Use of biometrics to evaluate intervention coverage and contamination in a cluster randomised trial in Zimbabwe

Use of biometrics to evaluate intervention coverage and contamination in a cluster randomised trial in Zimbabwe | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Use of biometrics to evaluate intervention coverage and contamination in a cluster randomised trial in Zimbabwe Victoria Simms, Tsitsi Bandason, Ethel Dauya, Chido Dziva Chikwari, and 11 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6758988/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Introduction Low intervention uptake and contamination can dilute effects of cluster randomised trials (CRTs) but can be difficult to assess. We investigated the feasibility of digital fingerprints to assess intervention coverage and contamination in a CRT of community-based integrated HIV and sexual and reproductive health services for youth (CHIEDZA). Methods 24 clusters in Zimbabwe were randomly allocated to intervention/control. In the 12 intervention clusters, services for youth aged 16-24 years were provided in community halls for 30 months. A population-based survey of youth aged 18-24 years (700/cluster) was conducted to ascertain impact on trial outcomes. Digital fingerprints were collected from service attendees and survey participants, and the two datasets were linked to assess intervention coverage at population level in intervention clusters, and contaminaton in control clusters. Multilevel logistic regression estimated the association of walking distance to the community hall with service uptake. Results Between April 2019-March 2022, 36,9991 clients attended the CHIEDZA service and 36,957 (99.9%) used biometric registration. In the survey 13,675/17,682 (77.3%) participants completed biometric registration: 1182 refused, 1235 bypassed registration and 1590 were unable to register. Database linkage showed that 23.1% of registered survey participants in the intervention clusters (coverage), and 3.7% of participants in control clusters (contamination) attended the CHIEDZA service. Sensitivity of self-reported service attendance against biometric registration match was 75.3% (95%CI 73.1-77.5) and specificity was 92.7% (95%CI 92.0-93.4). In intervention clusters, for every 1km increase in walking distance to the community hall, the odds of utilising the CHIEDZA service reduced by 52% (OR: 0.48 95%CI:0.44- 0.54). Conclusion Biometric identification was highly feasible and acceptable in a community setting with low time pressure. In population-based surveys additional technological challenges emerged. Biometrics enabled good estimation of intervention coverage and validated self-reported data. Community services for youth must overcome distance barriers and ensure communication. Trial registration https://clinicaltrials.gov/study/NCT03719521 registered 23 October 2018 Epidemiology biometrics cluster randomised trial community-based contamination coverage accessibility youth Zimbabwe Figures Figure 1 Key message What is already known on this topic: Digital biometrics have potential as a means of anonymous identification What this study adds: Biometric identification is acceptable and feaisble in a low-income community setting. Walking distance is a barrier to access to community services for young people. How this might affect research, practice or policy: Research in community settings may utilise biometric identification, particularly to assess intervention uptake in cluster-randomised trials Background Digital biometrics, such as fingerprint recognition, are increasingly utilised for precise identification and monitoring of individuals in targeted interventions within public health programmes to ensure efficient allocation of services 1 . Biometrics have potential for use among clients who do not wish to leave identifying information such as their name or address, and may be especially useful in settings where individuals might fear discrimination or might not have adequate documentation 1 . Digital fingerprint records have been used in clinical settings, for example in a group of 300 people living with HIV in Kenya 2 and for antenatal care in Malawi 3 . The technology has also been used in rural Ghana to link hospital and community records 4 . However, marginalised groups (including female sex workers in Zambia and people living with HIV) have expressed concerns about the use of biometric registration, such as infringement of privacy and exposure to risk of legal action 5 6 . A possible application of this technology is in the context of cluster-randomised trials (CRTs) of public health interventions. CRTs are particularly useful for evaluating interventions that can only be delivered at group or population level, and facilitate assessment of population-level efficacy 7 . Importantly, the efficacy of interventions will be driven not just by inherent effectiveness of an intervention but by its coverage. Contamination between intervention and control clusters can also undermine the efficacy estimated by a CRT. Intervention uptake is usually assessed by self-report which is vulnerable to desirability or observer bias. Digital biometrics offer an innovative and more objective method to assess intervention uptake in CRTs. The feasibility and acceptability of this approach is however not known. In Uganda, for example, digital fingerprinting was used for a study of TB contact tracing, but technological problems made implementation difficult 8 . We aimed to investigate the feasibility, accuracy and applicability of digital fingerprints in a CRT of community-based integrated HIV and sexual and reproductive health (SRH) services for youth in Zimbabwe. We hypothesized that provision of services which integrated both HIV and SRH, were youth-friendly and delivered outside a health facility setting would increase access, acceptability and uptake, and thus have population-level impact on health outcomes, which were assessed through a population-based survey 9 . We used digital fingerprints to register clients accessing the service (intervention) and to record each subsequent client visit. We also used digital fingerprints in the outcome survey, enabling us to link the datasets. As well as the feasibility and accuracy of the technology, we estimated uptake of the SRH service at population level in the intervention clusters (coverage) and also uptake in the control clusters (contamination). We investigated factors associated with intervention uptake. Methods Study design and setting CHIEDZA was a cluster-randomised trial conducted in three provinces in Zimbabwe. Details of the study design have been published 10 . Briefly, 24 clusters were randomised 1:1 to the trial intervention (community-based integrated HIV and SRH service delivered for 30 months) or to control (only existing, mainly clinic-based, services). The clusters were demarcated areas with a population of between 2500–4500 youth. All cluster residents aged 16–24 years were eligible to attend and receive services, which included HIV testing, and management including drugs, monitoring and adherence support, STI screening and treatment, contraception, condoms, menstrual health management and counselling, pregnancy testing and risk reduction and general health and relationship counselling. The service was operated from community halls equipped with entertainment facilities (sports, music etc) to create a welcoming environment, and small booth tents for privacy when providing services. A cluster was a demarcated area containing a community hall and a primary care clinic from where drugs e.g antiretroviral therapy could be obtained. Community mobilisers sensitised cluster residents on the availability of services. Biometric data collection and definition of cut-points Clients (those attending the CHIEDZA services) completed biometric registration after providing consent (recorded within the software), by taking prints from four fingers (left thumb, left index, right thumb and right index) using a fingerprint scanner, with an alternative option for manual registration using a paper-based form. Clients could attend the service more than once, and fingerprint identification was used to determine the date of registration and services previously received. Data were collected using Samsung Galaxy A7 10.4 (2020) wi-fi Android tablets, Bluetooth linked to the Simprints Vero fingerprint scanner with SurveyCTO for data collection and Simprints ID software to store the biometric data 1 . All fingerprint registrations were converted into a 32-character Globally Unique Identifier (GUID) which was stored in a SurveyCTO database. Actual fingerprint data were stored on the Simprints server and data were encrypted at every stage of data collection. To calculate the accuracy of the digital fingerprint system, a small dataset was created of 300 survey clients, each scanned twice. The biometric profiles of each pair of scans were compared to create an aggregated comparison score which was the mean comparison score of each of the 4 fingers. The false acceptance rate (FAR) and false rejection rate (FRR) were calculated at a range of different cutpoints of the comparison score. The optimal cutpoint was identified and then used in further analysis. Sensitivity analysis was conducted with a range of cutpoints. Population-based survey The primary outcome of the trial was prevalence of unsuppressed HIV viral load among youth living with HIV at population level, which was evaluated using a cross-sectional population-based survey of 700 youth aged 18–24 years per cluster, conducted after the end of the intervention period. Survey participants were also registered using Simprints in the same way as service clients. Survey participants were asked whether they had ever heard of CHIEDZA and, if so, whether they had ever accessed the service. Other variables collected included length of residence at current address ( 24 months), age, sex (male, female), marital status (never married, married or living together, divorced/widowed/separated), education level (no education, primary, secondary, tertiary), current main activity (none, in school, registered business/formal job, informal sector job) and ever had sexual intercourse (yes, no). For the survey, each cluster was mapped and then divided into sections of road of approximately 100-300m, except one cluster which had limited road infrastructure and was divided into smaller areas with an average perimeter of 3500m. Geographic Information Systems were used to randomly select road sections within each cluster which were overlaid on OpenStreet Map and satellite images. The size of sections was based on population distribution from the Zimbabwe 2012 Census. A sample of road sections were randomly selected within each cluster and all households in these sections were enumerated to identify all residents who were aged 18–24 years and hence eligible for the survey. If the sample was not adequate to reach a sample size of 700 youths, a further set of randomly selected road sections was taken and sections were visited sequentially to maintain the randomness until a sample size of 700 was reached. A median of 100 sections (IQR 88–148) were surveyed per cluster. Maps.me, a mobile application that provides offline maps using OpenStreetMap data was used for navigation to the exact location of the road section. All those eligible were contacted, if possible, and asked to participate in the survey and provide fingerprint registration. More details are provided in the trial protocol 10 . The coordinates of each household dwelling were collected using the in-built Global Positioning System (GPS) on the Android tablets. Initially, data collectors were unable to easily bypass fingerprint registration because all processes including bypass and refusals, were required to be processed through the Simprints ID application. On 21 April 2022 the bypass process was made easier in response to staff requests, due to challenges with scanners and the Simprints ID application. Data analysis Analysis was conducted using Stata v18.0 and R v4.3.3. The dataset of service clients was linked with the dataset of survey participants by fingerprint ID, within province. If a service client fingerprint was linked to more than one survey participant, these records were excluded because it was not possible to determine the correct identity. Feasibility was assessed from the proportion of survey participants who gave a digital fingerprint, the proportion of occasions when data collectors bypassed the process, the accuracy against self-reported attendance, and extent of duplicates. Chi-square tests were used to investigate whether demographic characteristics were associated with refusal of biometric registration. Intervention coverage was defined as the proportion of fingerprinted survey participants in the 12 intervention clusters who matched to the service client dataset. Self-reported CHIEDZA service attendance was compared with the fingerprint match to determine the accuracy of self-report. Contamination was defined as the proportion of fingerprinted survey participants in the 12 control clusters who matched to the service client dataset (Supplementary Fig. 1). Distance from each survey participant’s residence to the cluster community hall was calculated, both as Euclidean (straight line) distance and walking distance. Euclidean distance was calculated using the Stata vincenty module. Walking distance was calculated using DirectionFinder, a Google Apps script, based on the road infrastructure as recorded in Google Maps. Multi-level mixed-effects logistic regression was used to examine the association between utilisation of services and geographic walking distance and other covariates, accounting for cluster. Socioeconomic status was determined from asset ownership variables using factor analysis and divided into quintiles. Patient and public involvement The CHIEDZA service intervention was co-designed in collaboration with youth in two participatory workshops in Zimbabwe 11 . Digital biometrics were incorporated in the study design following discussion with youth. Participants were informed about the prevalence survey by watching a video which was designed in collaboration with youth. Results were disseminated to participant communities in collaboration with the youth advisory group. Results Feasibility and accuracy From 1 April 2019 to 31 March 2022, 36,991 unique clients attended the CHIEDZA service in the 12 intervention clusters, of whom 36,957 (99.9%) registered their fingerprints. Ninety-five clients (63 in Bulawayo, 17 in Mashonaland East and 15 in Harare) attended more than 1 CHIEDZA community centre; 93 attended 2 and 2 clients attended 3 centres. The start and end of service provision was staggered by province, and in each province the survey began within 1 month after the service closed. The survey took place from 4 October to 15 December 2021 in Harare, 4 January to 5 March 2022 in Bulawayo, and 4 April to 2 June 2022 in Mashonaland East. Data in Bulawayo were collected by 45 research staff who each worked for a median of 38 days (IQR 35–39, range 20–42). Another 45 staff collected the data in Harare and Mashonaland East, working for a median 82 days (IQR 76–85, range 18–89). Staff worked in pairs and each person enrolled a median of 3 clients per day (IQR 3–4, range 1–10). In total 17,682 participants were enrolled of whom 13,675 (77.3%) successfully registered their fingerprint (Fig. 1 ). In 4007 cases where the participant did not register their fingerprint the reasons were either that the researcher bypassed the process (N = 1235), the participant refused to give fingerprints (N = 975), or the registration process was begun but not completed (N = 1797). Of those who started the process but did not complete it, 207 refused, for a total of 1182 refusals (975 + 207). Out of the 207 participants who refused during registration, 95 refused for religious reasons, 66 because of data concerns, and 46 did not give a reason. Apart from refusal, the other causes of non-completion of biometric registration were: the scanner was not working (N = 723), the scanner did not connect to the tablet (N = 502), the scanner had low battery (N = 30), a software version incompatibility problem (N = 243), individual reasons which prevented fingerprint detection (e.g. the participant had grease on their hands) (N = 53), and 39 were for unknown reason. The version incompatibility problem occurred during the first 3 weeks of data collection in Harare province (Supplementary Table 1). Automatic software updates were disabled on the Android tablets, and as a result there were compatibility issues between newer scanners and the old software versions. The app recorded the event as ‘register biometrics complete’ although a biometric ID was not stored, so data collection staff did not realise that registration had failed. This problem was resolved by the data office on 23 October 2021 and did not re-occur. Until 20 April 2022 only 149 (1.1%) clients bypassed registration out of 13,367 recruited. From 21 April, when bypass was made easier, until the survey ended on 2 June 2022, 1086/4315 (25.2%) of survey participants bypassed fingerprint registration. Each data collector bypassed a median 7.4% of registrations (IQR 0–35%) during this time, and 4 data collectors bypassed more than 90%. Excluding those who bypassed, 667/6334 (10.5%) men refused vs 515/10,112 (5.1%) women(Table 1 ). There was no difference in refusal by age. Refusal among eligible participants was 9.9% in Bulawayo and Mashonaland East but 2.3% in Harare. In the intervention clusters, fingerprint refusal was 3.8% among survey participants who self-reported accessing CHIEDZA services versus 8.9% among participants who said they had not accessed services. Refusal by cluster ranged from 0.3–27.5% (median 6.7%, IQR 2.3%-10.0%). In the test dataset the optimal cutpoint for comparison score was determined to be 21.5. At this cutpoint the FAR was 0.01% and the FRR was 0.94% (Supplementary Fig. 2). At a cutpoint of 21.5, 128 service clients appeared to match > 1 survey participant. These records were excluded from the dataset of 36,991 service clients, leaving 36,863. Out of 1826 survey participants who linked to the service dataset, 10 participant fingerprints (0.5%) linked to > 1 client (in 2 cases to 3 clients, in 8 cases to 2 clients). On checking, these client pairs usually shared a birthdate and were at the same cluster, and on one occasion were seen on the same day. The most likely explanation is that these were visits by the same client, who was not recognised as a match during the biometric registration and was re-registered. In these cases, one record from each group was retained to link with the survey participant. Intervention Coverage Using a cutpoint of 21.5, coverage in the intervention arm was 23.1% (95%CI 22.1–24.1) (Table 2 ). Uptake was higher in women than men, and higher in Bulawayo than the other provinces. Overall, 10.5% of matches in the intervention arm were cross-cluster; the survey participant attended CHIEDZA services not in their cluster of residence, but in one of the other 7 intervention clusters within the province. Participants who had lived at their current address for a shorter time were less likely to have attended the CHIEDZA service, and those who had previously lived in a different area or city were more likely to have attended a CHIEDZA service in a different cluster (Supplementary Fig. 3). Table 2 Service uptake among prevalence survey participants who completed biometric registration (N = 13675); using cutpoint 21.5) N Intervention clusters N Control clusters Total 6806 1574 (23.1%) 6869 252 (3.7%) Sex Male 2403 460 (19.1%) 2625 54 (2.1%) Female 4403 1114 (25.3%) 4243 198 (4.7%) Age 18–20 3579 852 (23.8%) 3575 125 (3.5%) 21–24 3227 722 (22.4%) 3294 127 (3.9%) Province Harare 2592 471 (18.2%) 2350 73 (3.1%) Bulawayo 2390 790 (33.1%) 2481 122 (4.9%) Mash East 1824 313 (17.2%) 2038 57 (2.8%) Self-reported accessing CHIEDZA service No 5238 388 (7.4%) 6822 224 (3.3%) Yes 1568 1186 (75.6%) 47 28 (59.6%) Among survey participants with biometric registration, the positive predictive value of self-report in the intervention arm against fingerprint match was 75.6% (95%CI 73.4–77.7) and negative predictive value was 92.6% (95%CI 91.8–93.3). Sensitivity of self-report was 75.3% (95%CI 73.1–77.5) and specificity was 92.7% (95%CI 92.0-93.4). Contamination At a cutpoint of 21.5, uptake in the control clusters was 3.7% (95%CI 3.2–4.1). As with the intervention clusters, uptake was higher in women and in Bulawayo. In the control arm 47 participants self-reported accessing CHIEDZA services, of whom 28 (59.6%) had a fingerprint match. Among the 1574 survey participants in the intervention clusters with a fingerprint match to the dataset of CHIEDZA service attendees at a comparison score > 21.5, the median comparison score was 55.7 (IQR 41.1–70.5, range 21,5-142.7). Among the corresponding 252 control arm participants with a comparison score > 21.5, the median score was 22.2 (IQR 21.9–24.1, range 21.5-122.8) Results of sensitivity analysis are shown in Supplementary Table 2. At a cutpoint of 23, uptake was 21.3% in the intervention clusters (coverage) and 1.3% in the control clusters (contamination). Raising the cutpoint to 23 compared to 21.5 resulted in a 7.8% reduction in the number of service clients identified in the intervention arm, and a 65% reduction in the number of service clients identified in the control arm. Factors associated with service uptake In the intervention clusters 8885 participants were enrolled. Five participants did not have an accurate GPS reading for their location and 2805 did not have fingerprint data, leaving 6805. (Table 3 ). A higher proportion of females (25.3%) than males (19.1%) accessed CHIEDZA services. Univariable multilevel mixed-effects logistic regression showed that utilisation of CHIEDZA services was associated with being female (OR 1.55, 95%CI 1.37–1.76) and with being married or cohabiting (OR 1.21, 95%CI 1.21–1.26) versus never married (Table 3 ). Table 3 Utilisation of CHIEDZA services in intervention arm by sociodemographic characteristics Utilised CHIEDZA service Univariable crude OR adjusted for clustering (95%CI) p-value Characteristic All No, n (%) Yes, n (%) 6805 n = 5233 (76.9%) n = 1572 (23.1%) Euclidean distance from home to CHIEDZA location (km) 1 Mean (SD) 1.01 (0.48) 0.84 (0.31) 0.52 (0.46–0.59) < 0.0001 Walking distance from home to CHIEDZA location (km) 1 Mean (SD) 1.42 (0.55) 1.17 (0.39) 0.48 (0.44–0.54) < 0.0001 Duration of residence at current address < 12 months 1728 (25.4) 208 (12.0) 1 2 years 4376 (64.3) 1231 (28.1) 2.57 (2.18–3.03) Sex Male 2402 (35.3) 458 (19.1) 1 < 0.0001 Female 4403 (64.7) 1114 (25.3) 1.55 (1.37–1.76) Marital Status Never married 5037 (74.0) 1173 (23.3) 1 0.039 Married/living together 1479 (21.7) 335 (22.7) 1.21 (1.04–1.40) Divorced/widowed/separated 289 (4.3) 64 (22.1) 1.13 (0.85–1.52) Current main activity None 3343 (49.1) 790 (23.6) 1 0.625 Education 1927 (28.3) 446 (23.1) 0.93 (0.81–1.06) Registered business/Formal sector work 312 (4.6) 83 (26.6) 1.04 (0.80–1.37) Informal sector work 1223 (18.0) 253 (20.7) 0.94 (0.80–1.11) Level of education None, any primary or completed primary 365 (5.4) 75 (20.6) 1 0.0006 Any secondary or completed secondary 5892 (86.6) 1402 (23.8) 1.19 (0.91–1.55) Any post-secondary 548 (8.0) 95 (17.3) 0.76 (0.54–1.08) Ever had sexual intercourse No 2287 (35.1) 476 (19.9) 1 0.0002 Yes 4379 (64.3) 1082 (24.7) 1.28 (1.13–1.45) Do not want to say 39 (0.6) 14 (35.9) 1.96 (0.99–3.90) Age in years 18–20 3579 (52.6) 850 (54.1) 1 0.309 21–24 3226 (47.4) 722 (45.9) 0.94 (0.84–1.06) 1 increased odds per increased kilometre of distance For every km increase in walking distance, the odds of utilisation were reduced by 52% (OR 0.48, 95%CI 0.44–0.54) and a similar margin of effect was observed using Euclidian distance (OR 0.52, 95%CI 0.46–0.59). Although the mean walking distance to the CHIEDZA community halls was less than 3km (Supplementary Fig. 2), mean walking distance of those who utilised CHIEDZA services was on average 0.25km shorter than those in the same cluster who did not, and cluster-adjusted Euclidian distance was 0.17km shorter for service attendees than non-attendees. In multivariable sex-stratified multilevel mixed effects logistic regression models, utilisation of CHIEDZA services was associated with longer duration of residence both for males and females after adjusting for sexual debut, education and distance (Table 4 ). In males only, sexual debut was associated with increased odds of utilisation of CHIEDZA services (AOR 2.01, 95%CI 1.58–2.57), with statistically significant evidence for interaction by sex (p < 0.001). Females who had post-secondary education were less likely to have utilised CHIEDZA services (AOR 0.56, 95%CI 0.37–0.86) compared to females with primary level of education, but there was no evidence of an interaction effect by sex. Table 4 Association of utilisation of CHIEDZA service with walking distance and length of residence, for outcome survey male and female participants in the intervention arm Males Female Characteristic Adjusted OR (95%CI) p-value Adjusted OR (95%CI) p-value Walking distance (km) 0.48 (0.40–0.58) < 0.0001 0.49 (0.43–0.55) < 0.0001 Ever had sexual intercourse No 1 < 0.0001 1 0.401 Yes 2.01 (1.58–2.57) 1.15 (0.99–1.35) Do not want to say 3.14 (1.10–8.96) 1.71 (0.64–4.54) Level of education None, any primary or completed primary 1 0.0686 1 0.007 Any secondary or completed secondary 1.69 (0.93–3.09) 1.01 (0.74–1.39) Any post-secondary 1.08 (0.53–2.18) 0.56 (0.37–0.86) Duration of residence at current address < 12 months 1 < 0.0001 1 2 years 3.10 (2.05–4.69 2.99 (2.48–3.61) Discussion Our study showed that the use of digital fingerprints was feasible and accurate, and these biometrics can be applied to understand intervention coverage as well as contamination, two aspects that critically affect the measured efficacy of interventions. Digital fingerprints had < 0.1% refusal in the CHIEDA service, but we observed a higher refusal among survey participants. Notably there was generally good agreement of biometrics with self-reported attendance. There are several reasons why uptake of biometric registration was higher at the CHIEDZA services than during the prevalence survey. Each community centre had several tablets and scanners, so a nonfunctional piece of equipment could quickly be exchanged for a working one. Attendees may have been motivated to use biometric registration in order to access services. Manual registration was possible but was considered a tedious and time-consuming process, so staff preferred to use biometric registration whenever possible. Handwashing facilities were also readily available at community halls. By contrast, in the cross-sectional survey, there were operational challenges. The two-person teams moved from house to house, had no backup equipment with them, and were under time pressure to complete the survey. Scanner and tablet had to be reconnected frequently. New equipment was purchased as required but was not always immediately compatible with the older equipment. Survey participants had no particular motivation to consent to biometric registration. Participants might not be able to wash their hands, which could result in poor scans. The manual registration process was deliberately made easier, with the result that staff were more likely to opt for it when they experienced or anticipated technical challenges. Close observation of the data in real time is essential to identify and resolve problems. This was particularly evident at two periods in the prevalence survey; noticing and resolving the version compatibility problem which prevented storage of a biometric ID, and observing the difference in bypass rates by study staff ID. The bypass option was introduced because survey staff reported frustrations trying to complete the registration process when they were faced with challenges such as failure to connect and risked falling behind on their recruitment targets. However, once the bypass option was available, a small number of survey staff used it routinely rather than attempting biometric registration of participants. It is necessary to observe the data at sufficiently granular level and in real-time to identify these problems quickly and intervene. In the survey, men were more likely to refuse fingerprinting than women, and those in Harare were less likely to refuse than residents in Bulawayo or Mashonaland East. Refusal was most common in those with either less than primary education or postgraduate education, and in the lowest and highest socioeconomic quintiles. This bimodal distribution suggests two separate mechanisms may lead to refusal. However, because of its large sample size the study is highly powered to detect small differences in refusal rates between groups, and some observed effects may be due to chance. We demonstrate the applicability of fingerprints in understanding coverage of the intervention. As shown in other studies, females were more likely to use these services, and those who were not sexually active would have found these services less relevant to them. 12 13 . We observed much lower population-level coverage of the intervention than anticipated. Those who had been resident in the intervention clusters for more than two years were more likely to have accessed the intervention. Youth in this setting are a highly mobile group and out-migration from the clusters may explain the discrepancy between the large number of clients who attended CHIEDZA, as a proportion of residents, and the low coverage of the intervention among survey participants. Even in clusters that were small in area, we found a strong relationship between distance and uptake. The greater the distance from the hall, the lower the probability of awareness or utilisation of the service, consistent with the well documented inverse relationship between healthcare utilisation and distance 14 15 . Nearly two-thirds of survey participants did not know that the community-based youth-friendly SRH services existed in their area and, even among those who knew about their existence, nearly half did not utilise them. This finding underscores the need for intensive sensitisation to increase utilisation and is supported by findings from other studies where utilisation was low due to lack of knowledge 12 . While community mobilisers were used, it is possible that they either operated nearer the halls and/or clients living further away were less willing to come. Advertisement of services using radio, television and social media was not undertaken to reduce the risk of contamination across clusters by giving information about CHIEDZA services to those in control clusters. Three quarters (75.0%) of CHIEDZA clients were female, yet service uptake in the survey participants was only 32% higher in females than males (25.3% of females and 19.1% of males accessed services), not three times higher, as might be expected. This difference may partly be attributable to the cluster populations being skewed female (10,741/17,682 survey participants were female), and also to women being a more mobile population. Among all survey participants, 29.7% of women (3191/10741) and 15.2% of men (1054/6940) had resided at their current address less than 12 months, while less than half of women (49.3%) and more than two thirds of men (68.3%) had been resident more than 3 years. As a result, women in the survey had had less exposure time to CHIEDZA. The uptake results in the control arm were particularly sensitive to comparison score cutpoint. Most of the matches in the control arm had low comparison scores, so when the cutpoint was raised they were no longer categorised as matches. As such it may be more appropriate to use different cutpoints in the intervention and control arms, since prior knowledge that a match is more likely in the intervention arm was not taken into account for the probability assessment. However, such a step would require careful justification in an RCT. The main strength of this study was that it was based on data from a large randomly selected sample which was representative of the general population sample of youths from three provinces of Zimbabwe from both urban and peri-urban communities, increasing generalisability of the findings. The study also has limitations. There is a lack of information around what caused the ‘scanner not working’ or ‘scanner not connecting to tablet’ errors which prevented biometric registration of a large number of survey participants. Conclusions Digital fingerprinting offers a highly feasible and effective approach to evaluate intervention delivery within cluster randomised trials. There are additional challenges to using this technology in house-to-house surveys. Steps that could be taken to address them include carrying spare equipment, and moving biometric registration to the end of the questionnaire to prevent wasted time during connection. Declarations Author contributions RAF conceptualised the study. ED, CDC, CM, MT and CG assisted with management of fieldwork and data collection. OM and TA provided support and guidance on trial implementation, and RH and KK provided expertise on trial design and analysis. RA, DD and AS supported biometrics data management and analysis. VS and TB conducted statistical analysis and drafted the manuscript. RAF is the Principal Investigator of the trial. All authors contributed to interpretation of the results and have read and approved the final manuscript Competing interests RF, DD and AS are employed by Simprints. Data availability Data for this study are available upon reasonable request to the principal authors. Ethics approval and consent to participate Ethical approval for the trial including the survey was obtained from the Biomedical Research and Training Institute Institutional Review Board (AP149/2018), the Medical Research Council of Zimbabwe (MCRZ/A/2387) and the London School of Hygiene & Tropical Medicine ethics committee (16124/RR/11602). Consent for publications Not applicable Funding Financial support for this study was provided through the Wellcome Trust to RAF (Grant number: 206316_Z_17_Z). The funders had no role in study design, analysis and decision to publish or preparation of the manuscript. References Storisteanu DM, Norman TL, Grigore A, et al. Biometric fingerprint system to enable rapid and accurate identification of beneficiaries. Glob Health Sci Pract 2015;3(1):135-7. doi: 10.9745/GHSP-D-15-00010 [published Online First: 2015/03/07] Jaafa NK, Mokaya B, Savai SM, et al. Implementation of Fingerprint Technology for Unique Patient Matching and Identification at an HIV Care and Treatment Facility in Western Kenya: Cross-sectional Study. J Med Internet Res 2021;23(12):e28958. doi: 10.2196/28958 [published Online First: 2021/12/24] Bengtson AM, Kumwenda W, Lurie M, et al. Improving Monitoring of Engagement in HIV Care for Women in Option B+: A Pilot Test of Biometric Fingerprint Scanning in Lilongwe, Malawi. AIDS Behav 2020;24(2):551-59. doi: 10.1007/s10461-019-02748-6 [published Online First: 2019/11/28] Odei-Lartey EO, Boateng D, Danso S, et al. The application of a biometric identification technique for linking community and hospital data in rural Ghana. Glob Health Action 2016;9:29854. doi: 10.3402/gha.v9.29854 [published Online First: 2016/03/20] Abrams MP, Torres FE, Little SJ. Biometric Registration to an HIV Research Study may Deter Participation. AIDS Behav 2021;25(5):1552-59. doi: 10.1007/s10461-020-02995-y [published Online First: 2020/08/09] Wall KM, Kilembe W, Inambao M, et al. Implementation of an electronic fingerprint-linked data collection system: a feasibility and acceptability study among Zambian female sex workers. Global Health 2015;11:27. doi: 10.1186/s12992-015-0114-z [published Online First: 2015/06/28] Dron L, Taljaard M, Cheung YB, et al. The role and challenges of cluster randomised trials for global health. Lancet Glob Health 2021;9(5):e701-e10. doi: 10.1016/S2214-109X(20)30541-6 [published Online First: 2021/04/19] White EB, Meyer AJ, Ggita JM, et al. Feasibility, Acceptability, and Adoption of Digital Fingerprinting During Contact Investigation for Tuberculosis in Kampala, Uganda: A Parallel-Convergent Mixed-Methods Analysis. J Med Internet Res 2018;20(11):e11541. doi: 10.2196/11541 [published Online First: 2018/11/18] World Health Organisation. Making Health Services Adolescent Friendly: Devloping National Quality Standards for Adolescent Friendly Health Services. Geneva, 2012. Dziva Chikwari C, Dauya E, Bandason T, et al. The impact of community-based integrated HIV and sexual and reproductive health services for youth on population-level HIV viral load and sexually transmitted infections in Zimbabwe: protocol for the CHIEDZA cluster-randomised trial [version 2; peer review: 2 approved]. Wellcome Open Research 2022;7(54):https://doi.org/10.12688/wellcomeopenres.7530.2. doi: https://doi.org/10.12688/wellcomeopenres.17530.1 Mackworth-Young CR, Dringus S, Dauya E, et al. Putting youth at the centre: co-design of a community-based intervention to improve HIV outcomes among youth in Zimbabwe [version 2; peer review: 1 approved]. Wellcome Open Research 2022;7(53) doi: https://doi.org/10.12688/wellcomeopenres.17531.2 Ninsiima LR, Chiumia IK, Ndejjo R. Factors influencing access to and utilisation of youth-friendly sexual and reproductive health services in sub-Saharan Africa: a systematic review. Reprod Health 2021;18(1):135. doi: 10.1186/s12978-021-01183-y [published Online First: 2021/06/29] Sharma M, Khatri B, Amatya A, et al. Utilization of adolescent friendly health services and its associated factors among higher secondary students in mid-western Himalayan mountainous district of Nepal. PLOS Glob Public Health 2023;3(3):e0001616. doi: 10.1371/journal.pgph.0001616 [published Online First: 2023/03/25] Siedner MJ, Lankowski A, Tsai AC, et al. GPS-measured distance to clinic, but not self-reported transportation factors, are associated with missed HIV clinic visits in rural Uganda. AIDS 2013;27(9):1503-8. doi: 10.1097/QAD.0b013e32835fd873 [published Online First: 2013/02/26] Lankowski AJ, Siedner MJ, Bangsberg DR, et al. Impact of geographic and transportation-related barriers on HIV outcomes in sub-Saharan Africa: a systematic review. AIDS Behav 2014;18(7):1199-223. doi: 10.1007/s10461-014-0729-8 [published Online First: 2014/02/25] Table Table 1 is available in the Supplementary Files section. Additional Declarations The authors declare potential competing interests as follows: Supplementary Files SupplementaryTable.docx SuppFigure1.jpg Supplementary Figure 1: CHIEDZA Participant Fingerprinting Process SuppFigure2.png Supplementary Figure 2: False Acceptance Rate and False Reject Rate at different thresholds (cut-offs in red) SuppFigure3.png Supplementary Figure 3: Mean walking distance to CHIEDZA service location by uptake of service Table1.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6758988","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":463376746,"identity":"a9b0a3bf-b6da-4031-944b-5fd723e47fc4","order_by":0,"name":"Victoria Simms","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCUlEQVRIie2RMUsDMRTHXwhkepg1Bel9hdwiDnp+lR4Bp/oNRCLFjHY98UuIs0MkaJdq18A5dDrXc3EqYnq1BeHOWx3yg/fIg/eD/yMAkch/hDZdhCI6tKP95t2Ancp6daecIgjao/ysbnH9ipzR9+XHw2EC3F3d+9UCk9vp0x2cZyDntl1xLL0sKpFqkZvyzJQo3xzz8KxAvuhWZTBBMkEriBYkKDooQgWFWZCL9mBb5UTzR1OOV6+YFGvlq1vhdKPkGkKwMbMIPijE2M5gnLL0prBCmc0tCqVXBz6/VjjoOJ9xt6xre3E85bMqBMuGSZFXvv7MhnvzUXuynftrGv31kZFIJBLp5RuJPlwM5A5G3QAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-4897-458X","institution":"London School of Hygiene \u0026 Tropical Medicine","correspondingAuthor":true,"prefix":"","firstName":"Victoria","middleName":"","lastName":"Simms","suffix":""},{"id":463376747,"identity":"5a5f6120-76c1-47cf-bbda-63e5e3bcdead","order_by":1,"name":"Tsitsi Bandason","email":"","orcid":"https://orcid.org/0000-0003-0066-8345","institution":"Biomedical Research and Training Institute","correspondingAuthor":false,"prefix":"","firstName":"Tsitsi","middleName":"","lastName":"Bandason","suffix":""},{"id":463376748,"identity":"75b33632-0165-4ce0-a67e-d169eca6a58d","order_by":2,"name":"Ethel Dauya","email":"","orcid":"https://orcid.org/0000-0003-3241-2512","institution":"Biomedical Research and Training Institute","correspondingAuthor":false,"prefix":"","firstName":"Ethel","middleName":"","lastName":"Dauya","suffix":""},{"id":463376749,"identity":"1693e68a-2f85-48ae-9ad8-5658ab3175db","order_by":3,"name":"Chido Dziva Chikwari","email":"","orcid":"https://orcid.org/0000-0003-1617-3603","institution":"London School of Hygiene \u0026 Tropical Medicine","correspondingAuthor":false,"prefix":"","firstName":"Chido","middleName":"Dziva","lastName":"Chikwari","suffix":""},{"id":463376750,"identity":"c17d31f3-8767-4569-b5f6-7d08ef7b6012","order_by":4,"name":"Chris Grundy","email":"","orcid":"https://orcid.org/0000-0002-7042-2394","institution":"London School of Hygiene \u0026 Tropical Medicine","correspondingAuthor":false,"prefix":"","firstName":"Chris","middleName":"","lastName":"Grundy","suffix":""},{"id":463376751,"identity":"28200c83-85d6-46f9-bfb7-941f1eb1af48","order_by":5,"name":"Mandikudza Tembo","email":"","orcid":"https://orcid.org/0000-0002-4520-3317","institution":"London School of Hygiene \u0026 Tropical Medicine","correspondingAuthor":false,"prefix":"","firstName":"Mandikudza","middleName":"","lastName":"Tembo","suffix":""},{"id":463376752,"identity":"4e819020-f209-4cbd-bf80-6f95188d2ad2","order_by":6,"name":"Constancia Mavodza","email":"","orcid":"","institution":"London School of Hygiene \u0026 Tropical Medicine","correspondingAuthor":false,"prefix":"","firstName":"Constancia","middleName":"","lastName":"Mavodza","suffix":""},{"id":463376753,"identity":"d7c86570-d045-4874-aee3-93cc11a5da2a","order_by":7,"name":"Owen Mugurungi","email":"","orcid":"","institution":"Zimbabwe Ministry of Health and Child Care","correspondingAuthor":false,"prefix":"","firstName":"Owen","middleName":"","lastName":"Mugurungi","suffix":""},{"id":463376754,"identity":"d177542e-7561-43ab-b271-7a5c74206344","order_by":8,"name":"Tsitsi Apollo","email":"","orcid":"","institution":"Zimbabwe Ministry of Health and Child Care","correspondingAuthor":false,"prefix":"","firstName":"Tsitsi","middleName":"","lastName":"Apollo","suffix":""},{"id":463376755,"identity":"02590b1b-675e-48b1-b2d2-75045ff5ac06","order_by":9,"name":"Riffat Ashrafee","email":"","orcid":"","institution":"Simprints","correspondingAuthor":false,"prefix":"","firstName":"Riffat","middleName":"","lastName":"Ashrafee","suffix":""},{"id":463376756,"identity":"e0cd0fa8-1f03-42b9-9236-99d7baf2974b","order_by":10,"name":"Demetris Demetriou","email":"","orcid":"","institution":"Simprints","correspondingAuthor":false,"prefix":"","firstName":"Demetris","middleName":"","lastName":"Demetriou","suffix":""},{"id":463376757,"identity":"676a4ef8-630b-4424-8ace-fc408ab7821d","order_by":11,"name":"Anindya Sharma","email":"","orcid":"","institution":"Simprints","correspondingAuthor":false,"prefix":"","firstName":"Anindya","middleName":"","lastName":"Sharma","suffix":""},{"id":463376758,"identity":"0fca72f6-64e5-4867-b0e7-01135894544c","order_by":12,"name":"Richard J. Hayes","email":"","orcid":"https://orcid.org/0000-0002-1729-9892","institution":"London School of Hygiene \u0026 Tropical Medicine","correspondingAuthor":false,"prefix":"","firstName":"Richard","middleName":"J.","lastName":"Hayes","suffix":""},{"id":463376759,"identity":"7af640f7-990f-4380-82ec-9e3abd356044","order_by":13,"name":"Katharina Kranzer","email":"","orcid":"https://orcid.org/0000-0001-5691-7270","institution":"London School of Hygiene \u0026 Tropical Medicine","correspondingAuthor":false,"prefix":"","firstName":"Katharina","middleName":"","lastName":"Kranzer","suffix":""},{"id":463376760,"identity":"da643539-e9cb-43e1-a348-0a84c6997946","order_by":14,"name":"Rashida A. Ferrand","email":"","orcid":"https://orcid.org/0000-0002-7660-9176","institution":"London School of Hygiene \u0026 Tropical Medicine","correspondingAuthor":false,"prefix":"","firstName":"Rashida","middleName":"A.","lastName":"Ferrand","suffix":""}],"badges":[],"createdAt":"2025-05-27 11:38:19","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":true,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":true,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6758988/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6758988/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":83898909,"identity":"de789990-0140-4c3a-9b70-082677525872","added_by":"auto","created_at":"2025-06-04 09:09:32","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":230838,"visible":true,"origin":"","legend":"\u003cp\u003eSurvey enrolment flowchart\u003c/p\u003e","description":"","filename":"Figure11page00011.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6758988/v1/f080da0b6e14f4d1457c997f.jpg"},{"id":83901306,"identity":"c9c7a695-61be-4bc9-8b90-f16c168ff808","added_by":"auto","created_at":"2025-06-04 09:33:32","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1320246,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6758988/v1/307e42f2-a903-4866-a692-84ecbf41f0d0.pdf"},{"id":83898908,"identity":"80a6c4d3-a175-4fa3-9f41-196710d0bd38","added_by":"auto","created_at":"2025-06-04 09:09:32","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":18584,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable.docx","url":"https://assets-eu.researchsquare.com/files/rs-6758988/v1/98d41bdd715c4c6701500229.docx"},{"id":83899461,"identity":"b69db4a0-36f2-45d3-8c13-1df1069f3d23","added_by":"auto","created_at":"2025-06-04 09:17:32","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":167589,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 1: CHIEDZA Participant Fingerprinting Process\u003c/p\u003e","description":"","filename":"SuppFigure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6758988/v1/eee4a679c3a01ee7a0963771.jpg"},{"id":83899459,"identity":"9c4ced83-b2f1-4013-8c0f-512fed268d8b","added_by":"auto","created_at":"2025-06-04 09:17:32","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":17130,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 2: False Acceptance Rate and False Reject Rate at different thresholds (cut-offs in red)\u003c/p\u003e","description":"","filename":"SuppFigure2.png","url":"https://assets-eu.researchsquare.com/files/rs-6758988/v1/7078a68e4796b795f69f1429.png"},{"id":83900904,"identity":"30ca2424-0eb1-4dc5-bc91-2ea38994a3c1","added_by":"auto","created_at":"2025-06-04 09:25:32","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":65330,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 3: Mean walking distance to CHIEDZA service location by uptake of service\u003c/p\u003e","description":"","filename":"SuppFigure3.png","url":"https://assets-eu.researchsquare.com/files/rs-6758988/v1/bedf7bba4b89888df471f8db.png"},{"id":83899458,"identity":"e0867c43-9677-4d18-ae30-ad63a174c27b","added_by":"auto","created_at":"2025-06-04 09:17:32","extension":"docx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":17579,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.docx","url":"https://assets-eu.researchsquare.com/files/rs-6758988/v1/bbef9ae425f4003814c24f4b.docx"}],"financialInterests":"The authors declare potential competing interests as follows: ","formattedTitle":"Use of biometrics to evaluate intervention coverage and contamination in a cluster randomised trial in Zimbabwe","fulltext":[{"header":"Key message ","content":"\u003cp\u003eWhat is already known on this topic: Digital biometrics have potential as a means of anonymous identification\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhat this study adds: Biometric identification is acceptable and feaisble in a low-income community setting. Walking distance is a barrier to access to community services for young people.\u003c/p\u003e\n\u003cp\u003eHow this might affect research, practice or policy: Research in community settings may utilise biometric identification, particularly to assess intervention uptake in cluster-randomised trials\u003c/p\u003e"},{"header":"Background","content":"\u003cp\u003eDigital biometrics, such as fingerprint recognition, are increasingly utilised for precise identification and monitoring of individuals in targeted interventions within public health programmes to ensure efficient allocation of services \u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Biometrics have potential for use among clients who do not wish to leave identifying information such as their name or address, and may be especially useful in settings where individuals might fear discrimination or might not have adequate documentation\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Digital fingerprint records have been used in clinical settings, for example in a group of 300 people living with HIV in Kenya\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e and for antenatal care in Malawi\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. The technology has also been used in rural Ghana to link hospital and community records\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. However, marginalised groups (including female sex workers in Zambia and people living with HIV) have expressed concerns about the use of biometric registration, such as infringement of privacy and exposure to risk of legal action\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eA possible application of this technology is in the context of cluster-randomised trials (CRTs) of public health interventions. CRTs are particularly useful for evaluating interventions that can only be delivered at group or population level, and facilitate assessment of population-level efficacy\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Importantly, the efficacy of interventions will be driven not just by inherent effectiveness of an intervention but by its coverage. Contamination between intervention and control clusters can also undermine the efficacy estimated by a CRT. Intervention uptake is usually assessed by self-report which is vulnerable to desirability or observer bias. Digital biometrics offer an innovative and more objective method to assess intervention uptake in CRTs. The feasibility and acceptability of this approach is however not known. In Uganda, for example, digital fingerprinting was used for a study of TB contact tracing, but technological problems made implementation difficult \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWe aimed to investigate the feasibility, accuracy and applicability of digital fingerprints in a CRT of community-based integrated HIV and sexual and reproductive health (SRH) services for youth in Zimbabwe. We hypothesized that provision of services which integrated both HIV and SRH, were youth-friendly and delivered outside a health facility setting would increase access, acceptability and uptake, and thus have population-level impact on health outcomes, which were assessed through a population-based survey\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. We used digital fingerprints to register clients accessing the service (intervention) and to record each subsequent client visit. We also used digital fingerprints in the outcome survey, enabling us to link the datasets. As well as the feasibility and accuracy of the technology, we estimated uptake of the SRH service at population level in the intervention clusters (coverage) and also uptake in the control clusters (contamination). We investigated factors associated with intervention uptake.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy design and setting\u003c/h2\u003e \u003cp\u003eCHIEDZA was a cluster-randomised trial conducted in three provinces in Zimbabwe. Details of the study design have been published \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Briefly, 24 clusters were randomised 1:1 to the trial intervention (community-based integrated HIV and SRH service delivered for 30 months) or to control (only existing, mainly clinic-based, services). The clusters were demarcated areas with a population of between 2500\u0026ndash;4500 youth. All cluster residents aged 16\u0026ndash;24 years were eligible to attend and receive services, which included HIV testing, and management including drugs, monitoring and adherence support, STI screening and treatment, contraception, condoms, menstrual health management and counselling, pregnancy testing and risk reduction and general health and relationship counselling. The service was operated from community halls equipped with entertainment facilities (sports, music etc) to create a welcoming environment, and small booth tents for privacy when providing services. A cluster was a demarcated area containing a community hall and a primary care clinic from where drugs e.g antiretroviral therapy could be obtained. Community mobilisers sensitised cluster residents on the availability of services.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eBiometric data collection and definition of cut-points\u003c/h3\u003e\n\u003cp\u003eClients (those attending the CHIEDZA services) completed biometric registration after providing consent (recorded within the software), by taking prints from four fingers (left thumb, left index, right thumb and right index) using a fingerprint scanner, with an alternative option for manual registration using a paper-based form. Clients could attend the service more than once, and fingerprint identification was used to determine the date of registration and services previously received. Data were collected using Samsung Galaxy A7 10.4 (2020) wi-fi Android tablets, Bluetooth linked to the Simprints Vero fingerprint scanner with SurveyCTO for data collection and Simprints ID software to store the biometric data\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. All fingerprint registrations were converted into a 32-character Globally Unique Identifier (GUID) which was stored in a SurveyCTO database. Actual fingerprint data were stored on the Simprints server and data were encrypted at every stage of data collection.\u003c/p\u003e \u003cp\u003eTo calculate the accuracy of the digital fingerprint system, a small dataset was created of 300 survey clients, each scanned twice. The biometric profiles of each pair of scans were compared to create an aggregated comparison score which was the mean comparison score of each of the 4 fingers. The false acceptance rate (FAR) and false rejection rate (FRR) were calculated at a range of different cutpoints of the comparison score. The optimal cutpoint was identified and then used in further analysis. Sensitivity analysis was conducted with a range of cutpoints.\u003c/p\u003e\n\u003ch3\u003ePopulation-based survey\u003c/h3\u003e\n\u003cp\u003eThe primary outcome of the trial was prevalence of unsuppressed HIV viral load among youth living with HIV at population level, which was evaluated using a cross-sectional population-based survey of 700 youth aged 18\u0026ndash;24 years per cluster, conducted after the end of the intervention period. Survey participants were also registered using Simprints in the same way as service clients. Survey participants were asked whether they had ever heard of CHIEDZA and, if so, whether they had ever accessed the service. Other variables collected included length of residence at current address (\u0026lt;\u0026thinsp;12 months, 12\u0026ndash;24 months and \u0026gt;\u0026thinsp;24 months), age, sex (male, female), marital status (never married, married or living together, divorced/widowed/separated), education level (no education, primary, secondary, tertiary), current main activity (none, in school, registered business/formal job, informal sector job) and ever had sexual intercourse (yes, no).\u003c/p\u003e \u003cp\u003eFor the survey, each cluster was mapped and then divided into sections of road of approximately 100-300m, except one cluster which had limited road infrastructure and was divided into smaller areas with an average perimeter of 3500m. Geographic Information Systems were used to randomly select road sections within each cluster which were overlaid on OpenStreet Map and satellite images. The size of sections was based on population distribution from the Zimbabwe 2012 Census. A sample of road sections were randomly selected within each cluster and all households in these sections were enumerated to identify all residents who were aged 18\u0026ndash;24 years and hence eligible for the survey. If the sample was not adequate to reach a sample size of 700 youths, a further set of randomly selected road sections was taken and sections were visited sequentially to maintain the randomness until a sample size of 700 was reached. A median of 100 sections (IQR 88\u0026ndash;148) were surveyed per cluster. Maps.me, a mobile application that provides offline maps using OpenStreetMap data was used for navigation to the exact location of the road section. All those eligible were contacted, if possible, and asked to participate in the survey and provide fingerprint registration. More details are provided in the trial protocol\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. The coordinates of each household dwelling were collected using the in-built Global Positioning System (GPS) on the Android tablets.\u003c/p\u003e \u003cp\u003eInitially, data collectors were unable to easily bypass fingerprint registration because all processes including bypass and refusals, were required to be processed through the Simprints ID application. On 21 April 2022 the bypass process was made easier in response to staff requests, due to challenges with scanners and the Simprints ID application.\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eData analysis\u003c/h2\u003e \u003cp\u003eAnalysis was conducted using Stata v18.0 and R v4.3.3. The dataset of service clients was linked with the dataset of survey participants by fingerprint ID, within province. If a service client fingerprint was linked to more than one survey participant, these records were excluded because it was not possible to determine the correct identity.\u003c/p\u003e \u003cp\u003eFeasibility was assessed from the proportion of survey participants who gave a digital fingerprint, the proportion of occasions when data collectors bypassed the process, the accuracy against self-reported attendance, and extent of duplicates. Chi-square tests were used to investigate whether demographic characteristics were associated with refusal of biometric registration.\u003c/p\u003e \u003cp\u003eIntervention coverage was defined as the proportion of fingerprinted survey participants in the 12 intervention clusters who matched to the service client dataset. Self-reported CHIEDZA service attendance was compared with the fingerprint match to determine the accuracy of self-report. Contamination was defined as the proportion of fingerprinted survey participants in the 12 control clusters who matched to the service client dataset (Supplementary Fig.\u0026nbsp;1).\u003c/p\u003e \u003cp\u003eDistance from each survey participant\u0026rsquo;s residence to the cluster community hall was calculated, both as Euclidean (straight line) distance and walking distance. Euclidean distance was calculated using the Stata \u003cem\u003evincenty\u003c/em\u003e module. Walking distance was calculated using DirectionFinder, a Google Apps script, based on the road infrastructure as recorded in Google Maps. Multi-level mixed-effects logistic regression was used to examine the association between utilisation of services and geographic walking distance and other covariates, accounting for cluster. Socioeconomic status was determined from asset ownership variables using factor analysis and divided into quintiles.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePatient and public involvement\u003c/h3\u003e\n\u003cp\u003eThe CHIEDZA service intervention was co-designed in collaboration with youth in two participatory workshops in Zimbabwe \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Digital biometrics were incorporated in the study design following discussion with youth. Participants were informed about the prevalence survey by watching a video which was designed in collaboration with youth. Results were disseminated to participant communities in collaboration with the youth advisory group.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\n \u003ch2\u003eFeasibility and accuracy\u003c/h2\u003e\n \u003cp\u003eFrom 1 April 2019 to 31 March 2022, 36,991 unique clients attended the CHIEDZA service in the 12 intervention clusters, of whom 36,957 (99.9%) registered their fingerprints. Ninety-five clients (63 in Bulawayo, 17 in Mashonaland East and 15 in Harare) attended more than 1 CHIEDZA community centre; 93 attended 2 and 2 clients attended 3 centres. The start and end of service provision was staggered by province, and in each province the survey began within 1 month after the service closed.\u003c/p\u003e\n \u003cp\u003eThe survey took place from 4 October to 15 December 2021 in Harare, 4 January to 5 March 2022 in Bulawayo, and 4 April to 2 June 2022 in Mashonaland East. Data in Bulawayo were collected by 45 research staff who each worked for a median of 38 days (IQR 35\u0026ndash;39, range 20\u0026ndash;42). Another 45 staff collected the data in Harare and Mashonaland East, working for a median 82 days (IQR 76\u0026ndash;85, range 18\u0026ndash;89). Staff worked in pairs and each person enrolled a median of 3 clients per day (IQR 3\u0026ndash;4, range 1\u0026ndash;10). In total 17,682 participants were enrolled of whom 13,675 (77.3%) successfully registered their fingerprint (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). In 4007 cases where the participant did not register their fingerprint the reasons were either that the researcher bypassed the process (N\u0026thinsp;=\u0026thinsp;1235), the participant refused to give fingerprints (N\u0026thinsp;=\u0026thinsp;975), or the registration process was begun but not completed (N\u0026thinsp;=\u0026thinsp;1797). Of those who started the process but did not complete it, 207 refused, for a total of 1182 refusals (975\u0026thinsp;+\u0026thinsp;207). Out of the 207 participants who refused during registration, 95 refused for religious reasons, 66 because of data concerns, and 46 did not give a reason.\u003c/p\u003e\n \u003cp\u003eApart from refusal, the other causes of non-completion of biometric registration were: the scanner was not working (N\u0026thinsp;=\u0026thinsp;723), the scanner did not connect to the tablet (N\u0026thinsp;=\u0026thinsp;502), the scanner had low battery (N\u0026thinsp;=\u0026thinsp;30), a software version incompatibility problem (N\u0026thinsp;=\u0026thinsp;243), individual reasons which prevented fingerprint detection (e.g. the participant had grease on their hands) (N\u0026thinsp;=\u0026thinsp;53), and 39 were for unknown reason. The version incompatibility problem occurred during the first 3 weeks of data collection in Harare province (Supplementary Table\u0026nbsp;1). Automatic software updates were disabled on the Android tablets, and as a result there were compatibility issues between newer scanners and the old software versions. The app recorded the event as \u0026lsquo;register biometrics complete\u0026rsquo; although a biometric ID was not stored, so data collection staff did not realise that registration had failed. This problem was resolved by the data office on 23 October 2021 and did not re-occur.\u003c/p\u003e\n \u003cp\u003eUntil 20 April 2022 only 149 (1.1%) clients bypassed registration out of 13,367 recruited. From 21 April, when bypass was made easier, until the survey ended on 2 June 2022, 1086/4315 (25.2%) of survey participants bypassed fingerprint registration. Each data collector bypassed a median 7.4% of registrations (IQR 0\u0026ndash;35%) during this time, and 4 data collectors bypassed more than 90%.\u003c/p\u003e\n \u003cp\u003eExcluding those who bypassed, 667/6334 (10.5%) men refused vs 515/10,112 (5.1%) women(Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). There was no difference in refusal by age. Refusal among eligible participants was 9.9% in Bulawayo and Mashonaland East but 2.3% in Harare. In the intervention clusters, fingerprint refusal was 3.8% among survey participants who self-reported accessing CHIEDZA services versus 8.9% among participants who said they had not accessed services. Refusal by cluster ranged from 0.3\u0026ndash;27.5% (median 6.7%, IQR 2.3%-10.0%).\u003c/p\u003e\n \u003cp\u003eIn the test dataset the optimal cutpoint for comparison score was determined to be 21.5. At this cutpoint the FAR was 0.01% and the FRR was 0.94% (Supplementary Fig.\u0026nbsp;2). At a cutpoint of 21.5, 128 service clients appeared to match\u0026thinsp;\u0026gt;\u0026thinsp;1 survey participant. These records were excluded from the dataset of 36,991 service clients, leaving 36,863. Out of 1826 survey participants who linked to the service dataset, 10 participant fingerprints (0.5%) linked to \u0026gt;\u0026thinsp;1 client (in 2 cases to 3 clients, in 8 cases to 2 clients). On checking, these client pairs usually shared a birthdate and were at the same cluster, and on one occasion were seen on the same day. The most likely explanation is that these were visits by the same client, who was not recognised as a match during the biometric registration and was re-registered. In these cases, one record from each group was retained to link with the survey participant.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eIntervention Coverage\u003c/h3\u003e\n\u003cp\u003eUsing a cutpoint of 21.5, coverage in the intervention arm was 23.1% (95%CI 22.1\u0026ndash;24.1) (Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e). Uptake was higher in women than men, and higher in Bulawayo than the other provinces. Overall, 10.5% of matches in the intervention arm were cross-cluster; the survey participant attended CHIEDZA services not in their cluster of residence, but in one of the other 7 intervention clusters within the province. Participants who had lived at their current address for a shorter time were less likely to have attended the CHIEDZA service, and those who had previously lived in a different area or city were more likely to have attended a CHIEDZA service in a different cluster (Supplementary Fig. 3).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eService uptake among prevalence survey participants who completed biometric registration (N\u0026thinsp;=\u0026thinsp;13675); using cutpoint 21.5)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eN\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eIntervention clusters\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eN\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eControl clusters\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTotal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e6806\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1574 (23.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e6869\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e252 (3.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eSex\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2403\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e460 (19.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2625\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e54 (2.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4403\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1114 (25.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4243\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e198 (4.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e18\u0026ndash;20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3579\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e852 (23.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3575\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e125 (3.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e21\u0026ndash;24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3227\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e722 (22.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3294\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e127 (3.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003eProvince\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHarare\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2592\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e471 (18.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2350\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e73 (3.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBulawayo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2390\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e790 (33.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2481\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e122 (4.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMash East\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1824\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e313 (17.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2038\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e57 (2.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eSelf-reported accessing CHIEDZA service\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e5238\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e388 (7.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e6822\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e224 (3.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1568\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1186 (75.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28 (59.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eAmong survey participants with biometric registration, the positive predictive value of self-report in the intervention arm against fingerprint match was 75.6% (95%CI 73.4\u0026ndash;77.7) and negative predictive value was 92.6% (95%CI 91.8\u0026ndash;93.3). Sensitivity of self-report was 75.3% (95%CI 73.1\u0026ndash;77.5) and specificity was 92.7% (95%CI 92.0-93.4).\u003c/p\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003eContamination\u003c/h2\u003e\n \u003cp\u003eAt a cutpoint of 21.5, uptake in the control clusters was 3.7% (95%CI 3.2\u0026ndash;4.1). As with the intervention clusters, uptake was higher in women and in Bulawayo. In the control arm 47 participants self-reported accessing CHIEDZA services, of whom 28 (59.6%) had a fingerprint match.\u003c/p\u003e\n \u003cp\u003eAmong the 1574 survey participants in the intervention clusters with a fingerprint match to the dataset of CHIEDZA service attendees at a comparison score\u0026thinsp;\u0026gt;\u0026thinsp;21.5, the median comparison score was 55.7 (IQR 41.1\u0026ndash;70.5, range 21,5-142.7). Among the corresponding 252 control arm participants with a comparison score\u0026thinsp;\u0026gt;\u0026thinsp;21.5, the median score was 22.2 (IQR 21.9\u0026ndash;24.1, range 21.5-122.8) Results of sensitivity analysis are shown in Supplementary Table\u0026nbsp;2. At a cutpoint of 23, uptake was 21.3% in the intervention clusters (coverage) and 1.3% in the control clusters (contamination). Raising the cutpoint to 23 compared to 21.5 resulted in a 7.8% reduction in the number of service clients identified in the intervention arm, and a 65% reduction in the number of service clients identified in the control arm.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003e\u003cem\u003eFactors associated with service uptake\u003c/em\u003e\u003c/h2\u003e\n \u003cp\u003eIn the intervention clusters 8885 participants were enrolled. Five participants did not have an accurate GPS reading for their location and 2805 did not have fingerprint data, leaving 6805. (Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). A higher proportion of females (25.3%) than males (19.1%) accessed CHIEDZA services. Univariable multilevel mixed-effects logistic regression showed that utilisation of CHIEDZA services was associated with being female (OR 1.55, 95%CI 1.37\u0026ndash;1.76) and with being married or cohabiting (OR 1.21, 95%CI 1.21\u0026ndash;1.26) versus never married (Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). \u0026nbsp;\u003c/p\u003e\n \u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eUtilisation of CHIEDZA services in intervention arm by sociodemographic characteristics\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eUtilised CHIEDZA service\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eUnivariable crude OR adjusted for clustering (95%CI)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCharacteristic\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAll\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eNo, n (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eYes, n (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6805\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003en\u0026thinsp;=\u0026thinsp;5233 (76.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003en\u0026thinsp;=\u0026thinsp;1572 (23.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eEuclidean distance from home to CHIEDZA location (km)\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/strong\u003e\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMean (SD)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.01 (0.48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.84 (0.31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.52 (0.46\u0026ndash;0.59)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eWalking distance from home to CHIEDZA location (km)\u003c/strong\u003e \u003csup\u003e\u003cstrong\u003e\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/strong\u003e\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMean (SD)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.42 (0.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.17 (0.39)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.48 (0.44\u0026ndash;0.54)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eDuration of residence at current address\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;12 months\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1728 (25.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e208 (12.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"3\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12\u0026ndash;24 months\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e701 (10.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e133 (19.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.74 (1.37\u0026ndash;2.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026gt;\u0026thinsp;2 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4376 (64.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1231 (28.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.57 (2.18\u0026ndash;3.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eSex\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2402 (35.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e458 (19.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4403 (64.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1114 (25.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.55 (1.37\u0026ndash;1.76)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eMarital Status\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNever married\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5037 (74.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1173 (23.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"3\"\u003e\n \u003cp\u003e0.039\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMarried/living together\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1479 (21.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e335 (22.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.21 (1.04\u0026ndash;1.40)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDivorced/widowed/separated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e289 (4.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e64 (22.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.13 (0.85\u0026ndash;1.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"4\"\u003e\n \u003cp\u003e\u003cstrong\u003eCurrent main activity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3343 (49.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e790 (23.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"4\"\u003e\n \u003cp\u003e0.625\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEducation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1927 (28.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e446 (23.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.93 (0.81\u0026ndash;1.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRegistered business/Formal sector work\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e312 (4.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e83 (26.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.04 (0.80\u0026ndash;1.37)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eInformal sector work\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1223 (18.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e253 (20.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.94 (0.80\u0026ndash;1.11)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eLevel of education\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNone, any primary or completed primary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e365 (5.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e75 (20.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"3\"\u003e\n \u003cp\u003e0.0006\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAny secondary or completed secondary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5892 (86.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1402 (23.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.19 (0.91\u0026ndash;1.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAny post-secondary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e548 (8.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e95 (17.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.76 (0.54\u0026ndash;1.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eEver had sexual intercourse\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2287 (35.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e476 (19.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"3\"\u003e\n \u003cp\u003e0.0002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4379 (64.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1082 (24.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.28 (1.13\u0026ndash;1.45)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDo not want to say\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e39 (0.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e14 (35.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.96 (0.99\u0026ndash;3.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eAge in years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e18\u0026ndash;20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3579 (52.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e850 (54.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"2\"\u003e\n \u003cp\u003e0.309\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e21\u0026ndash;24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3226 (47.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e722 (45.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.94 (0.84\u0026ndash;1.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\"\u003e\u003csup\u003e\u003cstrong\u003e\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/strong\u003e\u003c/sup\u003e increased odds per increased kilometre of distance\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n \u003cp\u003eFor every km increase in walking distance, the odds of utilisation were reduced by 52% (OR 0.48, 95%CI 0.44\u0026ndash;0.54) and a similar margin of effect was observed using Euclidian distance (OR 0.52, 95%CI 0.46\u0026ndash;0.59). Although the mean walking distance to the CHIEDZA community halls was less than 3km (Supplementary Fig. 2), mean walking distance of those who utilised CHIEDZA services was on average 0.25km shorter than those in the same cluster who did not, and cluster-adjusted Euclidian distance was 0.17km shorter for service attendees than non-attendees.\u003c/p\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003cp\u003eIn multivariable sex-stratified multilevel mixed effects logistic regression models, utilisation of CHIEDZA services was associated with longer duration of residence both for males and females after adjusting for sexual debut, education and distance (Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e). In males only, sexual debut was associated with increased odds of utilisation of CHIEDZA services (AOR 2.01, 95%CI 1.58\u0026ndash;2.57), with statistically significant evidence for interaction by sex (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Females who had post-secondary education were less likely to have utilised CHIEDZA services (AOR 0.56, 95%CI 0.37\u0026ndash;0.86) compared to females with primary level of education, but there was no evidence of an interaction effect by sex.\u0026nbsp;\u003c/p\u003e\u0026nbsp;\u003ctable id=\"Tab4\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eAssociation of utilisation of CHIEDZA service with walking distance and length of residence, for outcome survey male and female participants in the intervention arm\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eMales\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCharacteristic\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAdjusted OR (95%CI)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eAdjusted OR (95%CI)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eWalking distance (km)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.48 (0.40\u0026ndash;0.58)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.49 (0.43\u0026ndash;0.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eEver had sexual intercourse\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"3\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"3\"\u003e\n \u003cp\u003e0.401\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.01 (1.58\u0026ndash;2.57)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.15 (0.99\u0026ndash;1.35)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDo not want to say\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3.14 (1.10\u0026ndash;8.96)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.71 (0.64\u0026ndash;4.54)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eLevel of education\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNone, any primary or completed primary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"3\"\u003e\n \u003cp\u003e0.0686\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"3\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAny secondary or completed secondary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.69 (0.93\u0026ndash;3.09)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.01 (0.74\u0026ndash;1.39)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAny post-secondary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.08 (0.53\u0026ndash;2.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0.56 (0.37\u0026ndash;0.86)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eDuration of residence at current address\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;12 months\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" rowspan=\"3\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12\u0026ndash;24 months\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.89 (1.03\u0026ndash;3.49)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1.71 (1.30\u0026ndash;2.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u0026gt;\u0026thinsp;2 years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3.10 (2.05\u0026ndash;4.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2.99 (2.48\u0026ndash;3.61)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur study showed that the use of digital fingerprints was feasible and accurate, and these biometrics can be applied to understand intervention coverage as well as contamination, two aspects that critically affect the measured efficacy of interventions. Digital fingerprints had\u0026thinsp;\u0026lt;\u0026thinsp;0.1% refusal in the CHIEDA service, but we observed a higher refusal among survey participants. Notably there was generally good agreement of biometrics with self-reported attendance.\u003c/p\u003e \u003cp\u003eThere are several reasons why uptake of biometric registration was higher at the CHIEDZA services than during the prevalence survey. Each community centre had several tablets and scanners, so a nonfunctional piece of equipment could quickly be exchanged for a working one. Attendees may have been motivated to use biometric registration in order to access services. Manual registration was possible but was considered a tedious and time-consuming process, so staff preferred to use biometric registration whenever possible. Handwashing facilities were also readily available at community halls. By contrast, in the cross-sectional survey, there were operational challenges. The two-person teams moved from house to house, had no backup equipment with them, and were under time pressure to complete the survey. Scanner and tablet had to be reconnected frequently. New equipment was purchased as required but was not always immediately compatible with the older equipment. Survey participants had no particular motivation to consent to biometric registration. Participants might not be able to wash their hands, which could result in poor scans. The manual registration process was deliberately made easier, with the result that staff were more likely to opt for it when they experienced or anticipated technical challenges.\u003c/p\u003e \u003cp\u003eClose observation of the data in real time is essential to identify and resolve problems. This was particularly evident at two periods in the prevalence survey; noticing and resolving the version compatibility problem which prevented storage of a biometric ID, and observing the difference in bypass rates by study staff ID. The bypass option was introduced because survey staff reported frustrations trying to complete the registration process when they were faced with challenges such as failure to connect and risked falling behind on their recruitment targets. However, once the bypass option was available, a small number of survey staff used it routinely rather than attempting biometric registration of participants. It is necessary to observe the data at sufficiently granular level and in real-time to identify these problems quickly and intervene.\u003c/p\u003e \u003cp\u003eIn the survey, men were more likely to refuse fingerprinting than women, and those in Harare were less likely to refuse than residents in Bulawayo or Mashonaland East. Refusal was most common in those with either less than primary education or postgraduate education, and in the lowest and highest socioeconomic quintiles. This bimodal distribution suggests two separate mechanisms may lead to refusal. However, because of its large sample size the study is highly powered to detect small differences in refusal rates between groups, and some observed effects may be due to chance.\u003c/p\u003e \u003cp\u003eWe demonstrate the applicability of fingerprints in understanding coverage of the intervention.\u003c/p\u003e \u003cp\u003eAs shown in other studies, females were more likely to use these services, and those who were not sexually active would have found these services less relevant to them.\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. We observed much lower population-level coverage of the intervention than anticipated. Those who had been resident in the intervention clusters for more than two years were more likely to have accessed the intervention. Youth in this setting are a highly mobile group and out-migration from the clusters may explain the discrepancy between the large number of clients who attended CHIEDZA, as a proportion of residents, and the low coverage of the intervention among survey participants. Even in clusters that were small in area, we found a strong relationship between distance and uptake. The greater the distance from the hall, the lower the probability of awareness or utilisation of the service, consistent with the well documented inverse relationship between healthcare utilisation and distance \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. Nearly two-thirds of survey participants did not know that the community-based youth-friendly SRH services existed in their area and, even among those who knew about their existence, nearly half did not utilise them. This finding underscores the need for intensive sensitisation to increase utilisation and is supported by findings from other studies where utilisation was low due to lack of knowledge \u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. While community mobilisers were used, it is possible that they either operated nearer the halls and/or clients living further away were less willing to come. Advertisement of services using radio, television and social media was not undertaken to reduce the risk of contamination across clusters by giving information about CHIEDZA services to those in control clusters.\u003c/p\u003e \u003cp\u003eThree quarters (75.0%) of CHIEDZA clients were female, yet service uptake in the survey participants was only 32% higher in females than males (25.3% of females and 19.1% of males accessed services), not three times higher, as might be expected. This difference may partly be attributable to the cluster populations being skewed female (10,741/17,682 survey participants were female), and also to women being a more mobile population. Among all survey participants, 29.7% of women (3191/10741) and 15.2% of men (1054/6940) had resided at their current address less than 12 months, while less than half of women (49.3%) and more than two thirds of men (68.3%) had been resident more than 3 years. As a result, women in the survey had had less exposure time to CHIEDZA.\u003c/p\u003e \u003cp\u003eThe uptake results in the control arm were particularly sensitive to comparison score cutpoint. Most of the matches in the control arm had low comparison scores, so when the cutpoint was raised they were no longer categorised as matches. As such it may be more appropriate to use different cutpoints in the intervention and control arms, since prior knowledge that a match is more likely in the intervention arm was not taken into account for the probability assessment. However, such a step would require careful justification in an RCT.\u003c/p\u003e \u003cp\u003eThe main strength of this study was that it was based on data from a large randomly selected sample which was representative of the general population sample of youths from three provinces of Zimbabwe from both urban and peri-urban communities, increasing generalisability of the findings. The study also has limitations. There is a lack of information around what caused the \u0026lsquo;scanner not working\u0026rsquo; or \u0026lsquo;scanner not connecting to tablet\u0026rsquo; errors which prevented biometric registration of a large number of survey participants.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eDigital fingerprinting offers a highly feasible and effective approach to evaluate intervention delivery within cluster randomised trials. There are additional challenges to using this technology in house-to-house surveys. Steps that could be taken to address them include carrying spare equipment, and moving biometric registration to the end of the questionnaire to prevent wasted time during connection.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor contributions\u003c/h2\u003e\n\u003cp\u003eRAF conceptualised the study. ED, CDC, CM, MT and CG assisted with management of fieldwork and data collection. OM and TA provided support and guidance on trial implementation, and RH and KK provided expertise on trial design and analysis. RA, DD and AS supported biometrics data management and analysis. VS and TB conducted statistical analysis and drafted the manuscript. RAF is the Principal Investigator of the trial. All authors contributed to interpretation of the results and have read and approved the final manuscript\u003c/p\u003e\n\u003ch2\u003eCompeting interests\u003c/h2\u003e\n\u003cp\u003e\u0026nbsp;RF, DD and AS are employed by Simprints.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eData availability\u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eData for this study are available upon reasonable request to the principal authors.\u003c/p\u003e\n\u003ch2\u003eEthics approval and consent to participate\u003c/h2\u003e\n\u003cp\u003eEthical approval for the trial including the survey was obtained from the Biomedical Research and Training Institute Institutional Review Board (AP149/2018), the Medical Research Council of Zimbabwe (MCRZ/A/2387) and the London School of Hygiene \u0026amp; Tropical Medicine ethics committee (16124/RR/11602).\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eConsent for publications\u003c/h2\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eFinancial support for this study was provided through the Wellcome Trust to RAF (Grant number: 206316_Z_17_Z). The funders had no role in study design, analysis and decision to publish or preparation of the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eStoristeanu DM, Norman TL, Grigore A, et al. Biometric fingerprint system to enable rapid and accurate identification of beneficiaries. \u003cem\u003eGlob Health Sci Pract\u003c/em\u003e 2015;3(1):135-7. doi: 10.9745/GHSP-D-15-00010 [published Online First: 2015/03/07]\u003c/li\u003e\n\u003cli\u003eJaafa NK, Mokaya B, Savai SM, et al. Implementation of Fingerprint Technology for Unique Patient Matching and Identification at an HIV Care and Treatment Facility in Western Kenya: Cross-sectional Study. \u003cem\u003eJ Med Internet Res\u003c/em\u003e 2021;23(12):e28958. doi: 10.2196/28958 [published Online First: 2021/12/24]\u003c/li\u003e\n\u003cli\u003eBengtson AM, Kumwenda W, Lurie M, et al. Improving Monitoring of Engagement in HIV Care for Women in Option B+: A Pilot Test of Biometric Fingerprint Scanning in Lilongwe, Malawi. \u003cem\u003eAIDS Behav\u003c/em\u003e 2020;24(2):551-59. doi: 10.1007/s10461-019-02748-6 [published Online First: 2019/11/28]\u003c/li\u003e\n\u003cli\u003eOdei-Lartey EO, Boateng D, Danso S, et al. The application of a biometric identification technique for linking community and hospital data in rural Ghana. \u003cem\u003eGlob Health Action\u003c/em\u003e 2016;9:29854. doi: 10.3402/gha.v9.29854 [published Online First: 2016/03/20]\u003c/li\u003e\n\u003cli\u003eAbrams MP, Torres FE, Little SJ. Biometric Registration to an HIV Research Study may Deter Participation. \u003cem\u003eAIDS Behav\u003c/em\u003e 2021;25(5):1552-59. doi: 10.1007/s10461-020-02995-y [published Online First: 2020/08/09]\u003c/li\u003e\n\u003cli\u003eWall KM, Kilembe W, Inambao M, et al. Implementation of an electronic fingerprint-linked data collection system: a feasibility and acceptability study among Zambian female sex workers. \u003cem\u003eGlobal Health\u003c/em\u003e 2015;11:27. doi: 10.1186/s12992-015-0114-z [published Online First: 2015/06/28]\u003c/li\u003e\n\u003cli\u003eDron L, Taljaard M, Cheung YB, et al. The role and challenges of cluster randomised trials for global health. \u003cem\u003eLancet Glob Health\u003c/em\u003e 2021;9(5):e701-e10. doi: 10.1016/S2214-109X(20)30541-6 [published Online First: 2021/04/19]\u003c/li\u003e\n\u003cli\u003eWhite EB, Meyer AJ, Ggita JM, et al. Feasibility, Acceptability, and Adoption of Digital Fingerprinting During Contact Investigation for Tuberculosis in Kampala, Uganda: A Parallel-Convergent Mixed-Methods Analysis. \u003cem\u003eJ Med Internet Res\u003c/em\u003e 2018;20(11):e11541. doi: 10.2196/11541 [published Online First: 2018/11/18]\u003c/li\u003e\n\u003cli\u003eWorld Health Organisation. Making Health Services Adolescent Friendly: Devloping National Quality Standards for Adolescent Friendly Health Services. Geneva, 2012.\u003c/li\u003e\n\u003cli\u003eDziva Chikwari C, Dauya E, Bandason T, et al. The impact of community-based integrated HIV and sexual and reproductive health services for youth on population-level HIV viral load and sexually transmitted infections in Zimbabwe: protocol for the CHIEDZA cluster-randomised trial [version 2; peer review: 2 approved]. \u003cem\u003eWellcome Open Research\u003c/em\u003e 2022;7(54):https://doi.org/10.12688/wellcomeopenres.7530.2. doi: https://doi.org/10.12688/wellcomeopenres.17530.1\u003c/li\u003e\n\u003cli\u003eMackworth-Young CR, Dringus S, Dauya E, et al. Putting youth at the centre: co-design of a community-based intervention to improve HIV outcomes among youth in Zimbabwe [version 2; peer review: 1 approved]. \u003cem\u003eWellcome Open Research\u003c/em\u003e 2022;7(53) doi: https://doi.org/10.12688/wellcomeopenres.17531.2\u003c/li\u003e\n\u003cli\u003eNinsiima LR, Chiumia IK, Ndejjo R. Factors influencing access to and utilisation of youth-friendly sexual and reproductive health services in sub-Saharan Africa: a systematic review. \u003cem\u003eReprod Health\u003c/em\u003e 2021;18(1):135. doi: 10.1186/s12978-021-01183-y [published Online First: 2021/06/29]\u003c/li\u003e\n\u003cli\u003eSharma M, Khatri B, Amatya A, et al. Utilization of adolescent friendly health services and its associated factors among higher secondary students in mid-western Himalayan mountainous district of Nepal. \u003cem\u003ePLOS Glob Public Health\u003c/em\u003e 2023;3(3):e0001616. doi: 10.1371/journal.pgph.0001616 [published Online First: 2023/03/25]\u003c/li\u003e\n\u003cli\u003eSiedner MJ, Lankowski A, Tsai AC, et al. GPS-measured distance to clinic, but not self-reported transportation factors, are associated with missed HIV clinic visits in rural Uganda. \u003cem\u003eAIDS\u003c/em\u003e 2013;27(9):1503-8. doi: 10.1097/QAD.0b013e32835fd873 [published Online First: 2013/02/26]\u003c/li\u003e\n\u003cli\u003eLankowski AJ, Siedner MJ, Bangsberg DR, et al. Impact of geographic and transportation-related barriers on HIV outcomes in sub-Saharan Africa: a systematic review. \u003cem\u003eAIDS Behav\u003c/em\u003e 2014;18(7):1199-223. doi: 10.1007/s10461-014-0729-8 [published Online First: 2014/02/25]\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[{"identity":"0efbc87b-f252-4fc3-811f-73f8ffc06344","identifier":"10.13039/100010269","name":"Wellcome Trust","awardNumber":"206316_Z_17_Z","order_by":0}],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"biometrics, cluster randomised trial, community-based, contamination, coverage, accessibility, youth, Zimbabwe","lastPublishedDoi":"10.21203/rs.3.rs-6758988/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6758988/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eIntroduction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLow intervention uptake and contamination can dilute effects of cluster randomised trials (CRTs) but can be difficult to assess. We investigated the feasibility of digital fingerprints to assess intervention coverage and contamination in a CRT of community-based integrated HIV and sexual and reproductive health services for youth (CHIEDZA).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e24 clusters in Zimbabwe were randomly allocated to intervention/control. In the 12 intervention clusters, services for youth aged 16-24 years were provided in community halls for 30 months. A population-based survey of youth aged 18-24 years (700/cluster) was conducted to ascertain impact on trial outcomes. Digital fingerprints were collected from service attendees and survey participants, and the two datasets were linked to assess intervention coverage at population level in intervention clusters, and contaminaton in control clusters. Multilevel logistic regression estimated the association of walking distance to the community hall with service uptake.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBetween April 2019-March 2022, 36,9991 clients attended the CHIEDZA service and 36,957 (99.9%) used biometric registration. In the survey 13,675/17,682 (77.3%) participants completed biometric registration: 1182 refused, 1235 bypassed registration and 1590 were unable to register.\u003c/p\u003e\n\u003cp\u003eDatabase linkage showed that 23.1% of registered survey participants in the intervention clusters (coverage), and 3.7% of participants in control clusters (contamination) attended the CHIEDZA service. Sensitivity of self-reported service attendance against biometric registration match was 75.3% (95%CI 73.1-77.5) and specificity was 92.7% (95%CI 92.0-93.4).\u003c/p\u003e\n\u003cp\u003eIn intervention clusters, for every 1km increase in walking distance to the community hall, the odds of utilising the CHIEDZA service reduced by 52% (OR: 0.48 95%CI:0.44- 0.54).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBiometric identification was highly feasible and acceptable in a community setting with low time pressure. In population-based surveys additional technological challenges emerged. Biometrics enabled good estimation of intervention coverage and validated self-reported data. Community services for youth must overcome distance barriers and ensure communication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTrial registration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ehttps://clinicaltrials.gov/study/NCT03719521 registered 23 October 2018\u003c/p\u003e","manuscriptTitle":"Use of biometrics to evaluate intervention coverage and contamination in a cluster randomised trial in Zimbabwe","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-04 09:09:27","doi":"10.21203/rs.3.rs-6758988/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"5466e7af-947f-4c35-bd63-550641796af1","owner":[],"postedDate":"June 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":49202576,"name":"Epidemiology"}],"tags":[],"updatedAt":"2025-06-04T09:09:27+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-04 09:09:27","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6758988","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6758988","identity":"rs-6758988","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}