Substance abuse treatment drop-out from client and clinician perspectives

Abstract

Reasons for premature termination of outpatient substance abuse treatment were evaluated from client and clinician perspectives using qualitative (focus groups) and quantitative (survey) methods in a pilot study (N = 44). The most frequently endorsed reasons were related to individual rather than program characteristics with heavy drug or alcohol use, transportation or financial problems, and ambivalence about abstinence being highly rated by both clinicians and clients. Survey results indicated that clinicians more frequently attributed treatment drop out to individual/client level factors than did clients. Focus group ratings indicated that clinicians felt client motivation and staff connection issues were primary reasons for drop-out, whereas clients indicated social support and staff connection issues. The findings suggest that the development of early therapeutic alliance and active problem solving of potential barriers to treatment attendance may influence treatment retention.

1. Introduction

The National Survey on Drug Use and Health reported approximately 3.9 million people over the age of 12 received some type of treatment for substance abuse in the past year (SAMHSA, 2006). However, early client attrition from substance abuse treatment is a major barrier for successful outcomes typically defined as reduction or cessation of substance use and improved psychosocial functioning or physical health. Rates of first-month attrition in outpatient (non-methadone) substance abuse treatment programs are approximately 30% and drop-out prior to 3 months can be 50% or more (Harris, 1998; Hubbard et al., 1989, Kang et al., 1991; Simpson, 1981; Simpson, Joe & Brown, 1997). Three months of treatment is considered the minimum to see symptom improvement (Katz et al., 2004; Simpson & Joe, 2004; Simpson, et al., 1997), and length of time in treatment has been associated with positive treatment outcomes (Hubbard et al., 1989; 1997; Simpson & Sells, 1982; Simpson et al., 1997; Zhang, Friedmann, & Gerstein, 2003). The current study evaluated reasons for remaining versus leaving treatment from both client and clinician perspectives.

A number of studies have examined demographic and clinical variables among clients who completed or stayed in different types of treatment for a longer period versus those who left early (Chou, Hser, & Anglin 1998; Doumas, Blasey, & Thacker, 2005; McCaul, Svikis, & Moore, 2001; Ross, Cutler, & Skylar, 1997). Other studies have examined drop-out, retention, and attrition within different types of substance abuse treatment such as long-term residential, outpatient, and methadone programs (Beardsley, Wish, Fitzelle, O’Grady, & Arria, 2003; Grella, Hser, Joshi, & Anglin, 1999; Joe, Simpson, & Broome, 1999), as well as focusing on the effect of specific drugs such as cocaine or methamphetamine (Maglione, Chao, & Anglin, 2000; Rawson et al., 2000). Specific demographic characteristics associated with drop-out from substance abuse treatment include being: a member of a minority group (Agosti, Nunes, & Ocepeck-Welikson, 1996; Veach, Remley, Kippers, & Sorg, 2000), unemployed (Veach, et al., 2000), and younger in age (Agosti, et al., 1996; Grella et al., 1999). Gender is an inconsistent predictor of treatment drop-out (Greenfield et al., 2007). Clinical predictors of dropout include: early onset of substance use (Agosti, et al., 1996), increased alcohol severity (Martinez-Raga et al., 2002), polydrug abuse (Fishman, Reynolds & Riedel, 1999; Wickizer, et al., 1994), impaired coping, as well as lower motivation and social support (Anderson & Berg, 2001; Dobkin, De Civita, Paraherakis, & Gill, 2002). Ball, Carroll, Canning-Ball and Rounsaville (2006) conducted a retrospective assessment of reasons for early attrition and found client motivation and conflicts with program staff were most frequently endorsed. To date, no research has prospectively evaluated reasons for client drop-out from the clinicians’ perspective and made comparisons to drop-out reasons from the clients’ perspective within an adult substance abuse outpatient treatment sample. Only one study has previously examined both clients’ and clinicians’ perceptions of barriers to attending treatment. Mensinger, Diamond, Kaminer, and Wintersteen (2006) utilized the Perceived Barriers to Treatment scale (Diamond & Kaminer, 1998) post-treatment among adolescents and their therapists to retrospectively assess factors which made it difficult to attend treatment within the Cannabis Youth Treatment project. They found few differences in perspectives, except that therapists’ ratings of the “treatment compatibility” (i.e., the therapist and the program) was related to treatment attendance whereas adolescents’ rating of compatibility was not.

Shared perspective between clients and clinicians with regard to agreement on goals, tasks, and reasons for treatment are considered by many to comprise the core components of a working or therapeutic alliance (Bordin, 1979). Several studies have suggested the importance of the therapeutic relationship and characteristics for treatment retention (De Weert-Van Oene, Schippers, De Jong, & Schrijvers, 2001; Joe, Simpson, Dansereau & Rowan-Szal, 2001; Meier, Donmall, McElduff, Barrowclough & Heller, 2006). However, these results have been inconsistent based on whether the alliance was rated by clients, clinicians, or observers (Barber, et al., 1999; 2001; De Weert-Van Oene, et al., 1999, 2001; Fenton, Cecero, Nich, Frankforter, & Carroll, 2001; Meier et al., 2006). Thus, further research on the discordance or concordance of client and clinician perspectives on goals, tasks and reasons for treatment may shed light on the quality of the therapeutic alliance which in turn could lead to improvements in retention or outcomes.

This study is the first to examine client and clinician perspectives of contributory factors to reasons for treatment drop-out and focused on both individual concerns and program/staff factors that may impact retention. We used both quantitative (survey) and qualitative (focus groups) methods to evaluate reasons for retention versus drop-out from treatment. Consistent with Ball et al.’s (2006) study of client perceptions, we predicted that the most common reasons for drop-out would be individual or personal factors such as motivation and limited support from family/friends, as well as program related factors such as limited connection with staff. We also predicted that clients and clinicians would differ in their opinions on why clients drop out. We expected that clients would report more staff/program-related issues (e.g., program rules) and logistical issues (e.g., transportation) than clinicians who would report more client/personal-related reasons (e.g., motivation).

2. Method

2.1 Participants

The client sample (n = 22) ranged in age from 20 to 57 with a mean of 39.1 (SD = 9.8). The majority (73%) were men, and 50% were African American, 41% Caucasian, 4% Hispanic/Latino, and 4% Multi-Ethnic. Clients reported initiating treatment an average of 4.8 (SD = 3.4) times in their lifetime. At the time of assessment, 36% had been in treatment approximately 1–4 weeks, 14% for 5–8 weeks, 23% for 9–12 weeks, 23% for 3–6 months, and 5% for 1–2 years.

The staff sample (n = 22) ranged in age from 25 to 76 with a mean of 43.8 (SD = 13.9). Approximately 64% were female, 52% were Caucasian, 24% African American, 10% Hispanic/Latino, 5% Asian/Pacific Islander, 5% Native American, and 5% Multi-Ethnic. Staff averaged 10.1 (SD = 7.9) years of experience in substance abuse counseling, 6.1 (SD = 7.7) years of supervisory experience, and 7.1(SD = 7.6) years working for the clinic. The average years of education was 15.2 (SD = 4.8), and 41% had obtained a Masters degree and 27% a Doctorate.

Both clients and clinicians were recruited from two outpatient programs located in the same building in New Haven, CT. The Connecticut Mental Health Center, Substance Abuse Treatment Unit (SATU) and The APT Foundation, Central Treatment Unit (CTU) both provide substance abuse counseling services primarily to low or no income residents of the Greater New Haven area. Each clinic receives a significant number of mandatory referrals for substance abuse evaluations from the Departments of Probation, Parole, and Children and Family Services. Both programs admit a wide range of patients including those with co-occurring, non-acute mental health conditions and a range of substances (alcohol, opiates, cocaine, marijuana). At the time of this study, the average daily census for the CTU clinic was approximately 68 patients and census for the SATU clinic was approximately 313 patients.

Standard outpatient care at these clinics begins with an initial evaluation and, if appropriate for outpatient treatment, patients are admitted into group or individual substance abuse counseling 1–2 times weekly. Both programs utilize a mixture of motivational enhancement, cognitive-behavioral, and 12 step counseling approaches. Patients at the clinics typically are expected to attend treatment for a minimum of 12 weeks, however, some continue in treatment for extended periods based on individual needs.

2.2 Measures

An 18-item version of the Reasons for Leaving Treatment Questionnaire (RLTQ; Ball et al., 2006) consists of two subscales: 1) individual level reasons, and; 2) treatment program related reasons for leaving treatment (see Table 1 for scale items). Clients and staff rated the same items, but with different instructions: 1) Staff -- “Consider all clients you may know that recently have left treatment early. Please estimate on a scale from 1 (not at all) to 5 (very much) how much the following reasons for leaving treatment early might have been true for these clients;” 2) Clients -- “The next page lists some problems or concerns that people have that may influence their decision to leave treatment early. Please rate on a scale from 1 (not at all) to 5 (very much) how much these reasons for leaving treatment are true for you.” The scale reliability was high (α = .93) for the full sample (n = 43)¹, but lower for staff (α = .72) than for clients (α = .95).

Table 1.

Responses from the Reasons for Leaving Treatment Questionnaire

Items	Clients (n = 21)	Clinicians (n = 22)
Individual	(α = .93)	(α = .55)
Physical or mental health problems	1.52 (1.21)	2.50 (.86)**
Transportation or financial problems	2.33 (1.39)	2.86 (.83)
Unsure if needed to stop using	2.00 (1.76)	2.81 (.93)
Family responsibility or problems	2.33 (1.43)	2.55 (.80)
Heavy drug or alcohol use	2.62 (1.75)	3.14 (.94)
Feel could get better on own	1.43 (1.47)	2.05 (1.05)
Unmotivated to keep appointments	1.43 (1.36)	2.64 (.73) **
Limited support from family/friends	1.62 (1.59)	2.45 (.91)*
Regret what said or did at program	1.67 (1.71)	.86 (.65)*
Little hope in ability to change	1.52 (1.54)	2.48 (.93)*
Total	18.47 (11.95)	24.05 (4.02)*
Program	(α = .87)	(α = .66)
Conflict with people at program	1.14 (1.23)	1.27 (.88)
Disagree with what staff expects	1.52 (1.40)	2.18 (.91)
Concern about personal privacy	1.62 (1.35)	1.77 (1.15)
Dislike or distrust people at program	1.67 (1.28)	1.73 (.94)
Staff not helpful or respectful	1.43 (1.36)	1.14 (.94)
Program hours or location a problem	1.05 (1.11)	2.14 (1.08)
Dislike program services or rules	1.33 (1.46)	1.86 (.89)
Need help of a different program	1.19 (1.21)	2.14 (.66)
Total	10.95 (7.61)	14.10 (3.97)
Total score	29.43 (19.21)	38.20 (6.51)*

Note. Values represent the mean (SD). α = coefficient alpha

^*p<.05,

^**p<.01.

2.3 Procedures

Participants were asked to complete a demographics form, an audiotape consent, and the RLTQ before participating in a focus group discussion. Participants were compensated $20 and provided with pizza and soda after the focus group. All study procedures were approved by the Yale Medical School Human Investigation Committee. A total of seven focus groups (4 clinician, 3 client) were conducted at the two sites. Appendix A provides the instructions and open-ended questions used to conduct the focus groups. The staff size at one clinic necessitated conducting a focus group with the supervisory staff and a separate group for the counselors. Clients were invited to participate in the focus groups via flyers posted in the waiting rooms and announcements made in therapy groups. Clinicians were recruited by announcements made in staff meetings. The focus groups lasted 90 minutes and were lead by two research staff with one person conducting the group and the other writing and summarizing participant responses.

Qualitative responses were rated for thematic content using a quantitative assessment developed by the senior author (SB) to rate participant reasons for leaving treatment. The focus groups were audio taped, fully transcribed, and then rated independently by the first three authors (RP, MM, AP) using the 12 rating categories. Cases of disagreements were resolved by the senior author (SB) (see Appendix B for focus group rating guidelines). The frequencies of endorsed reasons were used to create scores for two subscales and their respective items: 1) individual issues scale, consisting of two subscales: intrapersonal concerns and psychosocial concerns, 2) program/staff related issues scale.

2.5 Data Analysis

This report used summary statistics (means, frequencies, percents) for item endorsements and one-way ANOVAs to examine group (client, clinician) differences on both the survey (individual-level reasons; program-related reasons) and focus group (individual issues; program/staff-related issues) scales. We explored item-level differences only when there was a significant subscale difference to provide some control for multiple comparisons. Paired t-tests were used to assess differences between the individual-level and program related scales on the RLTQ and the focus group ratings within the groups (i.e., client and clinician). In addition, on the focus group ratings, a frequency count was conducted separately among clients and clinicians for the number of times a category was endorsed and a percentage was calculated based on the number of these responses relative to the frequency of all response categories. In addition, we included qualitative responses for descriptive purposes.

3. Results

3.1 RLTQ Scores and Items

Clients scored lower than clinicians on their total RLTQ scores, F (1,41) = 4.08, p < .05. Both clients, t (20) = 6.04, p < .001, and clinicians, t (21) = 10.07, p < .001, reported significantly more individual-level reasons than program-level reasons for drop-out. When the two subscales were examined between groups, clients reported significantly lower endorsements of the individual-level subscale than clinicians, F (1,41) = 4.28, p < .05, and there were no differences between clients and clinicians on the program-related subscale of the RLTQ. Further analyses of items within the individual-level subscale revealed that clinicians reported significantly higher ratings than clients for “physical or mental health problems,” F (1,41) = 9.38, p < .01, “unmotivated to keep appointments,” F (1,41) = 13.32, p < .001, “limited support from family/friends,” F (1,41) = 4.49, p < .05, and “little hope in ability to change,” F (1,41) = 5.90, p < .05. The only item where clients reported higher endorsements than clinicians was for “regret what said or did at program,” F (1,41) = 4.09, p < .05. An examination of the more highly endorsed individual-level reasons indicated that both clients and clinicians rated heavy drug or alcohol use, transportation or financial problems, and ambivalence about stopping as common reasons for drop-out (see Table 1).

3.2 Focus Group Ratings

Overall, clients and clinicians reported similar reasons for drop-out in the focus groups. Frequencies and percentages of endorsements within the individual issues category and the program/staff-related issues category were tested which indicated that the frequency of responses between clients and clinicians did not significantly differ. As well, within client and clinician groups the frequency of endorsements of individual versus program/staff related issues were tested indicating no significant differences (see Table 2).

Table 2.

Frequencies of Ratings for Focus Group Responses

Response Category	Client Groups (n = 3)	Clinician Groups (n = 4)
Individual Issues:	249 (65)	312 (66)
Intrapersonal Concerns	152 (40)	172 (36)
Negative Reactions	51 (13)	54 (11)
Health Concern	0 (0)	7 (1)
Motivation/Readiness to Change	56 (15)	76 (16)
Substance Use Recovery	45 (12)	35 (7)
Psychosocial Concerns	97 (25)	140 (30)
Limited Resources	4 (1)	42 (9)
Referral Sources	12 (3)	47 (10)
Life Stressors	7 (2)	14 (3)
Social Supports	74 (19)	37 (8)
Program/Staff-Related Issues:	134 (35)	162 (34)
Reactions and Relations with other Patients	34 (9)	10 (2)
Staff Limitations or Connection Issues	68 (18)	69 (15)
Confidentiality/Privacy Concerns	5 (1)	28 (6)
Program Services, Rules, or Expectations	27 (7)	55 (12)
Total number of responses rated	383 (100)	474 (100)

Note. Focus groups included clients (n = 21) and clinicians (n = 22). Values represent the percentage (%) of times a category was endorsed in each focus group.

^*p<.05

^**p<.01

Clients most frequently reported: social supports (19%), staff limitations/connection issues (18%), and motivation/readiness to change (15%). Clinicians differed minimally in their patterns of responses and frequently indicated: motivation/readiness to change (16%), staff limitations/connection issues (15%) and program services, rules or expectations (12%).

3.3 Focus Group Responses

We provide several examples of focus group responses to further illustrate some of the common and different perspectives of drop-out reasons among clients and clinicians reported in the rating scale analyses reported above.

3.3.1 Common perceptions among clients and clinicians

Both participant groups identified “staff limitations/connections” and “client motivation” as common reasons for drop-out.

3.3.1.1 Staff Limitations/Connections

Clients repeatedly emphasized the importance of a connection or working alliance with the program staff as important for retention. Clients conveyed a sense that they wanted clinicians to care and be invested in their recovery.

Try to connect more but I think empathizing with us is important. I mean we read a lot when we see people. What they can do really is get more involved, don’t just treat us like a child, or like it’s just a job that they do for eight hours and then their out. They need to listen more.

Staff also acknowledged the challenge of meeting the demands of their caseloads and how their interactions with clients can either add or detract from the connection they develop.

Quality before quantity, I mean I think we all provide quality care for our clients for the most part but sometimes find quantity getting more time. We need more time to not treat the clients like they’re “these people” or “those people” and just treat them like they’re somebody who’s trying to get treatment for whatever it is that they need treatment for.

3.3.1.2. Motivation/Readiness

Clients readily endorsed that they had experienced times when they had entered treatment and were not ready to engage in treatment and dropped out.

I dropped out early a couple times. Wasn’t ready, you know, didn’t think it was that bad, I could handle it. You know, felt I just wanted to keep running. Then over time, after running for a few years, I voluntarily went on a program, stayed clean, followed it till I came here.

Clinicians also frequently reported clients were not yet ready to change their substance use or see their use as a problem.

And they feel that other than getting arrested for buying it there’s no real negative consequences yet. So they don’t see it as a problem. Every once in a great while we get someone that’s mandated that comes in starts off saying, “I don’t have a problem,” and then part way through all of the sudden realize that well “yeah maybe it is” and they reach that awareness.

3.3.2 Different perceptions among clients and clinicians

Clients and clinicians differed in the emphasis placed on a couple of areas related to drop-out.

3.3.2.1 Program Services, Rules or Expectations

Clinicians reported that the treatment initiation process and clients lack of familiarity with treatment programs acted as barriers to remaining in treatment.

I think that a lot of clients come from unstable households, and they don’t understand what treatment is, what treatment means, so they drop out right away ‘cause they don’t understand the meaning of treatment.

3.3.2.2 Social support

Clients often reported that treatment retention was related to the support they received from family, friends, their church, and others in recovery.

I think listening, my family listens more. I don’t know, sometimes when you’re an addict you want to reach out and tell them you’ve got a problem but you just don’t know how and sometimes it’s tough to talk to your family about being, you know. Probably encouragement, if they encourage you to get help, you know, stay on top of things.

4. Discussion

Consistent with our prediction and prior findings (Ball et al, 2006), we found that the most commonly reported reasons for drop-out were individual or personal factors rather than program related factors. However, our expectation that clients would report more staff/program-related and clinicians who would report more client/personal-related reasons was only partially supported. Client and clinician responses to our quantitative survey and qualitative focus group methods indicated more similarities than differences. Overall, clinicians endorsed higher levels of RLTQ drop-out reasons than clients. Although both groups scored higher on individual than program related-reasons, clinicians endorsed significantly more individual-related reasons than clients. Item-level analyses of the individual level subscale highlighted interesting similarities as well as differences between the groups. Although both groups reported that substance use, motivation, ambivalence, transportation/financial difficulties, and staff connection issues (evident in both survey and focus group responses), they significantly differed on individual/client level reasons such as physical or mental health, motivation, limited support, regrets about behavior at the program, and little hope in ability to change. Our expectation that clients would report more staff/program-related issues (e.g., program rules) and logistical issues (e.g., transportation) was not supported.

Data from the current pilot study should be interpreted with caution as this study was limited by the small sample size, self report data, and the predominance of men within the client sample. In addition, almost half of the client sample had been in treatment for approximately 2–6 months, possibly making it difficult for them to consider potential reasons for dropping out. However, clients on average reported multiple prior treatments and so likely had prior experiences in reference to the treatment engagement and drop-out process. It also should be noted that client and clinician differences (both in subscale scores and internal consistency) on the RLTQ might be attributed in part to differences in the instructional set for these two participant groups. Specifically, clinicians were instructed to “consider your caseload,” thus covering a broad range of patient and a broader range of drop-out reasons than would be considered by one client completing the questionnaire about him/herself.

Despite these limitations, this study provides some insight into clinician and client perspectives on substance abuse treatment drop-out. The quantitative and qualitative measures highlighted both differences and similarities in their perspectives. Both clients and clinicians emphasized the importance of staff connection issues and motivation, and this may be an important early treatment goal given the high rates of drop-out within the first 3 months of treatment. Other important drop-out reasons noted by both groups could be influenced by the development of an early therapeutic alliance, with a specific focus on enhancing client trust, engagement, and motivation. In addition, clients may need assistance in considering potential barriers to attending treatment (i.e., transportation or financial difficulties) as well as assistance with coping. Therapists may consider utilizing techniques to rapidly build and maintain the therapeutic alliance in addition to problem solving practical barriers to promote better retention in treatment.

Appendix A

Participant instructions

“We should first offer some guidelines for how the group will be conducted. It is important for you to know that we feel there are no right or wrong answers; it is important for you to share your views and opinions, not what you may think we want to hear. We would appreciate if each of you took turns in answering the questions and try not to interrupt while others are speaking. We would like to emphasize that we are here to learn from you and would really like to know your thoughts on these questions. We are happy to answer any questions if something is unclear or if you would like a question repeated. Also, we want to respect everyone’s confidentiality so we ask that the things said in this room remain here. Does anyone have any questions? Before we start the group we would like to go through the consent form.”

Open ended questions used in the focus groups

1. Based on your own experience and people you’ve known, what are the most common reasons people drop out of substance abuse treatment early?

2. What could the person who dropped out have done differently to prevent dropping out?

3. What could the family or friends of the person who dropped out done differently to prevent treatment drop out?

4. What could the program or staff done differently to prevent this drop out?

Appendix B

Focus Group Rating Guidelines

Instructions: Read through the transcript and write one or more rating numbers for each sentence. If the next sentence is clearly identified as spoken by the same individual and simply elaborates upon the prior sentence, write the same rating number(s) above but in parentheses. If additional content is added into the next sentence, then write one or more new rating numbers above without parentheses the first time it occurs and with parentheses if repeated in subsequent sessions. Once a new speaker begins talking, then the process begins again with no parentheses on the first introduction of a category. When a speaker is interrupted by a full comment or question made by another group member or facilitator (discounting short phrases encouraging communication), the next statement should be rated without parentheses even if it obviously continues a point they were making before being interrupted.

Patient-Related Issues

• 1Negative Reactions: e.g., fear; distrust; anger; pessimism; avoidance; specific reaction to an event; negative attitude or expectancies; poor coping, planning, or problem solving; blaming; externalizing responsibility; overconfident with minimization of problems; anxiety about expectations or consequences of treatment
• 2Health Concern: e.g., psychological distress; psychiatric symptoms; medical illness
• 3Motivation/Readiness to Change: e.g., denial; resistance; ambivalence; treatment not a priority; hope or optimism about change; low confidence in treatment or ability to change
• 4Substance Use Recovery: e.g., relapse; recover is challenging; limited knowledge of addiction; need different or more intensive treatment; implementing recovery tools and making lifestyle changes at home; insufficient participation in treatment

Program/Staff-Related Issues

• 5Reactions and Relations with other Patients: e.g., negative reactions to group members; lack of connection or cohesion with fellow patients; reminders of substance use lifestyle; lack of safety with other patients; feeling different from other patients and don’t believe can help
• 6Staff Limitations or Connection Issues: e.g., weak working alliance; lack of positive reinforcement; lack of education about treatment and recovery; inflexibility or lack of individualization in counseling; unrealistic expectations; failure to follow-up; lack of caring; poor communication; problematic boundaries or limit setting; staff enabling; authority outweighs collaboration
• 7Confidentiality/Privacy Concerns: e.g., distrust of peer maintaining confidentiality; fear of exposure to those known in the community; worried about staff sharing information with referral source or significant others; worried about health and life insurance labeling as substance abuser; information sharing across groups
• 8Program Services, Rules, or Expectations: e.g., confusion about expectations; rule changes or inconsistencies; unresponsive procedures; feared consequences; services not matching needs; insufficient attention to other psychosocial needs; insufficient involvement of family in treatment; limited treatment options

External Issues

• 9Limited Resources: e.g., transportation, finances; child care; housing or employment instability; insurance
• 10Referral Sources: e.g., criminal justice; child protection; return to jail; work or health care referral
• 11Life Stressors: e.g., death, serious illness, or loss of significant others; loss of job; domestic violence; housing eviction; loss of basic services; arrest of self or others; trauma
• 12Social Supports: e.g., prefer to make changes without other’s help; lack of support from family or peers for recovery; active use in home; substance using peer group; seeking or failing to seek support from others including self-help groups