Background
Alcohol use is a leading risk factor for disease, disability and death globally. It is a causal factor in at least 200 disease and injury conditions [
1], and results in three million deaths per year globally [
2]. Socioeconomic status (SES) is also a significant factor in the association between alcohol use and risk of related harm [
3]. Lower SES is associated with an increased risk of alcohol-attributable conditions, even after accounting for risk and confounding factors for these conditions [
4]. An alcohol harm paradox exists, in which drinkers from lower SES groups are at greater risk of alcohol-related harm, despite drinking at the same level or less than drinkers from higher SES groups [
3,
4]. Evidence from a systematic review, examining the role of alcohol use and drinking patterns in socioeconomic inequalities in mortality, highlights that a pattern of heavy episodic drinking among individuals of lower SES may partly explain the association between SES and mortality [
5].
In the UK, those of lower SES are more likely to report heavy episodic drinking in the past week than those of higher SES [
6]. Yet, those of lower SES also report drinking less frequently than those of higher SES; 47% of lower SES individuals in England reported drinking alcohol in the past week compared to 79% of higher SES individuals [
7]. This alcohol harm paradox may have a greater impact in men than women. A systematic review examining gender differences in the distribution of alcohol-attributable mortality by SES indicates a three- to 10-fold higher risk of alcohol-attributable mortality in men of lower SES than women of lower SES [
8]. Supporting evidence indicates that younger males who live in deprived neighbourhoods are most likely to engage in heavy episodic drinking, however males aged 35 to 64 years old showed the greatest increase in heavy episodic drinking in deprived neighbourhoods [
6].
The disproportionate alcohol-related harm experienced by lower SES men requires investigation of this paradoxical relationship. However, lower SES individuals are under-represented in research investigating alcohol use, and in health research more generally, compared to higher SES individuals [
9]. Consequently, they are often considered to be “hard-to-reach” in terms of recruitment to and participation in research studies [
10]. Alternatively, this lack of representation may reflect that the researchers themselves are “hard-to-reach” due to their use of ineffective recruitment methods for this population, such as recruitment in university settings. Currently, researchers are not actively investigating how to recruit these participants through effective methods. Moreover, research may be limited by a lack of culturally appropriate information that is not relevant or tailored to the target population about participating in research [
11], and the time commitment required to participate [
12]. Developing research participation materials that are tailored to target populations may increase participation of under-represented populations in research studies [
11]. At present, there is limited evidence on effective strategies to recruit lower SES groups, which highlights the need to explore the use of different recruitment methods for different populations. Therefore, researchers may need to explore the feasibility of strategies to recruit under-represented groups to enable their inclusion in research. By failing to include such groups in research, there is limited evidence on the motivations and experiences which may underlie the alcohol harm paradox, and on the effectiveness of alcohol-related public health messages in male drinkers from lower SES backgrounds. Understanding the experiences and motivations for drinking in this group may enable the development of targeted health messages for individuals with the highest burden of illness. Some evidence suggests that community-based recruitment strategies may be more effective at recruiting male drinkers from lower SES backgrounds than recruitment through identifying participants from primary care registers [
13], which often fails to establish contact with these individuals to enable them to be invited to participate [
14]. Workplaces may represent an accessible and feasible community-based location for recruitment [
15]. Whilst recruitment through workplaces targets a working lower SES population, it may not capture all lower SES drinking males, such as those unemployed. However, those in routine and manual occupations experience high levels of alcohol-related harm [
16‐
18]. Moreover, hazardous physical working conditions and low job control may contribute to health inequalities among working populations [
19].
As male drinkers from lower SES backgrounds are at disproportionately high risk of alcohol-related harm [
20], and are under-represented in research investigating alcohol use, this study had the following aims.
The feasibility aim was:
1.
To explore whether workplace-based recruitment strategies are feasible for recruiting male drinkers from lower SES backgrounds
The pilot aims were:
2.
To examine the experiences and motivations for alcohol use in males from lower SES backgrounds
3.
To explore whether current alcohol-related public health messages are acceptable to male drinkers from lower SES backgrounds
Discussion
This study explored the feasibility of workplace-based strategies to recruit male drinkers from lower SES backgrounds to a survey. The recruitment of 84 male drinkers from predominantly lower SES backgrounds indicates that workplace-based recruitment strategies, including in person recruitment and a financial incentive, are feasible to recruit drinkers from this population. This study also investigated the experiences and motivations for alcohol use, and acceptance of health messages, in this population. More than half of the participants were at increasing risk of alcohol-related harm and approximately one fifth engaged in weekly heavy episodic drinking, which is higher than rates reported in the UK general population [
7,
28,
29]. Participation in campaigns aimed at reducing alcohol use, and accurate awareness of government alcohol consumption guidelines, were low. Health effects were the most common negative belief about drinking, and relaxation was the most common positive belief.
The feasibility findings extend the evidence base on the feasibility of methods to recruit male drinkers from lower SES backgrounds, and under-represented groups more generally. Previous evidence suggests that community-based recruitment strategies may be more effective at recruiting males from lower SES backgrounds than strategies such as primary care register recruitment [
13]. Our findings suggest that recruitment through workplaces may be an effective community-based recruitment strategy for this group. The feasibility of this method may be related to information provided to participants that the research was not driven by the employer, which may have increased participants willingness to participate in the research. The effectiveness may also have been influenced by the presence of the researchers, as participants could ask questions about the research [
30]. This recruitment strategy may be a useful alternative to those which have typically been less effective in recruiting perceived “hard-to-reach” populations, such as through primary care registers which often fails to make contact with such populations [
13]. It may be more likely to reach lower SES males who would not respond to recruitment through medical services. Recruitment was also facilitated by accessing workplaces through working with a local authority public health department. As the employers were the gatekeepers to the target population, it was important to explore how to best recruit the employers to take part.
In addition, higher recruitment to a survey with an incentive is consistent with previous findings [
31,
32]. This may be explained through social action theory [
33], in which survey completion depends on the associated rewards, costs, and trust. The recruitment rate may be due to perceived high rewards of a prize draw. Yet, the low uptake of the incentive may be due to reduced trust associated with entering an email address after answering sensitive questions. The delayed nature of the prize draw may also play a role. Surveys with delayed notification of prize draw results have lower response rates than those with immediate notice due to the immediacy effect, in which individuals tend to choose immediate rewards over delayed ones [
34]. Alternatively, the incentive used (a gift card) may have low acceptability in this population.
The pilot findings extend the evidence on alcohol use in lower socioeconomic status populations, which may improve understanding of the alcohol harm paradox. The rates of lower SES men at increasing risk (53.57%) and high risk (7.14%) of alcohol-related harm is higher than that reported for all men in the UK (24% and 3% respectively). Similarly, the rates of lower SES men at lower risk (39.29%) is lower than that reported for all men in the UK (72%) [
28]. However, these comparisons are limited by methodological differences between the surveys and the sample sizes. Yet, they provide preliminary indications of alcohol use in lower SES men in relation to the general population. In our sample, greater numbers of lower SES participants reported drinking alcohol at each measure of frequency compared to higher SES participants. This contrasts with findings that 10% of higher SES individuals in Great Britain drank on at least five days in the week prior to interview in 2017, compared to 7% of lower SES individuals [
29]. However, comparison across SES groups within our sample is limited by the greater number of lower SES participants recruited.
The higher rate of alcohol-related harm may be explained through differences in drinking patterns, as a higher rate of lower SES men had engaged in heavy episodic drinking on a weekly basis (20.24%) than individuals in the UK (16%) [
29]; (15%) [
7]. This is consistent with evidence highlighting an increased risk of heavy episodic drinking in men from lower SES backgrounds [
6,
18,
35]. Moreover, it is also consistent with a systematic review examining the role of alcohol use and drinking patterns in socioeconomic inequalities in mortality that identified 15 to 30% of socioeconomic inequalities in mortality as being explained by heavy episodic drinking, compared to 5 to 15% of mortality as explained by total alcohol consumption [
5]. Therefore, a pattern of heavy episodic drinking may explain the increased risk of alcohol-related harm in this population [
3]. Heavy episodic drinking may increase the risk of alcohol-related harm due to removing the protective effects of moderate consumption for ischemic heart disease [
36], stroke [
37] and unintentional and intentional injuries [
38]. Alternatively, the location in which participants were recruited may be a confounding factor, due to the relatively high rate of heavy episodic drinking in the south west of England (15%) compared to other areas in England, such as the south east (11%) [
29]. The south west is also the region of England with the lowest proportion of non-drinkers (12%) compared to the highest proportion in the west Midlands (25%) [
39]. Occupation may be another confounding factor, as workplace social norms about drinking predicts alcohol use amongst employees [
40].
The beliefs about alcohol use identified may motivate alcohol use, as attitudes towards alcohol correlate with alcohol use intentions [
41]. In terms of beliefs about relaxation, heavy episodic drinking occurs more frequently when experiencing stress, with a stronger effect of stress on heavy episodic drinking among lower SES individuals [
42]. Men from lower SES backgrounds may be more likely to engage in heavy episodic drinking in order reduce stress and increase relaxation. This supports a model in which stress motivates alcohol use in those who use avoidant coping [
43]. Therefore, relaxation may motivate alcohol use through a mechanism of coping. High rates of negative beliefs about health effects were reported despite comparatively higher rates of alcohol use than the general population. This supports evidence that knowledge of health effects alone does not affect behaviour change [
44]. Findings of low participation in campaigns aimed at reducing alcohol consumption and awareness of alcohol consumption guidelines is similarly low in the general population [
28,
45,
46]. However, it is not clear whether the underlying reasons are similar across SES groups, or whether this population may face unique barriers to increased participation in and awareness of public health messages.
Efforts to tackle alcohol-related inequalities could be informed by the development of targeted alcohol-related public health messages addressing the experiences and motivations identified in this high-risk population, such as a behavioural pattern of heavy episodic drinking. This supports recommendations for the development of public health strategies addressing alcohol-related inequalities by targeting patterns of heavy episodic drinking, rather than general alcohol consumption, among individuals from lower SES backgrounds [
5]. However, substantial variation in drinking patterns within this population exists, with just under half of participants engaging in heavy episodic drinking on a less than monthly basis. Therefore, public health messages aiming to tackle health inequalities may also need to account for variation in drinking behaviours within this population. This suggests that whilst a pattern of heavy episodic drinking may partly explain the alcohol harm paradox, other factors may also have a role in explaining the increased alcohol-related harm observed in this population.
To our knowledge, this is the first exploration of the feasibility of workplace-based recruitment strategies to recruit male drinkers from lower SES backgrounds to a survey. This study identifies feasible methods to improve the inclusion of lower SES males in research aiming to tackle inequalities in alcohol use. Yet, limitations exist. The method used to recruit workplaces resulted in a relatively low uptake of organisations to the study; only 26 workplaces agreed to take part out of 89 workplaces who were invited to participate. Whilst response rate is an important initial outcome indicating which recruitment strategies are effective, it is unclear why some workplaces and participants chose not to participate. This may be improved by recruiting workplaces face-to-face and working with organisations overtime to build relationships. Maintaining flexibility and persistence in recruitment may be key to identifying the most effective methods of recruiting the gatekeepers to enable access to the target population [
15]. The present sample may not be a representative sample of the target population given that some medium and higher SES participants were recruited in addition to lower SES participants. However, targeted sampling may not currently be feasible for recruiting lower SES working males due to relatively limited access to the population. Therefore, the recruitment strategy may require adaptation to improve the representativeness of the sample, eg: by recruiting across a larger geographical area and sample, or using methods such as snowball sampling and derived rapport. It is likely that a combination of strategies may be the best approach [
47]. Additionally, the selection of participant recruitment strategies by workplaces may have introduced selection bias. However, given the lack of evidence available on effective strategies to recruit this population, it was necessary to explore which strategies were feasible to recruit participants from this population. It is possible that fewer workplaces would have participated if we had randomly assigned them to a recruitment strategy. Similarly, whilst the recruitment strategy was costly in terms of time spent recruiting workplaces and participants, it is possible that fewer participants from the target population would have been recruited without exploring how best to recruit this population.
The use of workplace-based recruitment prevented recruitment of male drinkers from lower SES backgrounds who are not working due to alcohol use disorder, unemployment, or not being of working age. This potentially limits the generalisability of our findings to all groups of lower SES males as occupation is only one of many indicators of SES [
48]. Although, this study aimed to target those in routine and manual occupations due to their high levels of alcohol-related harm [
16‐
18]. Additionally, the impact of the location that the survey was completed in on the accuracy of responses is unclear. Whilst participant confidentiality was ensured, and accuracy of responses was aimed to be improved through the use of an image displaying the units of alcohol in different drinks; it is possible that participants recruited in person may be less likely to accurately report their alcohol use at their workplace, such as those with safety critical roles in which high alcohol use could have serious consequences.
Future research may wish to explore the feasibility of workplace-based recruitment for this population across different geographical locations. Future studies could also explore the acceptability of different types of incentives, including immediate versus delayed; and the feasibility of workplace-based recruitment for other under-represented groups in research. The pilot findings also provide directions for future research aiming to better understand alcohol use in this population and to inform the development of targeted health messages about alcohol. Future studies could conduct larger scale research to build upon these initial pilot results. The prevalence of heavy episodic drinking as a possible mechanism involved in the high alcohol-related harm seen in male drinkers from lower SES backgrounds would be useful to explore across a larger sample. It may be useful to explore what barriers and facilitators affect participation in campaigns aiming to reduce alcohol consumption in this population. This may be best explored through qualitative methods.
Workplace-based recruitment, with the use of an in-person recruitment strategy and a financial incentive, may be a promising strategy for future research aiming to tackle inequalities in participation in alcohol research. Improved understanding of the experiences and motivations of alcohol use in male drinkers from lower SES backgrounds may advance knowledge of the mechanisms of the alcohol harm paradox and inform the development of targeted alcohol-related public health messages.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.