Step 1: Preparation.

Before the translation and cultural adaptation of the original measures to the Tanzanian context, we obtained permission from the authors. Adaptation may include modifications of the measures based on the context of use.

Measures. We selected measures which contained a mix of items assessing experienced, anticipated, perceived and internalized stigma (Table 1). The HIV stigma measures included: 1) the Abridged Berger Scale [24] assessing personalized stigma, negative self-image, public attitudes, and disclosure concerns; 2) the Internalized AIDS-Related Stigma Scale (IA-RSS) [16] measuring internalized stigma; and 3) the HIV Stigma Mechanism Measure which measures internalized stigma, anticipated stigma and enacted stigma [25]. Drug use stigma measures included: 1) the Internalized Drug use Stigma Scale [19]; 2) the Substance use stigma mechanism model assessing enacted, anticipated, and internalized stigma [18]; and 3) the Perceived Stigma of Substance Abuse Scale (PSAS) [26].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Original measures of HIV stigma and drug use stigma that were translated and adapted.

https://doi.org/10.1371/journal.pone.0292642.t001

Step 2: Forward translation.

Three native Swahili speakers, fluent in English and experienced in conducting translation, independently conducted a forward translation of all the measures from English to Swahili.

Step 3: Reconciliation.

Reconciliation of the three forward translations was achieved through a panel discussion involving LBM, a psychiatrist experienced in treating PWUD, two people with a history of using drugs, a stigma expert and a linguist proficient in English and Swahili (S1 File).

Step 4: Back translation.

The reconciled Swahili translation was back translated (S1 File) into English by two new translators who were native speakers of Swahili and fluent in English. The back translators were blinded to the original English source measures.

Step 5: Back translation review.

LBM reviewed the back translated measures against the English source documents and prepared a discrepancy report highlighting the differences between the two back translations and the original English measures.

Step 6: Harmonization.

Harmonization was conducted by a panel comprised of a psychiatrist, two people with experience of using drugs, a stigma expert, a linguist proficient in English and Swahili, and LBM. Only the linguist and LBM had participated in Reconciliation. The panel reviewed the original English measures, back translations, and forward translations to make sure the items were translated conceptually and used clear and correct Swahili language that would be understood by PWUD. They addressed all discrepancies through discussion and consensus to develop the final Swahili version of the measures.

Step 7 (new): Expert review and development of new items.

LBM and a researcher with expertise in stigma who did not participate in earlier steps related to translation, reviewed the content of the drug use stigma and HIV stigma measures. During the expert review, we recognized that across the different measures, there were items which captured similar concepts (Table 1). To avoid duplication, we removed similar items and categorized the remaining items by type of stigma separately for HIV stigma and drug use stigma. We then developed new items reflecting stigmatizing actions known to exist in Tanzania based on the formative research findings by LN, LBM, and JKM (manuscript under review) that were not included in the original measures. We added these new items in all four stigma types (experienced, anticipated, perceived, internalized) for HIV stigma and drug use stigma. In addition, we phrased the new items differently based on the HIV status of PWUD and whether receiving medically assisted treatment or not. That is PWUD living with HIV (S2 File), PWUD not living with HIV (S3 File), people who are on medically assisted treatment and are living with HIV (S4 File) and people who are on medically assisted treatment and are not living with HIV (S5 File). All new items underwent the same process of translation and cultural adaptation that the original items underwent.

Step 8: Cognitive debriefing.

The final Swahili PROM including finalized and new items was then tested through cognitive debriefing. The aim of the cognitive debriefs was to assess the comprehension of the items of drug use and HIV stigma measures by PWUD.

Sample size and sampling strategy of participants in the cognitive debriefing. To assess the comprehension of the items comprising the drug use stigma and HIV stigma measures, we conducted cognitive debriefs (Step 8) with 50 PWUD. We used purposive sampling to recruit participants for the cognitive debriefing, by selecting participants with diverse characteristics including HIV status and MAT status. When conducting a cognitive debrief of measures, it is recommended that 10–15 participants should be debriefed for one item [27]. In this study, each item of the various types of stigma measures was tested among at least 10 participants during the cognitive debriefing. We first conducted cognitive debriefing with 40 PWUD. Out of 40 participants, 19 self-reported they were living with HIV and 21 self-reported a negative or unknown HIV status. Within each group, about half were debriefed using the HIV stigma measures and the other half debriefed with the drug use stigma measures. After reviewing the cognitive debriefing results from the initial 40 participants, we found that some items were not clearly understood. This determination was based on participants’ responses where some explicitly said that they did not understand certain words while others interpreted the items differently from their original intent. We therefore revised these items and conducted 10 additional cognitive debriefs among PWUD (half were HIV positive and half were HIV negative or of unknown HIV status) to address the identified issues.

Cognitive debriefing procedures. We recruited participants and collected data from April to October 2022. LBM and three trained research assistants (two female, one male) conducted the cognitive debriefing. The research assistants had prior experienced in qualitative data collection and were trained for a week on study objectives, procedures and ethical issues before data collection. We conducted cognitive debriefing in Swahili, either at community-based organization’s (CBO) offices or in the community at places frequented by PWUD. The duration of the cognitive interviews ranged from 40 minutes to two hours. We collected data using an interview guide, which consisted of all the items to be adapted. Following each item, four probing questions were asked to assess comprehension. The probing questions were i) Can you tell me in your own words what this question was asking? ii) Did you have any difficulty using the response options? Why? iii) How did you decide to answer this question? What was your thought process to arrive at your response? iv) Was there a particular experience or experiences that you reflected upon before/while answering the question? For some of the questions where a participant reported that they had experienced stigma, we asked for examples of what happened to him/her. During the initial interviews, we found that participants responded to the third and fourth probes with the same responses although the original intent of the questions were different. Consequently, we dropped the third probe because the fourth probe was more easily understood. The cognitive debriefs were audio recorded and transcribed. Drug use stigma items were asked to all participants. However, for HIV stigma items, participants who were not living with HIV were asked the items related to anticipated and perceived HIV stigma only. Anticipated HIV stigma items were phrased differently for PWUD living with HIV (e.g., How often do you worry that family members will avoid you because you are living with HIV?) and for PWUD not living with HIV (e.g., If you were living with HIV, how often would you be worried that family members would avoid you because you were living with HIV?).

Data analysis. At the end of each day of fieldwork, the study team convened to review the data obtained from the cognitive debriefs. We discussed the items, comprehension challenges encountered that day and strategies for addressing them. Following each cognitive debrief, a form was prepared by the research assistants. The feedback was shared with JKM and LN for review and decision. The process was iterative so that whenever we encountered a problem with the item, LBM, JKM and LN discussed that problem to find a solution and those items were adapted and tested in subsequent cognitive debriefs [28]. Authors had no access to information that could identify individual participants during or after data collection.

Step 9: Review of cognitive debriefing results and finalization.

After data analysis, we reviewed the results of the cognitive debriefing and finalized items and measures.

Step 10 and 11: Proofreading and final report.

LBM proofread the items and prepared final versions of the adapted measures. We are currently conducting research to determine the adapted measures’ psychometric properties and we will report the results separately.

English words that could be translated into more than one Swahili word.

All 54 original items and newly developed items covering 12 stigmatizing actions were translated into Swahili. Some English words capturing key dimensions of stigma had multiple possible translations into Swahili. For example: “avoided me” could be translated as “wamenikwepa” or “wameniepuka”. “Treated me differently” could be translated as “watanichukulia tofauti” or “watanitendea tofauti”. During reconciliation and harmonization, for all the items with multiple possible translation options, the panels reached a consensus to select the word that best fit in the context in which it would be used, that is the word that PWUD would understand easily and maintained the original intent of the item.

Removed items related to uncommon types of health care providers interacting with PWUD.

For the experienced HIV stigma items and anticipated HIV stigma items from the HIV Stigma Mechanism Measure [25] (Tables 4 and 5), we adapted six of the nine items in both categories of stigma because the remaining three items were about stigma actions PWUD experience from community or social workers whom PWUD do not interact with as much as they do with health care workers.

Changed the response options.

For the experienced HIV stigma items from the HIV Stigma Mechanism Measure [25], we changed the response options from “never”, “not often”, “somewhat often”, “often”, and “very often” to “never”, “once”, “a few times”, “often” because of translation challenge with the phrase “somewhat often” and participants had a clearer understanding of the response options never, once, a few times, and often as more concrete frequencies (Tables 4 and 5). This change was made for anticipated HIV stigma items from the HIV Stigma Mechanism Measure, and experienced and anticipated drug use stigma from the Substance use stigma mechanism model.

All items of the Internalized drug use stigma measure [19] were negatively phrased. Its response options were “strongly agree,” “agree,” “neither agree nor disagree,” “disagree,” and “strongly disagree”. We reversed the order of the response options to be in a direction similar to that of the other stigma measures. We adapted internalized HIV stigma items from IA-RSS [16]. In the original measure, the response options were “agree” and “disagree”. We added more response options in order to be similar with the Internalized drug use stigma measure. The revised response options were “strongly disagree”, “disagree”, “neither disagree nor agree”, “agree, strongly agree”. The same change was made for the response options of the Perceived Stigma of Substance Abuse Scale (PSAS) items from the response options of “very strongly disagree, strongly disagree, disagree, agree, strongly agree, and very strongly agree” to “strongly disagree, disagree, neither disagree nor agree, agree, strongly agree” (Table 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Perceived drug use stigma measures with original and culturally adapted items.

https://doi.org/10.1371/journal.pone.0292642.t006

Changing the phrasing of selected questions.

For the anticipated HIV stigma items from the HIV Stigma Mechanism Measure, we changed the questions from “How likely is it that people will treat you in the following ways in the future because of your HIV status?” to “How often do you worry that family members will (example avoid) you because you are living with HIV?” in order to assess the frequency of stigma and to be similar to the other items in the experienced HIV stigma measure (Table 4). We culturally adapted six items related to experienced drug use stigma from the Substance use stigma mechanism model. We changed the original items from a statement to a question format in order to assess the frequency of experiences of stigma, have a parallel style with other adapted measures, and use response options which are more concrete and easier to translate as occurred with the other types of stigma (Table 5). We culturally adapted eight perceived drug use stigma items from the Perceived Stigma of Substance Abuse Scale (PSAS). We changed the original items’ focus on issues related to substance use in general to issues related to drug use. The term drug use is specific for illicit drugs while substance use is a general term referring to various substances including alcohol (Table 6).

Changed wording of questions based on culture value placed on dating and marriage.

For the items of the Perceived Stigma of Substance Abuse Scale (PSAS), we changed the eighth item: “Most people would be willing to date someone who has been treated for substance use” to “Most people would be willing to marry someone who has been treated for drug use” as dating in the Tanzanian cultural context is not common (Table 6).

Challenges with interpretation.

In general, most of the original items (42/54, 78%) and response options and all items on the 12 new stigmatizing actions were understood as intended by the participants. Only a few items were not understood as intended by all participants. These included two items in the “Experienced HIV stigma” measure, (Item 3: Family members have treated me differently and Item 9: Healthcare workers have treated me with less respect); two items in the “Anticipated HIV stigma” measure which have the same stems as those in the experienced HIV stigma measure (Item 3: Family members will treat me differently and Item 9: Healthcare workers will treat me with less respect); one item in the “Perceived HIV stigma” measure (Item 1: Most people believe that a person who has HIV is dirty) and one item in the “Internalized HIV stigma” measure (2: Being HIV positive makes me feel dirty).

For drug use stigma measures, items that were not understood as intended by all participants included: one item in the “Experienced drug use stigma” measure (Item 3: Family members have treated me differently) and one item in the “Anticipated drug use stigma” measure which also has the same stem as the one in the experienced drug use stigma measure (Family members will treat me differently); two items in the “Perceived drug use stigma” measure (Item 1: Most people would willingly accept someone who has been treated for substance use as a close friend, Item 3: Most people would accept someone who has been treated for substance use as a teacher of young children in a public school); and one item in the “Internalized drug use stigma” measure (Item 2: Using drugs makes me feel dirty).

In the item “family members will treat me differently”, the words “treat me differently” which in Swahili means “kutendewa kwa namna tofauti”, was not understood by one participant, who queried as quoted below:

1. “Interviewer: how often have family members treated you differently because you are living with HIV?
2. Respondent: what does it mean to be treated?”

All of the other participants explained the meaning of “being treated differently” by giving examples of stigmatizing actions such as being avoided, gossiped about by others. After testing the item in the additional ten interviews, some participants provided the correct meaning of the words and others gave examples of stigmatizing actions to explain the meaning indicating that they understood the terminology. We concluded that the words “treated differently” are understood and decided to keep them as they were.

In the question “have healthcare workers treated you with less respect when you have gone to a health facility for general health care because you are living with HIV?”, the problem was with the words “less respect” which translates to “heshima ndogo” in Swahili which does not make much sense, as linguistically in Swahili, respect is either present or absent and is not often qualified in levels. We therefore decided to change the words “less respect” to “disrespect” which translated to “bila heshima” (English without respect) in Swahili for clarity.

1. “Participant: Aaaah..the way I have understood the question, it is less respect.
2. Interviewer: Okay, by less respect, how?
3. Participant: I mean they did not respect me because of drug use.”

A similar challenge was found in the item “Most people would willingly accept someone who has been treated for substance use as a close friend”. The problem was on the interpretation of the words “close friend” which translates to “rafiki wa karibu” in Swahili. Some participants interpreted “a close friend” as “a sexual partner”.

1. “Interviewer: What do you understand by the words close friend?
2. Participant: A close friend has two meanings, a sexual partner or a normal friend.”

We therefore, used the words “close friend”, and added the words “not a sexual partner” which translates to “sio rafiki wa kimapenzi” (English: “not a sexual partner”) to clarify that a close friend was not meant to include a sexual partner. Hence the item read ““Most people would willingly accept someone who has been treated for substance use as a close friend (not a sexual partner)”.

Another challenge we encountered was the difference in the interpretation of words in specific items among participants. In the item “most people would accept someone who has been treated for substance use as a teacher of young children in a public school”, we found out that, the words “young children” and “public school” were interpreted differently among participants. The words “young children” which translate to “watoto wadogo” in Swahili, were interpreted by participants as either children in kindergarten or children in primary school or all children below 18 years. We therefore revised those words to “children in primary school” as it was the closest meaning of the item. In addition, the words “public school” in Swahili mean “shule ya umma”. Some participants understood those words as schools owned by the government while other participants understood them as schools owned by both government and private individuals. We therefore decided to remove the word public and leave the word schools only to eliminate this problem of different interpretation.

1. “Interviewer: what does public schools mean?
2. Participant: Public schools mean government schools.”

The other items with a problem with interpretation had the word ‘dirty’. For example, item: “Using drugs makes me feel dirty”, in this item, the word dirty was interpreted differently among participants, some interpreted it as being dirty physically such as not taking a bath, or wearing dirty clothes, while others interpreted it as being someone who has made a mistake. We decided to leave it as it is and wanted to decide whether it fits in the measure during psychometric testing.