Determinants of Longer Job Tenure Among Home Care Aides

Abstract

An inadequate supply of direct care workers and a high turnover rate in the workforce has resulted in a “care gap” in our long-term care system. As people are increasingly choosing community-based care, retention of home care workers is particularly important. The mixed-method study described herein explored determinants of longer job tenure for home care aides (n = 261). Study participants were followed for 18 months, completing two mail surveys and one telephone interview each. Predictors of longer job tenure included older age, living rurally, lower physical function, higher wages, a greater sense of autonomy on the job, and less frequent feelings of personal accomplishment. Thematic analysis of telephone interviews revealed long-term stayers to be less concerned about low wages and inconsistent hours than those who left their jobs within a year; both groups of workers reported high levels of job satisfaction. Policy implications of study findings are discussed.

Keywords

home care aides long-term care workforce job retention

Introduction

There is little doubt that, even in the midst of a recession, the nation faces a labor force shortage among direct care workers in its long-term care system. As the U.S. population ages and the need for personal care services increases, this shortage is expected to reach crisis proportions (Institute of Medicine, 2008). An inadequate supply of workers ultimately means compromised care for our elders. Thus, understanding how to recruit and retain skilled caring individuals to do the important hands-on, day-to-day care that these paraprofessionals carry out is of paramount concern.

Four of five elders needing long-term care live in the community, and most individuals prefer to remain at home as long as possible (Stone, 2011). Consequently, the need for home care workers will be particularly severe; among all direct care workers, home care aides receive the lowest wages and face the most inconsistent hours, factors which have consistently been shown to increase job turnover (Morris, 2009; Seavey & Marquand, 2011). Home care aides (also referred to as personal care attendants [PCAs], personal assistant workers, and personal support specialists) provide social support and assistance with Activities of Daily Living (ADLs) and Instrumental Activities of Daily Living (IADLs) for their clients. The study reported on here examines factors contributing to retention of these valuable workers. As Mittal, Stiehl, and Leanna (2009) noted, factors associated with turnover and retention are not necessarily the same: what promotes retention may be different from what prevents turnover among direct care workers. Hence, it is important to examine not only predictors of termination, but also those factors that lead these valuable workers to remain in their jobs, in order to fully understand how to solve the workforce crisis we are facing.

The Home Care Aide Job: Challenges and Rewards

Although it is difficult to get an accurate count of the total number of personal care aides working in the community—given that so many are employed privately by families—conservative estimates suggest that there are currently about 2.5 million individuals in the workforce (Seavey & Marquand, 2011). Recent projections by the Bureau of Labor Statistics predict personal care aides to be the fastest growing occupation between 2010 and 2020, increasing 70.5% (Bureau of Labor Statistics, 2012). This growth is fueled by the increasing number of people needing personal assistance; the demand for personal care services is expected to double in the first half of this century, from 13 million in 2000 to 27 million by 2050 (Kaye, Chapman, Newcomer, & Harrington, 2006). While most personal assistance services are provided by unpaid family caregivers, the availability of these informal caregivers is decreasing due to rising employment rates among women, smaller family sizes, increased childlessness, and higher divorce rates (Johnson & Wiener, 2006). Ultimately this translates in a greater reliance on paid personal care.

While many—if not most—home care aides find the work they do very rewarding (Ashley, Butler, & Fishwich, 2010; Butler, Wardamasky, & Brennan-Ing, 2012; Faul et al., 2010; Stacey, 2011), there are also challenges which ultimately contribute to a high rate of turnover, estimated to range from 44% to 65% annually (Seavey & Marquand, 2011). First, there is an array of hazards that can affect the safety and health of home care aides. These include musculoskeletal issues related to lifting and transferring clients, risks involved with physically demanding housekeeping, and hazards related to transportation and negotiating snow and ice covered entrances to homes (Seavey & Marquand, 2011; Stacey, 2011). Emotional stressors also play into the nature of this work including having too much responsibility with inadequate training (Stacey, 2011), dealing with deterioration and death of their clients (Butler, Simpson, Brennan, & Turner, 2010), or working within bureaucratic rules that do not always jibe with what workers believe is in the best interest of their clients (Stacey, 2011).

The low level of compensation and lack of benefits is likely the primary challenge to this work. In 2007, Forbes magazine included personal and home care aide jobs among the 25 worst paying occupations in the country (Stone, 2011). In 2010, home care aides had a mean wage of US$9.44 per hour and 31% lacked health insurance—nearly twice the rate of people lacking insurance among all Americans under age 65 (Seavey& Marquand, 2011). To compound the problem of low wages, personal care workers in the home face inconsistent hours with more than 50% of jobs being part-time (Scala, 2008).

Despite such difficult conditions, many workers remain in the field for many years, priding themselves on the fine care they provide to their clients. Stacey (2011), in her comprehensive study of the work experiences of home care aides, argues that the workers who stay construct an identity she labels “the caring self,” which allows workers to “establish dignity and social worth in an invisible, devalued and ‘dirty’ job” (p. 136). Further understanding the factors that contribute to job retention is crucial as we attempt to address the existing and growing care gap in home- and community-based long-term care. The Home Care Worker Retention Study (HCWRS) described herein sought to advance understanding regarding factors and the experienced context that contribute to home care aides remaining in their jobs despite the often difficult job conditions. The specific research questions were:

What factors predict longer job tenure among home care aides?

How do the reported experiences of short-term “leavers” and long-term “stayers” compare?

Based on prior research, three categories of factors were predicted to affect job tenure: personal characteristics, job characteristics, and worker experiences of their jobs. As shown in Figure 1, personal characteristics include demographic factors such as age, health, and rural versus urban location; job characteristics include factors such as wages, hours, and occupational injuries; and individual experiences of the job include feelings of burnout (e.g., Gellis & Kim, 2004), job satisfaction (e.g., Lambert, Hogan, & Barton, 2001), and perceived empowerment (e.g., Barry, Brannon, & Moor, 2005; See Figure 1). The overall hypothesis of the study was that older age, along with more favorable employment factors (i.e., higher wages) and job experience factors (i.e., greater job satisfaction) would predict longer job tenure for home care aides. Based on the work of Feldman (1990), which found older age to be a predictor of greater job satisfaction and intent to stay on the job in a sample of 1,284 home care workers, an exploratory pilot study (Butler, 2009) was conducted to examine the experiences of older home care workers. While participants in this pilot study described problems with the work, they also indicated advantages of the job for older workers. The HCWRS built on this pilot study to more fully explore how age along with other factors contributed to a worker staying with a home care employer.

Figure 1.

Factors related to length of employment.

Method

Data Collection

After approval by the Institutional Review Board of the University of Maine, the longitudinal HCWRS was implemented in October 2008. A mixed-method design was chosen to address the stated research questions based on the current state of the literature on the topic. Eleven home care agencies/companies participated in the study. At the time of data collection, Maine was estimated to have about 6,201 home care aides (Seavey & Marquand, 2011) and about 122 home care agencies (Morris, 2009); thus the HCWRS sample was composed of about 4.2% of all home care aides in the state who collectively were employed by about 9.0% of all state home care agencies. All counties of the state were served by the 11 agencies in the study whose Personal Support Specialists (PSSs, the title of personal assistance workers in Maine) provided personal care to individuals in their homes. Clients receiving care were predominantly frail elders, but also included some younger adults with physical disabilities. Agencies agreed to send a mail survey (Time 1) to all their PSS workers (n = 496) in October 2008; over half of the workers completed the survey (n = 261), giving a response rate of 52.6%. Participating agencies provided the study team with regular reports of study participant terminations in the 18 months after completion of the first survey. Workers who left their employment received a second mail survey and completed a telephone interview lasting between 15 and 30 min soon after their termination (n = 90); workers who remained employed throughout the 18-month study period (n = 171) received the second survey and completed the telephone interview during the spring and summer of 2010 (Time 2); 252 interviews were completed giving a study retention rate of 96.2%. Study participants received US$20 grocery gift cards at both Time 1 and Time 2.

Dependent Variable

The first research question examined factors predicting longer job tenure. Job tenure was measured in days from the date of hire to either the date of termination (for those who left during the study period) or 18 months after completing the Time 1 survey (for those who stayed on the job).

Survey Instrument and Interview Questions

The two survey instruments assessed demographic information and current home care jobs, and included several standardized scales measuring health status and job experience factors such as job satisfaction, burnout, and empowerment. Demographic variables examined at Time 1 included age, gender, race, family income, education, marital status, whether the study participant cared for dependent children or adults, and whether or not they lived in the most rural regions of the state. Participants’ overall health—both physical and mental—was measured at Time 1 by the SF-36 of the Medical Outcomes Study (Ware, 1993). Job factors such as hours, wages, unreimbursed mileage, and occupational injuries were measured both at Time 1 and Time 2.

The job experience measures were also included on both the Time 1 and Time 2 survey instruments. The 17-item Grau Job Satisfaction Scale was used to measure both intrinsic and extrinsic job satisfaction (Grau, Chandler, Burton, & Kilditz, 1991); items included statements such as “I see the results of my work” (from very true to not true at all) and in this study sample, Cronbach’s α was .86. Burnout was measured by the Emotional Exhaustion (nine items, e.g., “I feel emotionally drained from my work”) and the Personal Accomplishment (eight items, e.g., “I deal effectively with the problems of my clients”) components of the Maslach Burnout Inventory (Maslach & Jackson, 1981). Each of these components of burnout was measured for frequency (seven response categories, from never to daily) and intensity (five response categories, from very mild to very strong). Internal consistency scores for the Emotional Exhaustion scale in this sample were .89 for frequency and .91 for intensity; the Personal Accomplishment scale had a Cronbach’s α of .75 for frequency and .93 for intensity. Empowerment—including the constructs of autonomy, responsibility, and participation—was measured through the 15-item Perceptions of Empowerment Instrument (Kiefer et al., 2005) which had a Cronbach’s α of .88 in the study sample. This scale included statements such as “I have the freedom to change how I do my job” for which respondents indicated level of agreement (from strongly agree to strongly disagree).

The telephone interview was composed of 18 open-ended questions. The questions of particular focus in this article asked participants about what led them to do home care aide work; what aspects of their PSS work brought them satisfaction; what parts of their job they found draining; and why they had left, if terminated, or what might make them consider leaving, if still employed. We also examined how they described the home care agency/company for which they worked.

Analysis

Quantitative Analysis

Data from the Time 1 and 2 survey instruments were entered into the statistical software program SPSS 17.0 for analysis. Several demographic factors were converted to dummy variables for analysis. Due to the homogenity of the sample, race was recoded to “White”, with all other racial categories forming the reference group. Two dummy variables were created for marital status: “widowed” and “married” with single/divorced as the reference group. Five other demographic factors were also treated as dummy variables measuring whether or not the study participants were Franco-American, had health insurance, cared for dependent children, were responsible for dependent adults, or lived rurally. Descriptive statistics were performed on demographic data. Bivariate statistics—correlations and independent sample t-tests—were used to explore relationships between length of employment and variables of interest as outlined in our conceptual model. Only those variables that were significantly related to length of job tenure at the bivariate level were retained in the multivariate analysis in order to develop a parsimonious regression model. Ordinary least squares regression was utilized to identify the primary factors predicting increased job tenure.

Qualitative Analysis

Telephone interviews were tape-recorded and transcribed verbatim. Using the constant comparative method of grounded theory, the project staff (the first and third authors) performed a systematic thematic analysis on interview transcripts of two groups of study participants: those who had worked at their agency for 7 years or more (long-term stayers, n = 30) and those who had left before completing 12 months (short-term leavers, n = 29). Transcripts were analyzed line-by-line or phrase-by-phrase to inductively identify emerging themes with regard to how study participants experienced their work (Boyatzis, 1998; Strauss & Corbin, 1998). After broad themes were identified by the two readers through open coding, the first author fine-tuned the analysis to identify key areas of overlap and difference between the two groups as a means of further understanding factors that lead to longer job tenure. Themes were identified under the four questions listed in previous section and both readers coded for the existence of these themes. Each reader also assessed how the study participants described their home care agency/company employer: positively, neutrally, or negatively. Percentage agreement was calculated after the open coding (ranging from 83% to 96% across the four questions and worker perception of agency categorization) to assess interrater reliability (Boyatzis, 1998).

Differences were easily negotiated through discussion between the two coders and full consensus was obtained. Chi-square analysis was performed to distinguish differences between short-term and long-term workers on coded themes. Both observer triangulation (multiple interviewers) and data triangulation (telephone interviews and narrative survey data) were used to increase the trustworthiness of the analysis (Padgett, 2008). Trustworthiness was further enhanced through peer debriefing when the first and third author met regularly to discuss emerging patterns in the data.

Findings

Sample Description

The sample of 261 PSS workers was compared to the population of workers receiving the survey (n = 496) on the variables of age, gender, and tenure at the agency; there were no statistically significant differences. The sample of workers ranged in age from 19 to 82, with an average age of 46. The vast majority was female. Reflecting the racial composition of Maine (97% White; U.S. Census Bureau, 2010), the sample was largely White (93.5%; see Table 1). Twenty-nine study participants identified as Franco-American, the largest ethnic group in the state. Although unusual for an occupation which attracts many immigrant women on a national level, this state sample only had nine individuals who had been born outside the United States; again, this reflects the population of the state which has a very small immigrant population (3.3%; U.S. Census Bureau, 2010). The study participants reported very low household incomes (see Table 1).

Table 1.

Demographic Characteristics of the Sample.

Variable	N	%
Gender
Female	247	94.6
Male	14	5.4
Age
< 20	3	1.1
20-30	39	14.9
31-40	52	19.9
41-50	64	24.5
51-60	64	24.5
> 60	39	14.9
Race
White	244	93.5
Native American	9	3.4
African American	2	0.8
Multiracial	6	2.3
Ethnicity
Franco-American	29	11.1
Latino/a	3	1.1
Born in the United States
Yes	252	96.6
No	9	9.0
Family status
Married/partnered	127	48.7
Divorced/separated	68	26.1
Widowed	15	5.7
Single	51	19.5
Education
< HS graduation	10	3.8
HS grad or GED	132	50.6
Some college	98	37.5
4-yr degree	13	5.0
Post college	8	3.1
Household income
< US$10,000	67	25.9
US$10,001-US$20,000	95	36.7
US$20,001-US$30,000	30	11.6
US$30,001-US$40,000	37	14.3
>US$40,001	30	11.6

Note: N = 261. HS = High school. GED = General educational development.

Over a quarter of the study participants lacked any type of health insurance (28.0%); a similar number (28.7%) received Medicaid and 8.4% received Medicare; and only 14 (5.4%) study participants received health insurance through their home care agency/company. The PSS workers had mean scores that were higher (indicating better health) than U.S. population norms on all but one of the eight subscales of the SF-36. These included General Health, Vitality, Role-Emotional, Role-Physical, Physical Functioning, Social Functioning, and Mental Health. Only on the Body Pain subscale did the average score fall below the average U.S. norms—meaning greater body pain (Ware, 1993). Looking specifically at national norms for women between 45 and 55 years of age—a comparable group to those in the HCWRS sample—study participants similarly had slightly higher scores (i.e., better health) on the same seven components of the SF-36, but not the body pain component, for which their scores were slightly lower (71.6 vs. 72.1) indicating more reported body pain than for women of similar age in the general population.

Employment and Job Experience Factors

When the study participants were first surveyed (Time 1), their wages ranged from US$7.50 to US$13.50 with an average wage of US$9.05; at the second survey (Time 2), the average wage was slightly higher (US$9.20). The number of hours worked in the preceding 4 weeks ranged from no hour to 504 hr, with an average of 18.0 hr per week at Time 1. At Time 2, the average number of hours worked per week was somewhat lower at 15.4. Study participants reported driving many unreimbursed miles in order to carry out their work in their clients’ homes: at Time 1, this averaged 46.1 miles (74.2 km) per week and at Time 2, 40.9 miles (65.8 km) per week. Forty-one study participants (15.7%) had injured themselves at least once while working for the employer through which they had received the study survey. At both Time 1 and Time 2, study participants, on average, scored well above the scale midpoint on job satisfaction, perceived empowerment, and personal accomplishment, and well below the scale midpoint on emotional exhaustion.

Predictors of Longer Job Tenure

To keep the regression model parsimonious, bivariate relationships were examined between covariates of interest (demographic factors at Time 1; employment and job experience factors at Time 1 and Time 2) and the dependent variable of job tenure (ranging from 2 months to 13 years; see Table 2.) Among the demographic variables, six were highly correlated to job tenure. Older workers had longer job tenure, as did widows, individuals without dependent children, and people living in more rural regions of the state. (For this study, six of Maine 16 counties were defined as rural; these “rim” counties are sparsely populated and situated far from the state’s main north/south highway.) Interestingly, individuals who had lower scores on the physical functioning scale of the SF-36 and those indicating more body pain also had longer job tenure. Two employment variables were also correlated to job tenure: longer tenure was correlated to having had more occupational injuries at Time 1 and higher hourly wages at Time 2. In addition, four job experience variables were correlated to job tenure. Somewhat surprisingly, study participants scoring lower on both the frequency and intensity of the Personal Accomplishment scale at Time 1 had longer job tenure, as did those scoring higher on the Autonomy subscale and the Extrinsic Job Satisfaction subscale at Time 2.

Table 2.

Correlation Matrix of Covariates and Length of Employment.

Demographics		Employment		Job experience	r
Variable	r	Variable	r	Variable	r
Age	.455^**	Wage Time 1	.086	Client behavior problems Time 1	.080
White	−.051	Wage Time 2	.133^*	Client behavior problems Time 2	.007
Franco-American	.116	Hours Time 1	−.081	Job satisfaction intrinsic Time 1	.109
Income	.035	Hours Time 2	−.032	Job satisfaction intrinsic Time 2	.084
Health insurance	.098	Unreimbursed miles Time 1	−.038	Job satisfaction extrinsic Time 1	.029
Education	−.020	Unreimbursed miles Time 2	.017	Job satisfaction extrinsic Time 2	.132^*
Dependent children	−.190^**	Injuries on job	.160^**	Emotional exhaustion frequency Time 1
Dependent adults	.008			Emotional exhaustion frequency Time 2	−.025
Widowed	.200^**			Emotional exhaustion intensity Time 1	−.021
Married	.078			Emotional exhaustion intensity Time 2	−.078
Rural	−.154^*			Personal accomplishment frequency Time 1	−.184^**
Physical function	−.249^**			Personal accomplishment frequency Time 2	.011
Body pain	−.156^*			Personal accomplishment frequency Time 1	−.152^*
General health	−.047			Personal accomplishment frequency Time 2	−.013
Mental health	.072			Empowerment Time 1	−.013
Vitality	−.013			Empowerment Time 2	.112
Role emotional	−.025			Autonomy Time 1	.042
Role physical	−.081			Autonomy Time 2	.156^*
Social function	.054			Participation Time 1	−.025
				Participation Time 2	.092
				Responsibility Time 1	−.069
				Responsibility Time 2	−.037

Note. N = 261.

p < .05. ^**p < .01.

These 12 variables were entered into a multivariate linear regression model with length of employment as the dependent variable. Six variables had predictive value as shown in Table 3. Among the demographic variables, older age, living rurally, and having lower physical function predicted longer job tenure. Higher wages, a greater sense of autonomy, and less frequent experiences of personal accomplishment were also predictors (see Table 3).

Table 3.

Predictors of Increased Job Tenure.

Variable	β	SE	t-value
Age	.341	4.84	5.50^**
Rurality	.174	126.87	3.08^**
Widowed	.069	239.54	1.25
Dependent children	−.066	125.69	−1.14
Number of injuries	.080	98.61	1.48
Physical function	−.140	3.75	−2.18^*
Body pain	−.040	2.99	−0.63
Personal accomplishment-frequency	−.149	68.70	−2.59^*
Personal accomplishment-intensity	−.014	67.84	−0.24
Hourly wage	.157	59.98	2.87^**
Autonomy	.157	77.97	2.81^**
Extrinsic job satisfaction	.066	75.72	1.18

Note: Dependent variable Job Tenure is measured as days from date of hire to either job termination or end of data collection (18 months after completing Time 1 survey).

N = 261.

R² = .363 F = 11.26 (12,237) p < .001.

p < .05. ^**p < .01.

The two subgroups examined in the next section—long-term stayers and short-term leavers—were compared on these six variables, and they differed significantly on four. The long-term stayers were older (p < .001), had lower Personal Accomplishment frequency scores (p < .01), had lower physical functioning (p < .001), and were more likely to live rurally (p < .05). They did not differ significantly on wages or scores on the Autonomy subscale.

Perceptions of Work for Long-Term Stayers and Short-Term Leavers

In the initial review of the 30 interview transcripts of long-term stayers, 28 themes were identified, which were grouped into three overarching constructs: rewarding work, challenges of the job, and the job flexibility works for me. The 39 themes emerging from the narrative data from the interviews with the 29 short-term leavers were similarly grouped under three constructs: rewarding work, job not worthwhile, and fed up with agency. There was considerable overlap in themes between the long-term stayers and short-term leavers including “the importance of helping people stay at home,” “have always been a nurturer,” “good relationships with clients,” “enjoy making a difference in people’s lives,” “like caring for elders,” and “ability to do the work despite disabilities.” Differences arose between the two groups around sustainability of the work. For long-term stayers, the part-time schedule generally worked and they often had other income in the household beyond their home care work employment. Short-term leavers spoke about the problems of low pay, lack of mileage reimbursement, inconsistent hours, and poor communication with their agencies. Other than the low pay, these issues had not emerged in the regression analysis.

To further explore similarities and differences between these two groups of workers, themes were identified under responses to four specific interview questions, and the first and third authors independently coded for their existence in each transcript. Table 4 presents these questions and the themes most frequently cited by study participants. We also coded each respondent as having either a positive, negative, or neutral overall impression of her or his employing agency (see Table 4).

Table 4.

Most Frequent Themes From Telephone Interviews.

	Long-term		Short-term		Total
Variable	N	%	N	%	N	%
What was initially appealing?
Fell into it	11	36.7	6	20.7	17	28.8
Like helping	8	26.7	7	24.1	15	25.4
Structure of job	4	13.3	10	34.5	14	23.7
Like elders	8	26.7	5	17.2	13	22.0
What is satisfying about work?
Fulfilling need	13	43.3	10	34.5	23	40.0
Bringing joy	5	16.7	9	31.0	14	23.7
Appreciation	5	16.7	7	24.1	12	20.3
What is draining about work?
Nothing	8	26.7	6	20.7	14	23.7
Physically hard	6	20.0	3	10.3	9	15.3
Difficult clients	2	6.7	5	17.2	7	11.9
Not able to meet needs	3	10.0	3	10.3	6	10.2
Long days	2	6.7	4	13.8	6	10.2
Why did/would you quit?
Not enough hours^**	1	3.3	13	44.0	14	23.7
Wouldn’t quit^**	12	40.0	0	0	12	20.3
Compensation^*	1	3.3	7	24.1	8	13.6
Perception of agency (one perception/worker)
Positive^**	20	66.7	6	20.7	26	44.1
Neutral	5	16.7	9	31.0	14	23.7
Negative^*	5	16.7	14	48.3	19	32.2

Note. N = 59.

p < .05. ^**p < .01.

Long-term stayers and short-term leavers did not differ significantly on their responses to the first three questions examined, but did on four responses to why they did quit or might quit their job. While 12 long-term stayers said they would never quit, no short-term leavers did (p < .01). Short-term leavers were more likely to say they had quit due to not enough hours (p < .01), poor wages (p < .05) or family needs (p < .05) than long-term stayers. Not surprisingly, long-term stayers were significantly more likely to have a positive view of the agency/company (p < .01), while short-term leavers were more likely to describe their employers in more negative (p < .05) or neutral ways (see Table 4).While higher pay was a predictor of longer job tenure in the regression analysis, the concern about insufficient hours, outside family needs, and conflicts with employers had not emerged in the quantitative analysis. Differences between long-term stayers and short-term leavers are further explored through quotes from the telephone interviews in the next section.

Long-Term Stayers

A 62-year-old worker, who had been with her agency 7 years, described the age discrimination she faced in looking for additional employment to supplement the wages from her home care job: “Well the older you get, you have it hard in certain places, and I don’t have a high school diploma, so I guess they are getting stricter.” But neither her age, nor her lack of a high school education prevented her from securing her Personal Support Specialist job, work she enjoyed very much and which she planned to continue as long as possible despite her advancing age. Several long-term stayers indicated that they were at a point in their lives that the part-time hours and flexible job structure worked in their lives. One woman who was 57 years old and had worked at her agency for 9 years remarked, “I’m semi-retired and my husband says I work to buy stuff for the grandkids.”

These long-term stayers generally described their employment in a very positive light, enumerating the rewards of the work. They also were realistic about the job and spoke about needing to take the bad with the good. For example, when asked about what she found draining about the work, one 11-year employee, age 48 reflected:

I guess, I mean, everything can be draining. It all depends on how you want to go about it. I mean I get drained, but not really in the sense of what people think that I’m drained. I may be tired, maybe just a little bit.

She could not really understand why people would leave the work she loved so much; like others, she reiterated that she wasn’t in the job to make money, but rather for the people: “If you enjoy making people happy, and letting them stay home, home health is the perfect job you can get. If they’re looking to make a million bucks, then don’t do it.”

Short-Term Leavers

In general, the short-term leavers appeared to be less able to sustain the low wages and inconsistent hours. As shown in Table 3, inconsistent hours and low compensation were primary reasons for these workers to leave their jobs. For example, one 23-year-old study participant who left her agency after only 5 months said:

I didn’t get the hours that I was told that I was going to get. When I first started they told me that they had all kinds of clients and, you know, it was gong to equal out to full-time and different stuff. And that wasn’t the case at all. Sometimes I only get like eight hours a week . . . I mean I liked working there, but I couldn’t afford to.

She left her home care aide position to take a full-time job working with people with brain injuries and developmental disabilities.

Another worker who was 26 years old and had prior experience in the health care field left her home care company after nine months because she wasn’t getting enough hours and she took issue with how the agency provided client care. She stated:

They just didn’t really communicate anything with us, or really have an open forum so that we could discuss things or have input in things. We were just kind of told that “This is changing. This is how it’s going to be. Change it now!”

Although this study participant loved working with elders, she decided to move back into work in mental health and developmental disabilities, a field in which she could get more full-time work.

A 42-year-old study participant, who quit working for her company after 8 months when she obtained another job that paid more, stated that she was burned out from being underpaid and undervalued:

It is hard work, underpaid, and you’re always doing the same things. It is repetitive, it is very repetitive. And you don’t get a lot of respect. I don’t feel like I did . . . [I don’t mean] from the client, but from the agency.

Discussion

As hypothesized, older age was a predictor of longer job tenure in this study. In some respects this is not remarkable as older age brings the potential to stay on a job for more years, and with age comes the higher likelihood of being widowed, not longer having dependent children, having decreased physical function, and greater body pain—all factors that were significantly correlated with longer job tenure at the bivariate level. Age and physical function, two predictors of longer job tenure in the regression model, were highly correlated in this data set (p < .001), and thus as the job appears to be more appealing to older workers than younger workers (who may need more consistent hours and higher wages), it is perhaps not surprising that those remaining on the job are less physically strong. Despite the physical demands of the job, home care work is flexible enough for workers to refuse work that might be beyond their capacity, something that was supported in both the quantitative and qualitative analyses. Unlike the more rigid structure of nurse’s aide work in nursing homes, home care workers have more control over their hours, their clients and the nature of their tasks, making it a job that people can continue to do into their 60s and even 70s, even as their physical capacity diminishes. Faul and colleagues (2010) also found older age to be a predictor of longer job tenure among home care workers in Kentucky. Furthermore, for some study participants, there was a welcomed sense of occupational autonomy that came from working alone with little supervision; those study participants scoring higher on questions measuring autonomy (i.e., “I am given a chance to do the things I do best”) had longer job tenures. This finding likely differs from factors predicting retention for direct care workers in nursing facilities (e.g., Dill, Craft Morgan, & Konrad, 2010; Donoghue, 2010) where nursing aides generally have far less autonomy.

On the other hand, younger workers in the study and workers who more frequently felt personal accomplishment in their work did not stay as long with their home care employers. These workers may have felt very effective in their work as measured by Personal Accomplishment scale questions such as “I have accomplished many worthwhile things in the job,” but, the lack of consistent income appears to have been more difficult to sustain for workers in their 20s, 30s, and 40s. Indeed, though there was not a wide range in the wages earned by study participants, higher wages did predict longer job tenure, further indicating the importance of adequate compensation. It is also possible that some individuals frequently felt “exhilarated after working closely with clients” (another item on the Personal Accomplishment scale) in part because they were new to their jobs. Over time such workers may feel this personal accomplishment somewhat less frequently as they become long-term stayers, or they may burn out and leave. In fact, scores on the Personal Accomplishment frequency scale decreased significantly from Time 1 to Time 2 for both study participants who stayed in their jobs throughout the data collection period (p < .05) and those who terminated (p < .01). Moreover, in comparing long-term stayers and short-term leavers on this variable, it is noteworthy that the scores for the stayers do not drop significantly from Time 1 to Time 2 (5.7 to 5.6), while the scores for the short-term leavers do drop significantly (6.4 to 5.7 [p < .01]). Those that stay on the job appear to have slightly lower, but steadier, feelings of personal accomplishment. The quote cited earlier from a long-term stayer who stated that any job can be draining but that “it all depends on how you want to go about it,” supports the idea that while scores may decrease over time on some items of the Personal Accomplishment scale, those who stay on the job may know how to pace themselves so as not to experience burnout.

Living in the more rural regions of the state was the final predictor of longer tenure. While all of Maine is considered nonmetropolitan by national standards, there is no question that some parts of the state are far more rural than others. These more isolated areas have experienced particularly high levels of unemployment. In such a context, with fewer job opportunities to attract them away, it is not surprising that home care workers in rural areas would be more likely to stay in their home care jobs.

The narratives from the telephone interviews further illuminated differences between short-term and long-term workers. Interestingly, the two groups did not significantly differ on what motivated them to work in the field, what they found satisfying about the job, nor what they found draining. Nonetheless, the lack of hours and poor compensation led many to quit in their 1st year of employment. These short-termers also tended to have more difficult relationships with their employing agency/company than their counterparts who stayed on the job, citing issues such as broken promises, poor communication, and a sense of not being adequately valued. The qualitative analysis provided a context for understanding the predictors of longer job tenure in the quantitative analysis and surfaced several factors that had not been captured in the regression model. In particular, insufficient hours and poor communication with employing agencies emerged in the qualitative analysis as problematic for those who left their jobs within the 1st year of employment, providing a more complete picture regarding the issue of job retention in home care.

Although predictors of turnover and job tenure among home care workers have not been examined as closely as those for direct care workers in nursing homes, a few recent studies have occurred to which these findings can be compared. The factors of older age and higher wages as predictors of job tenure were also found by Faul and colleagues (2010). Morris (2009) also found higher wages to predict retention in her study of Maine home care workers. McCaughey and colleagues (2012) found job injuries to predict intent to leave in their study of home health aides, something not found in the HCWRS described here. In fact, at the bivariate level, longer job tenure was related to having had more job injuries in the Maine HCWRS. Higher education and greater intrinsic job satisfaction were found to be predictors of retention by Faul and colleagues (2010), neither of which were significant in our model. Morris (2009) found more hours and higher mileage reimbursement predicted retention, neither of which were significant in our regression model, but which were spoken about frequently in the telephone interviews. Short-term leavers, in particular, expressed that lack of hours and mileage reimbursement contributed to their decisions to terminate. Morris also found that lack of health benefits predicted turnover, something we also found in analyses of predictors of termination in the HCWRS (Butler et al., 2010; Butler et al., 2012), though it was not a predictor of longer job tenure described here.

Implications

Generalizations from this study must be made with caution due to a number of limitations. First, neither the home care agencies nor workers in this study were randomly selected, but rather volunteered their participation, so generalization to all home care aides in the state must be done with caution. Nonetheless, workers who chose to join the study were compared to other workers in the participating agencies who did not complete the survey on the variables of sex, age, and job tenure, and no statistically significant differences were found. Second, the study took place in a state that is very rural, largely White, and with a particularly small immigrant population. The national direct care worker population is composed of many more immigrants (23%) and people of color (53%; PHI, 2011). Undoubtedly, concerns and experiences of workers in more urban environments and who work with clients from different racial and cultural backgrounds might differ from those expressed by the home care aides in this study. Another study limitation was the difficulty in measuring average net income for workers whose hours and mileage varied considerably week to week. A further limitation was the exclusion of broader economic factors, such as unemployment rates and availability of other job opportunities, that certainly impact workers’ decisions to stay or leave their employment. Wiener, Squillace, Anderson, and Khatutsky (2009), for example, found higher county unemployment rates predicted longer tenure among certified nursing assistants (CNAs) in nursing homes. Such an analysis would clarify whether rurality predicted longer job tenure due to limited alternative job options or if other factors are involved (e.g., nature of employing agencies). The study also had a number of strengths including its prospective design, a fairly large sample (especially given the small population in the state of Maine), its multivariate model, and the inclusion of both quantitative and qualitative data. Thus some implications are worth noting based on the findings presented.

As the U.S. population ages, not only is the need for personal care at home increasing, but the number of older workers is also growing substantially. Direct care workers tend to be older on average than the overall workforce (PHI, 2010a) and it is a job that may in fact appeal to workers in their later years due to its flexibility and potential for meaningful human connections (Kleyman, 2008). Many of the older study participants noted that the work kept them active and involved with people. As stated by a 67-year-old study participant, “I do it to get out and I enjoy working with people.” At the same time she was very critical regarding compensation and the lack of recognition by her employer. She said she would not do the work if she were younger: “If I was younger I wouldn’t be doing this job. I would go where there’s benefits because we have no benefits. We haven’t had a pay increase in five or six years.” As home care positions may be increasingly filled by older women, this could be a win-win situation given that both home care employers and clients tend to appreciate the reliability and experience of older workers (Rix, 2001). Nonetheless, it could also be fertile ground for exploitation unless wages are increased (Butler, 2013). This study reinforces the findings of Hwalek, Straub, and Kosniewski (2008) that older workers—particularly those who are low-income—are a promising pool of direct care workers and should be tapped in efforts to fill the care gap in long-term care. The older workers in their study (over 1,000 low-income participants in Operation ABLE employment and training organizations) indicated a high level of interest in direct care work (45%), particularly in home health, though some were unable to pay for the necessary training. Employers in Hwalek et al.’s study generally had very positive perceptions of older workers, though reported that they were less willing to learn new technologies. As discussed by Hwalek and colleagues, Title V funds from the Older Americans Act could be targeted toward specialized training for older adults to prepare them for the field of direct care work.

Both quantitative and qualitative data from this study indicate that it is neither job satisfaction nor burnout that determines how long a home care aide will stay on the job; generally, workers had high levels of job satisfaction and low levels of burnout whether they stayed or left, and those with shorter tenure even reported higher levels of personal accomplishment on the job. Ultimately, the qualitative data in particular allow us to see that low wages, lack of benefits, and inconsistent hours made this job “not worthwhile” and eventually caused many workers—predominantly the younger ones—to leave. Clearly, the complementary qualitative and quantitative findings from this study reinforce what others have found: the economic viability of this work is hard to maintain unless the worker has other income in the household. The low wages, unreimbursed mileage, the lack of benefits, and the inconsistent hours that characterize home care work ultimately lead many workers to terminate employment.

Previous quantitative studies investigating predictors of turnover and retention among direct care workers have uncovered different aspects of compensation as being most significant. For example, higher wages were significant in predicting job retention among home care aides by Faul and colleagues (2010), while Morris (2009) not only found wages to be significant in reducing turnover among home care workers, but also found increased hours and mileage reimbursement to be additional significant factors. In examining job tenure among certified nursing assistants (CNAs) in nursing homes, Wiener and colleagues (2009) found higher wages, paid time off, and access to pensions to be significant determinants of longer job tenure, while access to health insurance was not. Interestingly, absence of health insurance was significant in a later study of turnover among CNAs in nursing homes by Rosen, Stiehl, Mittal, and Leana (2011), but wages were not. Moreover, analysis of turnover among the home care workers in the HCWRS revealed absence of health insurance as a key predictor (Butler et al., 2010; Butler et al., 2012), but in the regression analysis of predictors of longer job tenure presented here, neither health insurance nor number of hours were significant, but higher wages were. The qualitative analysis of interviews with both those workers remaining on the job for many years as well as those who left after only a few months underscores that it is a combination of these compensation issues that make the job difficult to sustain, particularly for workers without other income in their households and who cannot survive on part-time hours, at low wages, and who do not have access to health insurance through a spouse or government program.

Unless home care is paid for privately, personal care services are reimbursed through Medicaid, and in Maine, the state’s General Fund. Reimbursement rates are low and must cover administrative and capital costs, as well as the wage to the workers (Pohlmann, 2003). So increasing compensation requires increased public funding, something that is difficult in these tight budgetary times. This situation has resulted in stagnant wages for home care workers in Maine and across the nation (PHI, 2008) at a time when inflation has affected gas prices and general living costs. Nonetheless, to promote retention, particularly among workers who depend on the job to support themselves and their families, compensation for personal care work must be increased.

Recently there has been considerable attention focused on compensation for home care aides. In 1938, the Fair Labor Standards Act provided wage and overtime protections to workers, although the jobs women traditionally filled were excluded. A quarter of a century later, in 1974, Congress expanded the Fair Labor Standards Act to provide protection to domestic workers (such as cooks and maids), but exempted individuals who provided companionship services. At the time, the Department of Labor, charged with writing regulations to implement the new law, interpreted the “companionship exemption” quite broadly. The intention of Congress had been to exempt the babysitter or neighbor who provided occasional support, but the Labor Department chose to also include professional home care aides, employed by third party agencies, under the exemption umbrella (Martin, 2009).

On December 15, 2011, Hilda Solis, former Secretary of the Department of Labor, announced a new rule eliminating the companionship exemption; as of this writing the rule has not yet been finalized.

The Patient Protection and Affordable Care Act (ACA), which was passed in 2010, also has a number of provisions that could well improve the work situation for home care workers. It raises the eligibility level for Medicaid to 133% of the poverty line in 2014, an income level under which many home care aides live, thereby assisting the significant proportion of the workforce currently without health insurance. Based on the Supreme Court ruling in August 2012, states may opt out of expanding Medicaid; each state that makes that choice will miss an opportunity to provide health care insurance to these key workers in our long-term care system. The ACA also makes providing health insurance benefits more affordable for small companies, such as those employing home care workers in this study (Center for Children and Families, 2010). Furthermore, the Act established a Personal Care Attendants Workforce Advisory Panel which has been given the mandate to examine issues such as wages and benefits of personal care workers and to advise the Secretary of Health and Human Services and Congress (PHI, 2010b).Hopefully the recommendations put forward by this panel will carry weight for policy makers and the final rule to eliminate the companionship exemption will be submitted so that more home care workers will be able to afford to continue to do the work they love.

Until such changes come about, the findings of this study suggest home care agencies may be confronting a two-part workforce. On one hand, the current structure and compensation levels of the home care job appear to work well for some older workers who may be happy with the part-time hours and autonomous nature of the work, and may be less dependent on the wages or less in need of health insurance benefits. On the other hand, younger workers may desire more supervision and support from their employers and may require more dependable hours, higher wages, and health insurance in order to support themselves and their families. Until such time that wages can be raised and more hours guaranteed, an interim goal could be for home care employers and job developers to design this job to be as attractive and accessible as possible for older workers and to emphasize its opportunities for younger workers to be a stepping-stone to other health care work with more consistent hours, higher pay, and employment benefits such as paid sick leave and health care insurance.

Footnotes

Acknowledgements

The authors wish to acknowledge Drs. Nancy Fishwick, Mary Ball and Winston Turner for their assistance on this project and the following students, who worked as research assistants: Nan Simpson, Julia Hewes, Meghan Wyman, Dawn Ravan, Diane Helder, Stephanie Allard, and Meredith Wurpel.

Author’s Note

IRB of University of Maine approval number: 2208-03-09

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by a National Institute on Aging Academic Research Enhancement Award: 1 R15 AG029599-01A2.

Author Biographies

Sandra S. Butler, PhD, is a professor and MSW coordinator at the School of Social Work, University of Maine.

Mark Brennan-Ing, PhD, is a senior research scientist at AIDS Community Research Initiative of America in New York City.

Sara Wardamasky, LSW, MSW, recently completed her masters degree in social work, prior to which she worked for over 10 years in the field of aging.

Alison Ashley, LMSW-CC, is a social worker with the Community Care Team of Eastern Maine HomeCare.

References

Ashely

Butler

S. S.

Fishwick

(2010). Home-care aides’ voices from the field: Job experiences of Personal Support Specalists—The Maine Home Care Worker Retention Study. Home Healthcare Nurse, 28, 399-405.

Barry

T. T.

Brannon

Moor

(2005). Nurse aide empowerment strategies and staff stability: Effects on nursing home resident outcomes. The Gerontologist, 45, 309-317.

Boyatzis

R. E.

(1998). Transforming qualitative information : Thematic analysis and code development. Thousand Oaks, CA: SAGE.

Bureau of Labor Statistics. (2012). Employment projections: Fastest growing occupations. Retrieved from www.bls.gov/emp/ep_table_103.htm

Butler

S. S.

(2013). Older women doing home care: Exploitation or ideal job? Journal of Gerontological Social Work, 54, 299-417.

Butler

S. S.

Simpson

Brennan

Turner

(2010). Why do they leave? Factors associated with job termination among personal assistance workers in home care. Journal of Gerontological Social Work, 57, 665-681.

Butler

S. S.

Wardamasky

Brennan-Ing

(2012). Older women taking care of older women: The rewards and challenges of the home care aide job. Journal of Women & Aging, 24(3), 1-22.

Center for Children and Families. (2010). Summary of Medicaid, CHIP, and low-income provisions in health care reform. Washington, DC: Georgetown University Health Policy Institute. Retrieved from http://phitnational.org/policy

Dill

J. S.

Craft Morgan

Konrad

T. R.

(2010). Strengthening the long-term care workforce. Journal of Applied Gerontology, 29(2), 196-214.

10.

Donoghue

(2010). Nursing home staff turnover and retention. Journal of Applied Gerontology, 29, 720-739.

11.

Faul

A.C.

Schapmire

T. J.

D’Ambrosio

Feaster

Oak

C. S.

Farley

(2010). Promoting sustainability in frontline home care aides: Understanding factors affecting job retention in the home care workforce. Home Healthcare Management & Practice, 22, 408-16.

12.

Feldman

P. H.

(1990). Who cares for them? Workers in the home care industry. New York, NY: Greenwood.

13.

Gellis

Z. D.

Kim

J. C.

(2004). Predictors of depressive mood, occupational stress, and propensity to leave in older and younger mental health case managers. Community Mental Health Journal, 40, 407-421.

14.

Grau

Chandler

Burton

Kilditz

(1991). Institutional loyalty and job satisfaction among nurse aides in nursing homes. Journal of Aging and Health, 3(1), 47-65.

15.

Hwalek

Straub

Kosniewski

(2008). Older workers: An opportunity to expand the long-term care/direct care labor force. The Gerontologist, 48(Special Issue), 90-103.

16.

Institute of Medicine. (2008). Retooling for an aging America: Building the health care workforce. Washington, DC: The National Academic Press.

17.

Johnson

R. W.

Wiener

J. M.

(2006). A profile of frail older Americans and their caregivers. Washington, DC: The Urban Institute. Retrieved from http://www.urban.org/UploadedPDF/311284_older_americans.pdf

18.

Kaye

H. S.

Chapman

Newcomer

R. J.

Harrington

(2006). The personal assistance workforce: Trends in supply and demand. Health Affairs, 25, 1113-1120.

19.

Kiefer

K. M.

Harris-Kojetin

Brannon

Barry

Vasey

Lepore

(2005). Measuring long-term care work: A guide to selected instruments to examine direct care worker experiences and outcomes. Washington, DC: U.S. Department of Health and Human Services.

20.

Kleyman

(2004). Boomers to redefine work place. Aging Today, 25(6), 7-10.

21.

Lambert

E. G.

Hogan

N. L.

Barton

S. M.

(2001). The impact of job satisfaction on turnover intent: A test of a structural measurement model using a national sample of workers. Social Science Journal, 38, 233-250.

22.

Martin

(2009, August 10). Evelyn Coke, home care aide who fought pay rule, is dead at 74. New York Times. Retrieved from http://www.nytimes.com

23.

Maslach

Jackson

S. E.

(1981). The measurement of experienced burnout. Journal of Occupational Behaviour, 2, 99-113.

24.

McCaughey

McGhan

Kim

Brannon

Leroy

Jablonski

(2012). Workforce implications of injury among home health care workers: Evidence from the National Home Health Aide Survey. The Gerontologist, 52, 493-505.

25.

Mittal

Rosen

Leanna

(2009). A dual-driver model of retention and turnover in the direct care workforce. The Gerontologist, 49, 623-634.

26.

Morris

(2009). Quits and job changes among home care workers in Maine: The role of wages, hours and benefits. The Gerontologist, 49, 635-650.

27.

Padgett

D. K.

(2008). Qualitative methods in social work research (2nd ed.). Los Angeles, CA: SAGE.

28.

Paraprofessional Healthcare Institute. (PHI, 2008). State chart book on wages for personal and home care aides, 1999-2006. Retrieved from http://www.directcareclearinghouse.org/download/PHI_State_Chartbook_PHCA_Wages_99-06.pdf

29.

Paraprofessional Healthcare Institute. (PHI, 2010a). Older direct-care workers: Key facts and trends. Retrieved from http://www.directcareclearinghouse.org/download/PHIOlderDCWAnalysisApril2010.pdf

30.

Paraprofessional Healthcare Institute. (PHI, 2010b). Direct care workforce and long-term care provisions as enacted in Patient Protection and Affordability Care Act and Health Care and Education Reconciliation Act of 2010. Retrieved from http://phinational.org/policy/wp-content/uplaods/DCW-LTC-Provisions-in-Health-Reform.pdf

31.

Paraprofessional Healthcare Institute. (PHI, 2011). Who are direct-care workers? Retrieved from http://www.directcareclearinghouse.org/download/PHIFacts3.pdf

32.

Pohlmann

(2003). Without care: Maine’s direct care worker shortage. Augusta: Maine Center for Economic Policy.

33.

Rix

(2001). The role of older workers in caring for older people in the future. Generations, 25(1), 29-34.

34.

Rosen

Stiehl

E. M.

Mittal

Leana

C. R.

(2011). Stayers, leavers, and switchers among certified nursing assistants in nursing homes: A longitudinal investigation of turnover intent, staff retention, and turnover. The Gerontologist, 51, 597-609.

35.

Scala

(2008). Home and community-based services: Workforce and quality outcomes. In Scala

Hendrickson

Regan

(Eds.), A compendium of three discussion papers: Strategies for promoting and improving the direct service workforce: Application to home and community-based services (pp. 3-45). New Brunswick, NJ: Rutgers Center for State Health Policy. Retrieved from http://www.cshp.rutgers.edu/Downloads/7740.doc

36.

Seavey

Marquand

(2011). Caring in America: A comprehensive analysis of the nation’s fastest growing jobs: Home health and personal care aides. New York, NY: Paraprofessional Healthcare Institute. Retrieved from www.directcareclearinghouse.org/download/caringinamerica-20111212.pdf

37.

Stacey

C. L.

(2011). The caring self: The work experiences of home care aides. Ithaca, NY: Cornell University Press.

38.

Stone

R. I.

(2011). Long-term care for the elderly. Washington, DC: The Urban Institute Press.

39.

Strauss

Corbin

(1998). Basics of qualitative research: Grounded theory procedures and techniques. Newbury Park, CA: SAGE.

40.

U.S. Census Bureau. (2010). American Community Survey 2006-2008 3-year estimates. Retrieved from http://factfinder.census.gov

41.

Ware

J. E.

(1993). SF-36 Health Survey: Manual and interpretation guide. Boston, MA: The Health Institute, New England Medical Center.

42.

Wiener

J. M.

Squillace

M. R.

Anderson

W. L.

Khatutsky

(2009). Why do they stay? Job tenure among certified nursing assistants in nursing homes. The Gerontologist, 49, 198-210.