Design and setting
The study was carried out in two Ethiopian cities Addis Ababa [27] and Hawassa. According to the 2007 census, and projected to 2017, Ethiopia has a total population of 103.8 million. The majority of the population (86%) is known to live in a rural community. In terms of health and welfare, Ethiopia ranks among the world’s poorest nations, and agriculture is the backbone economy of the country. This report is a study finding from Hawassa, a capital city of the Southern Nations, Nationalities, and Peoples’ Region. Hawassa is located about 285 km south of Addis Ababa, on Lake Hawassa in the Great Rift Valley, with 258,808 people (for 2017). Hawassa lies on the Trans-African Highway extending from Cairo to Cape Town, and it is a city with a high tourist flow [28].
A rapid assessment technique was employed to supplement the combination approach that included the nomination and multiplier methods. The combination method had data collection from people who inject drugs. It used interviews and focus group discussions on enhancing the estimation of the population who inject drugs.
Study population
The source population for the study group was a mix of people linked to people who inject drugs. To estimate the number of injecting drugs, the source population was people who inject drugs, fulfilling the following inclusion criteria. A person who is at least 18 years old and uses an injection drug for non-medical purposes stayed in the study city for the last six months. The drug injection should be at least in the last three months was considered a case of people who inject a drug. Study subjects who fulfilled the above criteria and captured by the respondent-driven sampling method were further assessed for using injection drugs by confirmatory checklists. The confirmatory inventory included; 1) Knowledge of proper place of injection site on the body; 2) Knowledge of place of finding the injection rigs; 3) Knowledge of the size of syringe and needle and on how it is injected 4) Assessing scar of the point of the last injection on the body. During the assessment time, if the subjects’ responses suggested the presence or lack of familiarity, eligibility was made to be assessed by a coordinator who has a high understanding of people who inject drugs.
The respondent-driven sampling (RDS) approach included people who inject drugs for a face-to-face interview until a saturation time was achieved. Respondent-driven sampling is a kind of chain referral sampling that uses peers to attract participants. RDS may give relative estimates typical of the networks of the populations studied when properly implemented and evaluated [29, 30]. Injecting drug users, syringe suppliers, and individuals trying to minimize the issue were among the people who contributed to the rapid situational evaluation. It also included health providers, networks of governmental and non-governmental organizations (NGOs) that deal with drug users.
Sampling and recruitment
The sample size determination was not made in evaluating a research topic since the study approached the entire population of people who inject drugs. However, maximizing inclusion in the study of people who inject drugs is necessary. In the combined methods, for the nomination and multiplier methods, all people who inject drugs and are captured by the respondent-driven sampling in the given time were included. Using the respondent-driven sampling approach, three persons who inject drugs were selected in each enumeration site. The three recruiters who agreed with the rapid evaluation were considered seed recruits representing various individuals who inject drugs. Each seed was personally informed about the study’s goals and incentives to attract three eligible volunteers. Each source received three coupons after registering and answering the behavioural questions.
The coupon is helpful to recruit other peers with the same seed behaviour as the first wave of recruitment. The second wave of recruitment was people who inject drugs, which come with a recruitment coupon provided by recruits included in the first recruitment trend. Each recruit in succeeding instructions was interviewed to estimate questions related to behaviours and distribution of the problem in PWIDs. All recruits were given three coupons favorably advised to recruit other peers who inject drugs. The successive recruitment trend was ensured in long recruitment chains of people who inject drugs. It went on until it was impossible to locate PWID recruits. The research participants’ recruitment was tracked using the unique number-coded coupons given on each participant’s recruiting vouchers.
Data collection process
Participants were further questioned by asking two questions for the main objective: to estimate the population and prevalence of individuals who inject drugs using the nomination technique. The questions were, broadly, of the following two sorts of questions: 1) ‘How many friends do you have that have used drugs regularly in the past year?’ 2) ‘How many of these, you know, received treatment for drug-related problems at the specific health facilities in the past year.’ The multiplier was calculated using these two responses as an estimator for health facilities. Finally, as a benchmark, the multiplier was multiplied by the total number of PWIDs served at the chosen health institutions that utilized the previous year’s service.
The research group assessed the number of people who used treatment facilities due to drug use problems during a year to set the benchmark for the city. The total number of PWIDs was counted using data extracted from the logbooks of a hospital and a health center in Hawassa and serve as a benchmark. Multiplying the two sets estimates the likely size of the population who inject drugs in the city. During data collection, descriptive information was gathered from research participants identified during the evaluation.
Within each enumeration site, two data collectors having data collection experience were assigned for a face-to-face interview. Unique identifiers, socio-demographic data, the kind of drugs injected (with frequency and recency), and the geographical and social locations of injecting drug users are all included in a questionnaire (village level). Safety and how to approach individuals who inject drugs were taught to the data collectors. The researcher’s group educated data collectors about counseling services, giving a coupon for cases, safety, ethical problems, and the contents of the questionnaire because of its sensitive nature. Repeated data collection of eligible research participants was avoided with extreme care. The questionnaire was translated into the local language and tested in a comparable environment. The researchers have compensated the study subjects for their travel expenses and the cost of bringing three PWIDs to the research.
Data analysis
Data were collected to estimate the number of people who inject drugs into a computer using an EPI-DATA version 3.0 software package. A double-entry scheme was employed using a programmed entry template, having a unique identity and validation daily. Data analysis was done based on the objectives of the study. In addition, descriptive analysis was made to assess demographic characteristics and assess some problems encountered among people who inject drugs. For estimating individuals who inject drugs, both the multiplier and nomination approaches utilized data from two independent sources to evaluate the overlaps between the two sources. The first input was a count of PWIDs from the abstracted logbook handled in the past year from two health facilities in Hawassa as a benchmark. The second source was the percentage of PWIDs. They were deemed to utilize the health facilities in Hawassa in the past year.
The total number of PWIDs utilizing the health facilities was divided by the multiplier to determine the overall number of individuals who inject drugs using either nomination or multiplier approaches. Thus, the total number of PWIDs using the health facilities, although called the benchmark, is similar for each scenario. The difference lies in the estimate of the multiplier. Furthermore, the 95% CI of the multiplier was produced by calculating the 95 percent confidence level and considering the entire sample’s multiplier [27].
The multipliers were pooled estimate, computed with the fixed model effect, giving weight to the size of individuals in the survey [multiplier technique]. Participants in the research knew the total number of nominations PWIDs to synthesize a single estimate