiveness. METHODS: Before vaccine introduction, rotavirus surveillance was established at two mainland hospitals; children admitted for acute diarrhea were eligible for enrollment and stools were tested for rotavirus antigen. We compared the rotavirus positivity rate in the pre-vaccine period (Tanga Hospital, 2009 and 2011; Bugando Medical Centre, 2012) to that from post-introduction years, 2014-2015. In 2013, surveillance was established at 9 additional hospitals. We examined rotavirus positivity among infants at these sites for 2014-2015. We obtained vaccine records and calculated vaccine effectiveness at 3 sites using case-test-negative control design. RESULTS: At Tanga Hospital, the rotavirus positivity rate among infants was 41% (102/251) pre-vaccine and 14% (28/197) in post-vaccine years (rate ratio: 0.35 [95% CI 0.22-0.54]). At Bugando, the positivity rate was 58% (83/143) pre-vaccine, and 18% (49/277) post-introduction (rate ratio 0.30 [95% CI 0.210.44]). Results were similar among children <5 years. At the new sites, the median site rotavirus positivity rate among infants was 26% in 2014 (range 19-44%) and 18% in 2015 (range 16-33%). The effectiveness of ≥1 RV1 dose against rotavirus hospitalization among children 5-23 months was 53% (95% CI: -14, 81), and 66% (95% CI: 9-87) against hospitalization with intravenous rehydration. Following introduction, peak rotavirus activity occurred later in the year and appeared more concentrated in time. CONCLUSION: Rotavirus surveillance data from Tanzania indicate that the rotavirus positivity rate among children hospitalized with diarrhea that were enrolled was substantially reduced after vaccine introduction. Low positivity rates among infants were detected at hospitals across the country. Overall, the data support that rotavirus vaccine has been successfully introduced and is effective in Tanzanian children.

Monovalent Rotavirus Vaccine Effectiveness and Impact on Rotavirus Hospitalizations in Zanzibar, Tanzania: Data From the First 3 Years After Introduction.

Background: Low-income settings challenge the level of protection provided by live attenuated oral rotavirus vaccines. Rotarix (RV1) was introduced in the United Republic of Tanzania in early 2013, with 2 doses given at the World Health Organization-recommended schedule of ages 6 and 10 weeks, along with oral poliovirus vaccine. Methods: We performed active surveillance for rotavirus hospitalizations at the largest hospital in Zanzibar, Tanzania, from 2010 through 2015. Using a case-test-negative control design, we estimated the vaccine effectiveness (VE) of 2 RV1 doses in preventing rotavirus hospitalizations. Results: Based on 204 rotavirus case patients and 601 test-negative controls aged 5-23 months, the VE of 2 RV1 doses against hospitalization for rotavirus diarrhea was 57% (95% confidence interval, 14%-78%). VE tended to increase against hospitalizations with higher severity, reaching 69% (95% confidence interval, 15%-88%) against the severity score for the top quarter of case patients. Compared with the prevaccine period, there were estimated reductions of 40%, 46%, and 69% in the number of rotavirus hospitalizations among infants in 2013, 2014, and 2015, respectively, and reductions of 36%, 26%, and 64%, respectively, among children aged <5 years. Conclusions: With data encompassing 3 years before and 3 years after vaccine introduction, our results indicate that successful delivery of RV1 on the current World Health Organization schedule can provide substantial health benefits in a resource-limited setting.