Overcoming Recruitment Challenges in Nursing Home Research with Nurses and Health Care Aides

Background Recruiting busy health care providers into research can be challenging. Yet, the success of a research project can hinge on recruitment response rates. This article uses a case study to demonstrate how qualitative researchers creatively readjusted their methods when standard methods were not yielding enough recruitment response with the aim of supporting other researchers with their recruitment. Methods Case Example – Interest was expressed but response rates were low among nurses and health care aides in a research project on person-centred health care in a personal care home research site. The research team reconceptualized the participation design, creating a research ‘event’, which accommodated the time constraints and work culture of the respondents. The research event was much better attended than standard interview recruitment. The research event approach overcame barriers to participation. An 80% response rate resulted. Standard response rates for research interviews tend to be well under 20%. Successful recruitment hinged the researcher’s willingness the recruitment approach through, when recruitment difficulties were encountered. The high response could be attributed to the methods’ alignment to the available and work culture ways. The results suggest that attending to aspects of the work culture can increase recruitment, improve the chances of successful data collection, and reduce the likelihood of research ‘stall’.


Introduction
Health care professionals are busy and can be difficult to recruit as respondents in research studies.
Discussions of these difficulties has been written in large part on either the recruitment of nursing home facilities into studies 1,2,3 or specifically, recruiting physicians into research 4,5,6,7 . Research on recruitment barriers and facilitators of nurses is scant in the literature 8,5 and rare on health care aide recruitment. Discussion about recruitment in nursing and other health fields focuses mainly on patient populations and clinical studies 9, 10,5,8 . Much less is written on the strategies that may be most successful in recruiting busy health care providers into research studies using qualitative methods 5 , which may be perceived as demanding more of participants' time.
It is interesting to note that while the success of a research project may hinge on recruitment response rates, the specifics of recruitment are not often given much space for discussion in articles and research textbooks 5,11,12,13,14,15,8 Recruitment is often seen as the technical background of research studies and the motivation of respondents to participate is not examined a great deal in the literature 16. The decisions that researchers make and the basis for those decisions, particularly midproject when barriers to recruitment are met, can sometimes be obscured in the ways research methods need to be condensed into publications. The messiness of the research process is made neat and linear for the sake of readability, word count, and to conform to ideas of how the research process is supposed to look 17 . Therefore, it becomes unknown by reading research articles and reports how often researchers encounter barriers and may need to shift recruitment methods.
Research recruitment does not always proceed according to design, and in fact may often not proceed at all according to the original design 16 . Snowball sampling may stop rolling, mediators can make or break access, potential respondents may not see the posters, respond to the email, return the phone call, or pay attention to the memo 18 . Researchers may not always be able to anticipate the conditions in the field and although we can try to make the best plan for the participant population in advance of fieldwork, it is not possible to plan for every possible scenario that may influence the recruitment and participation. A researcher will likely have a stronger knowledge of what hinders or improves participation after starting the research with the exact study participants in the time period in which the study takes place (those nurses, for example, that are being interviewed/surveyed/etc. and the conditions at the time). As a result, there is often a need to reconsider recruitment methods after the research has already commenced. Otherwise there is a risk of the research stalling or ultimately, failing to collect any data.
Reconsidering and innovating on recruitment methods, when the plan is not yielding the necessary saturation, should be thought of as being responsive to the research environment, rather than a failure of research design 19 . Sampling and recruitment approaches are meant to serve the research 19 , and so it is not unusual for qualitative sampling and recruitment approaches to be somewhat emergent 20,21 . Modifications most often require approval from ethics bodies, so the ability to move a modification quickly through ethics approval facilitates the researcher's ability to innovate on recruitment.
In the case of our research study, barriers to recruitment necessitated creative thinking and innovating on recruitment methods after the research commenced. We had to rethink our recruitment to try to make it more inviting for busy health care providers working in a nursing home facility, to attend and share their knowledge. Upon reflection after the research event, we were able to draw out some factors that other researchers may consider when designing or redesigning their recruitment.

Background
The project had just moved from the observation phase into the interview recruitment phase at the first site of a research project on excellence in person-centred care at a long-term care home in an urban centre in Canada. It was puzzling why staff were not volunteering for interviews in larger numbers, even though administration had allowed staff to conduct the interview on work time.
Several recruitment approaches had been invoked: presentations at nursing rounds and staff meetings, posters in staff areas and public areas such as hallway bulletin boards, encouragement to participate from administration and senior staff, email calls to participate, and in-person on-site requests from researchers. When asked about their non-participation, they cited lack of time as the main reason; too busy while on shift and too busy with home life to have the researcher come to their home during the off days. To accommodate respondents' busy schedules, an interviewer was made available during all eight-hour work shifts, during breaks, and after-shift, in the day, evening, or night, mindful to provide options, given that shifts are very busy and to accommodate the different hours of shift work. The length of interview was made flexible, fifteen minutes to an hour and a half, whatever time respondents had to share. Accommodations were made for interviewing off-site, even to go to the respondent's home. Still very few staff volunteered for the formal interviews; they responded that they were still 'too busy'.
Given that it is not unusual for people to say, 'I don't have time' rather than give more complicated reasons for not participating in something, it is not always clear if the 'lack of time' response from a potential participant means just that or something else. However, we could not pinpoint any hesitations other than time constraints. We had ensured that comfortable, trusting relationships had been established. Several months had been spent building trust and familiarity with staff, residents and families and gaining an in-depth understanding of the routines and institutional cultures. The project had a regular presence at the nursing home. There were posters on the walls and a research assistant chatted with residents and staff for a few hours daily over a period of several months in common areas and hallways, even helping pass snacks and attending recreational activities. The project had just completed an observation phase where the research assistant recruited sets of staff (e.g., health care aides, registered nurses) and residents, with appropriate permissions, to observe and asked questions of staff and residents as they provided intimate personal care to residents, such as toileting and other activities of daily living. Residents, staff and families were relaxed and forthcoming during the short, informal interactions. When we checked confidentially with staff, they did not indicate that fear of speaking out was an issue, nor had there been any previous bad experiences with researchers. No other barriers, such as cultural differences or communication difficulties seemed to be present. Staff seemed to find the project interesting and were not hesitant to participant in the observational portion of the study. After checking other potential reasons for not participating we concluded that time was indeed the factor preventing staff participation.
The research team had to get creative, given that the standard interview recruitment was not yielding the response rates required. The study had gathered only seven interviews with health care aides (HCAs) and two with nurses. The aim was approximately 25 interviews. Recruitment required a new approach. The team had to rethink and anticipate what might encourage the busy staff to participate meaningfully in the study was undertaken. It was imperative to keep participation time short and during staff break times so that responding to the study would not impinge on resident care or personal time outside of work. The team considered the needs of the staff and conditions of the field environment and then sent a redesigned recruitment event through ethics approval.

Methods -Overview of the Research Event
The redesign involved drop-in events where staff could take a bit of time on their breaks to participate in the research study. The research team organized three drop-in sessions in the nursing home for HCAs and nursing staff to respond to seven research questions; questions that had originally formed the interview guide to be used in the face-to-face interviews. The drop-in sessions were organized to correspond with break periods on all three shifts: day (1-3 pm), evening (8-11 pm), and night (1-3 am). Potential participants were made aware of the research events for a month prior to the events via event posters throughout the care home in areas accessible to the staff but not residents or visitors (i.e. behind the nursing desks, in staff rooms, stairwells, etc.). One of the research assistants and a number of enthusiastic staff members announced verbal reminders leading up to the event and during the first hour of each event when participation was slower. To provide extra incentives to participate, to make the event festive, and to thank staff for participating, the project provided pizza, snacks, refreshments, small promotional items (pens, hand sanitizer, etc.) and gift cards.
Coincidentally, the timing of the event was in December and the recreation room on the main floor of the care home where the event was held had Christmas decorations up, which helped to add a festive atmosphere.
Staff on each shift came to the event during their breaks. They sometimes came with colleagues, in twos or threes. It had been conveyed in advance from the researchers through the senior staff who promoted the event, that this was an optional event. When potential participants arrived, the research team emphasized the voluntary nature of the study along with the consent form, that if they 7 wished at any point in the event, participants could simply take some food and choose not to participate, to end their participation in the middle, or to withdraw the answers they had already provided.
Staff who agreed to participate completed the consent and demographic forms, received seven 4"x6" Two research assistants were available to provide direction and support to staff as needed throughout the event. The research questions are provided in Table 1. 2) What makes a good nurse? 3) What kinds of things do you do that makes a resident feel cared for? 4) How do you manage your feelings when you see or smell things in the course of providing care that some people might consider smelly or disgusting? 5) What frustrates you about your job? 6) Often when providing care, there will be several residents who require your attention all at the same time. How do you decide whom to respond to first? 7) What would be on your "wish list" if you could change anything to enhance the care of residents at (this PCH) or to enhance the workplace?
Most participants took 2-3 minutes per question to read and prepare a response. At times, staff were

Results
There was a very high level of participation; out of a possible 88 staff there were a total of 71 8 participants (43 on day/evening shifts and 28 on evening/night shifts). That is an 80% response rate, unusually high for most field research. Staff even reported they were experiencing unusually busy shifts over the two-day period of the study, however they voluntarily participated in the study on their breaks (completing resident care prior to attending). Staff expressed that it was important to attend the research event and encouraged their coworkers to come with them to the study. Some staff said they covered the co-worker's assigned residents' so that the co-worker could go on a break and participate in the study. Participants did not seem to rush when composing their answers. Responses were more in depth than they would have been in surveys, although less so compared to the interviews conducted. Results of this research event allowed researchers to complement the rich detail of the interviews with a greater number of short responses. This allowed the researchers breadth in the data, and to check how common some of the issues raised in the interviews were.
Unique responses provided the basis for follow-up questions, asked either informally to staff and management, in the few follow-up interviews at that facility, or at the next research site.

Confluence between data collection and work culture
Researchers learned that one of the most significant reasons that staff participated in the research event was that they were already used to filling out surveys for care home and the parent company of the facility as part of their jobs. It was already part of their 'work culture' and endorsed by administration, a factor other's have noted as increasing successful recruitment 2 . Staff were used to the idea of taking a break from the unit to fill in surveys, in fact they were previously expected to do this as a part of their work. They were accustomed to covering each other's work to facilitate their coworkers to leave the unit to complete surveys and checking to ensure other staff had a chance to do so.
We learned only after the recruitment redesign the ways that making data collection parallel their usual work culture improved recruitment. However, this points to questions that researchers can ask beforehand about the participants and the research field when setting up the data collection, perhaps even before writing the grant when possible.
What experiences do the participants in this population already have with research?
With responding to questions as part of their work?
With having discussions with others?
Or any other engagement that is similar to the data collection methods used by the research that is planned?
How does this engagement usually work? What are the general conditions and factors for this engagement? How can we make the research like what they are already used to?
When viewed as simply a part of the regular work that each staff member had to do in a day, they encouraged other staff members to participate. They asked each other if they had taken time out to participate, if not, when would they do that, and that they would 'cover' each other's work while they were at the research event. This was a very quick type of chain-referral recruitment, akin to multiple snowball sampling, that proved to be a boon to our recruitment numbers, resulting in the 80% response rate from the population. Matukaitis Broyles et al 5 discuss how some nurses act as 'anchors', identifying and recruiting other colleagues. Of course, if other studies required a greater degree of anonymity of participation, factors of anonymity in chain-referral recruitment would have to be considered 5,21 . In this case, a higher response rate actually strengthened anonymity of responses, given that researchers paid careful attention to the factors needed to anonymize the data.
Are there ways to get staff to encourage each other to take a break to participate, as long as anonymity of participation is not integral to the study?

Part of participants' workday
Researchers need to consider whether the research will occur on work time or not 5 . In our study, when the research was part of their workday and staff did not have to use their own time, it made them feel strongly that they could make time to participate. This was true for the research event as well as individual interviews. Although there was some initial concern from staff that they would be too busy while on shift to participate, once they knew a co-worker would be able to 'cover' them for them for a short period of time, they were eager to participate. Framing the research as 'getting paid to talk to the researcher instead of the usual care work' was also found to pique the interest of 10 potential respondents.

Questions a researcher might ask beforehand include:
What are the routine activities and events they participate in? Special events?
Can data collection be planned to parallel or coincide with these events?
What goes on at these events that the participants are used to?
Can researchers arrange for the facility to allow staff to respond to research on work time? Or to compensate the facility for the respondents' time during a shift?
Can it be arranged so that missed work can be performed by someone else if the time out for the research will impact the work site?

Planning flexible timing into data collection
In our experience as nurses and researchers we have observed that workers on evening and night shifts in health facilities often miss out on events and professional development, which usually take place during the day shift. Staff expressed appreciation that there was a research session for each shift, that the study recognized that perspectives from evening and night shifts as well as day shift are important and experiences on these different shifts might have unique characteristics. Offering schedule flexibility is an important recruitment strategy that has been previously identified 5 . We were pleasantly surprised by the enthusiasm of staff on evening and night shifts to participate. Five nurses and twenty HCAs on evening and night shifts participated.

Framing perception of research duration
Even though participants were offered interview times of equal duration, they perceived the drop-in event as a shorter duration. We hypothesize that one of the reasons was that they felt had control over the research duration because they were free to come and go without having to disengage from an interview situation. Drop-in events also had the benefit over standard interviewing for researchers of gathering a greater number of responses in a short period of time.

Making participation a treat and a joy -food and small gifts "swag"
Everyone enjoys getting free stuff. Food is also a draw at research events, as most seasoned researchers know. For staff half-way through a shift, often with limited or deferred break times, having a chance to grab some food was a joy for respondents. Along with standard fare (pizza and soft drinks), researchers provided food items that the nurses could grab and take with them to consume later in the shift, like granola bars and bottled water. Small gifts, such as gift cards for coffee, small Kleenex packs, sticky note pads, pens and hand sanitizer also were much appreciated. It is the experience of the researchers that gifts and food show appreciation for the time and expertise respondents offer the research.

Conclusion
Instead of following recruitment and data collection methods by rote, a little creative inquiry about the field and redesign of recruitment strategies can go a long way to improving recruitment in nursing research. In summary, we found five areas we found in our study for researchers to examine, either in the initial research design or in redesign when recruitment is encountering barriers: Confluence between data collection and work culture Making research participation part of participants' workday or not Planning flexible timing into data collection Framing the perception of participation duration

Making participation a treat and a joy
Attention to these details of the recruitment process greatly improved recruitment numbers in our study, and the breadth and of data collection. Researchers may want to consider these five areas when planning or redesigning their studies with busy health care workers. GT was the principle researcher; she conceived of and designed the study. SP and LS gathered the data. GT, SM, SP, and LS analysed the data. SP was a major contributor in writing the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate The research was approved by the nursing ethics committee of the University of Manitoba. Written informed consent was obtained from all participants. If residents were cognitively impaired, informed consent by proxy was obtained from the guardian of the resident.

Availability of Data and Materials
The qualitative data supporting this article is not deposited into a repository. However, requests to use this data may be considered by the lead researcher: Genevieve.thompson@umanitoba.ca

Consent for publication
Not applicable.