Introduction
When researchers use multiple methods in social science research, often their goal is to triangulate findings. Denzin (Reference Denzin1978) described triangulation as the “combination of methodologies in the study of the same phenomenon” (p. 291). While data convergence is frequently the aim of triangulation strategies, Jick (Reference Jick1979) has pointed out that multiple methods are especially powerful when used to “[elicit] data and [suggest] conclusions to which other methods would be blind” and also to “enrich our understanding by allowing for new or deeper dimensions to emerge” (p. 604).
Mixed-method studies, a distinct form of multiple-method studies, are classified as between (or across) method designs, that incorporate quantitative and qualitative data collection (Creswell & Plano Clark, Reference Creswell and Plano Clark2007; Creswell, Plano Clark, Gutmann, & Hanson, Reference Creswell, Plano Clark, Gutmann, Hanson, Tashakkori and Teddlie2003). Mixed-method designs are gaining popularity in health and social care research (Stewart, Makwarimba, Barnfather, Letourneau and Neufeld, Reference Stewart, Makwarimba, Barnfather, Letourneau and Neufeld2008). For instance, researchers have increasingly used mixed methods within single studies as a pragmatic strategy to help address the complexity of the health care environment, resulting in what O’Cathain (Reference O’Cathain2009) has referred to as “a quiet revolution” (pp. 3–4). Part of this revolution certainly emanates from what Johnson and Turner (Reference Johnson, Turner, Tashakkori and Teddlie2003) referred to as the mixed-methods fundamental principle: the understanding that there are inherent limitations to all methods. Thus, researchers have used mixed methods not only to achieve data convergence or corroboration of findings but also to eliminate alternate explanations about the data, and to “elucidate the divergent aspects of a phenomenon” (p. 299).
In studies of long-term care (LTC) facilities, mixed methods are less commonly used, dominated by “small n” ethnographic studies (see, for example, Diamond, Reference Diamond1995; Lopez, Reference Lopez2006), or single country/jurisdiction statistical studies (e.g., Berta, Laporte, Zarnett, Valdmanis, & Anderson, Reference Berta, Laporte, Zarnett, Valdmanis and Anderson2006; Harrington, Woolhandler, & Mullan, Reference Harrington, Woolhandler and Mullan2001). Comparative and mixed-method designs are required to expand our cross-national understandings of this women-dominated care work sector.
There is consensus within the literature about the general lack of societal value ascribed to care work, tied in part to its highly gendered context because it is women dominated as a workspace and as a place of care (Armstrong, Armstrong, & Scott-Dixon, Reference Armstrong, Armstrong and Scott-Dixon2008; Kittay, Reference Kittay1999). A few studies have focused on work organization issues, such as work overload, which result from a work structure that compartmentalizes caring into individual tasks and ignores caring’s relational aspects (James, Reference James1992; Lopez, Reference Lopez2006). Other studies have addressed conditions of work including staffing intensity (McGregor et al., Reference McGregor, Cohen, McGrail, Broemeling, Adler and Schulzer2005), public, non-profit or for-profit ownership patterns (Harrington et al., Reference Harrington, Woolhandler and Mullan2001), stress (Morgan, Stewart, D’Arcy, Forbes, & Lawson, Reference Morgan, Stewart, D’Arcy, Forbes and Lawson2005), and job satisfaction (Castle, Degenholtz, & Rosen, Reference Castle, Degenholtz and Rosen2006). There is a noticeable lack of international comparative work focused on the everyday realities of care workers’ experiences (Daly & Szebehely, Reference Daly and Szebehely2011).
Other work organization issues such as health and safety are regulated by governments to protect workers and residents, but rules are frequently circumvented within facilities for reasons of profit making, cost reduction, or efficiency (Harrington & Pollock, Reference Harrington and Pollock1998; McGregor et al., Reference McGregor, Tate, McGrail, Ronald, Broemeling and Cohen2006). Lopez (Reference Lopez2007), while noting that managers, care workers, and residents in facilities engage in a “mock routinization” of the work, has documented how workplace rules that were intended to safeguard workers’ and residents’ health and safety are routinely ignored. A handful of studies have demonstrated how care workers across settings have been subjected to high levels of injury, illness, and violence (Armstrong et al., Reference Armstrong, Banerjee, Szebehely, Armstrong, Daly and LaFrance2009; Baines, Reference Baines2006; Banerjee et al., Reference Banerjee, Daly, Armstrong, Armstrong, LaFrance and Szebehely2008; Boyd, Reference Boyd1995; Gates, Fitzwater, & Meyer, Reference Gates, Fitzwater and Meyer1999; Gates, Fitzwater, & Succop, Reference Gates, Fitzwater and Succop2005; Hellzen, Asplund, Sandman, & Norberg, Reference Hellzen, Asplund, Sandman and Norberg2004; Levin, Reference Levin, Hewitt, Misner and Reynolds2003; Shaw, Reference Shaw2004). Fewer studies have reported specifically on the ways in which violence is normalized by workers or how it is an “expected, tolerated, and accepted” part of care work (Gates et al., Reference Gates, Fitzwater and Meyer1999).
Violence is frequently ascribed to the behaviour of some residents with dementia and Alzheimer’s disease (Kolanowski & Whall, Reference Kolanowski and Whall2000). Better trained care workers and improved practice guidelines have been used as the key managerial containment strategies (Robinson & Tappen, Reference Robinson and Tappen2008). However, factors such as lower than optimal staffing levels and higher than optimal worker turnover levels have been identified as contributing to the creation of conditions for resident aggression towards staff (Robinson & Tappen, Reference Robinson and Tappen2008). Violence occurs frequently during basic body care activities: activities that take place when workers and residents interact, including when care workers turn, change, bathe, dress, or feed a resident (Miller, Reference Miller1997). Although much attention has been paid to the timing of activities and related behaviours that elicit violence, little attention has been paid to the structural organization of the work, including how workload, worker autonomy, managerial support, and work content affect levels of violence in care facilities.
Aim
In this article, we aim broadly to show how an iterative mixed-methods strategy can contribute immensely to a program of research in the area of aging, health, and social care. This article’s specific focus is on the iterative methodology we employed in studying working and living conditions in facilities that provide long-term residential care for older adults. To highlight the contributions of a mixed method approach, we discuss two of our findings: working conditions that prevented Canadian workers from having the time to properly care for residents; and the comparative frequency with which Canadian and Scandinavian workers were subjected to physical violence and to unwanted sexual attention by a resident or by the resident’s relative. These findings emerged from analysis of cross-jurisdictional open and closed survey data from surveys with Canadian and Scandinavian workers, and verbatim focus group interviews with Canadian workers.
In Canada, LTC facilities are dwellings for chronically ill older adults and for people with disabilities who require on-site medical care and social supports. As has traditionally been the case, far more women than men live in LTC facilities (see Table 1). Canadian research has documented that a person’s having Alzheimer’s disease or dementia is one of the key factors in nursing home institutionalization (Trottier, Martel, Houle, Berthelot, & Legare, Reference Trottier, Martel, Houle, Berthelot and Legare2000).
Table 1: Percentage of residents, by sex, living in homes for the aged in Canada and territories, and in select provinces (2004–2008)
Source: Statistics Canada. Table 107-5504 – Residents on books in residential care facilities, by age group, sex, principal characteristic of the predominant group of residents and size of facility, Canada, provinces and territories, annual (number) (table), CANSIM (database); http://estat.statcan.gc.ca/cgi-win/cnsmcgi.exe?Lang=E&EST-Fi=EStat/English/CII_1-eng.htm. (Accessed: July 23, 2010).
In 2007–2008, Canada had 2,182 facilities operating as homes for the aged and funded by government (Statistics Canada, 2010). They were primarily operated by regional or municipal governments, charitable or religious organizations, or commercially or privately run for-profit organizations, although a few homes (Statistics Canada, 2010) were operated by the provinces/territories and federal government. Canada had a higher proportion of commercially owned homes than did the Scandinavian countries (National Board of Health and Welfare, 2010). Long-term care was primarily a unionized work environment at all of the Canadian and Scandinavian jurisdictions we studied, with about 8 in 10 workers belonging to a union. Despite high levels of unionization, the work was at the bottom of the remuneration scale, involved little autonomy, and reflected what Karasek and Theorell (Reference Karasek and Theorell1990) referred to as high-demand, low-control, and low-support environments.
In Canada and in Scandinavia, women continue to be the majority of direct care workers. In Canada, most are unlicensed personal support workers (care aides), with a few dietary workers and housekeepers. Only a few licensed or registered nursing staff work in each facility, and many facilities are without therapists of all types. The workforce appears more regulated in the Canadian context. For instance, a study conducted by Gustafsson and Szebehely (Reference Gustafsson and Szebehely2005) found that there were fewer than 6% of care workers with an RN designation in Sweden, with most working as assistant nurses or as care aides.
Since 2004–2005,Footnote 1 Canadian reported full-time equivalent staffing levels have risen (see Table 2). Women are thus the majority of both care providers and care recipients, making these facilities highly gendered working and living environments.
Table 2: Full-time equivalent (FTE) Footnote (*) personnel in residential care facilities classified as “homes for the aged”, in Canada and the territories, and in select provinces (2004–2008)Footnote (**)
Source: Statistics Canada. Table 107-5505 – Full-time equivalent personnel in residential care facilities, by principal characteristic of the predominant group of residents and size of facility, Canada, provinces and territories, annual (number) (table), CANSIM (database), http://estat.statcan.gc.ca/cgi-win/cnsmcgi.exe?Lang=E&EST-Fi=EStat/English/CII_1-eng.htm. (Accessed: July 23, 2010)
(*) “Full-time equivalent” is the total number of paid hours divided by 1,950 hours, which corresponds to a 37.5-hour workweek.
(**) “Homes for the aged” refers to nursing homes, homes for the aged, and other facilities providing services and care for the aged. Not included are homes or lodges for senior citizens in which no care is provided.
In this article, we briefly discuss our theoretical approach and our methods, and present two examples of our key research findings illustrating how our iterative approach, in which we asked workers similar questions using a quantitative and then a qualitative method, yielded convergent and divergent results. The first example involves comparative data on work overload, showing how the methods resulted in data triangulation and convergence, and enabled analytical extension about issues related to (a) workers’ time to perform care activities; (b) working short-staffed; (c) important aspects of caring activities that are left undone; and (d) the norms for use of incontinence products. The second example demonstrates the divergent and contradictory responses obtained when we asked focus group respondents to comment on whether our survey results reflected their experiences of violence and of unwanted sexual attention.
Theoretical Approach and Methods
Our research has been grounded in feminist political economy (Andrew, Armstrong, Armstrong, Clement, & Vosko, Reference Andrew, Armstrong, Armstrong, Clement and Vosko2003; Armstrong, Armstrong, & Coburn, Reference Armstrong, Armstrong, Coburn, Armstrong, Armstrong and Coburn2001; Doyal & Pennell, Reference Doyal and Pennell1979), which focused our inquiry on care work in the highly gendered (Armstrong & Kits, Reference Armstrong and Kits2001; James, Reference James1992; Kittay Reference Kittay1999; Watson & Mears, Reference Watson and Mears1999) and racialized (Armstrong et al., Reference Armstrong, Armstrong and Scott-Dixon2008) working and living conditions in LTC facilities in Canada. Our research has also been grounded in transformative assumptions that good working conditions create good living conditions for residents, and in feminist assumptions that the whole range of workers in care facilities knows and best understands their working conditions.
Our research findings exemplify how the iteration of the research instruments enabled us to interpret our results, which were based on LTC facility workers’ responses to three studies – a study in Ontario, a Scandinavian study, and a Canadian study – we conducted. The findings from the Ontario study and the questions asked in the Scandinavian study informed the Canadian study, which included workplaces in Ontario, Nova Scotia, and Manitoba (see Figure 1). The resident populations were similar between the studied jurisdictions in that women were the vast majority of residents, they had multiple health and/or disabling conditions, and they had high rates of Alzheimer’s disease and other forms of dementia.
Figure 1: Research flow.
Table 3 reports each of the study’s characteristics. With approval from York University’s Research Ethics Review Committee, the first survey entitled the “Nursing Home Workplace Study” – herein referred to as the Ontario study – was distributed to 2,322 workers recruited from one union and 18 randomly selected workplaces in Ontario. The Institute for Social Research (ISR) at York University aided in the survey’s design, distributed the survey to workers, and entered data between March and May 2004. A total of 917 surveys were returned representing a response rate of 39.5 per cent. Three of the survey’s open-ended questions prompted workers to identify whether they faced particular problems with meeting their responsibilities as a result of their work schedules; to identify what other influences their job placed on their personal lives; and to provide other comments.
Table 3: Studies and study characteristics
CAW = Canadian Auto Workers (Canada); CFNU = Canadian Federation of Nurses Unions (Canada); CUPE = Canadian Union of Public Employees (Canada); NUPGE = National Union of Public and General Employees (Canada); SEIU = Service Employees International Union (Canada); FOA = Trade and Labour Denmark; JHL = The Trade Union for the Public and Welfare Sectors (Finland); SuPer (Finnish Union of Practical Nurses (Finland); Tehy The Union of Health and Social Care Professionals (Finland); Fagforbundet (Norwegian Union of Municipal and General Employees (Norway); Kommunal (The Swedish Municipal Workers’ Union (Sweden).
The Scandinavian study, headed by Dr. Marta Szebehely, included fixed and open-ended questions collected as part of the “NORDCARE: The Everyday Realities of Care Workers in the Nordic Welfare States” project. It was distributed in 2004 to a random sample of 5,000 unionized care workers in private homes and residential care facilities in four countries. Unlike in Canada, where the sample was conducted at the level of workplaces, the Scandinavian team was able to send the survey directly to workers’ homes because we had address information supplied by the care workers’ unions in each country.Footnote 2 The overall response rate was 72 per cent (Denmark, 77%; Finland, 72%; Norway, 74%; and Sweden, 67%).
Like the Ontario study, the Canadian study received approval from York University’s ethics review process, and the ISR was responsible for the sample design, distribution, and data entry. We worked with our Scandinavian partners to replicate as many of their survey items as possible, and prior to distribution, a separate Canadian consultation elicited feedback from our collaborating organizations in the five major Canadian health care unions.Footnote 3 In addition to many of the same questions that were in the Scandinavian instrument, the Canadian instrument also included items from the Ontario study. Thus, like the Ontario study, the Canadian survey instrument explored employment qualifications, patterns, and workload, and like the Scandinavian survey, it also asked workers more-detailed questions about their working conditions, health and safety risks, work and family life issues, and the acuity of residents’ needs.
The ISR piloted the Canadian study in November 2005 with 11 (8 female and 3 male) LTC workers in a variety of positions working within the same union in different workplaces across the jurisdiction of Ontario. After completing the survey, respondents were asked about its clarity, each question’s readability and importance, and the amount of time needed for its completion. We made revisions on the basis of this feedback, and the ISR distributed the survey to contacts at 81 randomly selected, unionized LTC workplaces between January and August 2006; workers from 71 (87.6%) of the facilities participated.Footnote 5. Our sampling strategy was designed at the organizational level because, unlike our Scandinavian partners, we did not have a list of workers with home addresses to work with.Footnote 4 A total of 948 surveys were returned from housekeepers (n = 101), dietary aides (n = 73), personal support workers (n = 415), licensed practical nurses (n = 139), and registered nurses (n = 141), representing 13 per cent of all workers at these workplaces (n = 8,149). Quantitative data were entered into Statistical Package for the Social Sciences (SPSS), and descriptive statistics, cross-tabulations, and tests of significance were applied. The open-ended survey responses were entered into a Microsoft Excel spreadsheet and analysed for thematic content. We compared Canadian results with Scandinavian results to find areas of commonality as well as those of difference.
With the threefold objective of validating, elaborating, and explaining the Canadian survey findings, we conducted nine qualitative focus groups post-survey in 2006 as part of the Canadian study. Each group had between three and eight workers drawn from each participating province and union. The purpose of conducting the focus group interviews was to expand upon and triangulate the survey data. Following participant consent, we read a survey question to focus group participants, presented descriptive statistical results, and asked them: “Does this reflect your experience?”.Footnote 6 Workers were also asked select, open-ended questions about issues that had surfaced in previous focus group sessions.Footnote 7 The results were professionally transcribed verbatim. These qualitative data were thematically analysed by individual team members and then collectively analysed by the team.
According to Doyle, Brady, and Byrne (Reference Doyle, Brady and Byrne2009), there are numerous reasons for conducting a mixed-methods research project. Our Canadian study was designed to triangulate quantitative closed-ended survey responses with structured qualitative open-ended focus group responses. We conducted the survey, followed by the focus groups, and then completed an integrated analysis. This format corresponds to the full, sequential, and equal mixed-methods design described by Leech and Onwuegbuzie (Reference Leech and Onwuegbuzie2009) as well as the convergence-model mixed-methods triangulation design described by Creswell and Plano Clark (Reference Creswell and Plano Clark2007). Whereas these two designs establish the sequence of data collection and the degree of analysis integration, we refer to our study as an iterative mixed-method design because it also involved integration and iteration of the survey and focus group instruments.
Findings
Despite jurisdictional similarities in terms of unionization levels and resident characteristics, our findings indicate that the violence reported in Canadian LTC homes was ubiquitous and persistent compared with much lower levels of violence reported in Scandinavia. This led us to question why, given these similarities, violence was more pervasive in Canadian facilities. The quantitative data showed that, when working conditions in Canada were compared with Scandinavia, staffing levels were much lower in Canada, where workers cared for higher numbers of residents and experienced more physical, sexual, and verbal forms of violence at work. Some of these findings were confirmed and extended by the qualitative data, which revealed how working conditions – including working time-crunched and working short-staffed – made the frequency of violence of all types – resident to worker, worker to resident, and resident to resident – a more likely and, in many respects, a normalized part of the workday.
Here, we show how the different methods enabled us to triangulate some of our findings, while also revealing to us results that diverged or were contradictory. We report on two specific findings – about under-staffing that contributed to poor working conditions, and about violence workers experienced during their shifts – and illustrate them with selected examples from our results.
Working Time-Crunched and Short-Staffed
Working under serious time pressure while a facility simultaneously lacked adequate numbers of staff was a key issue that emerged in the Canadian surveys and Canadian focus group interviews. The daytime shift is generally considered the most intensive (Lopez, Reference Lopez2006) because it involves the morning dressing ritual, more toileting than at other times of the day, and meals and snacks. This shift, therefore, is most representative of the workloads experienced and described by workers. We found that each Canadian direct care worker working on the most recent weekday shift, and having hands-on responsibility for residents, was responsible to care for an average of 19.6 residents. This far surpassed the number of residents cared for by each worker in the Scandinavian countries: Danish workers were each responsible for an average of 6.2 residents, which increased to an average of 7.7 in Norway and 8.5 in Sweden. Finnish workloads, averaging 15 residents for each day shift weekday worker, were most similar to those of Canadian workers. Canadian workers who reported caring for more residents also reported feeling that they cared for too many residents. For instance, 8 in 10 Canadian workers (80.5%) reported they had too many residents to care for on the daytime shift, compared with less than half of Finnish (48.3%) workers, about one in three Norwegian (33.4%) and Swedish (28.6%) workers, and even fewer Danes (25.3%). The difference between the Scandinavian and Canadian workers was significant (p = .001).
Canadian workers were also more likely than the Scandinavian workers to report that they worked short-staffed – that is, they regularly worked without the facility’s full complement of workers. Nearly half of Canadian direct care workers (46.2%) reported that they worked short-staffed almost every day. By way of comparison, an average of only 15.4 per cent of Scandinavian workers reported being short more or less every day, but almost as many Scandinavian workers (30%) as Canadian workers (34.4%) reported being short on a weekly basis. The difference between the Scandinavian and Canadian response patterns was statistically significant: p = .001. In addition, approximately three quarters of Canadians working a weekday shift (78.7%), an evening shift during the week (74.6%), or a weekend shift (75.5%) reported a lack of time to properly care for residents’ needs. In line with the fact that the most active work does not take place during the night, time constraint issues were less pronounced for night-shift workers, with just under half (47.3%) reporting a lack of time to properly care for residents. Likewise, Canadian personal support workers were twice as likely as their Swedish counterparts to report that they had “too much to do all or most of the time” (60.3% compared with 36.4%; p = .001). Some personal support workers described fellow workers as “sick, stressed, and burned out”. Some noted that the normal workweek was usually four, not five, days in a row because people experienced too much burnout when working more than four consecutive days.
In the Canadian survey, we found that when a facility was short-staffed, workers could not complete some physical aspects of caring tasks in a timely manner. The three tasks most frequently cited as being left undone were foot care, tooth brushing, and toileting; almost one quarter (22.2%) of Canadian workers reported that foot care (a critical issue for anyone with diabetes) was “often” not done, with another quarter (25.6%) noting that they were “sometimes” unable to do it. More than 1 in 10 workers (13.7%) reported that tooth brushing was often not done while 2 in 10 workers (22.7%) were sometimes unable to complete the required grooming. Six per cent of workers reported that toileting was “often” left undone while another 13.5 per cent reported that it was “sometimes” left undone, which raised questions about the circumstances within which care was provided.
The survey’s open-ended written comments helped to illuminate workers’ experiences of being time-crunched and short-staffed, and they frequently underscored what solutions could ease these conditions. The word “more” appeared frequently in worker comments: for instance, they cited the need for more staff, more nursing hours, more time to care, and more activities for residents. Similarly, the word “better” emerged in relation to perceptions of the need for better attention being paid to residents, better remuneration, better facilities, and better food. Although the written comments provided a window into the work, the Canadian focus group discussions brought to life the sense of being hurried and the lack of time that workers experienced. Workers described how residents were “processed” in too little time. For example, when LTC homes were short-staffed, “Beds don’t get made. Rooms don’t get tidied. Residents are washed, dressed, and fed”.
It was the focus group interviews that provided analytical insights. In the first focus group, workers linked their sense of time crunch directly to a managerial focus on budgets; workers provided adult diaper policies and procedures as examples of how they were rushed to care for residents. When we asked if residents receive enough assistance to meet their needs with toileting, for instance, we heard that “… there’s a lot of residents that don’t need to have briefs on but they do [have them] because there’s not enough staff [to toilet them properly].” A worker discussed how management was “really push[ing] the use of incontinence products … and now they’re limiting us to how many [diapers] that we can put on these residents”. She described how residents were not being toileted on a regular basis, and how they were left sitting in the diapers until the saturation point. She talked about how “there’s only so many [diapers] that are sent to each unit. It’s one [diaper per resident] per shift. It’s unbelievable”.
We asked questions about diaper policies in each of the subsequent focus groups and learned how incontinence products were used, in effect, to stretch the number of staff available: in describing how the diapers have a blue coloured line that appeared at the top of the diaper once it was three quarters wet, some respondents indicated that their facilities had policies that diapers were not to be changed until that line was visible to workers. Other respondents indicated they were “supposed” to put diapers back onto residents, even following the twice-weekly bath, if it was not 75 per cent full. One staffer referred to management as the “diaper police”. Another described how they were supposed to sign their names on diapers as a form of accountability and quantity control for managers to monitor how many diapers workers used on residents.
Resistance to these practices figured prominently in workers’ narratives about diapers, as we heard not only about the rules but also about how workers circumvented them in the interests of providing better care for their residents. Workers resisted by hiding unused diapers in closets, drawers, and above ceiling tiles. One woman said that, in effect, “We have to steal them”. Others spoke of ignoring the diaper policy or of being frustrated by constantly having to ask for appropriate resources. Some workers reported that they could stand up to management on such issues, although this was not a universal sentiment.
The Violence Veil
In trying to better understand the issue of violence, and to trouble the notion popularized in media accounts that violence is primarily abuse of residents by workers, part of our Ontario survey focused on the prevalence of resident-to-worker and resident-to-resident violence. Our results showed that violence is common and occurs frequently within LTC homes. Almost all workers (96.3%) indicated that some type of violent incident had occurred in their facility in the previous three-month period. In fact, the majority of workers (54.9%) reported that some form of violence occurred 11 or more times in the three-month period preceding the survey’s distribution. In total, over 80 per cent (82.6%) of workers indicated having to deal with violence towards a co-worker initiated by one or more residents. Overall, the results from the Ontario survey showed that nearly 10 per cent of workers (8.4%) reported that violence was a daily incidence in their LTC facilities (Armstrong & Daly, Reference Armstrong and Daly2004).
When we asked about violence towards workers perpetrated by a resident or resident’s family member in the Canadian survey, nearly 4 in 10 workers (38.2%) reported experiencing violence “more or less everyday”. Interestingly, Canadian workers were more likely to report violence than those in the Nordic countries where an average of only 6.6 per cent of workers reported experiences of violence or the threat of violence more or less every day (p = .001): Finnish workers represented the highest proportion (8.1%), and Danish workers represented the lowest (5.0%). The Canadian survey may reflect a further underestimation of the problem’s true extent because the question did not include the “threat of violence” as it did in the Scandinavian version; however, a further 2 in 10 Canadian workers (22.8%) reported at least weekly exposure to physical violence, meaning that almost two thirds of the Canadian workers we interviewed (61%) reported that it was at least a weekly occurrence, compared with fewer than 2 in 10 Scandinavian workers (18%).
During the focus groups, we reiterated the Canadian survey’s results (as described in the foregoing paragraph) and asked workers “Does this reflect your experience?”. The workers recounted stories that made it clear: violence was a persistent, even an expected part of their work with seniors. In some cases, it appeared so normalized that many excused it as an ordinary part of working with seniors. One woman observed, “[i]f it happened on the street we would report it because we take it for granted that that person on the street knows what they’re doing. When it becomes seniors in a long-term care facility they might know what they’re doing, but they’re senile and … we’ve been told, ‘it’s part of [our] job’”. The focus group data indicated that workers were subjected daily to physical violence such as scratching, kicking, spitting, swinging, and pinching: “somebody has always got a bruise or a bump” inflicted by a resident or resident’s relative.
“I’ve been punched in the face several times. I’ve been punched in the jaw several times. Getting hit. Having your wrists twisted. That’s the big thing. They’re constantly twisting your wrists. Pulling and shoving at you. I mean that’s a day-to-day thing. I work on the locked unit. Violence is an everyday occurrence.”
We heard about the ways in which violence against workers was frequently unacknowledged by workers or that its impact was diminished by managers. Typical comments in the focus groups included these: “‘He’s just an old man. He’s 97 years old. What can he do to you?’ “‘Well you’re a big girl … Lighten up.’” In the open-ended comments, an Ontario female nurse, with 11 to 15 years’ experience on the job noted: “[r]esidents have all the rights and when it comes to staff being abused, nothing is being done about it. What about our rights?” Another worker commented, “it is a possibility on my floor but it’s not – I wouldn’t classify it as violence. Basically [it’s] like groping or if you happen to get them on a bad day when maybe their pain control isn’t met through medication, they strike out at you”.
Interestingly, a publicly available Ontario provincial data set that documents complaints in LTC homes (Ontario Ministry of Health and Long Term Care, 2009) contained 611 complaints logged between 1996 and 2007. Yet, there were no resident-to-staff cases of abuse listed, suggesting that violence against workers perpetrated by residents and their families occurs under the policy radar. This normalization of violence escaped us in our quantitative analysis, not because workers didn’t report violence in the survey, but rather, because we were unable to capture workers’ under-reporting that resulted from their minimizing events that, to us, were unquestionably examples of violence but which to them amounted to a normal part of their workdays. They simply did not see how things could be any different.
Violence in these facilities involved much more than the physical and psychological interactions between two people, and workers directly related it to the organization and conditions of work within the facilities. Without the focus group discussions, the structural conditions breeding violence and its normalization would have remained veiled to us as researchers in the same way that workers operate behind this “violence veil”. If workers had not talked about the conditions of their work that contributed to violence, or if they had not excused some aspects of violence as normal parts of working with seniors, and had this finding not been reinforced by the high rates of violence reported in the Canadian survey data (far exceeding rates reported in the Scandinavian jurisdictions that were part of the study), we would have missed the ways in which violence was systemic and structural in Canadian long term residential care. In effect, the combination of the qualitative and quantitative methods, allowed us to document that violence happened and to understand what contributed to its being systemic in Canada.
Workers’ responses about resisting violence were more sublimated compared to their responses about their facilities’ diaper policies. Front-line workers could handle the latter issue with far less difficulty: they described managers as being unwilling to get their hands dirty and they thus had some autonomy at the bedside to make micro-decisions in the interests of their residents. In contrast, workers described managers as being unsupportive in the attempt to moderate violence against workers because managers did not take staff “seriously”. In many cases, workers described managers who blamed the staff. The workers were “sometimes” able to have a violent resident transferred elsewhere, but this was not a strategy they could rely upon.
Another control mechanism was worker training to diffuse violence and to follow tactics such as “back away”, or to “leave a resident alone” by retreating. However, in the words of one worker, this did not prevent what was described as “a daily part of our day dealing with some form of violence in the workplace”. In many respects, strategies of backing away and retreating easily fit into a workplace culture of overwork and time crunch, in which other residents also urgently require care, and “regular” cases cannot be adequately staffed or accommodated, thus leaving little time for the “hard to handle” cases. But as one worker noted, the situation with respect to violence had moved beyond hard-to-handle cases, as the buildings were full, and more residents had multiple conditions, had higher levels of acuity, and exhibited higher proportions of Alzheimer’s disease and other forms of dementia, all requiring more intense care work.
Stewart et al. (Reference Stewart, Makwarimba, Barnfather, Letourneau and Neufeld2008) noted that when they used qualitative and quantitative data in the same study, they sometimes achieved divergent results, which corresponded with our own findings. For instance, a clear majority of the Canadian survey respondents (72%) reported that they either “never” or “less than every month” received unwanted sexual attention from a resident or their relative. On the basis of the survey, we could have easily concluded that unwanted sexual attention was not a problem in LTC facilities. Some workers in the focus groups laughed in disagreement when presented with our survey data. They noted how unwanted sexual attention was a regular part of their experience, increasing in severity depending on the floor they were on, the level of cognitive impairment amongst the residents, and how many men were in a home, although some noted that women residents did things too: “I’ve had women grab me by the breast: ‘Oh, aren’t you big?’, you know, and you don’t expect it and [yet] it happens”. It was only in the focus groups that respondents reported “constant” and unwanted sexual attention in combination with the verbal, physical, sexual, and moral violence and assault: “there’s a lot of inappropriate sexual behaviour by the residents. You’re doing cares [sic] on them and they’ll make quite crude comments”. As was the case with the violence experienced, the workers had few strategies other than to laugh off the advances or to complain to fellow workers. In one case, a worker indicated that when she went to management to complain, she was told that “perhaps she shouldn’t be so friendly with the male residents”, suggesting that she had brought the incident upon herself.
Discussion and Conclusions
Lifting the Violence Veil and Improving Working Conditions
The finding that care workers in various settings confront high levels of injury, illness, and violence has been demonstrated in previous studies (Armstrong et al., Reference Armstrong, Banerjee, Szebehely, Armstrong, Daly and LaFrance2009; Banerjee et al., Reference Banerjee, Daly, Armstrong, Armstrong, LaFrance and Szebehely2008; Baines Reference Baines2006; Boyd, Reference Boyd1995; Gates et al., Reference Gates, Fitzwater and Meyer1999, Reference Gates, Fitzwater and Succop2005; Hellzen et al., Reference Hellzen, Asplund, Sandman and Norberg2004; Levin, Reference Levin, Hewitt, Misner and Reynolds2003; Shaw, Reference Shaw2004). Our findings, which corroborate those from the earlier studies, show high levels of reported violence in the Canadian context but also reveal other ways in which workers normalize violence as an intrinsic part of their care work, and show the links between violence and the structural conditions of work, including how having too little time to properly care can contribute to making violence more likely. Our research also identified how violence, though normalized, is not a normal part of care work experienced elsewhere. Although it is pervasive in Canadian LTC facilities, the rates reported far exceed those reported by surveyed workers in four Scandinavian countries. Our findings indicated that Canadian workers toiled under conditions that were often lacking in adequate supports or appropriate recourse. As a result, violence was a veiled aspect of the care work culture in Canada, mostly hidden from view and frequently under the radar of policy makers who are in positions to improve working conditions.
In the Scandinavian countries, working conditions in LTC facilities were shown to be far superior: caring was more readily funded by the state and valued as a right of citizenship (Szebehely, Reference Szebehely, Armstrong, Boscoe, Clow, Grant, Haworth-Brockman and Jackson2009), and our survey respondents cited a higher staff-to-resident ratio, more-frequent baths for residents, and having more time for everything from chatting to toileting. In these countries, workers also reported less-frequent occurrences of violence, a finding also borne out in several other quantitative Scandinavian studies (e.g., Menckel & Viitasara, Reference Menckel and Viitasara2002; Sharipova, Borg, & Hogh, Reference Sharipova, Borg and Hogh2008). Note, however, that one study that focused explicitly on violence in care homes found about double the Scandinavian rate reported in our study (Astrom, Bucht, Eisemann, Norberg, & Saveman, Reference Astrom, Bucht, Eisemann, Norberg and Saveman2002), although this is still much lower than the amount of violence reported in our Canadian study. Another qualitative study (Åkerstrom, Reference Åkerstrom2002) found widespread violence in Scandinavia that remains unacknowledged due to the extent of normalization that also occurs there. As a result, the conditions that contribute to violence and to its normalization in the Scandinavian context need to be better understood by conducting more research using mixed methods.
Farmer (Reference Farmer, Nizeye, Stulac and Keshavjee2006) noted that structural violence involves “social arrangements that put individuals and populations in harm’s way” (p. 1686). Upon reflection, we were struck with how the qualitative and quantitative methods revealed different but reinforcing aspects of structural violence. The quantitative data revealed lower staffing levels, heavier workloads, and higher levels of reported violence in Canada compared with Scandinavia. Our use of an iterative approach, literally asking focus group participants if the survey results reflected their experiences, prompted workers to elaborate on our survey results and tell us personal stories that became key to our data analysis and interpretation. Without a mixed-method design, we could not have learned how violence was experienced, normalized, and seldom resisted by Canadian workers, or about the lack of managerial support to prevent violence, or, finally, about the feeling of time crunch that resulted in inadequate care. Our findings coincide with those of Donna Baines (Reference Baines2006) who noted that the ways in which care work is organized “draw on notions of the endlessly stretchable capacity of women to provide care work in any context, including contexts involving violence and overwork” (p. 130). The time crunch that workers described confirmed others’ findings (e.g., Lopez, Reference Lopez2006), and highlighted how lower than optimal levels of funding for LTC facilities may contribute to conditions of work overload with workers who, with less time, performed more tasks and cared for more residents.
Using an Iterative Mixed Method
Although some of the foregoing issues might have emerged in the qualitative semi-structured key informant interviews, the quantitative survey data allowed us to begin the analysis, to more easily compare Canadian and Scandinavian conditions, and to note significant differences in workers’ responses about exposure to violence and their assessment of time available to care. As a result, using a mixed-methods approach allowed numerical data to be qualified, workers’ rich descriptions to be enumerated, and unexpected findings to surface, all of which contributed to a more nuanced and contextualized understanding of violence and of work organization.
Importantly, these international data revealed that high levels of violence need not be the norm: whereas the quantitative data showed that, in Canada, violence occurs to a degree that vastly exceeds the other jurisdictions under investigation, the degree of normalization and the reasons for it were only revealed after further in-depth qualitative examination, leading us to conclude that violence may be ubiquitous and systemic in Canadian LTC homes. Over half (61%) of the workers surveyed reported being subjected to violence at least on a weekly basis. Yet, the written survey comments together with the focus group responses revealed that workers normalized their experiences of violence, particularly if management support or acknowledgement was lacking. This suggests that workers might excuse instances that in another setting would unquestionably be identified as violence. In this way, violence is likely an under-reported, largely invisible, veiled, and accepted part of the job in Canadian LTC facilities. Thus, a more nuanced understanding of violence was revealed during the iteration between the qualitative and quantitative instruments and developed as a result of the integrated analytical approach.
In reflexively analysing how our method was instrumental in yielding these insights, we were able to consider the sequence of our qualitative and quantitative data collection and our iterative approach to asking workers about their experiences. Whereas much of the methods literature focuses on the sequence of qualitative and quantitative data collection and the analytical interaction between the two data sets, our contribution to this literature has been to expand the discussion of the interaction between qualitative and quantitative instruments in terms of what and how questions are asked.
For instance, there is ongoing work that articulates how qualitative and quantitative methods mesh in terms of the data collection sequence and whether the respective data results are treated separately, compared, or analysed together as a whole (Creswell, Reference Creswell2009; Creswell et al., Reference Creswell, Plano Clark, Gutmann, Hanson, Tashakkori and Teddlie2003; Creswell & Plano Clark, Reference Creswell and Plano Clark2007; Leech & Onwuegbuzie, Reference Leech and Onwuegbuzie2009). In contrast, our iterative method focused on repetition of the content of the instruments as opposed to a focus on the data collection sequence. Our approach was iterative because we literally restated our questions and survey findings to participants in each focus group, and collectively their insights became a central part of our analysis and interpretation. We found that this approach yielded rich, detailed qualitative data and encouraged participants to disagree with, elaborate on, and/or extend our understandings. In areas not readily defined, such as “violence”, this approach was particularly helpful, because it revealed how workers focused less on the violent act (hitting, punching, kicking, or verbal abuse) and instead normalized the acts to the setting; some noted that they would not tolerate such treatment in other settings, for instance if it “happened on the street”. Our iterative method was also very useful for helping to discern symbolic issues: for example, as workers noted, the diaper policies adopted in many LTC homes were emblematic of larger problems associated with the organization of their work and the lack of time available to properly perform good care.
We conducted the survey and focus groups sequentially primarily to triangulate findings to improve the study’s validity; however, we did not always achieve data convergence. In some instances, we grappled with data divergence: although the qualitative data sometimes reinforced our survey findings, at other times it revealed the ways workers minimized violent circumstances because of their expectations that violence was routinely part of their work with seniors in long-term residential care. Thus, it was our sequential combining of methods and our use of the same questions to elicit results that afforded us deeper meaning, context, and insights.
The iterative method also revealed limitations of our survey questions. For instance, we mirrored common notions of violence in LTC facilities as dyadic relationships between two individuals – resident-to-resident, resident-to-worker, or worker-to resident. In contrast, the qualitative data directly opposed interpretations of straightforward dyadic violence by revealing the structural and systemic conditions of work organization, such as heavy resident-to-staff workloads and a lack of time afforded to provide good care – that is, sufficient time to perform all expected activities involved in, which produce a culture of accepted violence within the LTC context.
Our use of this iterative approach in which we recounted survey questions and results to focus group participants and asked if this reflected their experiences had the intended effect of encouraging workers to elaborate on our survey results and tell us personal stories that became key to our data analysis and interpretation. Without a mixed method, we could not have learned how violence was experienced, normalized, and seldom resisted by Canadian workers. It was important that we ask ourselves what the quantitative results revealed and what they concealed. Would we have understood work in long-term care if we did not use the survey? Would we have realized the significance of the context of care, such as the average number of residents cared for by each worker in each of the settings? How would our understanding of workloads and of violence have differed without the focus groups? We would not have known that workers were expected to limit their use of incontinence products, for example, which contributed to worker stress and to creating a site of worker resistance, nor would we have known that workers’ experiences of violence were routine to the point of being invisible and were thus accepted as just a part of the job. We would have missed the ways that violence co-exists with poor working conditions that impinge on these workers’ autonomy, and affect the ways they negotiate their work. We did not always hear consensus from our focus group participants: it took exceptional stories, such as those telling about diapers in our first focus group, to reveal care workers’ limited agency, and which allowed us to explore linkages and to ask about this issue at each subsequent focus group.
Priorities for Future Research
The study results discussed here contribute to our understanding of comparative working conditions for low-paid women care workers. Given that the demand for LTC facility spaces will continue, combined with the reality that these facilities employ and care for our most vulnerable citizens, more research is needed so that we understand how work organization affects working conditions in facilities and thus how to ameliorate poor working conditions. International comparisons can provide promising practices for work organization, for models of care and for funding policies in Canada, including practices that involve having a lower resident-to-staff ratio. In addition, more mixed-method studies must be done in this sector to better understand difficult and hard to quantify issues. Specifically, more studies are needed that explore the interactions between working conditions and structural conditions of violence, as are studies that explore caring well as a form of resistance to policies and practices that create poor working conditions.
Two themes – lack of time to properly care (as exemplified by workers’ using diapers on residents in situations where they otherwise would not if staffing were sufficient), and high levels of violence experienced and accepted by workers – dominated our focus groups. Thus, beyond an academic research agenda, lessening the conditions that allow violence to flourish is a key issue that policy makers, unions, workers, and employers must address. In that regard, improving conditions for residents and for workers in LTC homes is a necessary first step
This article contributes to debates about the utility of using mixed methods in aging, health, and social care research, and to debates about the insights gained from data convergence and data divergence. At the same time, in highlighting how violence in Canadian LTC homes is structural, this article contributes to policy debates and the academic literature about working and living conditions in LTC facilities. This study thus provides a compelling illustration of how mixing qualitative and quantitative inquiry can result in research that contributes to clarifying struggles involving basic human rights, such as the right for workers to perform care in safe conditions and for seniors to receive quality care.
