The general study objective was to determine the epidemiology of leprosy and its protective and risk factors through active and early case detection approaches using three active case detection approaches and to estimate their yield and direct costs. The study was conducted between October and December 2021. The study was performed in three provinces in Nepal: Madhesh Province, Lumbini Province and Bagmati Province. The Siraha and Rauthat districts of Madhesh province and the Banke and Bardia districts of Lumbini Province have been selected as these districts have not eliminated leprosy as a public health problem. These districts have clusters of leprosy cases with ongoing transmission. Only prisons of Bagmati Province were included in study as it was assume that prisoner populations could be at-risk populations. Full details of each study are given below in the "Active case detection" section, Approaches 1–3. Except for the prison population, age and gender were only recorded for cases identified and not other contacts. All cases diagnosed by trained health workers were examined and validated by dermatologists for this study.

House-to-house visits in communities with high-risk groups and vulnerable populations, such as marginalized habitants of Dalit, Mushhar, and Chamar groups, were undertaken in Rautahat District of Madhesh Province and Banke District of Lumbini Province. In Rautahat, four rural municipalities (Palika), namely, Dewahi Gonahi, Rajpur, Ishnath and Rajdevi, were selected in close coordination with district health authorities. These municipalities were considered to have inhabitants from more vulnerable populations. From the four municipalities, 24 sites (wards) covered by 24 health facilities were selected. The same process was followed in Banke, where 27 sites (wards) covered by 27 health facilities from four municipalities, Baijnath, Narainapur, Janaki and Nepalgunj, were selected. A total of 60 to 100 households with inhabitants of marginalized people living in overcrowded houses made of soil or mud, which favored leprosy transmission, were used for the census. Trained local health workers and local female community health volunteers (FCHVs) visited the selected sites and performed house-to-house visits, examining all the members present in the household for any signs of leprosy. In total, 13,420 and 13,049 individuals were examined in Rautahat and Banke, respectively (Additional file 1: Table S1). Local trained health workers examined male individuals, and FCHV examined female individuals present in the household. Simultaneously, demographic and epidemiological variables were collected by a trained local health worker.

Members of the households were informed 2 days before the survey and asked to be present at their own household at the time of the survey via the local FCHV. All suspected cases identified by local health workers and FCHV were invited to health facilities, and cases were confirmed by a dermatologist. After diagnosis confirmation, leprosy cases were treated as per the national protocol.

Household and neighboring contacts of previously identified confirmed leprosy cases in the previous 2–5 years were examined in the Siraha district of Madhesh Province and Bardiya of Lumbini Province by trained local health workers. The cases diagnosed between the last 2–5 years in the respective districts were selected randomly in planning meetings conducted before the implementation of field work. Local trained health workers and FCHVs examined 106 and 177 confirmed leprosy case contacts, respectively, in Siraha and Bardiya. A total of 7608 contacts were screened during the case–contact survey (Table 1).

Siraha (n = 449), Rautahat (n = 360), Banke (n = 826), Bardiya (n = 319), Lalitpur (n = 251) and Kathamandu (n = 2223) prisons were used as screening sites for active case detection using convenience sampling among 4428 prisoners to assess the transmission status of leprosy in prisons. The prisoner population comprised 4229 males and 199 females.

In all approaches, participants were requested to give verbal informed consent. As this study was part of regular surveillance of epidemiology and disease control division (EDCD), written informed consent was not obtained. Approval for data publication was obtained from EDCD, and exemption from ethical review (347/2022) was obtained from the ethical board of the Nepal Health Research Council.

Data collected on paper-based questionnaires developed by the Leprosy Control and Disability Management Section (LCDMS)/EDCD were entered in Excel® spreadsheets. Consistency was checked, and data analysis was performed in IBM SPSS statistics 22 (IBM Corp, Armonk, NY, USA) and R version 4.2.0 (R Foundation for Statistical Computing, Vienna, Austria) [20]. Differences in the yields (cases per 100,000 people screened) for all methods (Approaches 1–3) were tested first using prop.test in R and then pairwise using an exact test with a Poisson distribution using the poisson.test in R. The attack rate (AR) with respect to different demographic variables and types of leprosy cases were calculated with 95% confidence intervals (CIs) using binomial models in R, where the attack rates are the case per contact calculated from the case-contact survey (Approach 2). Chi-squared tests (χ2) and odds ratios (ORs) using Fisher’s exact test with 95% CIs were calculated for associations between attack rates with respect to different demographic variables and between BCG scar presence and leprosy using R’s chisq.test or fisher.test. We used a Poisson generalized linear regression for testing the significance of age classes and gender of being a case from all the case data (Approaches 1–3, see Additional file 1: Table S2), where:

$$\mathrm (E\left(_\right))=_+__+__$$

where \(_\) is the intercept, \(_\) the coefficient for the age class i and \(_\) the coefficient for the gender j, using R’s glm function. To adjust for the screened population at risk and index cases present in the case-contact survey (Approach 2), we also use simple Poisson regression with an offset for index cases in the population present to assess risk [21], where for district i:

$$\mathrm (E\left(_\right))=_+__+\mathrm\left(_*_\right)$$

where \(_\) is the intercept, \(_\) the coefficient for the district i. We also simply offset this with population alone, where the offset was \(\mathrm\left(_\right)\) to test the sensitivity of the results to this assumption.

The direct medical and non-medical costs for each approach were calculated and comprised expenses related to training, orientation, health worker per diems, dermatologist’s fees, expenses for monitoring and supervision and data management [22]. The total direct cost was divided by the total number of patients identified or diagnosed by the approach and was derived per unit cost for the leprosy cases identified. Finally, for discussion, we converted costs from national currencies to US dollars for comparison. We used the date in publications and adjusted to 25 December 2021 rates using Google’s default currency convertor provided by Morningstar at 119.11 Nepalese rupee (NPR) per US dollar (USD).