Determinants of violence towards care workers working in the home setting: A systematic review

Abstract Background Home care is a rapidly growing industry. Violence towards home care workers is common, while also likely underreported. This violence adversely affects the physical and mental health of both workers and care recipients. The current study aims to identify and appraise recent evidence on the determinants of violence towards care workers working in the home setting. Methods Six electronic databases: the Cumulative Index to Nursing and Allied Health Literature (CINAHL), EMBASE, Informit, Medline, PsycINFO, and Web of Science, were systematically searched. A systematic review was conducted in accordance with the Joanna Briggs Institute manual for evidence synthesis. Results A total of 18 papers met the inclusion criteria. All were cross‐sectional surveys. The majority of studies were from the United States. The most commonly investigated associations were those between the medical history of clients, workers' apprehension of violence, worker‐client relationship, or care plans, and any form of violence or verbal abuse. Conclusion Violence was common in clients with cognitive disorders, substance abuse disorder, and limited mobility; toward workers who feared that violence might happen; toward those who had very close or very distant worker‐client relationships; and when care plans were not inclusive of clients' needs. The current review highlights a gap in evidence on determinants of violence towards care workers working in the home setting, and suggests potential areas to be addressed to reduce such violence.

social support, nursing care, and allied health services, 4 the workers providing home care services varied widely.
Violence in the workplace was defined by the International Labor Organization (ILO) as "a range of unacceptable behaviors and practices, or threats thereof, whether a single occurrence or repeated, that aim at, result in, or are likely to result in physical, psychological, sexual or economic harm, and includes gender-based violence and harassment." 5 The definition and scope of workplace violence differed across countries and based on research purposes. [5][6][7][8] A few examples are as follows: • Verbal violence: yelling, using abusive words, making racial slurs or other types of discrimination, and verbal threats of harm, etc.
• Sexual violence: unwanted sexual comments, unwanted body touch, and sexual assault, etc.
• Damage or loss of personal or work-related materials: vandalism, robbery, or theft, etc.
• Income-related violence: exploitation, fewer working hours, and loss of job, etc.
• Attacks by domestic animals.
Similarly, the classification of workplace violence differed across institutes or reports. The 2021 ILO report classified the violence as vertical-by employers, horizontal-by peer workers, and third parties-by clients. 5 The 2001 workplace violence report by the Injury Prevention Research Center categorized the violence as Type I-with criminal intent, Type II-by clients, Type III-by fellow workers, and Type IV-by perpetrators who had personal relationship with workers. 8 Violence towards care workers by clients or their families, Type II violence, is common in the home setting. [9][10][11][12] A meta-analysis of 21 studies of home care workers from the United States (n = 12), Israel (n = 4), Japan (n = 2), Australia (n = 1), Canada (n = 1) and Ireland (n = 1) estimated the violence prevalence of 22.3% over the 12 months before the survey, and 30.2% over the carer's career. 9 A review of 21 other studies, mostly from the United States (n = 8), looked at violence towards care workers both in home (n = 10) and institutional settings (n = 11).
Between 33% and 87% of home care workers experienced verbal abuse from patients over the workers' career or while doing fieldwork. 10 Both these reviews noted the diversity of the survey participants in terms of frequency of home visits, duration of each visit, nature of care, and interaction with care recipients. 9,10 Sexual abuse and sexual harassment was reported by 4% and 12%, respectively, from home care workers in a review of 14 studies of 6014 workers mostly from the United States (n = 5). 12 Despite the high prevalence of violence against caregivers in the home setting, literature suggests that these incidents may be underreported. [13][14][15] Reasons home care workers may not report violence include growing tolerance to violence, concerns of being blamed for the violence, losing working hours or their jobs, holding a temporary work visa, and unfamiliarity with legal system. 14,[16][17][18][19] Overall, it was not uncommon that home care workers were subject to different forms of violence perpetrated by clients.
Violent incidents have been shown to affect the workers' physical, mental and emotional health and to also impact the care recipients adversely. [20][21][22][23][24] A survey of 1214 female home care workers in the United States reported a statistically significant association between exposure to any form of workplace violence and stress, depression, sleep disturbance and burnout. 20 A review of nine studies that examined workers in home and institutional settings found a statistically significant association between physical violence and workers' mental health in nine out of 11 associations examined. 21 Similarly, of the 13 associations examined between psychological violence and mental health, 11 statistically significant associations were reported. 21 Furthermore, two studies included in the same review concluded that physical violence was significantly associated with musculoskeletal pain in workers. 21 In addition to the direct impact of workplace violence on workers' health, there was an indirect impact on the care recipients. Among 823 home health aides in the United States, low job satisfaction and retention was associated with workplace violence. 22 The poor health and high turnover rate of the workforce can adversely impact on the quality of care being provided. 23,24 Violence can lead to shortened or missed care visits and changes in care plans and this can have a negative impact on care outcomes. 23,24 Several determinants have been put forward by other authors as being associated with the poor safety of home care workers. 25,26 These include working alone in noninstitutional settings, inadequate health policies around home care, insufficient or a lack of record keeping and staff training, gaps between care planned for or received and the recipients' desire, and miscommunication among care recipients, providers and managers. 25,26 In addition, building designs not being conducive to care delivery, a lack of patient moving or handling equipment, and the location of client's homes in unsafe neighborhoods have also been suggested as factors associated with violence. 25,26 However, this is not a comprehensive list of possible determinants of violence as these studies have considered the safety of home care workers and not specifically determinants of the violence toward them. 25,26 In addition, there is limited experimental research around intervening factors that might reduce or prevent violence towards home care workers. 10 Several literature reviews on the risk factors for workplace violence in institutional settings already exist. However, much less is known about those risk factors that are unique to carers working in their clients' homes. [27][28][29][30] In the home setting, clients and their families are in a position of power. Substance use cannot be banned, or weapons cannot be removed from the home setting. Care workers usually have to work alone in clients' homes with minimal or no support or protection by colleagues or managers. There is often a lack of appropriate equipment, for example, slide sheets, shower chairs, patient lifters and hoists, and this lack has been associated with patient violence. Moreover, domestic animals and unsafe neighborhood could impose threats to care workers.
Given the numerous and varied range of adverse consequences following violence towards home care workers, and the unique risk factors of that violence, the aim of this study is to conduct a systematic literature review to identify determinants of violence that can be tested in an intervention study to reduce or prevent violence against home care workers at work.

| MATERIALS AND METHODS
A systematic review was undertaken according to the Joanna Briggs Institute (JBI) Manual for Evidence Synthesis. 31 Six electronic databases including Cumulative Index to Nursing and Allied Health Literature (CINAHL), EMBASE, Informit, Medline, PsycINFO, and Web of Science were searched systematically. The search was performed in August 2020 and updated in August 2021. The key words related to care workers working in the home setting and violence were used as search terms. The search terms were "home," "home care," "home healthcare," "home health aide," "direct care," "personal care," "nurse," "nurs* assistant," "social worker," "violen*," "workplace violence," "client violence," "safety," "workplace safety," "abuse,", "harass*" "aggress*," and "assault." The inclusion criteria were primary research studies conducted in high-income countries, published in the peer-reviewed literature, that examined factors associated with violence toward caregivers in the home setting, published in English during the years 2000 to July 2021. Only the studies from high-income countries were included for the following reasons: having an established home care industry, an ageing population, a high number of women in workforce, and nuclear family structures in these countries. 32 Home care workers were defined as those who (1) assist care recipients with activities of daily living, e.g., personal care attendants, home care aides, (2) provide health care, e.g., certified health care aides, home care nurses, or (3) provide emotional support in the homes of clients, e.g., counselors, chaplains.
Home setting referred to the homes of clients. Violence towards care workers by clients or their friends or family members, that is, Type II violence, was included, regardless of the form of violence such as verbal, physical, sexual, property damage or loss, and exploitation, etc. Only quantitative studies which reported determinants of violence towards home care workers were included.
The exclusion criteria were care workers who did not provide hands-on care to care recipients, family caregivers, or carers in aged care homes, nursing homes or long-term care homes. Specifically excluded were studies that examined violence perpetrated on clients, violence caused by co-workers, or that which took place during travel to clients' homes or in the neighborhood of clients' homes. Studies in low-and middle-income countries and published before the year 2000 were also excluded, as were secondary analyses and reviews.
The search results were screened against inclusion and exclusion criteria by one of the two authors, and 10% of the search results were screened by the second author. The papers with relevant titles and abstracts were retrieved for full text check. A manual search of the bibliographies of relevant papers was undertaken to identify further studies not found in the literature search.
The methodological quality appraisal of the included studies was performed using the JBI critical appraisal instrument for systematic reviews of prevalence and incidence. 33 The instrument assessed the presence of bias in study design, implementation, and data analysis.
The instrument included nine criteria and each criterion was judged as "yes," "no," "unclear," or "not applicable." The quality assessment was performed by one of the two authors, and 10% of the included studies was appraised by the second author. Discordant appraisal comments were discussed, and agreement reached. When more than one paper reported the same study, the first published paper was appraised for methodological quality.
Data extraction was performed according to the JBI data extraction form for prevalence studies. 34 The data extracted were the citation details including authors, title, journal, year, volume, and issue; generic study details such as study design, country, setting, timeframe for data collection, participant characteristics, violence types and prevalence, factors examined for associations with violence and descriptions of main results.
The determinants of violence towards the care workers working in the home setting were grouped into three groups: client factors, worker factors, and organizational factors. Due to the limited number of studies which examined associations between each factor and each type of violence, no meta-analysis or sub-group analysis was performed. The relevant findings of 18 included papers were critically appraised, narratively summarized, and a provisional conclusion was presented.

| Methodological quality of the studies reviewed
Lists of the target population, for example, staff registers of home care agencies, were used as sampling frames in all (n = 14) but one study, in which recruitment was performed at the places frequented by Filipino migrants working in Israel 39 ( Figure 2). Convenience PHOO AND REID | 449 sampling from staff meetings and training sessions was performed in nearly half of the studies (n = 7); the remaining (n = 7) performed random sampling by inviting all care workers listed in the registers via mails or emails, or by performing systematic randomization; one study from Japan did not clearly report the sampling strategy. 40 The required sample size was reported in only three studies, 11,40,41 and the sample size reached was adequate in two of them. 11,41 All the studies described their study subjects and settings. The response rates for sub-groups of participants were reported in only two studies, 16,42 and the rates were different across the groups with F I G U R E 1 Search results of a systematic review of determinants of violence towards care workers working in the home setting, 2000-2021 F I G U R E 2 Methodological quality appraisal of the studies included in a systematic review of determinants of violence towards care workers working in the home setting, 2000-2021 different socioeconomic status (SES) in the study from Australia, with 36% response rate in low SES capital city, 50% in high SES capital city and 54% in mixed-SES noncapital city. 42 Majority of the studies (n = 14) used valid methods to measure the conditions of interest.
Magin et al. 42 did not report details about the validity of the study questionnaire. The studies reported details about the validity of the questionnaire, and definitions of the variables used in the questionnaire. The majority of the studies (n = 10) applied the same measurement method for all the participants, that is, same recruitment approach, similar support to survey respondents regardless of survey modes, and trained interviewers. All the studies performed appropriate statistical analysis. The response rates were reported in the majority of the studies (n = 14), and it was less than 30% in three studies. 7,40,43 Overall, nearly two-thirds of the studies (n = 9) scored positively in six to eight out of nine quality appraisal criteria ( Figure 2 and Table 1).

| Characteristics of the studies reviewed
The study characteristics of the 18 papers reviewed from 15 studies are summarized inTable 1. All 15 studies were cross-sectional surveys. Nearly half of the studies (n = 7) were conducted in the United States. There was a total of 10,332 participants in 15 studies, ranging from a minimum of 130 participants to a maximum of 3377. The response rate was reported in 14 studies, and it ranged from 17% to 84%.

| Characteristics of the study participants
More than 80% of the study participants were females in the majority of the studies (n = 13). There were more male participants than females in the two studies which recruited medical doctors. 11,42 The mean age of the participants was around 45 years. The education level of the study participants was reported in eleven studies. More than half of the participants completed high school or less in five studies, 16,35,39,43,45 and more than 70% of the participants attained college or higher education in another five studies. 11,42,44,46,47 The remaining study reported that 64% of study participants held home care certificates. 6 Among the five studies with a majority of participants completing high school or less, three reported the race of the participants, and the percentage of African American was considerably larger than that of Caucasian (80% vs. 17%, and 42% vs. 14%) in the two studies, respectively. 35,43 In contrast, in the study of which 88.5% of participants had a college degree or higher education, almost 80% were Caucasian. 44 Race was reported in six of seven studies from the United States, and the proportion of African American ranged from 21% to 80%, Caucasians from 14% to 79.2%, and Asian 4% to 12%. 6,16,28,37,38,43 Among the eight papers from countries other than the United States, only one paper from Israel reported that all the study participants were from the Philippines. 39 The job titles of study participants varied widely and included direct care workers in homes, home care aides, home care attendants, personal care attendants, personal care homemakers, companions, certified health care aides, nursing assistants, hospice aides, home care nurses, visiting nurse managers, home health and hospice care providers, general practitioners, physiotherapists, speech therapists, social workers, social work assistants, chaplains, and bereavement counselors. Collectively, 63.6% delivered assistance for daily living, 35.8% provided health care such as nursing care, treatment, speech therapy and physiotherapy, and 0.2% undertook social work. The mean duration of home care experience reported in five studies was 9.7 years. [38][39][40][41]46 The mean working hours per week in three studies was 24.5 h per week. 6,28,39 Among the five studies which reported the proportion of study participants who had full-time or part-time jobs, less than half of the participants had full-time jobs in three studies. 11,41,47 3.4 | Factors associated with violence towards care workers working in the home setting

| Client factors
The most commonly assessed client factor was the medical history of clients (Table 2). Associations between client's illnesses or disorders and any form of violent events were examined in four studies, 7,28,38,39 and five effect sizes were reported. Associations between illnesses and verbal abuse was examined in one study, 36  Each methodological quality criterion was appraised as one of the four categories: "yes," "no," "unclear," or "not applicable." Each "yes" score was counted as "1" and summed for reporting purpose. impact on the quality of care received by clients. 51,52 Another determinant of violence against home care workers was when the care plan did not meet clients' expectations, or when the time allocated for care was inadequate. Unclear care plans could lead to misunderstandings between workers and clients and result in violence. 36 Moreover, there could be violence even with clear care plans when such plans did not meet the clients' needs, and when workers or clients did not adhere to the plans. 30

CONFLICTS OF INTEREST
The authors declare no conflicts of interest.

DISCLOSURE BY AJIM EDITOR OF RECORD
John Meyer declares that he has no conflict of interest in the review and publication decision regarding this article.

AUTHOR CONTRIBUTIONS
Alison Reid: identified the gaps in literature relating to the topic of the current study. Under her supervision, Nang Nge Nge Phoo: conducted a systematic review and drafted the manuscript. Reid cross-checked each phase of the review process, and finalized the manuscript.

DATA AVAILABILITY STATEMENT
All the papers reviewed in this study are available.