Suchergebnisse

AbstractThe units at the early stages of multi-stage area samples are generally sampled with probabilities proportional to their estimated sizes (PPES). With such a design, an overall equal probability (EP) sample design would yield a constant number of final stage units from each final stage cluster if the measures of size used in the PPES selection at each sampling stage were directly proportional to the number of final stage units. However, there are often sizable relative differences between the measures of size used in the PPES selections and the number of final stage units. Two common approaches for dealing with these differences are: (1) to retain a self-weighting sample design, allowing the sample sizes to vary across the sampled primary sampling units (PSUs) and (2) to retain the fixed sample size in each PSU and to compensate for the unequal selection probabilities by weighting adjustments in the analyses. This article examines these alternative designs in the context of two-stage sampling in which PSUs are sampled with PPES at the first stage, and an equal probability sample of final stage units is selected from each sampled PSU at the second stage. Two-stage sample designs of this type are used for household surveys in many countries. The discussion is illustrated with data from the Population-based HIV Impact Assessment surveys that were conducted using this design in several African countries.

AbstractIntroductionPopulation‐based biomarker surveys are the gold standard for estimating HIV prevalence but are susceptible to substantial non‐participation (up to 30%). Analytical missing data methods, including inverse‐probability weighting (IPW) and multiple imputation (MI), are biased when data are missing‐not‐at‐random, for example when people living with HIV more frequently decline participation. Heckman‐type selection models can, under certain assumptions, recover unbiased prevalence estimates in such scenarios.MethodsWe pooled data from 142,706 participants aged 15–49 years from nationally representative cross‐sectional Population‐based HIV Impact Assessments in seven countries in sub‐Saharan Africa, conducted between 2015 and 2018 in Tanzania, Uganda, Malawi, Zambia, Zimbabwe, Lesotho and Eswatini. We compared sex‐stratified HIV prevalence estimates from unadjusted, IPW, MI and selection models, controlling for household and individual‐level predictors of non‐participation, and assessed the sensitivity of selection models to the copula function specifying the correlation between study participation and HIV status.ResultsIn total, 84.1% of participants provided a blood sample to determine HIV serostatus (range: 76% in Malawi to 95% in Uganda). HIV prevalence estimates from selection models diverged from IPW and MI models by up to 5% in Lesotho, without substantial precision loss. In Tanzania, the IPW model yielded lower HIV prevalence estimates among males than the best‐fitting copula selection model (3.8% vs. 7.9%).ConclusionsWe demonstrate how HIV prevalence estimates from selection models can differ from those obtained under missing‐at‐random assumptions. Further benefits include exploration of plausible relationships between participation and outcome. While selection models require additional assumptions and careful specification, they are an important tool for triangulating prevalence estimates in surveys with substantial missing data due to non‐participation.

AbstractIntroductionLogistical complexities of returning laboratory test results to participants have precluded most population‐based HIV surveys conducted in sub‐Saharan Africa from doing so. For HIV positive participants, this presents a missed opportunity for engagement into clinical care and improvement in health outcomes. The Population‐based HIV Impact Assessment (PHIA) surveys, which measure HIV incidence and the prevalence of viral load (VL) suppression in selected African countries, are returning VL results to health facilities specified by each HIV positive participant within eight weeks of collection. We describe the performance of the specimen and data management systems used to return VL results to PHIA participants in Zimbabwe, Malawi and Zambia.MethodsConsenting participants underwent home‐based counseling and HIV rapid testing as per national testing guidelines; all confirmed HIV positive participants had VL measured at a central laboratory on either the Roche CAP/CTM or Abbott m2000 platform. On a bi‐weekly basis, a dedicated data management team produced logs linking the VL test result with the participants' contact information and preferred health facility; project staff sent test results confidentially via project drivers, national courier systems, or electronically through an adapted short message service (SMS). Participants who provided cell phone numbers received SMS or phone call alerts regarding availability of VL results.Results and discussionFrom 29,634 households across the three countries, 78,090 total participants 0 to 64 years in Zimbabwe and Malawi and 0 to 59 years in Zambia underwent blood draw and HIV testing. Of the 8391 total HIV positive participants identified, 8313 (99%) had VL tests performed and 8245 (99%) of these were returned to the selected health facilities. Of the 5979 VL results returned in Zimbabwe and Zambia, 85% were returned within the eight‐week goal with a median turnaround time of 48 days (IQR: 33 to 61). In Malawi, where exact return dates were unavailable all 2266 returnable results reached the health facilities by 11 weeks.ConclusionsThe first three PHIA surveys returned the vast majority of VL results to each HIV positive participant's preferred health facility within the eight‐week target. Even in the absence of national VL monitoring systems, a system to return VL results from a population‐based survey is feasible, but it requires developing laboratory and data management systems and dedicated staff. These are likely important requirements to strengthen return of results systems in routine clinical care.

With the highest HIV incidence and prevalence globally, the government of Eswatini started a substantial scale-up of HIV treatment and prevention services in 2011. Two sequential large population-based surveys were conducted before and after service expansion to assess the impact of the national response. Cross-sectional, household-based, nationally representative samples of adults, ages 18 to 49 years, were sampled in 2011 and 2016. We measured HIV prevalence, incidence (recent infection based on limiting antigen ≤1.5 optical density units and HIV RNA ≥1000 copies/mL), viral load suppression (HIV RNA <1000 copies/mL among all seropositive adults) and unsuppressed viremia (HIV RNA ≥1000 copies/mL among all, regardless of HIV status) and assessed for temporal changes by conducting a trend analysis of the log ratio of proportions, using a Z statistic distribution. HIV prevalence remained stable from 2011 to 2016 [32% versus 30%, p = 0.10]. HIV incidence significantly declined 48% [2.48% versus 1.30%, p = 0.01]. Incidence remained higher among women than men [2011: 3.16% versus 1.83%; 2016: 1.76% versus 0.86%], with a smaller but significant relative reduction among women [44%; p = 0.04] than men [53%; p = 0.09]. The proportion of seropositive adults with viral load suppression significantly increased from 35% to 71% [p < .001]. The proportion of the total adult population with unsuppressed viremia decreased from 21% to 9% [p < .001]. National HIV incidence in Eswatini decreased by nearly half and viral load suppression doubled over a five-year period. Unsuppressed viremia in the total population decreased 58%. These population-based findings demonstrate the national impact of expanded HIV services in a hyperendemic country.

AbstractIntroductionThe global target for 2020 is that ≥90% of people living with HIV (PLHIV) receiving antiretroviral therapy (ART) will achieve viral load suppression (VLS). We examined VLS and its determinants among adults receiving ART for at least four months.MethodsWe analysed data from the population‐based HIV impact assessment (PHIA) surveys in Eswatini, Lesotho, Malawi, Zambia and Zimbabwe (2015 to 2017). PHIA surveys are nationally representative, cross‐sectional household surveys. Data collection included structured interviews, home‐based HIV testing and laboratory testing. Blood samples from PLHIV were analysed for HIV RNA, CD4 counts and recent exposure to antiretroviral drugs (ARVs). We calculated representative estimates for the prevalence of VLS (viral load <1000 copies/mL), nonsuppressed viral load (NVL; viral load ≥1000 copies/mL), virologic failure (VF; ARVs present and viral load ≥1000 copies/mL), interrupted ART (ARVs absent and viral load ≥1000 copies/mL) and rates of switching to second‐line ART (protease inhibitors present) among PLHIV aged 15 to 59 years who participated in the PHIA surveys in Eswatini, Lesotho, Malawi, Zambia and Zimbabwe, initiated ART at least four months before the survey and were receiving ART at the time of the survey (according to self‐report or ARV testing). We calculated odds ratios and incidence rate ratios for factors associated with NVL, VF, interrupted ART, and switching to second‐line ART.ResultsWe included 9200 adults receiving ART of whom 88.8% had VLS and 11.2% had NVL including 8.2% who experienced VF and 3.0% who interrupted ART. Younger age, male sex, less education, suboptimal adherence, receiving nevirapine, HIV non‐disclosure, never having married and residing in Zimbabwe, Lesotho or Zambia were associated with higher odds of NVL. Among people with NVL, marriage, female sex, shorter ART duration, higher CD4 count and alcohol use were associated with lower odds for VF and higher odds for interrupted ART. Many people with VF (44.8%) had CD4 counts <200 cells/µL, but few (0.31% per year) switched to second‐line ART.ConclusionsCountries are approaching global VLS targets for adults. Treatment support, in particular for younger adults, and people with higher CD4 counts, and switching of people to protease inhibitor‐ or integrase inhibitor‐based regimens may further reduce NVL prevalence.