Study design.

Cross-sectional, quantitative survey data assessing demographic factors, alcohol use, and alcohol-related consequences were collected from KCMC ED and RHC patients. As discussed below, preliminary quantitative analyses were used to guide our qualitative sampling strategy.

Sample size estimation.

Quantitative sample size was calculated based on the difference in average Alcohol Use Disorder Identification Test (AUDIT) scores between our RHC and ED patient populations, as the prevalence of alcohol use disorder (AUD) was a primary point of interest in the broader study [42]. Based on existing literature and the experience of local research team members, we hypothesized the prevalence of AUD to be 10% among RHC women, 15% among ED women, and 30% among ED men. Thus, at 80% power and with 90% confidence, a sample size of 587 participants was calculated to sufficiently estimate the prevalence of risky drinking between all three study subgroups. Given that the expected difference in the prevalence of AUD was narrower between the two female populations than between either male and female groups, the study team initially prioritized enrolling more women to ensure sufficient statistical power to detect differences between the two female populations.

Preliminary data analysis revealed that the prevalence of AUD was much higher than anticipated (40% in ED women and 45% in ED men). Recalculations of sample size based on these updated prevalence estimates revealed that 1,200 patients would need to be enrolled to determine a difference in proportions between these two groups. While the study timeline and funding could not accommodate the newly estimated sample size of 1200 participants, with IRB approval, the study’s target sample size was increased to enroll as many male and female participants as feasible within the study timeline.

Sampling technique.

Over the course of the study timeline, 676 patients were enrolled through a systematic random sampling strategy, with 655 participants fully completing survey questionnaires. A systematic random sampling strategy was employed to acquire a representative sample of our three patient populations, male ED, female ED patients, and female RHC patients, and better describe the characteristics of each. Except for ED females, every third patient was approached for potential study participation using the ED triage or RHC intake registries. For women seeking ED care, every eligible individual was approached as significantly fewer women than men were ED patients, and more women were needed to accurately determine the prevalence of AUD.

Data collection.

KCMC ED and RHC patients were approached for participation in this study by a Kiswahili-speaking, Tanzanian research assistant of the same gender (i.e., female research assistants approached female patients) trained in good clinical practices and specific study procedures.

After the research assistants ensured patients were medically stable, they approached patients in a quiet, private location and introduced the study. All patients had the opportunity to decline participation, but if willing to proceed, an in-depth discussion of the risks and benefits of study participation was provided. If agreeing to participate, written consent was obtained, and survey questions were delivered while within the same private, secure location. All questions were administered orally by a research team member to encourage participation from all literacy levels.

Instruments and variable measurements.

Quantitative surveys consisted of three major components: (1) collection of basic demographic data, such as participants’ age and gender; (2) assessment of typical alcohol use practices, including frequency and quantity of consumption as well as typical expenditure on alcohol; and (3) administration of validated instruments, specifically the AUDIT and Drinker Inventory of Consequences (DrInC) tools. The AUDIT is a 10-item screening tool developed by the WHO to identify hazardous, harmful, and dependent patterns of alcohol use [43,44]. Scores range from 0 to 40, with higher scores indicating more hazardous consumption and scores of 8 or greater signaling a clinically significant risk for adverse alcohol-related health effects [45–48]. The AUDIT has undergone extensive psychometric and clinical validation across diverse populations and has been specifically adapted and validated in Tanzania in KiSwahili, ensuring cultural and linguistic appropriateness for this context [49,50].

The DrInC was used to quantitatively assess the consequences of alcohol use among participants. This 50-item, yes/no questionnaire has been previously cross-culturally adapted and clinically validated at KCMC [51]. The tool evaluates alcohol-related consequences across five domains: Physical, Social Responsibility, Interpersonal, Intrapersonal, and Impulse Control (Table 1) [52,53]. Of the 50 items, 45 assess negative consequences and are positively scored, while 5 serve as control items, reflecting perceived benefits of alcohol use (e.g., enjoying the taste or drinking without problems), and are reverse scored. Responses were collected for both the past three months and across the participant’s lifetime. Prior confirmatory factor analysis conducted in the study setting demonstrated the construct validity of the tool, confirming its five-domain structure as reflective of distinct alcohol-related consequences [51]. Reliability was also supported by high internal consistency, with Cronbach’s alpha exceeding 0.8 across all domains [51]. These findings validate the tool’s suitability amongst KiSwahili-speaking participants, ensuring cultural and linguistic relevance in this context.

Download:

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

Table 1. Consequence Domains and Corresponding Items of the Drinkers’ Inventory of Consequences.

https://doi.org/10.1371/journal.pone.0331490.t001

DrInC scores range from 0 to 50, with higher scores overall indicating more affirmative responses to various alcohol-related items as shown in Table 1 below. Although the DrInC tool quantifies a range of alcohol-related consequences, it does not measure each item’s severity or frequency, and no universally accepted clinical cutoff score exists. Designed to capture the broad spectrum of alcohol-related issues, the tool has not been associated with specific cutoff scores in previous studies [51–53]. Consistent with this approach, we did not apply a cutoff score in our study, and instead, DrInC scores were analyzed in relation to alcohol use patterns as described below.

Data analysis procedure.

AUDIT and DrInC scores for the three patient populations were assessed through means and standard deviations for overall and category-specific scores. The relationship between patient DrInC and AUDIT scores was measured through a linear regression analysis to identify if an association between increased alcohol consumption and more consequences existed, with the independent variable being DrInC scores and the dependent variable being AUDIT scores. Incomplete surveys were still included in the analysis. All quantitative data were analyzed in R Studio version 1.4.

Study design.

Using a purposive sampling strategy guided by preliminary quantitative data analyses, 19 individuals who had previously been enrolled in this study and completed all study questionnaires were contacted either at the end of their survey or later by phone to gauge their interest in participating in a semi-structured IDI. IDIs were selected as the study’s qualitative method due to the highly sensitive and stigmatizing nature of the study topic, particularly for women participants. This one-on-one format facilitated open discussions, ensuring participants felt comfortable sharing their experiences while preserving privacy and confidentiality. All contacted patients agreed to participate. IDIs were conducted at KCMC and later analyzed using a grounded theory methodology.

Sample size estimation.

Of the 676 individuals initially enrolled in this study, all 655 patients who had fully completed quantitative surveys were eligible for IDI enrollment. The study team aimed to enroll 20 IDI participants — ten male participants from the ED group and ten female participants, five each from the ED and RHC groups — or until data saturation was reached, whichever occurred first. This approach was chosen to provide equal representation from both genders, while also prioritizing thematic saturation as the primary determinant for concluding data collection.

This original estimated sample size of 20 participants was informed by commonly accepted recommendations in qualitative research, which suggest 15–20 interviews as a general guideline for reaching saturation in relatively homogenous groups [54]. Saturation was also explicitly defined as the point at which no new themes emerged after three consecutive interviews for each subgroup (ED men, ED women, and RHC women). To account for the iterative nature of qualitative research, the team regularly monitored emerging themes throughout data collection and concluded after 19 interviews when saturation was reached across all subgroups. This approach helped balance initial pragmatic estimates of sample size with the flexibility needed to capture all relevant themes.

Sampling technique.

A purposive sampling strategy was employed to ensure representation across key demographic and experiential characteristics from the 655 participants who completed the initial quantitative survey. Participants were purposively selected to ensure diversity across key factors, including age, marital status, education level, occupation, tribe, and religion. Additionally, IDI participants were chosen to reflect varying personal experiences with and perspectives on alcohol use, from patients who only reported neutral or positive past experiences with alcohol, used to drink but are now abstinent, have a close friend or relative that drinks heavily, or suffered significant negative consequences from alcohol. Preliminary quantitative findings also informed participant selection to better explore notable patterns identified in the data. For example, early quantitative analysis indicated that divorced or widowed women had above-average alcohol intake, prompting the inclusion of a recently divorced woman with high alcohol consumption to explore this trend in greater depth.

To ensure a representative sample and minimize potential bias, the characteristics of IDI participants were reviewed monthly by the study lead. Any imbalances or gaps in representation were addressed through adjustments to subsequent participant selection. This iterative process ensured that diverse demographic and experiential perspectives were captured while aligning with the study’s overarching focus.

Data collection and instruments.

Qualitative data collection began approximately one month after initial survey collection in alignment with our sequential, explanatory study approach. During their initial quantitative survey, if research assistants identified participants perceived to be excellent candidates for IDI based on the previously described selection parameters, these patients were approached for IDI participation at the conclusion of their initial survey or in a later phone call. If expressing interest in participating, the same research assistant who had completed their quantitative survey with them set up a time to meet and conducted the interview in a private room at KCMC. All interviews were conducted in Kiswahili using a semi-structured interview guide with open-ended questions and probes. Research assistants were encouraged to ask their own clarifying questions when the information participants provided was unclear, conflicted with previous statements, or warranted further explanation.

The interview guide was piloted with the study lead and three KiSwahili-speaking research team members. These team members, who were not part of the final study sample, had professional backgrounds in nursing, medicine, and qualitative research, and were familiar with the target population and healthcare setting. Their feedback informed refinements to ensure the cultural and linguistic validity of the guide prior to initiating interviews with participants. Pilot interviews were conducted solely for refinement purposes and were not included in the study dataset. With the knowledge and consent of the participant, all IDIs were audio recorded using a handheld recording device for later transcription and translation, with a snack and drink provided midway through the interview. Participants were informed prior to participation that they would be given a small stipend (5000 TSH, equivalent to approximately $2 USD) to reimburse their travel expenses.

Data analysis procedure.

A grounded theory approach was chosen as this topic remains poorly understood in the local context, with little prior research foundation from which to draw. Unlike thematic analysis, which is often used to organize and interpret pre-existing frameworks or themes, grounded theory is designed to generate new, theory-driven insights directly from the data itself [55–57]. This approach allowed us to iteratively learn from participants’ responses and develop a holistic understanding of alcohol use perceptions specific to this setting, rooted in the data rather than external frameworks.

US and Tanzanian study team members co-developed the initial codebook from the first several IDIs. This codebook was inductively and deductively generated, using broad categories from the interview guide as an initial framework while also enabling organic generation of codes without preconceived notions, in alignment with grounded theory methodology [55–58]. The evolving codebook facilitated the creation of content memos for each emerging theme and sub-theme, which formed the basis of our qualitative findings.

Interviews were coded by both Tanzanian and US research team members who independently coded initial interviews and then met to discuss and resolve any coding discrepancies. This process was repeated until an 80% agreement was reached. At that point, with approval from the research team, the lead analyst completed the coding of all remaining interviews, updating the codebook accordingly. After the preliminary quantitative analyses, qualitative and quantitative data collection occurred in parallel for the remainder of the study period. All qualitative analysis was conducted in NVivo software version 12 and approved by the entire research team prior to reporting these results.

Demographics.

Six-hundred and seventy-six patients started, and 655 patients completed the survey questionnaire. The average age of all ED and RHC participants was 41.7 years old (SD = 18.7) (Table 2). Most participants identified as Christian (80%), married (50%), and employed (57%) at the time of survey delivery. The income distribution varied, with 32% earning ≤ 50,000 TZS per month and 39% earning > 200,000 TZS per month. Education levels ranged widely, with 29% reporting college-level attainment and 8.7% reporting no formal education. Further details on demographics, including tribal affiliation, are reported in Table 2.

Download:

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

Table 2. Study Population Demographics.

https://doi.org/10.1371/journal.pone.0331490.t002

Alcohol consumptions and related consequences.

Descriptive analyses revealed significant variation in alcohol consumption and related consequences across populations. ED men had the highest average AUDIT scores (mean = 6.76, SD = 8.16), followed by ED women (mean = 3.07, SD = 4.76) and RHC women (mean = 1.86, SD = 3.46). Similarly, DrInC scores showed that ED men experienced the most alcohol-related consequences (mean = 16.4, SD = 19.6), followed by ED women (mean = 9.11, SD = 13.1) and RHC women (mean = 5.47, SD = 9.33) (Table 3). DrInC subscores for physical, interpersonal, social responsibility, and impulse control consequences followed the same trend, where ED men followed by ED women and RHC women faced a higher burden of alcohol-induced physical, interpersonal, intrapersonal, impulse control, and social responsibility harms (Table 3, Fig 1). Of note, ED women showed the largest variability in DrInC scores in several subcategories, including physical and intrapersonal consequences (Fig 1).

Download:

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

Table 3. Reported Average AUDIT and DrInC Scores across Patient Populations.

https://doi.org/10.1371/journal.pone.0331490.t003

Download:

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

Fig 1. Box-Plot Distribution of DrInC Scores and Subscores across Patient Populations.

https://doi.org/10.1371/journal.pone.0331490.g001

Linear regression analyses demonstrated a significant linear association between AUDIT and DrInC scores across all groups (Fig 2). The association was strongest for ED women (R² = 0.76, p < 0.001), reflecting that higher alcohol use strongly correlates with more alcohol-related consequences in this group. The association was slightly weaker for ED men (R² = 0.65, p < 0.001) and RHC women (R² = 0.53, p < 0.001), indicating population-specific differences in the impact of alcohol consumption (Fig 2).

Download:

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

Fig 2. Association of AUDIT and DrInC Scores across Patient Populations.

https://doi.org/10.1371/journal.pone.0331490.g002

Inability to fulfill family responsibilities.

Most participants mentioned that alcohol was an inhibitor in one’s ability to fulfill their responsibilities to their family. Those who drank in excess would “forget [their] responsibilities” (IDI #11, Male), become too “weak” (IDI #3, Female), “lazy” (IDI #19, Female), or simply, “fail to perform their daily duties” (IDI #6, Female). One participant elaborates on their own experience:

“I see alcohol is not such a good thing…It can make one forgetting everything about his/her family, even children can go to bed hungry without you having an idea because you’re busy drinking. Children left home by drunkard parents can even experience assaults of different kinds just because the parents are not there to protect them…I advice my fellow parents and those who have families to care for to stop taking alcohol” (IDI #7, Female)

Stigma and discrimination.

All interviewees noted alcohol-related stigma to be present to some extent in their communities, but to varying degrees and concentrated in certain circumstances. One participant reported that they “haven’t seen any stigma among men or women…unless if after drinking you cause problems like fighting or abusing other people” (IDI#12, Male). As another explains, stigma could also arise when one fails to accomplish preset responsibilities or drinks unsafe alcoholic beverages:

“Alcohol intake in Tanzania is like a ritual. Anyone can drink. The stigma arises when someone fails to control their drinking habit, like when someone drinks all the time and fails to fulfill his or her obligations, or if someone drinks those homemade beers (local brews) that are not tested and not standardized by the bureau of standards. People who drink Gongo and Dadii are really stigmatized in the community. Because once you drink those, you may even fail to control your bowels, which leads to shame and embarrassment” (IDI #2, Male).

In addition to the quality of liquor, another notable facet of alcohol-related stigma is its interaction with gender. Participants noted that “women who drink are stigmatized more than men” (IDI #7, Female) and that their “community strongly hate women who drinks alcohol” (IDI #19, Male). This arose in part because of aforementioned gender roles, where because “women stay with the children most of the time” (IDI #19, Male) there is concern that their alcohol habits will have greater negative repercussions on their children.

Sexual consequences.

Interviewees reported that alcohol use caused or contributed to significant sexual consequences uniquely experienced by each gender. The first of these was the increased risk of acquiring a sexually transmitted disease – “when you drink too much it can put you a risky of even practice unsafe sexual intercourse” (IDI #4, Male), which generally stemmed from having more sexual partners:

“Men who drink lacks discipline, they don’t respect themselves or their wives, they engage in sexual immorality without considering the fact that they are aged, married and with families.” (IDI #7, Female)

A couple of men and women participants noted that alcohol could hinder their ability to perform sexual acts –

“A person cannot perform sexual activity well as before, to men they may encounter erectile dysfunction and women may also not be able to involve and perform sexual activity properly. Alcohol can make a person really desire the sexual intimacy but just not able to perform it because of the effect of alcohol in the brain.” (IDI #16, Female)

This, though, was only mentioned by a few participants and stands in sharp contrast to the common theme of alcohol-related sexual assault. This final sexual consequence was described almost exclusively as men being violent against women – “there is someone I know she was invited to a man house to drink alcohol it was her first time to drink alcohol she got drunk the man raped her” (IDI #9, Female). Participants disagreed on whether or not men incurred sexual assault when drinking, one saying “men…cannot experience sexual violence because it is not a common practice in Tanzania” (IDI #2, Male), while another reported that “men…when they are drunk, they have been sexually abused by the fellow men” (IDI #13, Female).

Interpersonal conflict and violence.

Sexual assault was not the only product of the increased violent and aggressive behavior that alcohol use incited. Interpersonal conflict and violence was a theme noted by three-quarters of participants. IDI #19 states,

“Alcohol causes misunderstandings and fights in society, I have witnessed many times that people get drunk and start to say abusive words, become very aggressive and arrogant.” (IDI #19, Female)

Respondents described how alcohol use instigates emotional and physical conflict and distress within the family – “I know many families that when the father is drunk, the mother and children become so anxious and afraid to sit with him closely. The most common violence is physical and emotional abuse” (IDI #17, Male). As seen in this and the next quote, aggression was described as being more common in men – “it’s very difficult to see women fighting after drinking. But men are always aggressive and violent” (IDI #3, Female), with the ensuing conflict leading to family disintegration.

Moreover, several participants (4 of the 19 participants) discuss how a child’s mental health is significantly affected by witnessing their parents “fighting and shouting” (INT#12, Male), a form of emotional conflict, due to alcohol consumption. As a result, many pre-existing interpersonal relationships were harmed or broken – “I have witnessed many times that people get drunk and start to say abusive words to one another...I can say that alcohol can truly bring people together but can also destroy people’s relationship” (IDI #19, Female).

Injuries and chronic disease.

While all the categories of harms thus far discussed can lead to significant physical, mental, and emotional problems, a plethora of other long-term health issues stemming from alcohol use like injuries and road traffic incidents, illness and chronic disease were also discussed by most participants.

“Alcohol can lead to organ damage like liver so there is a lot of complications...I had a friend of mine who was drinking every day, and he was not eating well so he ends up dying. Because of alcohol intoxication. But before that he was having anemia (low blood), low body weight, and muscle wasting” (IDI #12, Male)

Health issues that were attributed to alcohol use by participants included the following: liver cirrhosis, cancer, kidney failure, “puffy cheek” (water retention resulting from alcohol use), neurological insults, and even death. Some participants even recount how local brews —which are “not tested and not safe” (IDI #4, Male) and have been characterized as “strong alcohol” (IDI #8, Male)— may incite even greater health consequences in regular consumers.

Financial instability.

Lastly, almost half of the participants mentioned economic loss and financial instability, encouraged via a variety of different mechanisms, as a major source of personal and community harm. First, participants proposed that economic loss and financial instability seem to arise from the perceived difficulty of drinkers to find or maintain employment. That is, participants agreed that employed individuals who drink experience a loss of productivity. IDI #17, for example, stated, “Alcohol-related behaviors may lower the individual performance at (the) workplace which leads to inadequate or inefficient production” (IDI #17, Male).

In addition, if an individual spends “a lot of money buying alcohol” (IDI #12, Male) for themselves, this can take a toll on personal savings –“I lost my business because of drinking too much” (IDI #12, Male). One participant describes the monetary loss alcohol-related spending could cause:

“ Alcohol affected my uncle so badly, he was so rich, and helped people in his family and relatives but once he became addicted with alcohol everything turns to zero. He lost everything, even closer friends. He is unable to take care of his family even paying school fees for his children!” (IDI# 5, Male)

As men especially were seen to be the primary breadwinners for their families, a father’s excess spending on alcohol could “decrease in family income” (IDI #17, Male). This created negative repercussions for a man’s wife and children and permanently cripple their financial holdings:

“Alcohol is a reason for many families here to fall into extreme poverty because of men as family providers spend a lot of resources in alcohol while they just earn little to help their families survive as most of them work as drivers of tricycles and motorcycles. So, when a man has no control of his drinking habits and spends all of his income into drinking and definitely the family get into a vicious cycle of poverty.” (INT #19)

While increased alcohol consumption was seen as contributing to personal or familial financial distress for some, as will be discussed in the next section, several participants noted a financial benefit to increased alcohol use for sellers or the government.

Economic growth.

In contrast to the last sub-theme presented, IDI #2 described the monetary benefit alcohol brought to his community, saying it

“contributes to the economic growth of an individual who is selling, a family or a national income. The government collects taxes from the brewing industry, which helps to increase national income” (IDI# 2, Male).

Interestingly though, while the government benefited financially from alcohol sales, several respondents discussed how “local brew” (IDI #4, Male) is often not effectively tested for alcohol content levels and results in harmful consequences for their community. One participant explained this lack of government monitoring:

“The community views [alcohol] as a burden promoted by the government, it is the government that legalized it and allow people to manufacture alcohol. And sometimes the government fails to monitor these industries and see what kind of alcohol they are producing. It is very unfortunate that some of the businessmen are not faithful and manufacture alcohol which are very strong and damage people’s brain.” (IDI #12, Male)

Importantly, while the production of alcohol was seen to bring economic benefit for some, even this perceived boon carried with it the potential for harm to community members.

Upholding cultural practices.

As mentioned in the first quote in this manuscript, alcohol use was tied to cultural and tribal events – “Many people consider alcohol to be part of their social culture and way of life” (IDI #2, Male). Some respondents found it appropriate to consume alcohol at cultural events or special occasions, using alcohol during rituals as a way to “signify respect and honor to our ancestors” (IDI #2, Male) and “honor their ancestors’ way of living, social customs that have been practiced for generations and generations” (IDI #3, Female). One woman summarized these sentiments:

“There are social events that cannot be done without alcohol, for example, if you go to pay for dowry you have to bring alcohol to the event and even in some rituals, they must brew alcohol during the ritual service. So, there’s a positive way alcohol affects our society as a whole…many people drink alcohol to spend time together and enjoy” (IDI #9, Female)

Social unity.

Finally, alcohol’s positive influence on the social aspect of community life was a common theme in IDIs – “alcohol helps people interact and have time to discuss different issues…it strengthens unity among people” (IDI #4, Male) and

“[Alcohol] helps to bring people together in unity. You may find friends or just random people sitting together, making stories about a thing or two and celebrating their achievement over a glass of beer. In my society, you cannot just gather around grown-up people to discuss about something and just give them soda or juice, you must offer them alcohol so that they will listen to you properly”(IDI #19, Female).

“The good results of alcohol, to me personally it has made my social life very great and I have been meeting a lot of people when I go for a drink, for my friends as we enjoy drinks together and talk about our lives but I have also met important and influential people to discuss important matters in business and work as we enjoy 1 or 2 glass of beer” (IDI #10, Female)

Not only was alcohol use noted as bringing people together, but alcohol was also mentioned as having a key role in resolving social conflict and social gatherings:

“If you have misunderstandings with friends or family members, and you want to resolve those disputes…you need to offer them alcohol while you talk! You can’t offer them soft drinks! Even when you want to get married, you need to buy a lot of alcohol for elderly people to drink and enjoy” (IDI #2, Male).