Examining Direct Service Worker Turnover in Ohio

Transcription

1 Examining Direct Service Worker Turnover in Ohio Farida K. Ejaz, Ph.D., LISW-Benjamin Rose Institute on Aging Ashley M. Bukach, B.S.-Benjamin Rose Institute on Aging Nicole Dawson, P.T., M.A., Graduate Student-Benjamin Rose Institute on Aging, Cleveland State University, The University of Akron Robert Gitter, Ph.D.-Ohio Wesleyan University Katherine S. Judge, Ph.D. -Cleveland State University

2 Acknowledgements This work is made possible through a subgrant from the Ohio Office of Medical Assistance (OMA) to the Ohio Colleges of Medicine Government Resource Center (GRC) as part of the Ohio Direct Service Workforce (ODSW) Initiative. Funding to support the ODSW Initiative, including this research project and report, comes from the federal Money Follows the Person (MFP) Demonstration Grant (CFDA ) to OMA from the U.S. Department of Health and Human Services, Centers for Medicare and Medicaid Services (CMS). However, the content of this report does not necessarily represent policy of OMA or the U.S. Department of Health and Human Services, and endorsement by OMA and the Federal Government should not be assumed. The project team would like to thank their Program Officers at GRC: 1.Barry Jamieson 2.Dushka Crane-Ross We would also like to thank the following Review Team members assigned to our project by GRC: 1. Teresa Applegarth, Ohio Department of Job and Family Services 2. Elizabeth Bragg, University of Cincinnati 3. Tiffany Dixon, Ohio Department of Aging 4. Kim Donica, Office of Medical Assistance 5. Lon Herman, Northeast Ohio Medical University 6. Katherine Kuck, Ohio Department of Health 7. John McAlearney, Wright State University 8. Troy McCollister, Ohio Department of Developmental Disabilities 9. Marc Molea, Ohio Department of Aging 10. Leslie J. Sawyer, Office of Medical Assistance 11. Michael Vallee, Ohio Valley Home Health Special thanks to Banita Giri, student at Ohio Wesleyan University for her help with Dr. Gitter s project.

3 Table of Contents Introduction... 1 Research Methods...1 Operational Definitions of DSWs...1 Sample...2 Provider Questionnaire...2 Procedures...3 Data Management...3 Data Verification and Cleaning...4 Data Analyses...4 Findings...4 Response Rates...4 Providers Participating in the Survey...4 Respondent Characteristics...5 Background Characteristics of Participating Providers...5 Number of Clients Served...5 Client Characteristics...6 DSW Characteristics...6 Hourly Wages...8 Benefits Offered by Provider Type...8 Best Practices by Provider Type...10 DSW Turnover for All Providers in the Study...11 DSW Turnover for Providers that Reported DSW Turnover in Impact of DSW Turnover...12 Predictors of DSW Turnover for All Providers in the Sample...13 Predictors of Turnover for Providers Who Reported DSW Turnover in Practice and Policy Recommendations...13 Improving Wages, Particularly Maximum Wages a DSW can Receive...13 Focus on Improving DSW Retention in Certain Geographic Regions of the State...14 Create a Repository of Best Practices and Evidence-Based Programs that have Demonstrated a Reduction in DSW Turnover...14 Test the Implementation of Selected Industry Best Practices and Evidence-Based Programs to Improve DSW Retention...14 Reward Providers with a Higher Reimbursement Rate for Low DSW Turnover...15 Next Steps...15 Limitations of the Current Project...15 Strengths of the Current Project...15 Future Research...16 References Appendix A: Organizational Survey-Nursing Home

4 Table of Contents Tables Table 1. Providers Participating in the Survey by Type...5 Table 2. Number of Clients and DSWs by Provider Type...5 Table 3. Percent of Providers Reporting Services Performed by their DSWs...7 Figures Figure 1. Minimum and Maximum Median Hourly Rates by Provider Type... 8 Figure 2. Benefits Offered to DSWs...9 Figure 3. Fully and Partly Paid Health Insurance Offered to DSWs...10 Figure 4. Best Practices Targeted at DSWs: Most Widely Utilized...10 Figure 5. Best Practices Targeted at DSWs: Least or Inconsistently Utilized...11 Figure 6. Overall Average DSW Turnover Rate by Provider and Region Type...12

5 Introduction Aside from family caregivers, direct service workers (DSWs) provide the majority of hands-on care to older adults as well as populations with physical and mental disabilities. Annual turnover among direct service workers (DSWs) in the nursing home industry is estimated at 71% (Seavey, 2004) and costs nearly $4.1 billion annually (Harris-Kojetin, Lipson, Fielding, Kiefer, & Stone, 2004). High DSW turnover rates also are prevalent among providers of home health, mental health, and developmental disabilities services at 25% (Seavey, 2004), 30% (Gitter, 2005), and 50% (U.S. Department of Health and Human Services, 2006) respectively. Turnover often results in service delivery failure, disrupts relationships with clients, and compromises quality of care (Seavey, ). Given the extent of turnover, gaining a better understanding of the reasons for DSW turnover is critical to address the needs of growing elderly and disabled populations. Another gap in the literature relates to the paucity of studies focusing on DSW turnover across different types of human service industries (Castle, 2006; Seavey, ). This project is the first of its kind to examine DSW turnover across four different types of human service industries: 1) nursing homes (NHs); 2) home health (HH) agencies; 3) mental health (MH) providers; and 4) providers of services to those with developmental disabilities (DD). Another unique feature of the project is that it reflects a collaboration of two separately funded proposals that focus on DSW turnover: one that focuses on turnover from the perspective of providers and the other that focuses on turnover from the perspective of DSWs. This collaboration includes the following three components: 1) Providers complete an initial organizational survey (Time 1). The design and implementation of this component is the responsibility of Dr. Ejaz from the Margaret Blenkner Research Institute of the Benjamin Rose Institute on Aging (MBRI) and Dr. Judge from Cleveland State University (CSU); 2) Participating providers at Time 1 distribute questionnaires to DSWs employed at their facilities (Time 2); and, 3) Participating providers complete a brief followup organizational questionnaire regarding the employment status of participating DSWs (Time 3). Components 2 and 3 are the responsibility of Dr. Gitter at Ohio Wesleyan University. In the spirit of the collaboration, Drs. Ejaz, Judge and Gitter have worked closely to ensure consistency of procedures, protocols and items in their respective organizational and DSW surveys to ensure long-term comparability of results across their respective projects. Dr. Gitter will submit a separate report focused on his components. The remainder of this report is focused on the first component of the project (Time 1). The specific aims of the first component (Time 1) of the project are to: 1) Examine DSW turnover across four human service industries in Ohio: nursing homes (NH), home health (HH), mental health (MH) and developmental disabilities (DD) providers in Ohio; 2) Examine whether DSW turnover rates differ across the five major geographical regions of Ohio (Appalachia, Metropolitan, Rural, Small City and Suburban); and 3) Examine statistically significant predictors of DSW turnover. Research Methods Operational Definitions of DSWs Our survey was focused on DSWs who are at the frontline of providing care to older adults and those with disabilities. In NH, HH and DD providers, a DSW was defined as a staff person who provides clients/residents* with hands-on care such as bathing, dressing, grooming and transferring (activities of daily living). Additional examples of support DSWs provided include: assisting with meal preparation, housekeeping, and budgeting and ensuring health and safety. In MH service providers, the employee most comparable to a DSW in Ohio is the community psychiatric supportive treatment (CPST) worker. Our survey included the following description of the CPST *From this point forward, the term client will be used to describe those served by DSWs in all provider types. 1

6 worker: the CPST worker provides services which are focused on an individual s ability to succeed in the community; to identify and access needed services; and to show improvement in school, work and family and integration and contributions within the community. Examples of support CPST workers may provide include, but are not limited to, care coordination, advocacy and outreach, facilitation of further development of daily living skills, and mental health interventions that address symptoms, behaviors or thought processes that assist an individual in eliminating barriers to seeking or maintaining education and employment. Sample Site Selection A sampling frame of all provider types was developed. The NH and HH provider lists were downloaded from the website; it includes only providers that are certified by Medicare and Medicaid, so it is likely that uncertified independent providers are excluded. The DD list was obtained from the Ohio Department of Developmental Disabilities (DODD). To be consistent with the NH and HH list, independent providers were excluded from the latter list. The list of MH providers was obtained from the Ohio Department of Mental Health (ODMH) and contained only providers certified by ODMH. Thus, all four provider types had certification from their respective governing bodies nationally and/or in Ohio. The four provider sampling lists were divided by the five geographic regions described earlier. Stratified random sampling procedures were used to draw a representative sample from each provider type and geographic region. Sample Size The initial goal was to collect data from approximately 100 providers across the four human service industries (i.e., 25 of each industry type, or 5 from each of the five regions). The goal was further expanded to reach an ambitious 160 providers (40 of each of the four provider types), if time permitted, during the limited data collection phase. Based on previous studies of providers, we anticipated a 33% response rate (Castle, Engberg, Anderson, & Men, 2007; Dill & Cage, 2010; Ejaz, Rose, & Bukach, 2011). Thus, we randomly selected three times the number of providers needed for a total of up to 120 providers per provider type (24 providers per region type). An exception was made in the MH sample as two geographic regions had fewer than 24 providers; in these regions all providers were included, resulting in a total of 467 providers across the four provider types. During the data collection period, 32 providers were determined to be ineligible for participation. They were excluded from the project and replaced with a provider randomly selected from the same provider type and region. The most common reasons for ineligibility included that the provider did not employ DSWs, the provider was new, or had not employed DSWs for a full year in order for us to calculate annual turnover rate. Respondent Selection Targeted respondents included management staff such as the Administrators/CEO and the Director of Human Resources. The survey was limited to one respondent per provider; however, more than one administrative staff member could assist with completion of the survey. Provider Questionnaire The questionnaire was developed based on a review of the existing literature regarding predictors of DSW turnover and job satisfaction (Ejaz, Noelker, Menne, & Bagaka s, 2008; Noelker, Ejaz, Menne, & Bagaka s, 2009; Menne, Ejaz, Noelker, & Jones, 2007). The project s Review Team served as experts to validate the initial drafts of the questionnaire. The Review Team was selected by the funders of this project and was comprised of key stakeholders representing the four human service industries in our study. When the initial draft of the questionnaire was finalized, research staff pre-tested it with a total of six providers (2 NH, 1 HH, 2 MH, and 1 DD provider). Following the pre-test and further discussions with the Review Team, refinements were made to the questionnaire. Two parallel provider questionnaires 2

7 were developed: 1) NH, HH and DD providers; and 2) MH providers. Developing a separate questionnaire for MH was considered necessary because DSWs were defined differently in MH (as CPST workers) compared to NH, HH, and DD providers. Based on the pre-test, a section on the cost of DSW turnover was dropped because of the length of the questionnaire. The final provider questionnaires included the following: a) Description of project, participation, and contact information; b) Information about person/s completing the survey, such as job title; c) Provider background information such as type of organization, auspice and years in operation; d) Clients/residents served in 2011, including demographic characteristics, and reimbursement sources; e) DSWs in 2011 including number on staff on January 1, 2011, demographic characteristics, types of services provided, pay and benefits offered; f) DSW turnover between January 1, 2011 and December 31, 2011; and, g) Impact of turnover on customer service and quality of care. Paper and electronic versions of the questionnaire were developed and providers were offered the choice of completing the survey by mail, fax, telephone (interview with a research assistant), or online using SurveyMonkey. (See Appendix A for the nursing home version of the questionnaire) Procedures Approval was obtained from the Institutional Review Board (IRB) at all three collaborating institutions (MBRI, CSU, and OWU) resulting in serious time delays. Initially, an introductory letter was mailed to the administrator or CEO/president of the selected provider. This letter explained the project, its purpose, its components, information on informed consent, the voluntary nature of participation in all components of the project, confidentiality of the data, and contact information for the project staff. The letter also informed participants that they could choose the mode for completing the provider questionnaire and included a link to the online version of the questionnaire. A printed copy of the questionnaire was included in the initial mailing. Approximately five to six weeks following the initial mailing, follow-up contact was made via telephone or (depending on available contact information) to those providers that had not completed the questionnaire. A research staff member re-introduced the project and inquired about the participant s preferred mode of completion. If a provider indicated that the organization did not wish to participate in the project, no further contact was made and the provider was removed from the participant list. Further contact was made with other providers on a periodic basis via and telephone. A postcard also was mailed to nonrespondents to encourage participation. In addition to the follow-up contacts to providers initiated by research staff, the project was advertised in a total of seven trade association newsletters, covering the four provider types, to encourage participation. Data Management An Access database was created in Access to track field operations including documenting the date of initial contact with each site, number of subsequent contacts, mode of contact (mail, , or telephone), completion/response rates, and reasons for refusal/non-participation. Tables were developed to generate the necessary information needed to guide field operations during the project period. Research staff met bi-weekly to troubleshoot and discuss any ongoing problems or issues that arose during data collection. Data from completed questionnaires received in the mail or by fax were entered manually into SurveyMonkey by research staff to facilitate aggregation of data across survey modalities. Data also periodically were downloaded from the online version in SurveyMonkey and exported to SPSS for analysis. 3

8 Data Verification and Cleaning Throughout the data collection phase, research staff routinely followed-up with providers if they left important questions blank or provided contradictory answers. Such discrepancies included the following issues: if the total number of DSWs on staff was less than the number that turned over during the year, or the total number of DSWs on staff did not match the number of staff who were listed as being full-time, part-time or other. Following initial analyses of the data, researchers conducted further data cleaning to verify answers to important questions that were missing information (e.g., turnover in 2011). This involved contacting 58 providers over a period of six weeks; resulting in a total of 133 additional contacts with providers during this data cleaning timeframe. Data Analyses Following data cleaning, descriptive statistics were run for all variables. Next, one-way analysis of variance (ANOVA) was conducted to examine statistically significant differences among provider and region types on interval-level measures, while chi-squares were run for categorical variables. Factor analysis also wasconducted for variables that were related such as the various types of benefits provided to DSWs, services provided by DSWs and best practices used by providers to help motivate and retain staff. In cases where the factor structures were clean and conceptually relevant, and had good reliability scores, scales were created. In the next phase, bivariate correlations were run to examine relationships between independent variables. In instances where high correlations between independent variables existed, one of the highly correlated variables was dropped in preparation for the multivariate analyses. A percentage rate of turnover also was calculated by examining the number of DSWs that turned over between January 1 and December 31, 2011, divided by the total number of DSWs on January 1, 2011 for each provider. This variable was treated as the outcome variable. Important factors correlated with the dependent variable and related to turnover from the literature (such as DSW wage rate and benefits offered), were entered as independent variables in the multiple regression equation. Findings Response Rates Of the 467 providers in our sampling frame, a total of 165 providers returned their questionnaires, 35 actively refused participation, and 267 passively refused participation (did not respond to our contacts and did not return the questionnaire). The 35% response rate was slightly higher than the 33% we had anticipated. Reasons given for refusing participation included competing surveys in the field, lack of time on the part of administrative staff, and the length and complexity of the data required. Of the 165 participating providers, 137 fully completed the questionnaire and were included in the data analysis. In order to be considered complete, a questionnaire had to meet all of the following criteria: 1) the questionnaire indicated that there was at least one DSW on payroll on January 1, 2011; 2) information was complete on the number that turned over between January 1 and December 31, 2011; and 3) the questionnaire overall was at least three-fourths complete. Of the 137 fully completed questionnaires, 86 providers completed the survey online using SurveyMonkey, 49 mailed their questionnaires, one sent the questionnaire by fax, and one completed the questionnaire by telephone. The remainder of this report focuses only on the 137 providers that were considered to have fully completed the questionnaire (see Table 1 below). Providers Participating in the Survey Providers participating in our study were fairly equally distributed across the four human service industries and across geographic regions. However, slightly more MH providers, and greater numbers of providers in Rural and Appalachian regions completed the survey. 4

9 Table 1. Providers Participating in the Survey by Provider Type and Region Respondent Characteristics On average 1.7 staff members completed each survey (range: 1-5 staff members). The majority of primary respondents were either Presidents/CEOs (37%) or Human Resource Directors (12%). The rest of the surveys were completed by Administrators or Directors of Clinical Programming. Background Characteristics of Participating Providers to a chain or franchise; 43% classified their provider as being For-Profit and 55% as Not-For-Profit with the remainder classifying themselves as both. Only 11% of all providers reported that their DSWs were required to work overtime, and only 5% reported that their DSWs belonged to a union. The remainder of this report concerns data provided by our respondents for Number of Clients Served The providers in our sample were in operation for a median* of 26 years (range: years). Only 16% of our sample classified themselves as belonging Table 2. Number of Clients and DSWs by Provider Type Providers varied greatly in the number of clients served from 1/1/2011 to 12/31/2011. The number of DSW staff on payroll on 1/1/2011 also varied greatly. Table 2 provides a summary of these data. Note: Percentages in some cells do not add up to 100% for each provider type because the data reported are median percentages. *In cases where responses were spread over a large range of values, we reported the median instead of the mean. The median is a better measure of central tendency than the mean in these instances because it is the middle value in a set of responses and unlike the average/mean is not skewed by very high or very low responses. 5

10 Unduplicated Clients Served in 2011 MH providers served the highest annual median number of unduplicated clients (2,061) while DD providers served the lowest median number of clients (20). NH and HH were fairly similar in terms of the median annual unduplicated number of clients served (206 and 282 respectively). DSWs on Staff on January 1, 2011 NH providers reported having the highest median number of DSWs on staff (72) on 1/1/2011. Full and Part-Time DSWs on January 1, 2011 MH providers had the highest median percentage of full-time DSWs on staff (96%), followed by NHs (62%), while HH providers had the lowest percentage (29%). HH providers reported having the highest median percent of part-time DSWs (57%), while MH was the lowest (3%). DD providers reported using full-time and part-time DSWs equally (50%). NH providers were the only ones to use PRNs (staff on an as-needed basis). Client Characteristics Demographic Information As expected, MH and DD providers primarily served a younger population (under 55) while NH and HH providers primarily served older adults (65 and older). Across settings, providers reported that 82% of clients were Caucasian (Non-Hispanic), 11% were African American, and 2% were Hispanic/Latino. percentage of Medicare utilization (53%), while 19% of NH, 8% of DD, and 5% of MH clients were reimbursed by Medicare. Other Sources: Private Pay: About 19% of NH clients were private pay followed by 8% in MH and 3% in HH and DD. Private Insurance: Fifteen percent (15%) of HH clients had services reimbursed by private insurance, followed by 10% in MH, 6% in NH, and 2% in DD. Local or County Funds: Seventeen percent (17%) of MH clients had services reimbursed through local or county funds; however, the percentage is less than 2% for all other provider types. State, Federal, Charity or Grant Funds: Were used to reimburse services for less than 3% of clients across provider types. DSW Characteristics Demographic Information Similar to the figures reported for clients, 82% of DSWs employed by providers in our study were Caucasian (Non-Hispanic), 13% were African American, and 2% were Hispanic/Latino. The percentage of Caucasian DSWs in our study (82%) is much higher than that reported in a national survey of DSWs (51%; Smith & Baughman, 2007). Services Provided Table 3 displays the percent of providers that reported the types of services performed by their DSWs. Reimbursement Sources for Services to Clients Medicaid: The primary source of reimbursement for client services in DD, NH, and MH providers was Medicaid, with averages of 91%, 58%, and 52% respectively. Only 25% of HH clients were reimbursed by Medicaid. Medicare: HH clients had the highest 6

11 Table 3. Percent of Providers Reporting Services Performed by their DSWs Services NH HH DD MH Bathing Toileting Dressing Grooming Feeding Meal Preparation Housekeeping Medication Reminders Transportation Case Management Conducting Assessments Crisis Intervention Development of Care Plans Positive Behavior Support Note: The percentages for the top services within each provider type are in larger font and appear in color. NHs: The services most frequently provided services by DSWs in NHs relate to Activities of Daily Living (ADLs; coded in red in Table 3). HH: The services most frequently provided services by DSWs in HHs include primarily ADLs followed by Instrumental Activities of Daily Living (IADLs; coded in green). The latter are more complex tasks necessary for independent living. DD: The services most frequently provided services by DSWs in DDs include IADLs (with transportation being the primary service) and ADLs. MH: The primary services offered in MH are different from the other providers and include case management, behavior support, crisis intervention, conducting assessments and development of care plans (coded purple). With a few exceptions, MH providers do not largely focus on ADL or IADL services. 7

12 Figure 1. Minimum and Maximum Median Hourly Rates by Provider Type Hourly Wages Reported hourly wages were low with MH providing the highest hourly wages. Low Hourly Wages Overall, the median minimum and maximum hourly rates for DSWs are low, at $9.17 and $14.00 per hour. MH provides higher wages because the majority of CPST workers have a college degree (Gitter, 2009). The overall average hourly rate for a DSW across providers is $11.87/hour. Thus, the average weekly salary for a DSW is approximately $475. This is in stark contrast to the average weekly salary of all Ohioans which is $873 (Bureau of Labor Statistics, 2012). Average hourly rates were significantly different for DSWs between the following provider types: Benefits Offered by Provider Type Providers that offered benefits indicated whether benefits were offered only to DSWs who worked a required number of hours, or whether all DSWs, irrespective of the number of hours worked, received benefits. In general, providers offered benefits to DSWs only if they worked a minimum number of hours per week (NH - 94%, MH - 88%, HH - 65%, DD - 41%). The median number of hours worked per week required for benefits eligibility was 27. Fewer providers offered benefits to all DSWs, irrespective of hours worked (DD - 31%, Other provider types - less than 14%). Benefits Offered to DSWs Figure 2 shows the percentage of providers offering certain key benefits to DSWs on payroll. Some benefits were more widely offered to DSWs than others, and varied greatly by provider type. -NHs ($11.25/hour) and DD ($9.25/hour); -MH ($14.10/hour) and HH ($10.25/hour); -MH ($14.10/hour) and DD ($9.25/hour). 8

13 Figure 2. Benefits Offered to DSWs Paid Vacation, Paid Holidays Off, and Paid Sick Leave: Overall, the vast majority of providers offered paid vacation (86%), paid holidays off (71%), and paid sick leave (71%) to their DSWs. However, there was variation across provider types with fewer DD providers offering such benefits. Career Ladder Programs: Overall, less than half of all providers offered career ladder programs to their DSWs. HH providers (19%) ranked lowest in terms of career ladder programs for DSWs. Traditional Pension Plan: Overall, less than 20% of all providers offered a traditional pension plan to DSWs. Twenty-seven percent (27%) of HH providers offered this service compared to only 7% of NH providers. Highest Number of Benefits Offered: Of the 4 provider types, MH providers offered the highest percentage of 10 of the 18 benefits included in our study. In addition to those listed in Figure 2, they offered fully paid health insurance for DSWs; sick leave for family illness; cost of living increase; IRA/403b/401k; mileage reimbursement; and Employee Assistance Program (EAP). Fully and Partly Paid Health Insurance We focused on health insurance because in other studies, this was a predictor of DSW retention and job satisfaction (Ejaz et al., 2008; Howes, 2008). See Figure 3. 9

14 Figure 3. Fully and Partly Paid Health Insurance Offered to DSWs Fully Paid Health Insurance for DSWs: Overall, fully paid health insurance was offered by only 10% of providers. Partly Paid Health Insurance for DSWs: In contrast, 3/4 of all providers offered partly paid health insurance to DSWs. The vast majority of NHs and MH providers offered this benefit compared to about 2/3 of HH and less than half of DD providers. Best Practices by Provider Type Our survey also examined the types of best practices and strategies that providers used for enhancing DSW job satisfaction. These practices include consistent assignment to clients and attendance bonuses for DSWs. Figures 4 and 5 reflect our findings regarding the best practices most and least widely adopted by providers in our survey. Figure 4. Best Practices Targeted at DSWs: Most Widely Utilized 10

15 Overall, 89% of providers paid for training sessions for their DSWs and offered consistent assignment to clients. Overall, 82% of all providers offered flexible scheduling; however, this practice was utilized by less than 2/3 of NH providers. Figure 5. Best Practices Targeting DSWs: Least or Inconsistently Utilized Paid Time to Attend Training beyond Annual Mandatory Requirements: More than 3/4 of MH and more than 2/3 of HH and DD providers offered paid time off to attend training beyond mandatory requirements. However, only 40% of NH providers utilized this practice. Performance Based Promotion: More than half of all providers offered performance-based promotions to their DSWs. DSW Involvement in Care Planning: MH providers (93%) are most likely to involve their CPST workers in care planning but this practice is less widely used by the other provider types. Attendance Bonus: Offering attendance bonuses is not a highly used practice. Slightly over half of NHs offered attendance bonuses vs. only 12% of MH providers. DSW Turnover for All Providers in the Study Figure 6 demonstrates the annual DSW turnover rate in 2011 by provider and region type. As mentioned earlier, turnover rate was calculated for each provider based on the number of DSWs on staff on January 1, 2011, and the number of those DSWs who turned over from January 1 to December 31,

16 Figure 6. Overall Average DSW Turnover Rate by Provider and Region Type Overall, the average DSW turnover rate for providers was 30% (median - 29%) in DSW turnover rate varied widely between providers (0-100%). Twenty percent of providers reported that no DSWs turned over in In general, these were smaller organizations with fewer DSWs. Four percent of providers did not know if they had DSW turnover. Respondents reported that a variety of reasons accounted for DSW turnover in their organization. Seventy-six (76%) of providers reported that their DSWs turned over voluntarily while 57% reported DSW turnover was involuntary (fired or terminated). Another 20% of providers reported that DSWs turned over because they were promoted, and 15% reported that DSWs had moved laterally. These percentages do not add up to a 100% because a provider could have reported various types of turnover. Statistically significant differences for the overall sample (providers with and without turnover): There was no significant difference among provider types in their overall rate of turnover. The lack of significance was probably due to the wide variation in turnover rate within and between provider types. With respect to region types, providers in metropolitan regions approached* a significantly higher turnover rate than those in suburban regions. DSW Turnover for Providers that Reported DSW Turnover in 2011 (Not reflected in Figure 6) The average turnover rate was 38% (range: 4-100%: no providers had between 1-3% turnover). Statistically significant differences for sample of providers that reported turnover: HH (52%) providers had a significantly higher turnover rate compared to MH providers (29%). Similarly, DD (46%) providers approached* a significantly higher turnover rate compared to MH providers (29%). No significant differences were observed across the five geographic regions. Impact of DSW Turnover Investigators asked providers to rate the impact DSW turnover had on their services using a scale from 0-10 with 0 being no negative impact and 10 being significant negative impact. The following areas are listed in order of their ranking: Interference with continuity of care (4.7) Quality of care (4.6) Impact on reputation/image in the commuity *Approached statistical significance is defined at <.10; all other statistical significance levels reported were <

17 (3.3) Prevention of new client admissions (2.2) Predictors of DSW Turnover for All Providers in the Sample Investigators conducted multiple regression analysis to examine factors associated with turnover for all providers in the sample including those that had and did not have turnover in 2011 (range in turnover rate: 0-100%). Turnover rate was considered as the outcome variable. When controlling for other variables in the model, the following factors were statistically significant predictors of turnover: Unsuitable fit for the position: When providers perceived that DSWs were not a suitable match, turnover rate was predicted to be higher by 15 percentage points. Work is too emotionally exhausting for DSWs: When providers perceived that DSWs found the work to be emotionally exhausting, the turnover rate was predicted to be higher by 14 percentage points. Policy and Practice Recommendations Regional Differences Providers in Metropolitan regions had a higher turnover rate of 24 percentage points compared to those in Suburban regions. The turnover rate of providers in Appalachian regions* was 13 percentage points higher than providers in Suburban regions. Maximum Hourly Wage An increase of $1 in the maximum hourly wages offered to DSWs is likely to decrease a provider s turnover rate by approximately 1 percentage point. Years in Operation* Providers who had been in existence 10 years longer than other providers were likely to have a turnover rate that was 1.6 percentage point lower turnover rate. Predictors of Turnover for Providers Who Reported DSW Turnover in 2011 Investigators focused on providers that had DSW turnover to examine additional factors that could explain DSW turnover. In addition to the maximum hourly wage* increases for DSWs and providers being in operation longer, two additional factors were statistically significant predictors of turnover. These were two factors related to the perceived causes of turnover from the perspective of respondents completing the survey: Our study findings suggest several practice and policy recommendations: Improving Wages, Particularly Maximum Wages a DSW can Receive It is interesting that our study found that the maximum wages a DSW can receive, rather than minimum starting wages, was a significant predictor of lower rates of DSW turnover. However, in another study, minimum starting wages were a predictor of DSW job satisfaction (Ejaz et al., 2008). Other studies have demonstrated that, overall, the issue of wages is linked to alarming rates of DSW turnover, particularly because such workers are primarily women from low-income households (Harris- Kojetin et al., 2004; U.S. General Accounting Office, 2001). Studies have also demonstrated that higher pay levels and benefits are related to DSW retention (Howes, 2004, 2008). In speculating about the finding that maximum wages are related to DSW turnover, there may be a variety of reasons supporting our results. We believe that when a DSW applies for a position in the human service industry, he/she is aware that it is an entrylevel position with low minimum wages. It is highly likely that the entry-level DSW could have found a similar, low-paying job in another industry, such as retail or fast food but consciously chose to enter the human services field. Many DSWs enter such fields for altruistic or religious or socially driven reasons (Howes, 2008; Noelker & Ejaz, 2001). Once in the job, some may become disenchanted and find the work to be emotionally and physically exhausting, *Approached statistical significance is defined at <.10; all other statistical significance levels reported were <

18 feel overworked and underpaid, realizing that they lack critical benefits such as health care coverage (Ejaz et al., 2008; Howes, 2008). One way to help retain DSWs who come motivated to work in the human service industry is to recognize their dedication and encourage them to stay on the job by offering them higher wages commensurate with their years of service. Focus on Improving DSW Retention in Certain Geographic Regions of the State Our study found that DSW turnover was higher in metropolitan and Appalachian regions compared to suburban areas. In speculating why certain regions had higher turnover rates, investigators believe that there are numerous competing opportunities for entry-level, low-paying jobs in metropolitan regions. In Appalachian regions, there may be less competition for low-paying jobs but the travel time and distance between where the DSW s home and the location of the provider or a client s home could be barriers to retention. Thus, providing mileage reimbursement to such workers may help increase retention. Once again, providing better wages and benefits is likely to attract and improve DSW retention especially in light of the availability of other low-paying, less physically and emotionally stressful jobs. Create a Repository of Best Practices and Evidence- Based Programs that have Demonstrated a Reduction in DSW Turnover Since DSW turnover in our study ranged from 0-100% with 20% of providers reporting no DSW turnover in 2011, it is important that providers with high turnover learn best practices and strategies for retention from those with little or no DSW turnover. Findings from across the nation suggest a variety of best practices are likely to improve DSW retention. Policy makers could fund researchers and/or academicians to: Conduct literature reviews and develop a series of white papers and videos to examine industry best practices and evidence-based programs that have led/are likely to lead to a reduction in DSW turnover; Examine the extent to which such programs are relevant to the four provider types included in our study; and, Support additional avenues for dissemination besides the white papers/videos, such as sponsoring statewide seminars, webinars and conferences focused on evidence-based programs and industry best practices. Test the Implementation of Selected Industry Best Practices and Evidence-Based Programs to Improve DSW Retention Some examples of best practices and evidencebased programs from across the nation could be implemented in Ohio to address some of the issues highlighted in our study. For example: Implement best practices to improve the hiring process and avoid hiring DSWs who are unsuitable for the position: Often, the human service industry is under time constraints to fill a DSW positions urgently (Institute of Medicine, 2008). However, this scenario is likely to result in high turnover, particularly within the first year on the job (Noelker & Ejaz, 2005; Paraprofessional Healthcare Institute, 2005). A recently conducted study in Northeast Ohio found that a variety of best practices were likely to reduce staff turnover in nursing homes. These strategies included improving the hiring process by using behavioral interviewing and skills assessment for potential hires and later customizing their orientation to the facility (Lavelle, Smudla & Ejaz, 2011). Encouraging the use of such best practices across the human service industry is likely to reduce DSW turnover. Provide work-life support programs to address the emotional demands of the job: Studies have found that DSWs find their work to be extremely emotionally demanding, which impacts their job satisfaction (Benjamin & Matthias, 2004; Ejaz et al., 2008; Geiger-Brown, Muntaner, McPhaul, Lipscomb, & Trinkoff, 2007; Stacey, 2005). Another study, found that providing work-life support programs was likely to help reduce staff turnover (Lavelle et al., 2011). In contrast, our 14

19 study found that only 30% of providers offered employee assistance programs to all of their DSWs. Thus, providing work-life support programs, and counselors or staff to help deal with the emotional demands of the job, are likely to reduce DSW turnover. Create career ladder programs: Studies have suggested that career ladder programs are likely to enhance DSW job satisfaction (Council for Adult and Experiential Learning, 2005; Lavelle et al., 2011; U.S. Department of Labor, 2008). Yet our study found that overall only 44% of providers offered a career ladder program to their DSWs. Ohio s Center for Education and Employment is working on developing core competencies and career lattices and ladders for DSWs (Lepicki, Austin, Adams, Johnson, & Lockett, 2012). The purpose of career lattices is to encourage DSWs who wish to remain in their positions get further training in certain speciality areas such as dementia care or medication management to create specialty areas to help them earn slightly higher wages. In contrast, the goal of career ladder programs is often to move a DSWs out of his/her current position. One example of this is to encourage a young and aspiring individual to start his/her career as a DSW and use this as a stepping stone to eventually move to a more lucrative, professional career in the health and human service industry. In this case, such DSW turnover should not be viewed negatively. Career ladder programs are likely to benefit the health and human services field at large, particularly where there are shortages in professional fields such as nursing and social work. In fact, creating loan forgiveness programs for DSWs who have advanced into professional careers and are willing to work in underserved regions and areas where there are widespread health care shortages could also benefit the field at large. Reward Providers with a Higher Reimbursement Rate for Low DSW Turnover Another strategy used by policy makers could be to create a reimbursement-based reward system for providers who have used evidence-based strategies to reduce their turnover rates. For example, the NH industry in Ohio already is participating in a Payfor-Performance system in which DSW retention is one of the quality incentives for increased reimbursement. This system can be expanded to other sectors such as DD, MH and HH. The rate of reimbursement should be adequate to attract and encourage providers to participation in such quality improvement programs. Next Steps Limitations of the Current Project Although our study gathered comprehensive data from 137 providers from a variety of provider types and geographic regions, there are certain limitations that need to be acknowledged. Primary among these is the issue of having a lower than expected response rate from providers. Response rates were negatively impacted by other competing surveys in the field, a lack of time on the part of Administrators/ Human Resource staff to complete our detailed questionnaire, the length and complexity of data required, and the lack of readily available electronic data on DSW turnover over a one year period. The limited timeframe and budget for conducting the study also impacted the response rate. Investigators believe that a longer timeframe, more resources, along with monetary or other incentives to providers would encourage participation and increase research staffing for more intensive follow-up with providers. Other limitations included the lack of a readily available, comprehensive list of providers of DD services, particularly those with DSWs on staff. Project research staff spent more than 100 hours cleaning this list. It is our understanding that DODD is in the process of creating a better online listing of these providers. Strengths of the Current Project The strengths of the study lay in the fact that four types of providers and various geographic regions were included in the study. This is the first study of its kind in the nation to include such a sample. Further, the investigative team was highly skilled in the subject matter and the collaboration between the Ejaz/Judge and Gitter projects led to a comprehensive approach to examining DSW turnover in Ohio. In addition, having an expert panel of stakeholders on the Review Team helped to guide investigators in various stages of the project 15

20 including development of a comprehensive survey. The survey too, underwent a series of revisions and pre-tests to ensure its reliability. Finally, conducting stratified, random sampling procedures allowed investigators to be confident about the reliability and validity of findings. Future Research Investigators believe it is important to build on the current studies in the following ways: More immediately, merge the data collected from this study on human service providers with that of Dr. Gitter s study on DSWs from the same providers. The combined dataset would allow for a more comprehensive picture of the causes of turnover, both from the perspective of providers as well as the DSW perspectives. Replicate the combined study with a larger sample of providers in Ohio to further enhance the generalizability of the findings. Conduct a similar study with a nationally representative sample. Conduct a follow-up survey of the same providers in the current study to examine the cost of DSW turnover in their facilities. This information will be invaluable to policy makers to help them develop strategies to reduce turnover costs. Conduct a qualitative study that includes site visits with Administrative staff with two groups of Ohio providers: 1) those with little or no DSW turnover in 2011 (0-10%) and 2) those with very high turnover (75% or more) in Ohio. Compare and contrast the types of benefits, and best practices/strategies used by these two types/groups of providers and use the findings to develop recommendations for improving DSW retention in Ohio. 16

21 References Benjamin, A. E., & Matthias, R. (2004). Work-life differences and outcomes for agency and consumer-directed home-care workers. The Gerontologist, 44(4), Bureau of Labor Statistics. (2012, November 9). County employment and wages in Ohio: First quarter 2012 (News release CHI). Retrieved from Castle, N. G. (2006). Measuring staff turnover in nursing homes. The Gerontologist, 46, Castle, N. G., Engberg, J., Anderson, R., & Men, A. (2007). Job satisfaction of nurse aides in nursing homes: Intent to leave and turnover. The Gerontologist, 47(2), Council for Adult and Experiential Learning. (2005). How career lattices help solve nursing and other workforce shortages in healthcare. Retrieved from Dill, J. S., & Cage, J. (2010). Caregiving in a patient s place of residence: Turnover of direct care workers in home care and hospice agencies. Journal of Aging and Health, 22(6), Ejaz, F. K., Gitter, R., Bukach, A., Dawson, N., Judge, K., Giri, B. (2012). The turnover crisis in Ohio s direct service workforce: Causes and potential remedies. Paper session presented at the Margaret Blenkner Research Institute of Benjamin Rose 50th Anniversary Conference, Aurora, Ohio. Ejaz, F. K., Noelker, L. S., Menne, H. L., & Bagaka s, J. G. (2008). The impact of stress and support on direct care workers job satisfaction. The Gerontologist, 48(1), Ejaz, F. K., Rose, M. S., & Bukach, A. B. (2011). The experience of long-term care organizations in seven states with older workers, volunteers and participants from the Senior Community Employment Program (SCSEP; Report to Senior Service America, Inc.). Retrieved from pdf Geiger-Brown, J, Muntaner, C., McPhaul, K., Lipscomb, J., Trinkoff, A. (2007). Abuse and violence during home care work as predictor of worker depression. Home Health Care Services Quarterly, 26(1), Gitter, R. J. (2005). The cost to employers of case manager turnover in Ohio s mental health system. New Research in Mental Health, 16, Gitter, R. J. (2009). The determinants of turnover and retention of case managers in Ohio s mental health system. New Research in Mental Health, 18, Harris-Kojetin, L., Lipson, D., Fielding, J., Kiefer, K., & Stone, R. I. (2004). Recent findings on frontline longterm care workers: A research synthesis Washington, D.C.: Institute for the Future of Aging Services. Howes, C. (2004). Upgrading California s home care workforce: The impact of political action and unionization. Retrieved from Howes, C. (2008). Love, money, or flexibility: What motivates people to work in consumer-directed home care? The Gerontologist, 48(1),

22 Institute of Medicine. (2008). Retooling for an aging America: Building the health care workforce. Washington, D.C.: The National Academies Press. Lavelle, C., Smudla, D., & Ejaz, F. K. (2011, November/December). Evidence-based sustainable employment. Leading Age Magazine, 1(6), Retrieved from Lepicki, T., Austin, J., Adams, J., Johnson, M., & Lockett, E. (2012). Core competencies for direct service workers. Paper session presented at Ohio s Direct Service Workforce Applied Research Conference, Columbus, Ohio. Menne, H. L., Ejaz, F. K., Noelker, L. S., & Jones, J. A. (2007). Direct care workers recommendations for training and continuing education. Training and education to address workforce recruitment and retention challenges in community-based and long-term care (Special Issue). Gerontology and Geriatrics Education, 28(2), Noelker, L. S., & Ejaz, F. K. (2001). Improving work wettings and job outcomes for nursing assistants in skilled care facilities. Final Report submitted to The Cleveland Foundation and The Retirement Research Foundation. Noelker, L. S., & Ejaz, F. K. (2005). Training direct care workers for person-centered care. Public Policy & Aging Report, 15(4), Noelker, L. S., Ejaz, F. K., Menne, H. L., & Bagaka s, J. G. (2009). Factors affecting frontline workers satisfaction with supervision. Journal of Aging and Health, 21(1), Paraprofessional Healthcare Institute. (2005). The role of training in improving the recruitment and retention of direct-care workers in long-term care. Retrieved from WorkforceStrategies3.pdf Seavey, D. (2004, October). The cost of frontline turnover in long-term care. (Practice & Policy Report: 1-3). New York, NY: Better Jobs Better Care. Seavey, D. (Winter ). Caregivers on the front line: Building a better direct-care workforce. Generations, 34(4), Smith, K., & Baughman, R. (2007). Caring for America s aging populations: A profile of the direct-care workforce. Monthly Labor Review, 130(9), Stacey, C. L. (2005). Finding dignity in dirty work: The constraints and rewards of low-wage home care labour. Sociology of Health and Illness, 27(6), U.S. Department of Health and Human Services. (2006, January). The supply of direct support professionals serving individuals with intellectual disabilities and other developmental disabilities. Report to Congress. U.S. Department of Labor. (2008). Registered apprenticeship trends in health care. Retrieved from doleta.gov/oa/brochure/2007%20health%20care.pdf U.S. General Accounting Office. (2001, May). Nursing workforce: Recruitment and retention of nurses and nurse aides is a growing concern (GAO T). Statement of William J. Scanlon, Director, Health Care Issues. Washington, D.C. 18

23 EXAMINING DIRECT SERVICE WORKER TURNOVER IN OHIO ORGANIZATIONAL SURVEY NURSING HOME Funded by: Ohio s Money Follows the Person Demonstration Project, Centers for Medicare and Medicaid* Investigators: Farida K. Ejaz, Ph.D., Margaret Blenkner Research Institute, Benjamin Rose Institute on Aging Robert Gitter, Ph.D., Ohio Wesleyan University Katherine S. Judge, Ph.D., Cleveland State University Additional Project Team Members: Ashley Bukach, B.S., Margaret Blenkner Research Institute, Benjamin Rose Institute on Aging Nicole Dawson, Ph.D. Graduate Student, Cleveland State University & University of Akron * More information on the Money Follows the Person Demonstration Project is available at: