## Abstract

Self-directed programs that allow individuals with intellectual and developmental disabilities to exercise greater control over their finances have become increasingly common in recent years. At the same time, challenges in the recruitment, retention, and training of direct support workers in the field have grown more acute. In this article, the authors investigate the status of the direct support workforce for people using self-directed supports in 1 Midwestern state, based on the results of a statewide survey of service users. Although additional research is needed, the results of this study suggest that people who use self-directed funding options are satisfied with their ability to direct staffing, though challenges remain. Among these challenges, the presence of higher than expected wages but lower than expected benefits provision compared with traditional services may have serious policy and staff retention ramifications that affect the long-term viability of self-directed funding options. In addition, staff training remains a challenge, with service users in this sample reporting low rates of training beyond a general skill set. Implications of these findings are discussed.

Self-directed supports have become an increasingly prevalent support option for individuals with intellectual and developmental disabilities and their families, with at least 34 states offering some measure of individualization through consumer control of financial benefits and/or individualized budgeting (Walker, Hewitt, Bogenschutz, & Hall-Lande, 2009). At the same time that this expansion of self-directed support has occurred, traditional challenges in the recruitment, training, and retention of direct support staff that have been present in the field for decades may be growing more acute (Hewitt & Larson, 2007). Although limited information about the effect of self-directed supports on direct support staffing has begun to emerge (Dale, Brown, Phillips, & Carlson, 2005; Matthias & Benjamin, 2008), much more data must be examined to gain a firm understanding of the role that self-direction may play in addressing the direct support workforce issues that continue to challenge the field, as well as how workforce challenges affect individuals with intellectual and developmental disabilities and their families who self-direct.

In this article we add to the current understanding of how self-directed supports affect outcomes in the direct support workforce supporting individuals with intellectual and developmental disabilities, using findings from a statewide survey of families accessing Minnesota's self-directed funding option, Consumer Directed Community Supports (CDCS).

## Literature Review

### Status of Self-Directed Supports

Self-directed supports is a program or service option in which people with disabilities, their families, or their allies manage and direct their services and supports. In this service option, people with disabilities and their families use individually set and controlled budgets to select, purchase, and manage their own supports within an established framework of guidelines (Alliance for Self Determination, 1999; Nerney & Shumway, 1996; Scala & Nerney, 2000). Furthermore, individuals and families exercise greater control in the recruitment, hiring, and management of support staff (Mosely, 2001).

Predicated on the principles of self-determination and normalization, the original efforts at promoting self-directed services and supports were funded by the Robert Wood Johnson Foundation (RWJ), which initially supported a small demonstration in New Hampshire in 1991 (Monadnock Developmental Services, 1996). Following this demonstration, in 2001, RWJ supported a series of larger scale programs: the Cash and Counseling Demonstration and Evaluation Projects in Arkansas, Florida, New Jersey, and New York. Although the New York project had difficulties in initiation (Sciegaj, Mahoney, & Simone, 2008), the other three states successfully implemented self-directed support models, which have provided valuable information to inform future efforts at implementation of CDS. Today, many service options are available in which people with disabilities and their families coordinate and direct their own services, and the trend toward self-directed support has been growing steadily (Greene, 2007).

Research reveals several common elements reflecting best practices in self-direction. These features include individual budgets, support coordination, and the use of fiscal intermediaries (Nerney, 2001; Scala & Nerny, 2000). In addition, Walker et al. (2009) have suggested that different levels of assistance should be built into the service structure, information technology resources need to be updated, and adequate training for consumers must be offered for self-directed initiatives to be optimally beneficial. These supports are critical in promoting true self-determination and achieving effective implementation of self-directed services.

Self-directed funds are most often used to pay workers who provide support to the service recipient. Even in situations where hiring and retaining staff are identified as challenges, consumers report that being able to exercise control over staffing decisions is a benefit (Young & Sikma, 2003). In a study of Cash and Counseling recipients in the three RWJ demonstration states (New Jersey, Arkansas, and Florida), Schore, Foster, and Phillips (2007) found that between 80% and 90% of beneficiaries used part or all of their Cash and Counseling allowance to pay workers, although rates were somewhat lower in Florida. In general, consumers have expressed high levels of satisfaction at being able to hire family members as paid caregivers (Matthias & Benjamin, 2008). Arkansas consumers who hired family members under Cash and Counseling also indicated similar or more positive health outcomes than consumers who hired nonfamily workers (Simon-Rusinowitz, Mahoney, Loughlin, & DeBarthe Sadler, 2005), suggesting the effectiveness of family staffing arrangements.

Workers who are related to recipients of self-directed services report high levels of satisfaction as well. These workers also indicate a desire for more respect from the person employing them, levels of emotional strain comparable with agency-based workers, and less formal training than agency staff (Dale, Brown, Phillips, & Carlson, 2005). As the traditional pool of direct support workers continues to narrow, some authors have suggested that the flexibility in hiring friends and family members that is afforded by self-directed support models may become a more necessary strategy in addressing the critical shortage of workers in the direct support workforce (Hewitt & Larson, 2007; Matthias & Benjamin, 2008) and in promoting the inclusion of individuals with disabilities in their communities (Walker, 1999).

Educating and informing individuals about self-directed models continue to be challenges in implementation. Although the support of traditional agency service providers is necessary for effective implementation of Cash and Counseling programs, educational outreach about self-directed programs works best when it is directly targeted to consumers and their representatives (Phillips et al., 2003). Additional outreach concerns have been raised about differentials in attitudes toward and use of Cash and Counseling among members of differing racial and ethnic communities (Sciegaj, Capitman, & Kyriacou, 2004; Simon-Rusinowitz et al., 2002). In a study of individuals who were elderly and were eligible to choose between self-directed or traditional services, Sciegaj et al. found a tendency among Chinese, Latino, and African American consumers to seek more control over their staffing, corroborating assertions by Simon-Rusinowitz and colleagues (2002), who suggested that Latino and African American families may prefer self-direction due to their strong access to large family and friend support networks. Such arguments support the necessity of increasing self-direction to help serve the needs of increasingly diverse communities, as such efforts may assist in alleviating recruitment strains on consumers who self-direct supports.

Critics of self-directed services sometimes cite concerns over the personal safety of recipients, who may be more vulnerable to financial exploitation and medical neglect with the less structured support system that is assumed to be present in self-directed models. These critics were more likely to be policy experts than consumers or their representatives (Simon-Rusinowitz et al., 2002). There has been evidence to suggest that these concerns may not be well grounded, as the personal security of individuals who self-direct is similar to those using traditional services (Schmitz, Luxenberg, & Eustis, 2005; Young & Sikma, 2003). Furthermore, Matthias and Benjamin (2003) reported that individuals who receive supports from family members or friends are less likely to be abused than those consumers receiving supports from strangers or agency-based staff. Despite such criticisms, self-directed models of service provision continue to grow (Greene, 2007), indicating that the desirability of such programs has outweighed the risks and challenges thus far.

### Direct Support Workforce

Shortages of direct support professionals (DSPs) to support the needs of individuals with intellectual and developmental disabilities have emerged since the early 1980s (Lakin & Bruiniks, 1981; Lakin et al., 1983) and have become increasingly acute as community residential arrangements for individuals with IDD have supplanted institutional settings (Hewitt & Larson, 2007; Prouty, Alba, & Lakin, 2008). A number of factors, including the continuing move toward community support, longer life expectancy for people with intellectual and developmental disabilities, and the aging of the population, have raised the projected demand for DSPs by 37% from 2003 to 2025 (Hewitt et al., 2008; Office of the Assistant Secretary for Planning and Evaluation, 2006). Further complicating the increased projected need for DSPs is the fact that the traditional pool of workers who support people with intellectual and developmental disabilities is shrinking (Hewitt & Larson, 2007; Larson, Hewitt, & Knoblauch, 2005). A number of more specific, interrelated factors are contributing to the current and looming shortages of DSPs, including wages and benefits, turnover, and training.

Wages for DSPs are generally low, and many DSPs are not offered fringe benefits. In the most recent study to compare wages of both community-based (private) and institutional (public) DSPs, Polister, Lakin, and Prouty (2003) found average national hourly wage rates to be $11.67 for DSPs in public settings and$8.68 for those DSPs working in the private sector, making full-time wages for work as a DSP roughly equal to the poverty line for a family of three (Lakin, Polister, & Prouty, 2003). As suggested by these findings, it is typical for DSPs in public settings to receive higher wages than their peers working in private settings, leading to additional concerns about wages as the move toward community-based residential settings continues.

Fringe benefits are available to DSPs with varying frequency, according to results of several studies compiled by Larson, Hewitt, and Knoblauch (2005). According to the 10 state-level studies compiled between 1998 and 2003, between 30% and 98% of organizations that employ DSPs offer health insurance, 26%–82% offer dental coverage, 21%–81% provide retirement savings programs, and 30%–87% provided DSPs with paid leave time. It is important to note, however, that many DSPs who work in agencies that offer benefits may not actually be eligible, because many agencies that offer benefits do so only for full-time employees, whereas many DSPs work on a part-time basis. Low wages and poor benefits may be primary reasons for high turnover rates among DSPs.

Turnover rates for DSPs working in community settings have not been systematically studied on a nationwide basis. However, state-level studies have emerged in the last 10 years to indicate that turnover rates range from 24% for full-time DSPs in Wisconsin to 109% among staff in North Carolina's adult care homes (Harmuth & Dyson, 2005). Adding further challenge, roughly 45% of DSPs who leave their positions do so within the first 6 months of employment (Larson, Lakin, & Bruiniks, 1998). Hewitt and Larson (2007) estimated that it may cost approximately $784 million annually to replace staff due to DSP turnover, based on the average national turnover rate and the mean cost of training a new DSP, which is estimated at$2,413. Thus, it is clear that turnover is not only a problem in terms of continuity of quality care to people receiving supports but a major financial burden on agencies and individuals that provide such services.

Last, training of DSPs has long been acknowledged as a challenge to direct support workforce development. As the move toward community support has continued, DSPs have been required to obtain more independence and a larger skill set to adequately support individuals with intellectual and developmental disabilities. The Community Support Skill Standards (Taylor, Bradley, & Warren, 1996) validated a set of 144 specific skills that are needed by DSPs, supporting the need to provide training beyond the 40 hr of preservice training that is generally offered to DSPs (Hewitt & Larson, 2007). DSPs and administrators sometimes disagree on which topics require additional training, as indicated by a North Carolina study in which administrators indicated the need for more DSP training in the areas of documentation, communication, and crisis intervention, whereas the DSPs surveyed preferred more training on crisis intervention, education, self-development, advocacy strategies, and community living skills (Test, Flowers, Hewitt, & Solow, 2004).

With the emergence of self-directed supports and the existence of serious challenges in the direct support workforce, the need exists to examine how the direct support workforce may look for users of self-directed services. Building on the knowledge above, in the remainder of this article, we present findings of a study of Minnesota's self-directed support option for individuals using the state's developmental disabilities waiver, which was known as the Mental Retardation and Related Conditions waiver at the time of data collection.

## Method

Data were collected via a mailed survey, which was distributed to both previous and current users of CDCS. Although the state requires that CDCS is available statewide, not all Minnesota counties had CDCS participants at the time of the study's initiation. A list of counties that had implemented CDCS for individuals on the Mental Retardation and Related Conditions (MRRC) waiver as of September 2006 was obtained from the Minnesota Department of Human Services. The research team reviewed the list to determine which counties had consistently implemented CDCS throughout the period 2004–2006. In total, the most populous urban county in the state, 5 additional metropolitan counties, and 16 nonmetropolitan counties constituted the sampling frame for this study.

It is important to note that the survey was conducted at a time of change within the CDCS program, as the state had recently taken control of setting individual budgets for recipients, a function that had previously resided at the county level. This change resulted in decreased budgets for some CDCS service recipients, leading to some reduction in the number of individuals using CDCS as well as some degree of confusion among county- and state-level program administrators (Schmitz, Luxenberg, & Eustis, 2005). Despite these changes in implementation and administration, however, Minnesota remains among the states with the highest number of recipients of self-directed funding (Walker et al., 2009). However, due to the changes that were occurring at the time of the survey, it is important to evaluate differences in perceptions of the CDCS program among current and past users, meaning that slight differences in the survey instrument were made to tailor it to current or past CDCS service users.

### Sampling

Twenty of the 22 counties in the sampling frame agreed to assist in the distribution of the survey to CDCS service recipients who were on the MRRC waiver. A sampling worksheet was distributed to the administrator in each of these counties to guide the sampling process. A stratified random sampling procedure was used to promote an even distribution of survey recipients from the urban, suburban, and nonmetropolitan parts of the state. Counties from all parts of the state were instructed to compile an alphabetized list of all individuals who had utilized CDCS since September of 2004, including those who currently used CDCS and those who no longer used it. Nonmetropolitan counties selected every 2nd person on their list, whereas urban and suburban counties selected every 10th person to produce the sample. The guardianship status and CDCS usage status (current or previously used) of individuals who would receive the survey was reported to the research team, who prepared surveys for mailing. Surveys were color coded to indicate location (urban, metropolitan, nonmetropolitan) and were prepared with a cover letter explaining the purpose of the study. In addition, the cover letter was translated into Spanish and Hmong to facilitate completion by potential participants without English proficiency, as identified by county administrators. These individuals were instructed to contact a contracted interviewer, who would conduct the interview via phone in the participant's preferred language. Follow-up postcards that were sent to these individuals were also translated into Hmong and Spanish.

The researchers sent the appropriate surveys to each county administrator, who addressed the envelope and mailed the survey. Recipients had 3 weeks to complete the survey. Follow-up postcards were mailed to all individuals in the sample 2 weeks after the initial mailing. Completed surveys and documentation of informed consent were returned via mail directly to the researchers. In addition, survey recipients were given the option of contacting a member of the research team via phone to verbally complete the survey. Excluding those returned as undeliverable, a total of 369 surveys were distributed across the state. One hundred twelve completed surveys were returned, for a total response rate of 30.4%. Although this response rate is in line with expectations for mail-based surveys (Dillman, Smyth, & Christian, 2008), it is possible that the addition of a Web-based survey option may have been successful in raising the overall rate of response. Based on the state's information at the time of sampling, 1,401 people were receiving CDCS under the MRRC waiver, meaning that survey results reflect the experience of almost 8% of all CDCS recipients in Minnesota.

It is important to note that, based on Medicaid regulations, CDCS users may only be defined as the individuals who are authorized to receive services, in this case individuals with intellectual and developmental disabilities. This survey, however, was completed by some individuals with intellectual and developmental disabilities and some family proxies for these individuals. This study, therefore, uses the terms recipient and family interchangeably, to reflect the variety in actual survey respondents.

### Sample

A total of 112 users of CDCS completed and returned the mailed survey. Of the respondents, 58.9% (n  =  66) were currently using CDCS, whereas the remaining 41.1% (n  =  46) had used CDCS in the past. Among those who answered a question about racial background, just over 87% (n  =  87) of the sample was Caucasian, with African Americans (4.6%, n  =  5) and Asian Americans (3.7%, n  =  4) composing the largest shares of other racial groups in the sample. Geographically, slightly over half the respondents came from the state's most populous urban county (54.5%, n  =  61), with the remainder from suburban (32.1%, n  =  36) and nonmetropolitan (13.4%, n  =  15) parts of the state, roughly reflecting the distribution of all CDCS users, as reported by the state, which suggests greater utilization in urban and suburban counties. Respondents predominantly lived with family members (80.7%, n  =  88). The demographic composition of the sample is summarized in Table 1.

Table 1

Respondent Demographics

### Data Source

An 8-page, mail-based survey was the means of gathering data for this study. The survey was developed by members of the research team, in collaboration with community-based partners, to elicit information on how users of CDCS utilized their funds, workforce outcomes that resulted from implementation of CDCS in Minnesota, and demographic information about respondents. The survey was piloted with a small sample of CDCS recipients and was revised based on feedback gained from the process. Data from telephone interviews and focus groups with CDCS users also informed the creation of the survey. The workforce outcomes portion of the survey provided the basis for this study and consisted of variables related to direct support worker training, workforce characteristics, and perceived benefits and challenges of using CDCS funds for staffing needs.

### Measures

#### Training

Respondents were asked if they had offered staff training to employees who had been hired with CDCS funds and if the self-directed program had made staff training easier. Through additional response items, we sought feedback as to the nature of training topics (including first aid, CPR, specific needs, etc.) as well the methods that respondents used to deliver training to their employees (e.g., outside agency, conference or workshop, self-administered training). Respondents were also afforded the opportunity to identify other types of training and methods of delivery that were not listed.

#### Benefits and challenges

Two multiresponse items in the survey asked study participants to identify the staffing-related benefits and challenges they experienced while using CDCS. Respondents selected as many responses as applicable among the listed benefits (including “staff are reliable,” “my staff were better trained,” “staff stay longer,” “I can set pay for my staff,” etc.) and were allowed to write in additional benefits.

Similarly, survey respondents selected applicable challenges among a list provided, with the opportunity to add additional write-in challenges. Categories of staffing challenge offered on the survey included “it's hard to find staff,” “payroll issues,” “cultural differences,” and “it's hard to keep staff,” among others.

#### Workforce characteristics

A grid was presented to gather information about the workforce status indicators that respondents encountered while using self-directed supports. The grid provided a line for each staff that had supported the person on the MRRC waiver during the time that CDCS was in use. For each direct support staff listed, information about the staff person's relationship to the service user, the date of hire, the length of service (in months), and the most recent rate of pay was obtained. The relationship, length of service, and wage categories were of particular importance to the analysis.

## Results

Data were analyzed in SPSS 16 in a multistep process. Input from surveys was entered into SPSS, checked for accuracy, and data were cleaned. Descriptive statistics and frequency distributions were produced, before a series of standard chi-square analyses were run on categorical variables to assess the relationships among key indicators of workforce development under CDCS.

### Staffing Benefits and Challenges with CDCS

A portion of the survey asked respondents to review a list of possible benefits of using CDCS to meet staffing needs. Participants indicated benefits that they experienced by marking a box that corresponded to each applicable benefit, with an additional space provided for the entry of benefits that were not identified. A similar process was followed for the identification of common staffing challenges.

Results suggested the presence of several distinct benefits and challenges that service users experienced with staffing while utilizing CDCS. The most commonly reported benefits of CDCS were the freedom to hire direct support staff who the individual knew (82.2%), including hiring family members (81.3%), and being able to set staff pay (82.2%). Of the staffing benefits listed on the survey, respondents were least likely to identify benefits related to better cultural match between the staff and the individual receiving supports (20.6%; possibly because more respondents were Caucasian) and with having better trained staff (34.6%).

The most commonly cited challenges selected by survey respondents were difficulties finding staff (51.9%) and trouble performing payroll duties when use of funds is self-directed (31.5%). Cultural differences (3.7%) and difficulties obtaining background screenings (9.3%) for potential direct support workers were the least commonly reported challenges. Frequencies of benefits and challenges are summarized in Table 2.

Table 2

Staffing Benefits and Challenges of CDCS

Interestingly, the individual's CDCS budget was not significantly related to most staffing benefits and challenges. However, the distribution of responses among individuals who stated that they found benefit in the flexibility to offer bonuses or raises to staff was significantly different than anticipated for individuals who currently used CDCS, χ2(2, N  =  52)  =  6.92, p  =  .032, as well as those who used CDCS in the past, χ2(2, N  =  77)  =  7.30, p  =  .026. Individuals with higher budgets were more likely to perceive benefits related to giving raises or bonuses in each case. This may indicate that wages under CDCS are stable, regardless of the size of the individualized budget that the consumer is granted to pay for his or her support. The fact that the ability to offer raises or bonuses is associated with higher budgets may indicate, however, that staff wages and benefits may stagnate for many recipients of self-directed supports and their direct support workers when individualized consumer budgets are low.

Also related to budgeting, results suggested that service recipients in nonmetropolitan counties may have had lower budgets than service users in suburban or urban counties prior to the state assuming control of the budgeting process in 2005, χ2(4, N  =  81)  =  15.68, p  =  .003. This suggests that administrators in different parts of the state may have applied differing budgeting methodologies or may have urged different types of service recipients (those with lower support needs) toward CDCS use. This difference in distribution was not observed for current CDCS users in differing parts of the state, suggesting that the state may apply a more consistent budgeting model.

Relationships between a number of staffing benefits and the ease of finding, training, and retaining direct support workers were also tested. The results of these chi-square tests are summarized in Table 3. Among the benefits tested, staff reliability and staff staying for longer amounts of time were two of the strongest when cross-tabulated with whether use of CDCS made finding, training, and keeping staff easier. People who reported benefits in staff reliability with CDCS use were more likely than expected to report that CDCS made finding, χ2(1, N  =  99)  =  16.83, p < .001; training, χ2(1, N  =  97)  =  5.43, p  =  .020; and keeping, χ2(1, N  =  103)  =  25.60, p < .001, DSPs easier. The respondents who reported benefits in retaining staff for longer were statistically more likely to report that CDCS made it easier to find, χ2 (1, N  =  99)  =  6.69, p  =  .010; train, χ2(1, N  =  97)  =  4.28, p  =  .039; and keep, χ2(1, N  =  103)  =  18.47, p < .001, their direct support staff.

Table 3

Staffing Benefits Reported Among Those Who Reported CDCS Made Finding, Training, and Keeping Staff Easier

In addition to the staffing benefits that were universally associated with making it easier to find, train, and retain staff, a number of staffing benefits were associated with more specific positive staffing outcomes. For example, those who found benefit in CDCS since it promoted cultural match with staff were also more likely to report that CDCS made finding staff easier, χ2(1, N  =  99)  =  4.13, p  =  .042. The benefits of having more skilled staff were associated with findings that CDCS made finding, χ2(1, N  =  99)  =  8.89, p  =  .003, and keeping, χ2(1, N  =  103)  =  7.73, p  =  .005, staff easier but not with greater ease in training staff, suggesting the possibility that staff hired under CDCS may have had greater skills coming into their work, thus necessitating less intensive training at the onset of employment.

### Training Patterns Among CDCS Users

Sixty two percent of respondents reported that they had offered staff training. The majority of survey respondents (63.4%) indicated that using CDCS made training staff easier than with traditional services. Despite this, having better trained staff was reported as a benefit by a minority of respondents (34.6%), indicating the possibility that greater ease of training may not have facilitated better staff training outcomes.

By a wide margin, the most common form of training was that which was designed to assist the DSP in addressing the specific needs of the person receiving supports. Among individuals who stated that they offered training using CDCS funds, 88.4% responded that the training included preparation to meet the specific needs of the CDCS consumer. More traditional training subjects like first aid, CPR, and medication administration, which are generally required of typical agency-based direct support professionals, were offered at a far lower rate under CDCS, suggesting the possibility that DSP roles may become more specialized and person centered under CDCS. Although it could be considered to be the topic of most applicability to CDCS specifically, only 14.5% of people who offered training provided training on fiscal support issues, an important function within the framework of CDCS.

Self-delivery of training by the person receiving CDCS or his or her legal representative was the most typical form of delivery for staff training, reported by 70.0% of respondents. Least common among the delivery methods suggested on the survey was training via consultants who specifically taught medication administration procedures, which was reported by only 10.0%. Other training delivery formats were generally reported in the range of 30%–40% and are summarized in Table 4, along with training topics.

Table 4

Types and Delivery Formats of Training Offered by CDCS Users

### Wages, Length of Service, and Worker Relationship

Because consumers are afforded the flexibility to recruit their own staff under CDCS, the survey solicited input about the sources that service users relied on to find staff. Immediate family (66.4%) and friends (71.3%) were the two most commonly cited sources of staff recruitment among members of this sample, whereas many people also relied on extended family members as sources of direct support workers (39.3%). Relatively few people used traditional staffing agencies (16.8%), advertisements (17.8%), or other people with disabilities (16.8%) to help them find DSPs.

Hourly wage rates reported by survey respondents ranged from $5.00 to$28.50, with a mean hourly wage of $12.76. Although wage rates were distributed normally on the whole, there were more observations than expected in the tails of the distribution, with 8% of wages reported both two standard deviations above and below the mean. Staff tenures ranged from 1 month to as much as 240 months, with the mean being 34.83 months. Wages departed from the expected distribution in a statistically significant way, χ2(12, N = 151) = 21.46, p = .044, when considering length of service for direct support staff. Direct support workers who had worked a longer period of time with the CDCS user were somewhat more likely to be paid at a higher hourly rate, according to the results displayed in Table 5, which also shows the categorical distribution of wages and tenure as percentages. There was a statistically significant trend for wages to be higher among DSPs for current users of CDCS than they were for employees of past users, χ2(3, N = 93) = 12.59, p = .006. Less than one fifth of survey respondents indicated that they provided fringe benefits such as health or dental insurance, paid leave time, or retirement savings plans to their direct support workers, a rate substantially below what is normally observed in traditional services. Table 5 Association Between Length of Employment and Worker Relationship With Staff Wages Using CDCS The survey solicited information about the relationships of the direct support workers who are hired by users of CDCS. Friends (34.7% of workers identified) composed the largest share of direct support workers, with siblings (16.3%), and extended family members (13.5%) of the CDCS user representing the family share of paid staff. Other individuals who were previously known to the CDCS user (e.g., school paraprofessionals, church acquaintances, parents' coworkers) were used 23.1% of the time, whereas people previously unknown to the CDCS recipient (e.g., workers from staffing agencies, advertisement respondents) constituted the remaining 12.4% of direct support workers reported by members of the sample. Although the association between staff relationship and opinions on whether CDCS made training or retaining staff easier was not significantly different than the distribution that would be expected, there was a significant difference in the distribution when considering staff relationship and whether CDCS makes finding direct support workers easier, χ2(4, N = 77) = 11.48, p = .022, with people who used staff who were known to them being more likely than expected to express that finding staff was easier than individuals who relied on previously unknown staff. Also significantly different from the expected distribution among variables was the association between the relationship of the direct support worker to the CDCS recipient and the wage that person was paid, χ2(12, N = 173) = 21.37, p = .045. Siblings tended to be paid at lower wages, and previously unknown individuals tended to be paid in higher wage categories than anticipated, with people of other relationships being closer to the expected distribution, generally. ## Discussion The primary purpose of this research was to examine how the direct support workforce may look for users of self-directed supports in one Midwestern state. The results above indicate both challenges and benefits among the respondents of this survey. These findings both overlap with existing knowledge about the direct support workforce and use of self-directed supports and generate new knowledge about the workforce trends with use of self-directed funding options for people with intellectual and developmental disabilities. The greatest benefits identified by survey respondents of using CDCS were those related to having the flexibility to select staff who were friends and family members to support the CDCS recipient. Nevertheless, difficulties in finding staff were the most frequently cited challenge with CDCS. This suggests that, although self-directed funding may add flexibility and trust to the hiring process, because individuals may hire people known to them, it may still be difficult to identify friends or family members who are willing to take on the role of direct support worker. Furthermore, questions may eventually arise about how the ability to pay family and friends for their service as direct support staff may impact the view of natural supports if self-directed initiatives are implemented more widely. For example, it may be interesting to see if free caregiving by natural support people, which is typically seen as part of the inherent social capital contract among families of individuals with disabilities, will remain in its current form if payment for direct support work becomes the norm among families. In this study, it was also not known what proportion of support for CDCS recipients was provided without payment, although the amount of unpaid support was assumed to remain large. This study indicates that family caregivers and other individuals previously known to the individual receiving CDCS generally tend to receive hourly pay that is less than agency-based staff or other individuals who were previously unknown, contradicting concerns about financial exploitation. At a mean of$12.76, wage rates of direct support staff in this sample are higher than the mean of staff in either public or private settings that are not self-directed, which were reported at a high of $12.53 (based on 2004 wage data) for public employees in reporting by Hewitt and Larson (2007). Although these wage data are still low compared with other professions, this increase represents an advance for the wages of direct support workers, especially those in the private sector, as well as for CDCS users, who report that one of the greatest benefits of using self-directed supports is the flexibility to set staff pay. However, a great deal of variation in wages existed in this survey, with the lowest wage ($5.00 per hour) being beneath the state's minimum wage of $5.25, and the highest ($28.50 per hour) being in excess of the hourly rates of many professional services. This variation may be explained to an extent, by the fact that higher wages could be paid to individuals who provide specialized services (e.g., occupational therapy, speech therapy) and the lowest wages may be paid either to minors (who may be exempt from the \$5.25 minimum wage) or to support services that can be paid at a flat rate, such as overnight time. Unfortunately, it is not possible to discern the actual reasons for this large wage range within the scope of this evaluation. Although flexibility in setting staff pay is a hallmark of self-direction, and one of the most commonly cited benefits of CDCS among users, this wide range of wages suggests the possibility that additional training and guidelines may be necessary to guide consumers in setting minimal wages while retaining the flexibility in setting wages that was viewed as a benefit by the majority of survey respondents. Although wages under CDCS are expected to fall within “community standards” (Walker et al., 2009), it may be questionable as to how community standards are defined, because each Minnesota county is charged with determining its own standard in this regard. For example, one may assume that sub–minimum wages are outside of the scope of community standards, although there appears to be little oversight in monitoring this. Other states have established more definite minimum and maximums for pay. With a good degree of latitude built into the range, such an approach may be workable in both providing the flexibility that self-directed supports users desire while increasing equity for direct support staff.

Also important to consider is the fact that less than 20% of this sample offered fringe benefits such as health insurance or paid time off to direct support employees. Although wage levels were higher than might be expected for DSPs in the private sector, the rate at which benefits were offered was substantially less than any of the studies cited by Larson, Hewitt, and Knoblauch (2005). This calls into question the overall value of direct support work under self-directed models of support. Although wage gains for part-time direct support staff may be significant because they are usually not offered benefits packages in agency-based settings, full-time DSPs who work in self-directed models may experience a net loss, because the gains in salary may not compensate for the value-added properties of health insurance, paid leave time, and dental care, which they are far more likely to receive from agency-based work. The exclusion of fringe benefits may also serve to limit the pool of potential workers to support users of self-directed funding options, because many potential DSPs may not be able to commit to work that does not provide health insurance.

The issue of health insurance may be one of particular importance to the long-term sustainability of self-directed models of support and is likely to require a multifaceted solution. One element of the solution is likely to be increased training for users of self-directed support options regarding the importance of the provision of fringe benefits, including health insurance, and how improvements to fringe benefit packages can lead to increased staff retention and better support. In addition, state-level interventions may be necessary in addressing this issue. For example, allowing individuals with intellectual and developmental disabilities and their family members to buy into the state health insurance system at a reduced rate would be a cost-effective and equitable way to reinforce the stability of the direct support workforce for individuals who self-direct supports.

Results indicate that training of direct support workers may be the greatest challenge for users of CDCS. Whereas most potential benefits were highly rated by respondents, having better trained staff was seen as a benefit by only a minority of participants. Only 63% of survey respondents reported that CDCS use made training their DSPs easier. Although this represents a majority of the sample, it is a smaller proportion than those who reported greater ease in finding or retaining staff with CDCS. However, users of self-directed services in this study also reported that their direct support workers were better skilled. This could be related to the more familiar relationship of most of the direct support workers, who were known by the individual receiving supports before they became a service provider about 88% of the time. This initial familiarity with the person receiving supports may be related to greater familiarity with his or her needs. It could also be the case that perceptions of CDCS users of family and friends as being more skillful and responsive may have influenced these results.

Regardless of skill level, however, consumers and their proxies reported relatively little training in categories except for self-delivered training on the specific needs of the individual CDCS user. It is possible that the development of more staff development options that are targeted to the needs of consumers who self-direct may be beneficial in easing training provision. Because training in traditional service delivery settings is usually provided by the agency employing the DSP, there may be more training resources in such settings than are available to individual CDCS users who must self-direct the majority of training. It is possible that locations that are highly invested in self-direction of services may need to offer a menu of training opportunities to users of self-directed supports and their staff to satisfy the training needs inherent in any relationship with a DSP. Training that is specifically targeted at how to find, choose, and retain quality DSPs would be of particular benefit, as would training that is based on the development of staff management skills, not just procedures for operating within the system. Electronic training formats, especially those that are competency-based and customizable, may also provide viable training solutions for self-directed supports.

Last, there were indications that budgeting may have been conducted differently at different points in the study. As stated previously, consumers in nonmetropolitan counties reported lower budgets than consumers in the state's metropolitan area prior to the state's assumption of budgeting in 2005. Although the reasons for this are unclear, the discrepancy did not exist after 2005, when the state took a more central role in budgeting, which had previously been left to the individual counties. This observation invites questions about the balance between increased flexibility, which may be present with higher degrees of local control, and more uniformity in budgeting methodology, which may be present when budgets are derived from the state level. Ultimately, this may be a challenge that will take time to resolve as self-directed models of support continue to gain acceptance, and policymakers will need to remain cognizant of the need for both equity and flexibility as self-directed policies are revised over time.

## Limitations and Future Directions

The current research is limited by a number of factors. Most obvious, the geographic scope of this study is limited to a single state. Although a growing number of states have adopted self-directed initiatives, such options have vastly different designs in differing locations, so the results of this study may not readily apply in states where self-direction is approached in different ways. In addition, the relatively small sample size, although it did include a significant proportion of CDCS users in Minnesota, does not lend itself to generalizability across all states, particularly because Minnesota's system of self-direction has had more enrollments and more complete implementation than is the case in many states (Walker et al., 2009).

Changes in CDCS rules within the time of data collection also may have complicated results. While the study was in progress, the state assumed substantial control over budgeting and developed a list of disallowable expenses for the use of CDCS funds. These developments likely had an impact on the results of this study, because some disallowable expenses directly affected direct support workers. Nevertheless, these results indicate the status of self-direction in Minnesota at this critical juncture in time.

Future studies related to the direct support workforce in self-directed models of support of individuals with intellectual and developmental disabilities are needed to build on current knowledge, as this area of study continues to be in formative stages. Among the foremost concerns in future research may be the need to investigate alternate models for training of DSPs. More in-depth knowledge about the training practices of individuals who self-direct services, as well as their needs in this area, will be necessary to shape enhanced training opportunities and structures. In turn, this knowledge may be used to produce training models that are specifically designed to meet the needs of DSPs, as well as individuals with intellectual and developmental disabilities and their family members who serve as employers. In the future, it will also be essential to update the Community Support Skill Standards (Taylor, Bradley, & Warren, 1996) to be consistent with the competencies required of DSPs in self-directed support settings, because these may then inform the development of competency-based training that is most relevant to this segment of community support services.

Financial aspects of self-direction should be examined in greater detail. The results in this study indicated higher hourly wage rates among direct support workers who were hired by consumers who self-directed, although the range of wages reported by survey respondents was large. These higher mean wages, however, are offset to some extent by much lower proportions of employers who offer health insurance and other fringe benefits than would be expected among traditional agency-based employers. A more thorough investigation into the costs and benefits of such wage and fringe benefit trade-offs may be necessary to determine the long-term suitability of this approach from the standpoint of consumers, DSPs, and policymakers. In addition, from a policy standpoint, it is essential to begin investigating ways to enable access to health insurance for DSPs working in self-directed services, as this would likely be a major step toward stabilizing the direct support workforce in the long term.

Thus far, there has been little systematic evaluation of workforce outcomes under self-directed models compared with the traditional service delivery systems that support individuals with intellectual and developmental disabilities and their families. Conducting such a comparison in future studies may be useful because it would provide data that may assist in isolating particular points of workforce development that need additional attention in each form of service delivery. It would also provide a strong source of data to investigate the relative contribution that self-directed supports may make in addressing the looming workforce crisis in community services for individuals with intellectual and developmental disabilities.

Last, research considering different perspectives will be important in building knowledge about the needs of DSPs who work in self-directed environments. This will be especially necessary to gain the insights of DSPs who are working in such conditions. Gaining an understanding of the training and support needs of DSPs may be helpful in designing systems that will promote greater retention, which may be key to the sustainability of self-directed services in the long term.

## References

References
Alliance for Self-Determination
1999
.
National leadership summit on self-determination and consumer directions and control: Summary of proceedings
.
Cheyenne, WY
Author
.
Dale
,
S.
,
R.
Brown
,
B.
Phillips
, and
B. L.
Carlson
.
2005
.
How do hired workers fare under consumer-directed personal care?
The Gerontologist
45
:
583
592
.
Dillman
,
D.
,
J.
Smyth
, and
L.
Christian
.
2008
.
Internet, main, and mixed-mode surveys: The tailored design method
.
Hoboken, NJ
Wiley
.
Greene
,
J.
2007
.
Issue brief: State approaches to consumer direction in Medicaid
.
Hamilton, NJ
Center for Health Care Strategies
.
Harmuth
,
S.
and
S.
Dyson
.
2005
.
Results of the 2005 national survey of state initiatives on the long-term care direct-care workforce
.
New York
National Clearinghouse on the Direct Care Workforce and the North Carolina Department of Health and Human Services, Office of Long Term Care
.
Hewitt
,
A.
and
S.
Larson
.
2007
.
The direct support workforce in community supports to individuals with developmental disabilities: Issues, implications, and promising practices.
Mental Retardation and Developmental Disabilities Research Reviews
13
:
178
187
.
Hewitt
,
A.
,
S.
Larson
,
S.
Edelstein
,
D.
Seavey
,
M.
Hoge
, and
J.
Morris
.
2008
.
A synthesis of direct service workforce demographics and challenges across intellectual/ developmental disabilities, aging, physical disabilities, and behavioral health
.
Minneapolis
University of Minnesota, Institute on Community Integration, Research and Training Center on Community Living
.
Lakin
,
K.
and
R.
Bruiniks
.
1981
.
Occupational stability of direct-care staff of residential facilities for mentally retarded people
.
Minneapolis
University of Minnesota, Center on Residential and Community Services
.
Lakin
,
K.
,
B.
Hill
,
R.
Bruiniks
,
F.
Hauber
, and
G.
Krantz
.
1983
.
Factors related to job stability of direct-care staff of residential facilities for mentally retarded people.
Journal of Community Psychology
11
:
228
235
.
Lakin
,
K.
,
B.
Polister
, and
R.
Prouty
.
2003
.
Wages of non-state direct support professionals lag behind those of public direct support professionals and the general public.
Mental Retardation
41
:
178
182
.
Larson
,
S.
,
A.
Hewitt
, and
B.
Knoblauch
.
2005
.
Recruitment, retention, and training challenges in community human services: A review of the literature.
In
Larson
,
S.
and
A.
Hewitt
. (
Eds.
).
Staff recruitment, retention, & training strategies for community human services organizations
.
pp.
1
18
.
Baltimore
Brookes
.
Larson
,
S.
,
K.
Lakin
, and
R.
Bruiniks
.
1998
.
Staff recruitment and retention: Study results and intervention strategies
.
Washington, DC
American Association on Mental Retardation
.
Matthias
,
R. E.
and
A. E.
Benjamin
.
2003
.
Abuse and neglect of clients in agency-based and consumer-directed home care.
Health and Social Work
28
:
174
184
.
Matthias
,
R.
and
A.
Benjamin
.
2008
.
Paying friends, family members, or strangers to be home-based personal assistants: How satisfied are consumers?
Journal of Disability Policy Studies
18
:
205
218
.
1996
.
Whatever it takes: Stories of self-determination from the Monadnock region
.
Durham, NH
University of New Hampshire, Institute on Disability
.
Moseley
,
C.
2001
.
Self-determination: From new initiative to business as usual.
Common Sense
8
:
1
7
.
Nerney
,
T.
2001
.
Filthy lucre: Creating better value in long-term supports
.
Ann Arbor, MI
Center for Self Determination
.
Nerney
,
T.
and
D.
Shumway
.
1996
.
Beyond managed care: Self-determination for persons with developmental disabilities
.
Durham
University of New Hampshire Institute on Disability
.
Office of the Assistant Secretary for Evaluation and Planning
2006
.
Supply of direct support professionals serving individuals with intellectual disabilities and other developmental disabilities: Report to Congress
.
Washington, DC
Author
.
Phillips
,
B.
,
K.
Mahoney
,
L.
Simon-Rusinowitz
,
J.
Schore
,
S.
Barrett
,
W.
Ditto
, et al
.
2003
.
Lessons from the implementation of cash and counseling in Arkansas, Florida, and New Jersey (Report 8349-107)
.
Princeton, NJ
Mathematica Policy Research
.
Polister
,
B.
,
K.
Lakin
, and
R.
Prouty
.
2003
.
Wages of direct support professionals serving persons with intellectual and developmental disabilities: a survey of state agencies and private residential provider trade associations. Minneapolis: University of Minnesota, Research and Training Center on Community Living, Institution on Community Integration.
Policy Research Brief
14
(
2
):
Prouty
,
R.
,
K.
Alba
, and
K. C.
Lakin
. (
Eds
).
2008
.
Residential services for persons with developmental disabilities: Status and trends through 2007
.
Minneapolis
University of Minnesota, Institute on Community Integration, Research and Training Center on Community Living
.
Scala
,
M. A.
and
T.
Nerney
.
2000
.
People first: The consumers in consumer direction.
Generations
24
(
3
):
55
59
.
Schmitz
,
C. C.
,
M. G.
Luxenberg
, and
N.
Eustis
.
2005
.
Second interim report: Evaluation of the Consumer Directed Community Supports Service (Contract A67221)
.
Minneapolis, MN
Professional Evaluation Services
.
Schore
,
J.
,
L.
Foster
, and
B.
Phillips
.
2007
.
Consumer enrollment and experience in the Cash and Counseling program.
Health Services Research
42
:
446
466
.
Sciegaj
,
M.
,
J.
Capitman
, and
C.
Kyriacou
.
2004
.
Consumer directed community care: Race/ethnicity and individual differences in preferences for control.
The Gerontologist
44
:
489
499
.
Sciegaj
,
M.
,
K.
Mahoney
, and
K.
Simone
.
2008
.
State experiences with implementing the Cash and Counseling Demonstration and Evaluation Project.
Journal of Aging and Social Policy
20
(
1
):
81
98
.
Simon-Rusinowitz
,
L.
,
K.
Mahoney
,
D.
Loughlin
, and
M. B.
.
2005
.
Paying family caregivers: An effective policy option in the Arkansas Cash and Counseling Demonstration and Evaluation.
Marriage & Family Review
37
:
83
105
.
Simon-Rusinowitz
,
L.
,
L.
Marks
,
D.
Loughlin
,
S.
Desmond
,
K.
Mahoney
,
B.
Zacharias
, et al
.
2002
.
(2002).
Implementation issues for consumer directed programs: Comparing views of policy experts, consumers, and representatives.
Journal of Aging and Social Policy
14
(
3/4
):
95
118
.
Taylor
,
M.
,
V.
, and
R.
Warren
.
1996
.
The community support skill standards: Tools for managing change and achieving outcomes
.
Cambridge, MA
Human Services Research Institute
.
Test
,
D.
,
C.
Flowers
,
A.
Hewitt
, and
J.
Solow
.
2004
.
Training needs of direct support staff.
Mental Retardation
42
:
327
337
.
Walker
,
P.
1999
.
From community presence to sense of place: Community experiences of adults with developmental disabilities.
Journal of the Association of Persons with Severe Handicaps
24
:
23
32
.
Walker
,
P.
,
A.
Hewitt
,
M.
Bogenschutz
, and
J.
Hall-Lande
.
2009
.
Implementation of consumer-directed services for persons with intellectual and developmental disabilities: A national study. Minneapolis, MN: University of Minnesota, Institute on Community Integration, Research and Training Center on Community Living.
Policy Research Brief
20
(
1
):
Young
,
H.
and
S.
Sikma
.
2003
.
Self directed care: An evaluation.
Policy, Politics, and Nursing Practice
4
:
185
195
.

## Author notes

Matthew Bogenschutz, PhD (E-mail: boge0021@umn.edu), Research Associate, University of Minnesota, Research and Training Center on Community Living, 204 Pattee Hall, 150 Pillsbury Dr., SE, Minneapolis, MN 55455. Amy Hewitt, PhD, Senior Research Associate, University of Minnesota, Research and Training Center on Community Living, 204 Pattee Hall, 150 Pillsbury Dr., SE, Minneapolis, MN 55455. Jennifer Hall-Lande, MA, Coordinator, University of Minnesota, Research and Training Center on Community Living, 204 Pattee Hall, 150 Pillsbury Dr., SE, Minneapolis, MN 55455. Traci LaLiberte, PhD, Director, Center for Advanced Studies in Child Welfare, University of Minnesota, School of Social Work, St. Paul, MN 55108.

Editor-in-Charge: David Helm