Risk factors for problematic alcohol use among male waste pickers and caddies in Johannesburg, South Africa: a cross-sectional study

Abstract Informal workers may be prone to problematic substance use due to many factors, including adverse working conditions and low income. The aim of this secondary analysis was to investigate problematic alcohol use risk factors among male informal workers in Johannesburg, South Africa. Alcohol use among the two groups of informal workers in the analysis y golf caddies and waste pickers was measured using the World Health Organization (WHO) Alcohol Use Disorder Identification Test (AUDIT) tool. The WHO self-reporting questionnaire (SRQ) for common mental health disorders (CMD) was used to assess mental health. A multivariate logistic regression model was used to explore the predictors of problematic drinking in informal workers. The study consisted of 514 participants, of which 48.4% were golf caddies and 51.6%, waste pickers. Most participants were younger than 40 years (50.9%). Over half of the participants (54.7%) were alcohol consumers and 74.1% were smokers. Over 60% of the participants who were alcohol consumers had a probable drinking problem. Unstratified regression results showed that common mental distress (aOR = 1.06; 95%CI: 1.01–1.09), age: 30–40 years (aOR = 2.17; 95%CI: 1.18–3.97), smoking (aOR = 2.25; 95%CI: 1.34–3.79), and other water sources (aOR = 0.2; 95%CI: 0.04–0.99) were associated with a probable alcohol problem. Waste pickers (aOR = 0.33; 95%CI: 0.20–0.70) were less likely to be problematic drinkers compared to golf caddies. Problematic drinking in this study was common in both caddies and waste pickers along with smoking. Problematic alcohol use was associated with caddying, mental distress, age, and smoking. Measures such as providing counseling services to informal workers and improvement of working conditions may help change the behaviors of these vulnerable groups.


Introduction
The International Labor Organization (ILO) defines an "informal economy" based on two methods. The first method is based on whether an enterprise in which workers are employed is registered with the authorities. The more recent definition of the second method covers all work that is not covered by formal arrangements-own account work, employment in the informal sector, and employment in the formal sector that is not regulated or protected. [1][2][3] The informal economy is often characterized by low-income opportunities, temporary employment, low productivity, and small-scale enterprises using labor-intensive activities. 4 South Africa (SA) is characterized by a relatively small economy, high levels of unemployment, high poverty levels, and high-income inequality leading to informal employment. 1,2 In 2018, the informal sector comprised approximately 2.5 million workers and business owners. 5 Golf is an important industry in SA. 6 In addition to carrying golfers SAros during their game, caddies are expected to clean golf clubs' wash the dirt off the golf balls and offer helpful advice to the golfers. 7 Caddies are independent contractors who work on the golf courses in South Africa, often the clubs provide a uniform to the caddies and a meal from the set payment. Fewer golf players use caddy services when carts are available, so often caddies only receive work a few times a week. Little research has been conducted on the health and lifestyles of this often unrecognized group of informal workers. 8 Waste picking and solid waste recycling are a prominent part of the urban landscape in many cities in the country. 9 Waste pickers collect waste from public places such as garbage dumps and streets and sell it to buy back centers; despite working full time they have low socioeconomic status. 10 A high proportion of South African waste pickers have been shown to consume alcohol (41%) in a study by Viljoen et al. 10 Workers in the informal employment sector are generally not protected by employment legislation 11 and lack guaranteed wages and benefits. 7 This lack of employment protection may lead to adverse mental and physical health. 11 In a study conducted among Brazilian formal and informal employees, common mental disorders (CMD) had a higher prevalence in informal employees before and after adjusting for sex, age, marital status, and migration (OR 2.16, 95% CI 1.3-3.7). 11 Mental disorders and social exclusion have been suggested to be drivers for substance abuse in waste pickers. 12 SA, as a low-middle income country has been reported to be one of the countries with high levels of alcohol consumption in a small proportion of citizens by the World Health Organization (WHO). 13 In 2016, the adult per capita (APC) alcohol consumption in SA was 9.3 liters (ranked 52nd in the world) of pure alcohol. 13 Amongst alcohol consumers in the African Region, SA has the third-highest APC after Eswatini and Namibia. 13 Between 2014 and 2015, South Africa reported a prevalence of heavy episodic drinking (binge drinking) of 43.0% in both sexes. 14 Amongst drinkers, males had a high prevalence of 48.2% of binge drinking. 14 Previous studies have linked informal work to smoking and drinking. 15,16 In American rural areas, alcohol consumption has been identified as an influence on one0% in both sexesiaol consumption in a small proportion of citizens by the World Healt. 12 There is evidence that both smoking and alcohol also contribute to the lifestyle associated health-inequalities in South Africa. 17 WHO defines a standard drink as a 10lg of pure alcohol, and advises that both men and women not to drink more than two standard drinks a day. 18 Problematic alcohol use is defined as heavy drinking or drinking that is accompanied by unpleasant consequences. 19,20 Alcohol use disorders include alcohol dependence, high intensity-drinking, difficulty cutting back, life interference, and or hazardous use such as binge drinking. 21,22 Depression, anxiety, difficulty sleeping, suicidal thoughts and attempts, as well as abuse of other notorious drugs, have been associated with alcohol dependency. 23 The aim of this paper was to investigate problematic alcohol use risk factors among male informal workers in Johannesburg, South Africa.

Primary studies
The waste pickers study The Johannesburg waste pickers study aimed to generate new knowledge by achieving a greater understanding of the nature and depth of potential health risks and outcomes posed to people informally working on landfills. There are three active landfill sites in Johannesburg; participants were selected from the two major landfill sites.
One of the study sites had approximately 600 waste pickers and the second one had approximately 3000 waste pickers. These landfill sites were chosen because of their proximity to the most densely populated area and they had the largest numbers of waste pickers.

The golf caddies study
This cross-sectional study aimed to investigate the working conditions (exposures) and health outcomes, as well as health care access, associated with working as a caddie at golf courses in Johannesburg. The participants were invited from six (randomly chosen) of 17 golf courses in central Johannesburg.
Golf caddies in Johannesburg do not report to work daily, thus we could not ascertain the total number of caddies working on each golf course. Convenience sampling was therefore used to include caddies in the study.
Both studies were conducted by the National Institute for Occupational Health (NIOH), Epidemiology and Surveillance Section. After the acquisition of informed consent, structured face-toface interviews using a structured questionnaire were performed by trained fieldworkers, with local language translation where possible, using electronic RedCap data processing software. Basic health screening was also conducted by trained nurses in the field offices.

Power of study
As this was a secondary analysis of two surveys, a power calculation was conducted to determine the power of the combined analysis. The power analysis was conducted using a two-sided test in STATA with a sample size of 218 for drinkers and the proportions of harmful drinking between the two groups, alpha was set at 0.05 and the power of the study was 93.4%. Convenience sampling was used to enroll participants in both studies and no participants were excluded based on their basic health screening results

Measures
This study used data from the NIOH, Epidemiology and Surveillance Section to determine risk factors for problematic alcohol use among male golf caddies and waste pickers in Johannesburg.

Demographic information
The demographic characteristics of the participants analyzed were age, nationality, socioeconomic proxies such as education level, cooking source (e.g. electricity, paraffin), water source, toilet type, and average income earned monthly. The average amount of money reported spent on lifestyle habits such as drinking and smoking was compared between the two groups of informal workers. Work stress was defined differently for caddies and waste pickers. In the golf caddies study, work stress was defined as feeling intimidated by golf players. In the waste pickers study, it was defined as experiencing physical violence at work with other waste pickers or with security guards. No female caddies were present at the visited golf courses; therefore, this secondary analysis only includes male participants.
For data analysis purposes, age was categorized into <30, 30-40, 41-50, and >503years old as few caddies were less than 30 years and few waste pickers were less than 20 years of age. However, for regression purposes ( Table 5), age was categorized differently for caddies and waste pickers due to the differences in distribution: caddies were categorized as gori-40syears" and "41 and above" and waste pickers were categorized as "<30 years," "31-40 years," and "41 and above." The nationality of the participants was considered as either South African or non-South African. The participants' education status was defined as "no schooling," "primary school," "secondary school," and "tertiary school levels." The water source was grouped into "household tap," "communal tap," and "other" (river, dam or reservoir). The toilet type was categorized into rized The toilet type was ," into rized The t," and rized The toilet type was to "household taptionthly income was classified using the South African rand currency (R) as "less than R1000," "R1000 to R2500," "more than R2500 to R5000," "and more than R5000." The average monthly amount spent on alcohol and smoking was organized into "less than R500," "500 to 1000" and "more than 1000." These were both approximate values as neither group earned or spent money monthly but rather as the money was received.

Alcohol use
Alcohol use amongst the two informal worker groups was measured using the WHO Alcohol Use Disorder Identification Test (AUDIT) tool, a screening tool for hazardous and harmful alcohol consumption. 24,25 The AUDIT questionnaire has 10 questions which are divided into three groups. Questions 1-3 (frequency of drinking, the typical quantity of drinking, and frequency of heavy drinking) assess hazardous alcohol use. Questions 4-6 (impaired control, increased salience of drinking, and morning drinking) assess dependence symptoms. The last four questions 7-10 (contained guilt after drinking, blackouts, alcoholrelated injuries, and concerns from others) assess the harmful effects of alcohol use. 23 The AUDIT score has a possible minimum score of zero for nondrinkers and a maximum score of forty. Based on previous research, two different scoring systems were used in this current study. The first scoring system used a score of > ¼8 to indicate a probable drinking problem and a score of > ¼13 to indicate probable alcohol dependence. 24,25 The second scoring system consists of four alcohol use levels: low risk (0-7 AUDIT score), high risk (8-15 AUDIT score), heavy drinking (16-19 AUDIT score), and hazardous drinking (20-40 AUDIT score). 25 However, in our analysis, ous drinking (20 alcohocategorized as an AUDIT score of 16-40. Drinking was defined as those who answered ohol use levels: low risk (0oring systems were used in this current study. The first scoring system used a score of > ¼8 to indicate a probable drinking proble Binge drinkers were categorized into rizeders were ined as those who answered ohol use levels: low risk (0oring systems were used in this current study. The first scoring system used a score of > ¼8 to indicate a probable drinking problem and a score of > ¼13 x or more standard alcohol drinks less than monthly. Frequent binge drinkers consisted of those participants who had had six or more standard alcohol drinks either monthly, weekly, or daily. Mental distress WHO self-reporting questionnaire for common mental health disorders (CMD) 26 was used as a case-finding instrument in both the caddies and the waste pickers studies. The SRQ20 consists of 20 questions that assess neurotic (depression, anxiety, psychosomatic) symptoms. 26 Each of the 20 items was scored 0 or 1. A score of 1 indicated that the symptom was present during the past month; a score of 0 indicates that the symptom was absent. The maximum score is therefore 20. A score of present during the past month; a score of 0 . 11,26 Data analysis All statistical analyses were conducted in STATA 16 (Stata Corp LLC, USA). Descriptive statistics were used to describe and summarize variables and compare the two informal worker groups including means, standard deviations, medians, frequencies, and percentages. Possible effect modifiers or confounding variables included were age, informal work type, education, housing, monthly income, source of cooking energy and water, smoking, and mental distress score (SRQ 20).
An independent student t-test was used to determine significant differences in the mean alcohol use scores between caddies and waste pickers. A proportions test was used to compare differences of proportions in demographic information and AUDIT risk levels between the two groups of informal workers. The Chi-square test (Fischer's exact test for expected values less than 10) was used to test for an association between AUDIT risk levels and symptoms associated with CMD.
Univariate analyses on all variables were conducted and a significance level of 95% was assigned to all tests. A multivariate logistic regression model was used to explore the predictors of problematic drinking in informal workers. The first scoring system of the AUDIT tool was used in the multivariate analysis. The AUDIT score was coded into a bivariate variable with those scoring less than 8, unlikely to have a drinking problem, and those with a score of 8 and greater were coded as having a probable drinking problem. The SRQ 20 score achieved by each respondent was used as a categorical variable. A backward regression method was used to determine the adjusted effect of different risk factors for problematic alcohol use in informal workers. Variables with p < 0.05 in the univariate analysis were kept in the final model if they produced a positive change in the pseudo R2, and improvement in the Akaike's Information Criteria (AIC) and Bayesian Information Criterion (BIC). Variables considered important risk factors for problematic alcohol use were kept in the model regardless of p-value and Akaike's Information Criteria and Bayesian information criterion scores. The variables fitted in the multivariable model were categorical except for "income" which was a continuous variable.

Population description
A total of 514 male informal workers were interviewed. The study group had a median age of 38byears. The waste pickers were significantly younger than the caddies, with 48.3% less than 30 years old, while only 2.0% of caddies were less than 30 years old (proportion test p < 0.0001, Table 1). Very few caddies (1.6%) and waste pickers (2.6%) had no formal education. Significantly, more waste pickers than caddies completed secondary school education, 84.5%, and 70.3%, respectively (proportion test p ¼ 0.0001). However, more caddies (6.8%) than waste pickers (1.1%) had tertiary education. Approximately 2.8% of caddies and 14.3% of waste pickers consenting to the survey were not South African but had come from neighboring countries including Zimbabwe, Lesotho, and Mozambique.

Socioeconomic status
Describing socioeconomic status, the majority of caddies and waste pickers used electricity as their cooking source, although more caddies used electricity than waste pickers (p < 0.0001). Approximately 74.7% of caddies and 65.3% of waste pickers had access to municipal water in their houses or yards. There was a difference in the proportion of caddies (61.5%) and waste pickers (53.2%) who used flushing toilets in their households; however, the difference was not significant. The larger proportion of caddies (48.2%) and waste pickers (54.3%) earned between R1000 to R2500 a month. However, on average caddies earned more than waste pickers, each group earning a mean of R2837.94 and R1906.59, respectively. Despite this low income, 27.3% of caddies and 9.1% of waste pickers reported spending approximately R500 to R1000 a month on alcohol. This figure is approximate as the caddies and waste pickers did not earn or spend monthly and they needed to calculate an amount.
Of the 514 participants, 392 (76.3%) provided answers to the AUDIT questionnaire. Of these, 137 participants (34.9%) reported never drinking. Of the 281 (54.7%) who reported ever drinking, 190 (67.6%) participants indicated ever binge drinking, and 127 (45.2%) were frequent binge drinkers monthly or more frequently (Table 2). Table 2 shows the mean total AUDIT score, which was higher in caddies compared to waste pickers (7.38 and 3.76 respectively, test p < 0.0001). Caddies showed significantly higher mean scores for all alcohol consumption behaviors (see Table 2). There was no significant difference in alcohol dependence symptoms between caddies and waste pickers, who both had low scores. However, the difference in total probable alcohol dependence scores between groups was marginally significant (AUDIT score !13, p ¼ 0.0596). Neither group was more at risk for a probable drinking problem (AUDIT score !8, p ¼ 0.6630). However, mean scores for alcohol-related problems such as blackouts, injuries due to alcohol, and loved ones/doctors concerned were significantly higher in caddies than in waste pickers.
Our results also showed that waste pickers who consumed alcohol (78.0%) were significantly more likely to binge drink than caddies (62.8%) (p ¼ 0.0051). Even though waste pickers participated more in risky and binge drinking than the caddies, harmful drinking (AUDIT score of 16-40) was more prominent among caddies than waste pickers, 23.8%, and 8.3%, respectively (p ¼ 0.0009).

Mental health and alcohol use
There was no significant difference in the prevalence of CMD between the two groups of informal workers (p ¼ 0.329) ( Table 1). In total, 26.2% of the participants had a score of 8 and greater and were therefore positive for CMD. Of the participants who experienced mental distress, 44.8% were caddies and 55.1% were waste pickers (not in table). Workers with a higher level of alcohol risk generally reported more mental health condition symptoms ( Table 3). Symptoms that were significantly different between the alcohol risk levels were shaking hands (p ¼ 0.015), feeling nervous (p ¼ 0.028), trouble thinking clearly (p ¼ 0.013), feeling unhappy (p ¼ 0.033), difficulty making decisions (p ¼ 0.015), loss of interest in things  (p ¼ 0.031), suicidal thoughts (p ¼ 0.008), and uncomfortable feelings in the stomach (p ¼ 0.032). Poor appetite and difficulty in enjoying daily activities were marginally significant, 0.078, and 0.099, respectively.
There was a significant relationship between CMD and risky and harmful drinking in this study (Chisquare test, p ¼ 0.005) (Table 4), with the difference in proportions increasing as the audit score increased. A significant association between smoking and mental distress was also found (Chi-square p ¼ 0.003) despite the high proportion of smokers.

Discussion
This study aimed to describe problematic alcohol use and risk factors among two groups of informal workers in South Africa, as few studies have looked at alcohol abuse and mental health simultaneously. Little work has been done on caddie's health, thus this study comparing the two groups provides novel information. 8 Identifying the risk factors that are associated with substance abuse is important to inform interventions for the communities. This study found that problematic alcohol use was linked to caddying, mental distress, age, source of water supply, and smoking among informal workers. Our analysis consisted of male caddies or waste pickers in Johannesburg, South Africa-two groups of informal workers. Both caddying and waste picking are considered low-skill jobs and have poor working conditions 10 and both groups receive no social benefits (e.g. sick leave days, pension funds). 27 Although caddies work for registered golf courses, they are not legally employed by the golf courses in South Africa. Caddies do not have guaranteed salaries, they may get to the golf course and not work a single round of golf and therefore, earning no money for the day. The precarious nature of salaries in the informal sector makes caddies and waste pickers vulnerable to poverty and substance abuse. 27 We ascertained from the results of the interviews that over half of the informal workers in the study were alcohol consumers, a higher proportion of caddies were alcohol consumers compared to waste pickers. However, drinkers in both groups were substantially higher than the figure for the South African general population of 31%. 12 Smoking was very common amongst both groups of informal workers although contrary to alcohol consumption a higher prevalence of smoking was seen in waste pickers compared to caddies. Both proportions were exceedingly higher than the national smoking prevalence (18%) in South Africa. 28 A smaller total of workers were positive for mental distress (based on the results of the SRQ 20). Our results also showed that amongst other factors common mental disorders were associated with problematic drinking among informal workers (Table 5). However, the relationship between alcohol use and mental distress is a complex one, and we cannot ignore the reverse causality that may occur between problematic alcohol use and mental health disorders. 29 This may call for a specific tool that uses symptoms independent of excess alcohol consumption in identifying mental distress in harmful drinkers.
In 2016, the WHO reported that SA had a relatively high alcohol abstention rate (abstaining from drinking for at least 12 months) (69%), however, the relatively smaller percentage of the population that consumes alcohol (31%), are mostly binge drinkers (consuming 60 grams or more of pure alcohol on at least one occasion over 30 days). 13 Similar to the WHO findings, the majority of alcohol consumers among informal workers in this study, 54%, which was higher than the national figure of 31% were binge drinkers (68.9%). This may be related to the precarious nature of their income.
This may be attributed to the difficult working conditions and financial stress they are subjected to. 15 The may be attributed to s on money wages, usually experiencing fluctuations and never having income security. 30 A majority (70%) of the informal workers in our study earned R2500 per month which is well below the legislated minimum wage of R3300 for 8-hour shifts placing them in the poor group. The work of Mukong et al. provides evidence that the high prevalence of smoking (43.4%) and alcohol consumption (43.4%) among the "poorest" and "poor" in South Africa contribute to the income-related health inequality seen in South Africa. 17 This indicates that the high levels of smoking and alcohol abuse in this study in both groups may add to their health outcomes.
As aforementioned, problematic alcohol use was associated with caddying, mental distress, age, source of water supply, and smoking among informal workers. We saw that the average income of the informal workers may be a possible explanation for the type of water source they had in their households. Informal workers who did not have running water in their households were less likely to be problem drinkers. This could be because they cannot afford to drink regularly and/or excessively.
When stratified by type of informal work, both mental distress and smoking were associated with a probable drinking problem in caddies. While in waste pickers, only age was associated with probable problematic use of alcohol. The caddies were significantly older than the waste pickers and their difference in alcohol consumption is suggestive of different coping mechanisms used to deal with the challenges and problems encountered in life.
Coping motives have been associated with problem drinking and drug abuse across diverse populations. [31][32][33][34] This means that the motives for substance use or abuse differ between populations and individuals. Coping skills may include behavioral disengagement, denial, mental disengagement, and substance use. 35 To cope with stresses, caddies were more likely to turn to drinking while most waste pickers turned to binge drinking, smoking, and the use of other illicit drugs (Table 1). These different coping mechanisms can be attributed to the difference in age, as there was no significant difference in the proportion of caddies and waste pickers with mental distress (Table 1). Because waste pickers were generally younger than caddies, they were more open to exploring with other coping mechanisms other than alcohol to deal with their problems. Even though in our study smoking was not significantly associated with problematic drinking amongst waste pickers this may be due to the high proportion of smokers in the sample. In another study conducted amongst waste pickers, results showed that smoking was significantly associated with mental distress. 36 Working conditions and the work setting may also influence the requirement of coping mechanisms. 11 Migrant informal workers in the United States of America stated that they often drank alcohol and smoked marijuana to escape feelings of sadness, loneliness, and work and life-related anxiety. 15 We found similar results in our analysis, informal workers with higher AUDIT scores showed more symptoms for mental distress, these included shaking hands, feeling nervous, feeling unhappy, difficulty in making decisions, and loss of interest in day-to-day activities. However, it is important to note that the symptoms used for mental distress are nonspecific and can be presented by individuals with alcohol-related problems. 12 This makes it difficult to differentiate individuals with alcohol-use disorders or mental distress symptoms as they present with similar general complaints such as fatigue, insomnia, anxiousness, and sadness. 23 Lack of security of employment, as well as less control of salaries, may contribute toward the risk of problematic drinking, substance use, and mental distress in informal workers. 11,37 Limitations Self-reported alcohol consumption and risk factors may well be subject to information bias. There was likely an underestimation of alcohol use and common mental disorders among these informal workers. Moreover, the absence of a clear alcohol-consumption use question in the waste pickers study could have led to lower responses to alcohol-related questions in the AUDIT questionnaire. Subsequently, this might have led to an underestimation of the number of alcohol consumers among waste pickers. Also, comparison of work stress between golf caddies and waste pickers may not be accurate due to the different definitions used in each respective study.
We acknowledge that the translation of the AUDIT and SRQ tools into native languages may have affected the reliability and validity of the tools. However, various versions of the AUDIT tool are reliable sub-Sahara Africa. 38 It was difficult to calculate the response rate of the participants since they were informal workers and there were no formal numbers of the number of workers on the landfill sites and golf courses.

Conclusion
This study enrolled male participants of which included golf caddies and waste pickers. Over half of the participants were alcohol consumers of which most had a probable drinking problem. Common mental disorders, type of informal work (caddying), age, and smoking were associated with a probable alcohol problem amongst two groups of male informal workers in Johannesburg. However, when stratified by type of job, we reported differences in risk factors for problematic drinking between caddies and waste pickers. Informal workers of different ages might use different coping mechanisms for dealing with problems. Overall, problematic alcohol among workers working outside the protection of employment legislation in low-and middle-income countries may lead to socioeconomic inequalities.

Recommendations
Future longitudinal studies could help better understand the relationship between alcohol use and mental distress in informal workers. Understanding problematic alcohol use and its risk factors in can help support intervention programs in informal workers. Policy interventions, health education/awareness, and support targeted at informal workers and those of lower socioeconomic status should be implemented to curb problematic substance use. Interventions to reduce and support mental distress may have an impact on substance abuse.

Consent for publication
All parties involved have given consent for publication.

Disclosure statement
There are no competing interests to disclose.

Ethics approval and consent to participate
The ethical clearance certificates for both the golf caddies (clearance number: M171120) and waste pickers (clearance number: M171120) studies were obtained from the Human Research Ethics Committee (Medical) of the University of the Witwatersrand.

Funding
The National Institute of Occupational Health funded the studies.