Adjusting for incomplete baseline covariates in randomized controlled trials: a cross-world imputation framework.

In randomized controlled trials, adjusting for baseline covariates is commonly used to improve the precision of treatment effect estimation. However, covariates often have missing values. Recently, Zhao and Ding studied two simple strategies, the single imputation method and missingness-indicator method (MIM), to handle missing covariates and showed that both methods can provide an efficiency gain compared to not adjusting for covariates. To better understand and compare these two strategies, we propose and investigate a novel theoretical imputation framework termed cross-world imputation (CWI). This framework includes both single imputation and MIM as special cases, facilitating the comparison of their efficiency. Through the lens of CWI, we show that MIM implicitly searches for the optimal CWI values and thus achieves optimal efficiency. We also derive conditions under which the single imputation method, by searching for the optimal single imputation values, can achieve the same efficiency as the MIM. We illustrate our findings through simulation studies and a real data analysis based on the Childhood Adenotonsillectomy Trial. We conclude by discussing the practical implications of our findings.

Partner Characteristics and HIV Outcomes among Partners Reached by Phone vs. In-person for Assisted Partner Services in Western Kenya.

BACKGROUND: Assisted partner services (APS) is an effective strategy for testing people with undiagnosed HIV but there is limited information on the relative reach and effectiveness of phone versus in-person contact. METHODS: We analyzed data from 31 facilities in Western Kenya providing APS to female index clients newly HIV-diagnosed, their male partners, and female partners of men newly HIV-diagnosed. APS providers attempted contacting partners using phone first and, if unsuccessful, in-person in the community. Using log-linear mixed models, we estimated relative risks (RR) between phone being the final contact method and partner characteristics and HIV outcomes. RESULTS: From May 2018-March 2020, 2534 female index clients named 7614 male partners, of whom 772 (10.1%) tested positive and named an additional 4956 non-index female partners. Of 11,912 (94.7%) partners reached, 5179 (43.5%) were via phone and 6733 (56.5%) in-person. Of 8076 testing-eligible partners, 99.7% tested and 11.2% first-time tested. Of those tested, 13.1% were newly-diagnosed, of whom 87.0% linked to care. Newly-diagnosed partners were less likely to have been reached by phone vs. in-person (9.8% vs. 15.9%; adjusted RR:0.61, 95%CI:0.53-0.70). Being reached by phone was not significantly associated with testing, first-time testing, or linkage to care. CONCLUSIONS: In an APS program that reached 94% of elicited partners, fewer than half were successfully contacted by phone. A combined phone and in-person approach is likely essential for equitable, effective APS.

Efficacy of the Dapivirine Vaginal Ring Accounting for Imperfect Adherence.

Product adherence is critical to obtaining objective estimates of efficacy of pre-exposure prophylactic interventions against HIV-1 infection. With imperfect adherence, intention-to-treat analyses assess the collective effects of complete, sub-optimal and non-adherence, providing a biased and attenuated estimate of the average causal effect of an intervention. Using data from the MTN-020/ASPIRE phase III trial evaluating HIV-1 efficacy of the dapivirine vaginal ring, we conducted per-protocol, and adherence-adjusted causal inference analyses using principal stratification and marginal structural models. We constructed two adherence cut offs of ≥ 0.9 mg (low cutoff) and > 4.0 mg (high cutoff) that represent drug released from the ring over a 28-day period. The HIV-1 efficacy estimate (95% CI) was 30.8% (3.6%, 50.3%) (P = 0.03) from the per-protocol analysis, and 53.6% (16.5%, 74.3%) (P = 0.01) among the highest predicted adherers from principal stratification analyses using the low cutoff. Marginal structural models produced efficacy estimates (95% CIs) ranging from 48.8 (21.8, 66.4) (P = 0.0019) to 56.5% (32.8%, 71.9%) (P = 0.0002). Application of adherence-adjusted causal inference methods are useful in interpreting HIV-1 efficacy in secondary analyses of PrEP clinical trials.

The Impact of Stigma and Sexual Identity on PrEP Awareness and Use Among At-Risk Men Who Have Sex With Men in Four U.S. Cities (HPTN 078).

Persistent pre-exposure prophylaxis (PrEP) use reduces the risk of HIV infection, yet uptake lags among those with the greatest need. Sexual identity stigma may be a significant barrier to PrEP awareness and use among high-risk communities. The primary aim of this study was to determine whether sexual identity was related to PrEP awareness and use. This multi-site HIV Prevention Trials Network (HPTN) study (HPTN 078) focuses on men who have sex with men (MSM) (n=335) who were HIV-negative at screening. The majority of participants were non-white (62.1%), younger than 35 (57.9%), single (79.1%), and aware of PrEP, yet had never taken PrEP (52.5%). Participants completed questionnaires including sexual history and identity; lesbian, gay, bisexual, transgender, queer (LGBTQ) community engagement; PrEP awareness and use; and several measures of sexual identity stigma including family and friend stigma, general societal stigma, and anticipated healthcare stigma. Univariate and multinomial logistic regression models helped to determine factors associated with PrEP awareness and use. There were stark disparities in PrEP awareness comparing Black and White participants; 50% of Black participants reported being PrEP unaware vs 11.8% of White participants. In this sample, gay sexual identity (compared to bisexual identity) was associated with increased PrEP awareness (AOR 6.66) and use (AOR 16.9). Additionally, 29% of the association between sexual orientation and PrEP use was mediated through internalized stigma. Given low PrEP uptake among MSM, interventions that address sexual identity stigma may motivate greater PrEP uptake.

Effectiveness of HIV self-testing when offered within assisted partner services in Western Kenya (APS-HIVST Study): a cluster randomized controlled trial.

INTRODUCTION: Assisted partner services (APS) is an effective strategy for increasing HIV testing, new diagnosis, and linkage to care among sexual partners of people living with HIV (PLWH). APS can be resource intensive as it requires community tracing to locate each partner named and offer them testing. There is limited evidence for the effectiveness of offering HIV self-testing (HIVST) as an option for partner testing within APS. METHODS: We conducted a cluster randomized controlled trial comparing provider-delivered HIV testing (Standard APS) versus offering partners the option of provider-delivered testing or HIVST (APS+HIVST) at 24 health facilities in Western Kenya. Facilities were randomized 1:1 and we conducted intent-to-treat analyses using Poisson generalized linear mixed models to estimate intervention impact on HIV testing, new HIV diagnoses, and linkage to care. All models accounted for clustering at the clinic level and new diagnoses and linkage models were adjusted for individual-level age, sex, and income a priori. RESULTS: From March to December 2021, 755 index clients received APS and named 5054 unique partners. Among these, 1408 partners reporting a prior HIV diagnosis were not eligible for HIV testing and were excluded from analyses. Of the remaining 3646 partners, 96.9% were successfully contacted for APS and tested for HIV: 2111 (97.9%) of 2157 in the APS+HIVST arm and 1422 (95.5%) of 1489 in the Standard APS arm. In the APS+HIVST arm, 84.6% (1785/2111) tested via HIVST and 15.4% (326/2111) received provider-delivered testing. Overall, 16.7% of the 3533 who tested were newly diagnosed with HIV (APS+HIVST = 357/2111 [16.9%]; Standard APS = 232/1422 [16.3%]). Of the 589 partners who were newly diagnosed, 90.7% were linked to care (APS+HIVST = 309/357 [86.6%]; Standard APS = 225/232 [97.0%]). There were no significant differences between the two arms in HIV testing (relative risk [RR]: 1.02, 95% CI: 0.96-1.10), new HIV diagnoses (adjusted RR [aRR]: 1.03, 95% CI: 0.76-1.39) or linkage to care (aRR: 0.88, 95% CI: 0.74-1.06). CONCLUSIONS: There were no differences between APS+HIVST and Standard APS, demonstrating that integrating HIVST into APS continues to be an effective strategy for identifying PLWH by successfully reaching and HIV testing >95% of elicited partners, newly diagnosing with HIV one in six of those tested, >90% of whom were linked to care. CLINICAL TRIAL NUMBER: NCT04774835.

Effectiveness of Using Additional HIV Self-Test Kits as an Incentive to Increase HIV Testing Within Assisted Partner Services.

BACKGROUND: Incentives have shown mixed results in increasing HIV testing rates in low-resource settings. We investigated the effectiveness of offering additional self-tests (HIVSTs) as an incentive to increase testing among partners receiving assisted partner services (APS). SETTING: Western Kenya. METHODS: We conducted a single-crossover study nested within a cluster-randomized controlled trial. Twenty-four facilities were randomized 1:1 to (1) control: provider-delivered testing or (2) intervention: offered 1 HIVST or provider-delivered testing for 6 months (pre-implementation), then switched to offering 2 HIVSTs for 6 months (post-implementation). A difference-in-differences approach using generalized linear mixed models, accounting for facility clustering and adjusting for age, sex, and income, was used to estimate the effect of the incentive on HIV testing and first-time testing among partners in APS. RESULTS: March 2021-June 2022, 1127 index clients received APS and named 8155 partners, among whom 2333 reported a prior HIV diagnosis and were excluded from analyses, resulting in 5822 remaining partners: 3646 (62.6%) and 2176 (37.4%) in the pre-implementation and post-implementation periods, respectively. Overall, 944/2176 partners (43%) were offered a second HIVST during post-preimplementation, of whom 34.3% picked up 2 kits, of whom 71.7% reported that the second kit encouraged HIV testing. Comparing partners offered 1 vs. two HIVSTs showed no difference in HIV testing (relative risk: 1.01, 95% confidence interval: 0.951 to 1.07) or HIV testing for the first time (relative risk: 1.23, 95% confidence interval: 0.671 to 2.24). CONCLUSIONS: Offering a second HIVST as an incentive within APS did not significantly impact HIV testing or first-time testing, although those opting for 2 kits reported it incentivized them to test.

Absenteeism and Health Behavior Trends Associated With Acute Respiratory Illness Before and During the COVID-19 Pandemic in a Community Household Cohort, King County, Washington.

Introduction: Longitudinal data on how acute respiratory illness (ARI) affects behavior, namely school or work participation, and nonpharmaceutical intervention (NPI) usage before and during the COVID-19 pandemic is limited. The authors assessed how ARIs and specific symptoms affected school, work, and health-related behaviors over time. Methods: From November 2019 to June 2021, participating households with children in King County, Washington, were remotely monitored for ARI symptoms weekly. Following ARIs, participants reported illness-related effects on school, work, and NPI use. Using logistic regression with generalized estimating equations, the authors examined associations between symptoms and behaviors. Results: Of 1,861 participants, 581 (31%) from 293 households reported 884 ARIs and completed one-week follow-up surveys. Compared with the prepandemic period, during the period of the pandemic pre-COVID-19 vaccine, ARI-related school (56% vs 10%, p<0.001) absenteeism decreased and masking increased (3% vs 28%, p<0.001). After vaccine authorization in December 2020, more ARIs resulted in masking (3% vs 48%, p<0.001), avoiding contact with non-household members (26% vs 58%, p<0.001), and staying home (37% vs 69%, p<0.001) compared with the prepandemic period. Constitutional symptoms such as fever were associated with work disruptions (OR=1.91; 95% CI=1.06, 3.43), staying home (OR=1.55; 95% CI=1.06, 2.27), and decreased contact with non-household members (OR=1.58; 95% CI=1.05, 2.36). Conclusions: This remote household study permitted uninterrupted tracking of behavioral changes in families with children before and during the COVID-19 pandemic, identifying increased use of some NPIs when ill but no additional illness-associated work or school disruptions.

Robust analysis of stepped wedge trials using composite likelihood models.

Stepped wedge trials (SWTs) are a type of cluster randomized trial that involve repeated measures on clusters and design-induced confounding between time and treatment. Although mixed models are commonly used to analyze SWTs, they are susceptible to misspecification particularly for cluster-longitudinal designs such as SWTs. Mixed model estimation leverages both "horizontal" or within-cluster information and "vertical" or between-cluster information. To use horizontal information in a mixed model, both the mean model and correlation structure must be correctly specified or accounted for, since time is confounded with treatment and measurements are likely correlated within clusters. Alternative non-parametric methods have been proposed that use only vertical information; these are more robust because between-cluster comparisons in a SWT preserve randomization, but these non-parametric methods are not very efficient. We propose a composite likelihood method that focuses on vertical information, but has the flexibility to recover efficiency by using additional horizontal information. We compare the properties and performance of various methods, using simulations based on COVID-19 data and a demonstration of application to the LIRE trial. We found that a vertical composite likelihood model that leverages baseline data is more robust than traditional methods, and more efficient than methods that use only vertical information. We hope that these results demonstrate the potential value of model-based vertical methods for SWTs with a large number of clusters, and that these new tools are useful to researchers who are concerned about misspecification of traditional models.

Evaluation of a novel university-based testing platform to increase access to SARS-CoV-2 testing during the COVID-19 pandemic in a cohort study.

OBJECTIVE: We aimed to evaluate the feasibility and utility of an unsupervised testing mechanism, in which participants pick up a swab kit, self-test (unsupervised) and return the kit to an on-campus drop box, as compared with supervised self-testing at staffed locations. DESIGN: University SARS-CoV-2 testing cohort. SETTING: Husky Coronavirus Testing provided voluntary SARS-CoV-2 testing at a university in Seattle, USA. OUTCOME MEASURES: We computed descriptive statistics to describe the characteristics of the study sample. Adjusted logistic regression implemented via generalised estimating equations was used to estimate the odds of a self-swab being conducted through unsupervised versus supervised testing mechanisms by participant characteristics, including year of study enrolment, pre-Omicron versus post-Omicron time period, age, sex, race, ethnicity, affiliation and symptom status. RESULTS: From September 2021 to July 2022, we received 92 499 supervised and 26 800 unsupervised self-swabs. Among swabs received by the laboratory, the overall error rate for supervised versus unsupervised swabs was 0.3% vs 4%, although this declined to 2% for unsupervised swabs by the spring of the academic year. Results were returned for 92 407 supervised (5% positive) and 25 836 unsupervised (4%) swabs from 26 359 participants. The majority were students (79%), 61% were female and most identified as white (49%) or Asian (34%). The use of unsupervised testing increased during the Omicron wave when testing demand was high and stayed constant in spring 2022 even when testing demand fell. We estimated the odds of using unsupervised versus supervised testing to be significantly greater among those <25 years of age (p<0.001), for Hispanic versus non-Hispanic individuals (OR 1.2, 95% CI 1.0 to 1.3, p=0.01) and lower among individuals symptomatic versus asymptomatic or presymptomatic (0.9, 95% CI 0.8 to 0.9, p<0.001). CONCLUSIONS: Unsupervised swab collection permitted increased testing when demand was high, allowed for access to a broader proportion of the university community and was not associated with a substantial increase in testing errors.

Cost-Effectiveness and Budget Impact Analysis of the Implementation of Differentiated Service Delivery Models for HIV Treatment in Mozambique: a Modelling Study.

INTRODUCTION: In 2018, the Mozambique Ministry of Health launched guidelines for implementing differentiated service delivery models (DSDMs) to optimize HIV service delivery, improve retention in care, and ultimately reduce HIV-associated mortality. The models were fast-track, 3-month antiretrovirals dispensing, community antiretroviral therapy groups, adherence clubs, family approach and three one-stop shop models: adolescent-friendly health services, maternal and child health, and tuberculosis. We conducted a cost-effectiveness analysis and budget impact analysis to compare these models to conventional services. METHODS: We constructed a decision tree model based on the percentage of enrolment in each model and the probability of the outcome (12-month retention in treatment) for each year of the study period-three for the cost-effectiveness analysis (2019-2021) and three for the budget impact analysis (2022-2024). Costs for these analyses were primarily estimated per client-year from the health system perspective. A secondary cost-effectiveness analysis was conducted from the societal perspective. Budget impact analysis costs included antiretrovirals, laboratory tests and service provision interactions. Cost-effectiveness analysis additionally included start-up, training and clients' opportunity costs. Effectiveness was estimated using an uncontrolled interrupted time series analysis comparing the outcome before and after the implementation of the differentiated models. A one-way sensitivity analysis was conducted to identify drivers of uncertainty. RESULTS: After implementation of the DSDMs, there was a mean increase of 14.9 percentage points (95% CI: 12.2, 17.8) in 12-month retention, from 47.6% (95% CI, 44.9-50.2) to 62.5% (95% CI, 60.9-64.1). The mean cost difference comparing DSDMs and conventional care was US$ -6 million (173,391,277 vs. 179,461,668) and -32.5 million (394,705,618 vs. 433,232,289) from the health system and the societal perspective, respectively. Therefore, DSDMs dominated conventional care. Results were most sensitive to conventional care interaction costs in the one-way sensitivity analysis. For a population of 1.5 million, the base-case 3-year financial costs associated with the DSDMs was US$550 million, compared with US$564 million for conventional care. CONCLUSIONS: DSDMs were less expensive and more effective in retaining clients 12 months after antiretroviral therapy initiation and were estimated to save approximately US$14 million for the health system from 2022 to 2024.

Remote surveillance and detection of SARS-CoV-2 transmission among household members in King County, Washington.

BACKGROUND: Early during the COVID-19 pandemic, it was important to better understand transmission dynamics of SARS-CoV-2, the virus that causes COVID-19. Household contacts of infected individuals are particularly at risk for infection, but delays in contact tracing, delays in testing contacts, and isolation and quarantine posed challenges to accurately capturing secondary household cases. METHODS: In this study, 346 households in the Seattle region were provided with respiratory specimen collection kits and remotely monitored using web-based surveys for respiratory illness symptoms weekly between October 1, 2020, and June 20, 2021. Symptomatic participants collected respiratory specimens at symptom onset and mailed specimens to the central laboratory in Seattle. Specimens were tested for SARS-CoV-2 using RT-PCR with whole genome sequencing attempted when positive. SARS-CoV-2-infected individuals were notified, and their household contacts submitted specimens every 2 days for 14 days. RESULTS: In total, 1371 participants collected 2029 specimens that were tested; 16 individuals (1.2%) within 6 households tested positive for SARS-CoV-2 during the study period. Full genome sequences were generated from 11 individuals within 4 households. Very little genetic variation was found among SARS-CoV-2 viruses sequenced from different individuals in the same household, supporting transmission within the household. CONCLUSIONS: This study indicates web-based surveillance of respiratory symptoms, combined with rapid and longitudinal specimen collection and remote contact tracing, provides a viable strategy to monitor households and detect household transmission of SARS-CoV-2. TRIAL REGISTRATION IDENTIFIER: NCT04141930, Date of registration 28/10/2019.

Determinants of Type-Specific Human Papillomavirus Concordance Across Anatomic Sites in Young Men Who Have Sex With Men and Transgender Women, 3 U.S. Cities, 2016-2018.

BACKGROUND: Among men who have sex with men (MSM) and transgender women (TGW), the dynamics of human papillomavirus (HPV) infections at different anatomical sites are not well understood. Information on HPV concordance between anatomic sites can inform the extent of autoinoculation, and susceptibility of different anatomic areas to HPV infection. We described and assessed correlates of HPV concordance across anal, oral, and genital samples. METHODS: We enrolled 1876 MSM and TGW aged 18 to 26 years in 3 US cities. Oral, genital, and anal samples were self-collected for type-specific HPV DNA testing (37 types). Demographics, sexual behaviors, and health history were self-reported. Kappa statistics based on percent positive agreement (kappa+) and generalized estimating equations were used to describe and identify correlates of HPV type-specific concordance between anatomic sample pairs. RESULTS: Any HPV was detected in 69.9%, 48.6%, and 7.4% of anal, genital, and oral samples, respectively. Detection of any HPV (concurrence) was most common in anal-genital pairs (40.9%) and uncommon in oral-genital and oral-anal pairs (3.4% and 6.5% respectively). Type-specific concordance was poor across all sample pairs (kappa+ <0.20). Younger age and older age at first sex were positively associated with type-concordant anal-genital infections. Sexual behaviors were unassociated with concordance. CONCLUSIONS: Poor oral/anogenital concordance suggests the oral mucosa has different susceptibility to HPV infection, differential clearance and/or autoinoculation between oral and anogenital sites is unlikely. There was some observed concurrence and concordance between anal and genital sites, unassociated with sexual behavior, suggesting autoinoculation. Longitudinal studies are necessary to further elucidate mechanisms of multisite infections.

The Relationship Between Insertive Oral and Anal Sex and Select Measures of the Composition of the Urethral Microbiota Among Men Who Have Sex With Men.

BACKGROUND: Sexual behavior may influence the composition of the male urethral microbiota, but this hypothesis has not been tested in longitudinal studies of men who have sex with men (MSM). METHODS: From December 2014 to July 2018, we enrolled MSM with nongonococcal urethritis (NGU) attending a sexual health clinic. Men attended 5 in-clinic visits at 3-week intervals, collected weekly urine specimens at home, and reported daily antibiotics and sexual activity on weekly diaries. We applied broad-range 16S rRNA gene sequencing to urine. We used generalized estimating equations to estimate the association between urethral sexual exposures in the prior 7 days (insertive oral sex [IOS] only, condomless insertive anal intercourse [CIAI] only, IOS with CIAI [IOS + CIAI], or none) and Shannon index, number of species (observed, oral indicator, and rectal indicator), and specific taxa, adjusting for recent antibiotics, age, race/ethnicity, HIV, and preexposure prophylaxis. RESULTS: Ninety-six of 108 MSM with NGU attended ≥1 follow-up visit. They contributed 1140 person-weeks of behavioral data and 1006 urine specimens. Compared with those with no urethral sexual exposures, those with IOS only had higher Shannon index ( P = 0.03 ) but similar number of species and presence of specific taxa considered, adjusting for confounders; the exception was an association with Haemophilus parainfluenzae . CIAI only was not associated with measured aspects of the urethral microbiota. IOS + CIAI was only associated with presence of H. parainfluenzae and Haemophilus . CONCLUSIONS: Among MSM after NGU, IOS and CIAI did not seem to have a substantial influence on measured aspects of the composition of the urethral microbiota.

Changing genomic epidemiology of COVID-19 in long-term care facilities during the 2020-2022 pandemic, Washington State.

BACKGROUND: Long-term care facilities (LTCFs) are vulnerable to disease outbreaks. Here, we jointly analyze SARS-CoV-2 genomic and paired epidemiologic data from LTCFs and surrounding communities in Washington state (WA) to assess transmission patterns during 2020-2022, in a setting of changing policy. We describe sequencing efforts and genomic epidemiologic findings across LTCFs and perform in-depth analysis in a single county. METHODS: We assessed genomic data representativeness, built phylogenetic trees, and conducted discrete trait analysis to estimate introduction sizes over time, and explored selected outbreaks to further characterize transmission events. RESULTS: We found that transmission dynamics among cases associated with LTCFs in WA changed over the course of the COVID-19 pandemic, with variable introduction rates into LTCFs, but decreasing amplification within LTCFs. SARS-CoV-2 lineages circulating in LTCFs were similar to those circulating in communities at the same time. Transmission between staff and residents was bi-directional. CONCLUSIONS: Understanding transmission dynamics within and between LTCFs using genomic epidemiology on a broad scale can assist in targeting policies and prevention efforts. Tracking facility-level outbreaks can help differentiate intra-facility outbreaks from high community transmission with repeated introduction events. Based on our study findings, methods for routine tree building and overlay of epidemiologic data for hypothesis generation by public health practitioners are recommended. Discrete trait analysis added valuable insight and can be considered when representative sequencing is performed. Cluster detection tools, especially those that rely on distance thresholds, may be of more limited use given current data capture and timeliness. Importantly, we noted a decrease in data capture from LTCFs over time. Depending on goals for use of genomic data, sentinel surveillance should be increased or targeted surveillance implemented to ensure available data for analysis.

Risk of Subsequent Respiratory Virus Detection After Primary Virus Detection in a Community Household Study-King County, Washington, 2019-2021.

BACKGROUND: The epidemiology of respiratory viral infections is complex. How infection with one respiratory virus affects risk of subsequent infection with the same or another respiratory virus is not well described. METHODS: From October 2019 to June 2021, enrolled households completed active surveillance for acute respiratory illness (ARI), and participants with ARI self-collected nasal swab specimens; after April 2020, participants with ARI or laboratory-confirmed severe acute respiratory syndrome coronavirus 2 and their household members self-collected nasal swab specimens. Specimens were tested using multiplex reverse-transcription polymerase chain reaction for respiratory viruses. A Cox regression model with a time-dependent covariate examined risk of subsequent detections following a specific primary viral detection. RESULTS: Rhinovirus was the most frequently detected pathogen in study specimens (406 [9.5%]). Among 51 participants with multiple viral detections, rhinovirus to seasonal coronavirus (8 [14.8%]) was the most common viral detection pairing. Relative to no primary detection, there was a 1.03-2.06-fold increase in risk of subsequent virus detection in the 90 days after primary detection; risk varied by primary virus: human parainfluenza virus, rhinovirus, and respiratory syncytial virus were statistically significant. CONCLUSIONS: Primary virus detection was associated with higher risk of subsequent virus detection within the first 90 days after primary detection.

Impact of differentiated service delivery models on 12-month retention in HIV treatment in Mozambique: an interrupted time-series analysis.

BACKGROUND: HIV treatment has been available in Mozambique since 2004, but coverage of, and retention in, antiretroviral therapy (ART) remain suboptimal. Therefore, to increase health system efficiency and reduce HIV-associated mortality, in November, 2018, the Ministry of Health launched national guidelines on implementing eight differentiated service delivery models (DSDMs) for HIV treatment. We assessed the effect of this implementation on retention in ART 12 months after initiation, and explored the associated effects of COVID-19. METHODS: In this uncontrolled interrupted time-series analysis, data were extracted from the Mozambique ART database, which contains data on individuals in ART care from 1455 health facilities providing ART in Mozambique. We included individual-level data from facilities that were providing ART at the beginning of the study period (Jan 1, 2016) and at the start of DSDM implementation (Dec 1, 2018). We compared the proportion of individuals retained in ART 12 months after initiation between the periods before (Jan 1, 2017, to Nov 30, 2018) and after (Dec 1, 2019, to June 30, 2021) implementation of the DSDMs, overall and stratified by sex and age. We applied a generalised estimating equation model with a working independence correlation and cluster-robust standard errors to account for clustering at the facility level. In a secondary analysis, we assessed the effect of COVID-19 response measures during the post-intervention period on ART retention. FINDINGS: The study included 613 facilities and 1 131 118 individuals who started ART during the inclusion period up to June 30, 2020, of whom 79 178 (7·0%) were children (age ≤14 years), 226 224 (20·0%) were adolescents and young adults (age 15-24 years), and 825 716 (73·0%) were adults (age ≥25 years). 731 623 (64·7%) were female and 399 495 (35·3%) were male. Introduction of the DSDMs was associated with an estimated increase of 24·5 percentage points (95% CI 21·1 to 28·0) in 12-month ART retention by the end of the study period, compared with the counterfactual scenario without DSDM implementation. By age, the smallest effect was estimated in children (6·1 percentage points, 1·3 to 10·9) and the largest effect in adolescents and young adults (28·8 percentage points, 24·2 to 33·4); by sex, a larger effect was estimated in males (29·7 percentage points, 25·6 to 33·7). Our analysis showed that COVID-19 had an overall negative effect on 12-month retention in ART compared with a counterfactual scenario based on the post-intervention period without COVID-19 (-10·0 percentage points, -18·2 to -1·8). INTERPRETATION: The implementation of eight DSDMs for HIV treatment had a positive impact on 12-month retention in ART. COVID-19 negatively influenced this outcome. FUNDING: None. TRANSLATION: For the Portuguese translation of the abstract see Supplementary Materials section.

Development of novel composite data quality scores to evaluate facility-level data quality in electronic data in Kenya: a nationwide retrospective cohort study.

BACKGROUND: In this evaluation, we aim to strengthen Routine Health Information Systems (RHIS) through the digitization of data quality assessment (DQA) processes. We leverage electronic data from the Kenya Health Information System (KHIS) which is based on the District Health Information System version 2 (DHIS2) to perform DQAs at scale. We provide a systematic guide to developing composite data quality scores and use these scores to assess data quality in Kenya. METHODS: We evaluated 187 HIV care facilities with electronic medical records across Kenya. Using quarterly, longitudinal KHIS data from January 2011 to June 2018 (total N = 30 quarters), we extracted indicators encompassing general HIV services including services to prevent mother-to-child transmission (PMTCT). We assessed the accuracy (the extent to which data were correct and free of error) of these data using three data-driven composite scores: 1) completeness score; 2) consistency score; and 3) discrepancy score. Completeness refers to the presence of the appropriate amount of data. Consistency refers to uniformity of data across multiple indicators. Discrepancy (measured on a Z-scale) refers to the degree of alignment (or lack thereof) of data with rules that defined the possible valid values for the data. RESULTS: A total of 5,610 unique facility-quarters were extracted from KHIS. The mean completeness score was 61.1% [standard deviation (SD) = 27%]. The mean consistency score was 80% (SD = 16.4%). The mean discrepancy score was 0.07 (SD = 0.22). A strong and positive correlation was identified between the consistency score and discrepancy score (correlation coefficient = 0.77), whereas the correlation of either score with the completeness score was low with a correlation coefficient of -0.12 (with consistency score) and -0.36 (with discrepancy score). General HIV indicators were more complete, but less consistent, and less plausible than PMTCT indicators. CONCLUSION: We observed a lack of correlation between the completeness score and the other two scores. As such, for a holistic DQA, completeness assessment should be paired with the measurement of either consistency or discrepancy to reflect distinct dimensions of data quality. Given the complexity of the discrepancy score, we recommend the simpler consistency score, since they were highly correlated. Routine use of composite scores on KHIS data could enhance efficiencies in DQA at scale as digitization of health information expands and could be applied to other health sectors beyondHIV clinics.

Sample Size Calculations for Stepped Wedge Designs with Treatment Effects that May Change with the Duration of Time under Intervention.

The stepped wedge design is often used to evaluate interventions as they are rolled out across schools, health clinics, communities, or other clusters. Most models used in the design and analysis of stepped wedge trials assume that the intervention effect is immediate and constant over time following implementation of the intervention (the "exposure time"). This is known as the IT (immediate treatment effect) assumption. However, recent research has shown that using methods based on the IT assumption when the treatment effect varies over exposure time can give extremely misleading results. In this manuscript, we discuss the need to carefully specify an appropriate measure of the treatment effect when the IT assumption is violated and we show how a stepped wedge trial can be powered when it is anticipated that the treatment effect will vary as a function of the exposure time. Specifically, we describe how to power a trial when the exposure time indicator (ETI) model of Kenny et al. (Statistics in Medicine, 41, 4311-4339, 2022) is used and the estimand of interest is a weighted average of the time-varying treatment effects. We apply these methods to the ADDRESS-BP trial, a type 3 hybrid implementation study designed to address racial disparities in health care by evaluating a practice-based implementation strategy to reduce hypertension in African American communities.