Engaging workplace representatives in research: what recruitment strategies work best?

Abstract

Background

Workplaces are key stakeholders in work and health but little is known about the methods used to recruit workplace representatives (WRs), including managers, occupational health advisers and colleagues, to externally funded healthcare research studies.

Aims

To detail the strategies used in recruiting WRs from three areas of the UK to a qualitative study concerning their experience of employees undergoing hip or knee replacement, to compare the strategies and inform recruitment methods for future studies.

Methods

Six strategies were used to recruit WRs from organizations of different sizes and sectors. Data on numbers approached and responses received were analysed descriptively.

Results

Twenty-five WRs were recruited. Recruitment had to be extended outside the main three study areas, and took several months. It proved more difficult to recruit from non-service sectors and small- and medium-sized enterprises. The most successful strategies were approaching organizations that had participated in previous research studies, or known professionally or personally to team members.

Conclusions

Recruiting a diverse sample of WRs to healthcare research requires considerable resources and persistence, and a range of strategies. Recruitment is easier where local relationships already exist; the importance of building and maintaining these relationships cannot be underestimated. However, the potential risks of bias and participant fatigue need to be acknowledged and managed. Further studies are needed to explore how WRs can be recruited to health research, and to identify the researcher effort and costs involved in achieving unbiased and representative samples.

Introduction

Workplaces are key stakeholders in workers’ health. Research studies involving employers are used to inform practice guidelines [1] and government policy [2]. Although a number of externally funded qualitative research studies have recruited workplace representatives (WRs) as participants (e.g. managers, supervisers, occupational health advisers, human resources personnel, colleagues, trades union representatives), many fail to describe their strategies. Researchers are urged to publish their recruitment methods and lessons learned [3,4].

This paper reports on the recruitment of WRs to a study funded by the National Institute for Health Research (NIHR) which is investigating the feasibility of developing a hospital-based occupational advice intervention for patients undergoing total hip replacement (THR) or total knee replacement (TKR) in order to facilitate a time-appropriate and successful return to work. To inform the development of the intervention, researchers conducted interviews with WRs from organ izations based in the geographical areas surrounding the three UK main study sites of Nottingham, Norwich and Middlesbrough. No previous studies have explored issues around low limb arthroplasty from the workplace perspective. To our knowledge, only one other study focusing on a specific health condition has detailed their employer recruitment strategies [3]; however, study data were collected from workers compensation claims rather than interviewing WRs.

Methods

Ethical approval was obtained from the University of Nottingham Faculty of Medicine and Health Sciences Ethics Committee.

The sampling frame included organizations of different sizes and sectors. Participants were eligible if they had direct experience of supporting one or more employees undergoing THR/TKR in the previous 12 months. We anticipated that a purposive sample of eight WRs from each study site would provide sufficient diversity.

Six strategies were used as detailed in Table 1. Data on the number of approaches made and the responses received were analysed descriptively.

Results

Twenty-five interviews were conducted. Data on the number of approaches made and number recruited are shown in Table 1.

Characteristics of the participants, their workplaces and recruitment strategy used are shown in Table 2.

Ten WRs were recruited from workplaces around Nottingham, five from Middlesbrough and three from Norwich. Seven interviews were conducted outside the study sites.

Eight participants, representing six workplaces, were recruited via organizations from previous research studies (strategy A). Fourteen organizations were approached; seven did not respond. Of the seven responders, one was ineligible; five responded positively and one declined but identified an alternative. In one case, three interviews were conducted within the same organization.

Seven organizations were recruited via contacts known to the researchers (strategy B). Sixteen individuals were approached. Of the remaining nine potential participant workplaces, two were ineligible, one declined, five did not respond and one circulated information which yielded no responses.

Three participant workplaces were recruited via e-mailing organizations listed on the open-access Chambers of Commerce databases covering the study sites (strategy C). From 2266 e-mails sent, 28 organ izations responded, six met the eligibility criteria, but of these, three did not respond further.

Five participant workplaces were recruited via 14 professional/business networks (strategy D). From the eight networks distributing information, around 7000 organ izations were reached. Three participant workplaces were from NHS trusts recruited through a national public health network.

Two participant workplaces were recruited through ‘cold calling’ letters, telephone or e-mails (strategy E). These were either large organizations known to researchers or identified through internet searching. Thirty-nine were contacted, three responded, one being ineligible.

A study Twitter account was set up (strategy F). Fifty-eight tweets were sent, 17 receiving Twitter-based responses. Responses did not appear to reach potential participant workplaces and instead engaged people who were interested in the topic.

Discussion

This study found that recruitment of WRs was difficult and it proved necessary to extend recruitment outside the study sites, and recruitment took several months, with non-service sectors and small- and medium-sized enterprises (SMEs) more difficult to recruit from. The most successful strategies were approaching organizations that had participated in previous research studies, or known professionally or personally to team members. Although not directly comparable to our study, Johnson et al. [3] also reported a wide variety of recruitment methods over a 2-year period in order to recruit 15 organizations.

The study criteria were a limiting factor. Using data from the Office for National Statistics [5] and National Joint Registry data for the same year for patients under 60 years [6] approximately 1.4% of registered VAT- and/or PAYE-based enterprises in England, Wales and Northern Ireland were likely to have been eligible for the study. Many organizations, particularly SMEs, would thus be unlikely to have experienced employees undergoing THR or TKR. In the UK, Clinical Research Networks (CRNs) facilitate the participation of patients/staff in NHS studies [7]. However, no such networks cover patients’ workplaces. This is a matter for urgent consideration, for example by the government’s Work and Health Unit, and the NIHR.

We used the online Chambers of Commerce databases, but were not members. To have joined the relevant chambers would have been costly but may have facilitated workplace contacts. Wainwright et al. [8] reported using local chambers for recruitment, but not their success rate. Databases make it difficult to know whether the appropriate person has been reached, and not all organ izations provide relevant contact details, requiring additional searches. However, this strategy did result in the recruitment of difficult-to-reach smaller organizations.

The most successful recruitment method was via organizations or individuals known to researchers, underlining the importance of personal contacts and building local research networks. In the study by Johnson et al. [3], more than half of the participating employers had prior working relationships or personal contact with the team. However, this approach limits study reach, and existing relationships may cease when contacts leave the organization. Participants might also develop ‘research fatigue’. Lysaght et al. [4] highlight the importance of identifying the best contact point within organizations and recommend the use and development of established connections.

A strategy of recruiting patients and then contacting their employer was included in the main study protocol, but rejected by the Health Research Authority. Although this method may have aided recruitment, it could have been influenced by the relationship between the patient and their employer and so increased the risk of bias in the sample.

No participants were recruited via Twitter. It was difficult to clearly describe the target group within the number of characters required. Prospective participants needed to have supported employees who had undergone surgery; however, it was common to attract interest from employees who had undergone surgery, rather than any individual who had supported them at work.

A study limitation is that data were not collected on resources given to the recruitment process, which was time and labour intensive. In a Canadian study of workers with back injuries [9], the authors estimated average costs of CDN$240 per participating worksite. It is suggested that future UK studies consider collecting similar data.

Recruiting WRs requires considerable resources, persistence and a range of strategies. Recruitment is easier where local relationships already exist; however, the risks of bias and participant fatigue need to be acknowledged and managed. Further studies are needed to identify the researcher effort and costs required to achieve unbiased and representative samples.

Funding

This project was funded by the National Institute for Health Research Health Technology

Assessment (HTA) programme (project number 15/28/02). The views expressed in this paper are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. Further information available at: https://www.journalslibrary.nihr.ac.uk/programmes/hta/152802/#/.

Competing interests

None declared.

Acknowledgement

The authors would like to thank the study participants.

References