Couple and family therapy is effective for resolving a variety of presenting problems (Blow et al. 2009; Lambert 1992; Shadish and Baldwin 2003; Sprenkle et al. 2009); though many clients therapy before they reach their goals (Bohart and Wade 2013). Though the individual psychotherapy literature has addressed premature termination (also called attrition, discontinuation, dropout, premature dropout, and early withdrawal), research on premature termination in relational therapy shows that clients who leave treatment prematurely tend to be less satisfied with therapy, (Lebow 1984) are less likely to improve, (Pekarik 1986; Saatsi et al. 2007; Westmacott et al. 2010) and are less likely to seek therapy with another treatment provider (Garfield 1986; Hamilton et al. 2011). Despite the benefits associated with therapy persistence, premature discontinuation is a frequent and problematic occurrence, with a mean dropout rate of 19.7%, and a range from 0 to 74% (Zimmermann et al. 2016; Barrett et al. 2009; Swift and Greenberg 2014; Kegal and Flückiger 2014; Westmacott et al. 2010).

Though there has been discussion about premature termination in individual therapy, research on premature termination in relational therapy (couple and family therapy) is limited (Bischoff and Sprenkle 1993). Premature termination is a particularly frequent and problematic in relational therapy. Hamilton et al. (2011) found that family therapy clients are 33.2% more likely to drop out of treatment than clients of individual therapy. Given the increased frequency of dropout in couple and family therapy, it may be possible that couple and family therapy clients face more barriers to treatment than individual therapy clients such as scheduling, level of motivation and childcare (Masi et al. 2003; Kazdin et al. 1997). They also found that clients who experience more barriers may be more likely to leave therapy prematurely.

Between 40 and 60% of couple and family therapy clients end therapy without discussing this decision with their therapist (Hamilton et al. 2011; Marchionda and Slesnick 2013; Mueller and Pekarik 2000; Wierzbicki and Pekarik 1993). The majority of clients who drop out of therapy do so after one session (Odell and Quinn 1998; Phillips 1987). These findings show that clients are dropping out of couple and family therapy often, and quickly. Premature dropout presents problems for clients who need treatment, but do not receive it, and therefore, may experience continued or worsening distress (Kazdin 1990; Bischoff and Sprenkle 1993).

To prevent premature discontinuation, couple and family therapists need a better understanding of why clients leave treatment before completing it (Westmacott et al. 2010; Kegal and; Flückiger 2014). Therefore, the purpose of this study is to develop an understanding of how clients make decisions about therapy persistence and premature discontinuation. To accomplish this aim, we interviewed current clients of MFTs who had discontinued therapy with a previous provider about their decision-making process of deciding to discontinue therapy.

Background

Gaps in Premature Dropout Research

Historically, there has been significant empirical attention devoted to identifying demographic factors associated with premature discontinuation and a lack of attention to the client perspective of premature discontinuation. Several client factors have been identified as dropout predictors, including ethnic minority status, race, low SES, and more severe symptoms (Kazdin et al. 1997), while gender and psychiatric diagnosis have not been found to predict dropout rates consistently (Garfield 1994; Masi et al. 2003). Similar to research literature on individual therapy, MFT research has primarily examined demographic predictors of dropout (Kazdin 1990; Pekarik and Stephenson 1988; Weisz et al. 1987). There has been a lack of qualitative research focused on premature dropout, as reaching the dropout client has been a longstanding challenge (Bischoff and Sprenkle 1993; Shadish et al. 1998). Research on therapist and process variables related to dropout is less common, and findings are inconsistent (Hamilton et al. 2011). Therefore, little is known about the in-session therapy processes that make therapy feel effective to clients, (Davis and Piercy 2007a, b) and support treatment retention (Marchionda and Slesnick 2013; Pinsof and Wynne 2000).

Defining Premature Discontinuation

Problems related to researching premature dropout are often connected to problems with how premature discontinuation is defined, and who should define it (Bischoff and Sprenkle 1993). The first of four commonly used methods uses a predetermined number of sessions to define discontinuation; clients who terminate therapy before reaching the predetermined number of sessions are identified as having dropped out of therapy (Bischoff and Sprenkle 1993; Hamilton et al. 2011). The second method relies on the therapists’ definition of a proper termination point, and clients who end therapy before this point are considered dropouts (Bischoff and Sprenkle 1993; Hamilton et al. 2011; Sledge et al. 1990). The third method combines a set number of sessions with the therapist’s assessment of the treatment case, while the fourth method involves identifying dropout as clients who present for a single session and fail to return (Bischoff and Sprenkle 1993; Garfield 1986; Hamilton et al. 2011). Each method poses unique benefits and challenges. Determining dropout based on a predetermined number of sessions as a cutoff is likely an arbitrary choice (Bischoff and Sprenkle 1993; Hamilton et al. 2011; Reis and Brown 1999) that does not accurately reflect the clients’ goals and experience of treatment. Therapist defined premature termination is likely to be subjective, because it omits the clients’ perspective of the treatment. Therapists’ ideas about the number of sessions necessary to reach the goal may differ from their clients’ ideas (Bischoff and Sprenkle 1993).

These methods omit the clients’ perspective of the treatment, which is problematic, as client perception of the therapeutic process is often the most reliable method of defining dropout (Horvath and Symonds 1991). It is critical to examine the client perspective of therapy, as clients control whether they continue therapy or stop therapy (Blow et al. 2007; Horvath and Symonds 1991; Oka and Whiting 2013). Premature discontinuation occurs unilaterally by the client, rather than through agreement with the therapist; (Swift and Greenberg 2014) therefore, using clients’ definitions of premature dropout would provide insight into their decision-making process when deciding to persist or discontinue therapy.

Treatment Goals

A useful way to measure therapy progress is to measure client progress toward achieving their therapy goals (Woodward et al. 1978). Research shows that progress toward treatment goals may contribute to clients’ perceptions of therapeutic effectiveness (Bohart and Wade 2013). Clients may become dissatisfied with therapy when they do not perceive progress toward their goals (Blow et al. 2009; Snyder et al. 1997). It is likely that during therapy, clients assess therapy effectiveness and use this assessment to weigh the costs and benefits of continuing therapy versus stopping therapy and continuing to experience the problem (Walitzer et al. 1999). Clients who discontinue therapy after one or two sessions have outcomes that are equivalent to people who never begin therapy (Masi et al. 2003). Clients who discontinue therapy before reaching their goals are often dissatisfied with treatment, and report that their therapist was not helpful, (Bohart and Wade 2013; Hamilton et al. 2011; Kegal and; Flückiger 2014; Lebow 1984; Pekarik 1986).

Little is known about the therapy processes that make therapy feel effective to clients and support therapy continuation (Davis and Piercy 2007a, b; Marchionda and Slesnick 2013; Pinsof and Wynne 2000). To develop a deeper understanding of clients’ experience of premature therapy termination, the aim of this study is to develop a theory of how clients make decisions about therapy persistence or discontinuation. The research question that guided the present study is: How do clients of marriage and family therapists decide to persist in therapy or decide to prematurely discontinue therapy?

Method

Grounded theory methods are indicated when the aim of the study is to construct theory (Glaser 1987; Glaser and Strauss 1967; Strauss 1987; Strauss and Corbin 1990a, 1998). Grounded theory was selected as the method for the present study because it supports the study aim of developing a theory about therapy continuation or discontinuation.

Participants

Participants were adult clients (over age 18) of marriage and family therapists (MFTs), currently being seen by an MFT for individual, couple or family therapy. The sample included 19 participants, ten men (53%) and nine women (47%) ranging from 20 to 77 years old (M = 41.3, SD = 15.62). Fourteen out of the 19 participants had previously been in therapy with another therapist prior to the current therapist. Thirty-seven percent (n = 7) of the participants were married, 26% (n = 5) were single, 16% (n = 3) were divorced, 16% (n = 3) were living with a partner, and 5% (n = 1) were widowed. The sample was evenly balanced between relational and individual therapy with 11 participants (58%) currently in relational therapy (nine in couple therapy, two in family therapy) and eight (42%) participants currently in individual therapy. Most participants had completed more than nine therapy sessions (n = 11, 58%) with their current therapist. Six participants (31%) completed between five and eight therapy sessions with their current therapist, while two participants (11%) completed less than four sessions with their current therapist (M = 2.23, SD = 0.87). See Table 1 for complete demographic information.

Table 1 Demographic characteristics of therapy client participants
Full size table

MFT researchers have called for research of therapy practiced in treatment-as-usual conditions (Wampler and Bartle-Haring 2015), which was the aim of this study. Consistent with this aim, we recruited a convenience sample (Patton 1990) of 19 participants from three community-based therapy centers that are owned and staffed by MFTs. In these centers, treatment is conducted by MFTs who treat individual, couple and family clients from a relational orientation. Therapists practiced a variety of therapeutic models and varied in experience level from student interns to therapists with over 10 years in practice. At the time of the study, American Association for Marriage and Family Therapy (AAMFT) Approved Supervisors supervised student therapists. We did not have opportunities to collect specific information from participants’ current or previous therapists in order to minimize disruption to clinic flow and participant concern that their therapist could access their responses. Though all participants were currently in therapy, several participants were eager to discuss their past therapy experiences. We allowed these discussions to emerge and reported on the results because these experiences had an impact on the participants’ current course of treatment.

Procedures

A recruitment poster, a locked box with a drop slot, blank note cards, and pens were displayed in therapy center waiting rooms. Participants left their first name and telephone number or email address to indicate interest in participating. During the initial contact, a member of the research team explained the study purpose and confirmed that the participant met inclusion criteria (over age 18, in therapy with an MFT). Individual interviews were conducted in a private room within the therapy center and lasted 1 h. After completing the consent form, the audio-recorded interview was conducted. After the interview, participants completed a brief questionnaire, and were given a $20 grocery store gift card and were informed that a de-identified copy of the findings would be available upon request.

Interview Protocol

The semi-structured interview protocol contained open-ended questions related to participant perception of the therapeutic process including therapy goals, therapeutic relationship, and the factors they considered when deciding whether to stop therapy, or persist.

If participants had dropped out of therapy with a previous provider, we asked them to describe how they arrived at the decision to drop out. We specifically asked them to discuss the factors they considered when making this decision. Participants described a time when they really looked forward to their next therapy session, as well as a time when they felt that they were making little or no therapeutic change. We also asked participants to describe their perception of progress toward their therapeutic goals, and their relationship with their therapist.

After the interview, we collected participants’ demographic information using a survey. Participants reported their age, ethnicity, gender, relationship status, number of sessions completed, and therapy constellation. We also asked participants to select the category that described how they currently felt about continuing therapy: “I will stop therapy because I’ve made change,” “I will continue coming to therapy because I am making good progress toward change,” “I will continue coming to therapy, but I am not making progress toward change,” “I will stop therapy because I am not making progress toward change” (Adams et al. 1991).

Data Analysis

Grounded theory is useful for investigating research questions about the therapeutic process because such questions usually refer to client meaning, perception, and understanding of therapy (Echevarria-Doan and Tubbs 2005). We analyzed interview data using Strauss and Corbin’s (1990) grounded theory triadic coding scheme. After each interview was conducted, the audio recording was transcribed, and coded line by line. First, each transcript was coded line by line using open coding. The coding process became increasingly focused around topics of theoretical interest (Charmaz 2006). Categories were worded using the participants’ language when possible, consistent with the concept of in vivo coding (Strauss and Corbin 1998). Following this step, related topics were grouped together to create themes (Strauss and Corbin 1998). Emerging themes were grouped together and labeled conceptually, and connections among themes were examined. The procedure of creating themes and sub-themes involved examining the connections between categories and their relationship to the phenomenon of therapy retention (Strauss and Corbin 1990). Finally, we identified interconnections among the identified categories to develop a theory about therapy retention.

We used Lincoln and Guba’s (1985) criteria (credibility, transferability, confirmability) for demonstrating trustworthiness. During the period of data collection and analysis, the first author met weekly with a peer debriefer. During peer debriefing sessions, we discussed that many of the participants’ experiences were similar to the peer debriefers’ experiences of working with clients in a clinical setting. This process supported credibility of the findings, as it allowed the researchers to become aware of inherent biases that may impact their view of the data (Lincoln and Guba 1985). Peer debriefing also provided the first author with an opportunity to test hypotheses with a person who did not have an interest in the results. Providing several quotations from participants supports transferability, allowing readers to experience the data first hand, and clarifies the connection between the data and the categories and themes (Lincoln and Guba 1985). Providing a thick description also allows readers to evaluate the extent to which the findings are transferable to other situations and participants (Lincoln and Guba 1985). To support confirmability, we employed analyst triangulation in the form of multiple second coders, which allowed the data to be analyzed from multiple perspectives, ensuring that multiple interpretations and explanations were considered and discussed (Lincoln and Guba 1985). The coders met to review the transcripts and codes and resolve any discrepancies in coding.

Findings

Participants’ decisions to persist or discontinue therapy were based on the intersection of three primary factors/categories: client motivation, the therapeutic relationship, and therapy productiveness. The outcome of this assessment resulted in the decision to persist in therapy or to drop out of therapy. In the current course of treatment, 15 participants (79%) felt the problem was “much better,” while four participants (21%) felt the problem was “somewhat better” since beginning therapy. Fourteen (74%) participants intended to continue therapy with the current provider until making the desired changes. Two participants (10%) intended to stop therapy because they had not made change, one (5%) participant was ready to end therapy because change had occurred, and two (10%) participants selected “not applicable” and did not further explain their response. Qualitative analysis led to three categories, with each category supported by themes and subthemes.

Category 1: Motivation to Change

Motivation to engage in the therapy process emerged as a factor in whether participants persisted in therapy or stopped therapy. Four participants (21%) reported beginning therapy expecting that it would not be productive in helping them resolve their presenting problem, and thus approached therapy with low motivation. One participant described feeling that change was not happening during a time when he was less motivated. He said, “I wasn’t that into [therapy] when I first started. I knew I had a problem, and I knew I needed to talk to someone, but I didn’t think it would work.” Four participants (21%) reported that they were initially ambivalent about going to therapy, and were unsure of whether they would return for additional sessions. For example, one participant said, “Every session I came I [thought] I’ll just go one more time. And just to one more time, one more time, one more time, and I really do love my therapist now.” This participant described how feeling that the session was helpful enhanced his motivation to return for subsequent sessions. Three participants who had discontinued therapy with a previous provider described the current course of therapy as more helpful because they are approaching it with increased motivation. For example, one participant said, “Before, I went [to therapy] thinking ‘this is not going to help, I don’t know why people do [go to therapy], this is silly’. Now I have a mindset of ‘I have a problem and I need to address it’.” Another participant said, “I think it was up to me. I had to be ready and until I was ready [a therapist] could have tried everything and I would have said forget it.”

Of the 14 participants who dropped out of therapy prior to their current course of treatment, two participants (14%) explained that they changed therapists when they felt that they were not making progress. Two participants (14%) described ending therapy because they lacked motivation for change. One participant said, “I have done therapy before but have only gone for two or three session each time and thought this was a waste of time, it’s not going to do anything.”

Category 2: Relationship with the Therapist

Participants explained that when they felt comfortable with the therapist they were motivated to continue therapy. Eleven participants (58%) indicated that a strong relationship with their therapist, led them to persist in therapy. In order to feel comfortable and connected with the therapist, participants needed to feel that their therapist would listen without judgement. With a non-judgmental therapist, participants could be vulnerable and able to disclose sensitive, even shameful details of the presenting problem. This ability to disclose sensitive details of the problem led participants to feel that the therapist had an accurate understanding of the presenting problem, and therefore, could facilitate effective therapy. All 19 participants said the therapist must understand the presenting problem and the therapeutic goals for them to persist in therapy. One participant said,

Oh it’s very important that [the therapist] understands. It would be a waste of my time if she didn’t. If the therapist didn’t understand the changes I want to make, it would be like talking to one of my friends. I’m going to see somebody who has been educated to see through [my] words to see what is really underlying.

Another participant explained that comfort with the therapist allowed him to become vulnerable enough to discuss the presenting problem. He said, “I’m nervous and anxious because I don’t like talking about [the problem] but I feel comfortable with her.” When evaluating the therapists’ understanding of the problem, the clients consider things the therapist does in session. For example, one participant described feeling understood by the way her therapist reflects what she says, “Sometimes being patient, reflecting and talking. She doesn’t just repeat back what I say. It’s the way it’s repeated that lets me know I was really heard.” Another participant explained that he knows his therapist understands the presenting problem when the therapist is active in structuring the session. This participant said, “[The therapist] repeats it back - then he let me talk for a bit, and then he stops me and we think about certain parts.”

The therapist’s understanding of the problem reinforced a strong connection with the therapist, which ultimately supported persistence. Conversely, when participants perceived irritation or hostility from the therapist, it they discontinued therapy. For example, one participant said, “I started therapy [with] someone else, and I did not like it. It felt that they were irritated that they had to listen to me. I didn’t feel connected. I didn’t feel like it was doing anything.” Four participants (29%) ended therapy because they felt the therapist did not understand the presenting problem. One participant said, “I felt like the therapist was making broad generalizations. [What she said] didn’t identify with us at all, but [we] couldn’t get her off that track.”

Category 3: Assessing Therapy Productiveness

Participants evaluated therapeutic process elements to determine whether therapy would be productive in resolving the presenting problem. Participants were particularly interested in their therapist’s frame of the problem. Participants wanted a frame of the problem that they could subscribe to, yet felt revolutionary, or moving beyond their initial understanding. Twelve participants (63%) perceived the therapist as having an accurate understanding of the presenting problem when the therapist offered feedback or asked a question that demonstrated their understanding. One participant explained, “I described something to her from my past leading up to stuff that’s going on now. She pin-pointed it and I never saw it the way that she saw it. That sealed the deal. I wanted to come back and see [her] again.” When participants perceived their therapist as having a thorough understanding of the problem, they also perceived the therapist as able to offer a revolutionary understanding of the problem that was above and beyond their own understanding. For example, one participant said, “I’m going to see somebody who is going through the education or has been educated to see through and sift through the words and the paragraphs of the things that I’m telling them to see what is really underlying.” Participants also discussed the importance of the therapist being willing and able to identify a large problem that is difficult to bring to the forefront. One participant said, “She called our [junk]. We were so busy at pointing fingers at each other, we didn’t see it.” Conversely, when the participant perceived the therapist as having an inaccurate understanding of the problem, participants regarded the therapists’ conceptualization as inaccurate. For example, one participant said, “I didn’t like being told ‘this is what you have to do.’ My therapist challenges me, but he does it in a way where he explains it, and why. It’s never like ‘Just do this it will help”.

Twelve participants (63%) intended to persist in therapy when their therapist offered a plan for resolving the problem. One participant described the process that unfolded in the first therapy session that she experienced as positive. She said,

My husband talked, then I talked and then she said, ‘I have some ideas.’ At the end she talked for five minutes and said, here’s what you can do right now to calm things down. Then, after three or four sessions she met us individually and listened. When we got back together she had a laundry list. She just nailed it, she pinpointed okay we need to work on a, b, c, and d.

Above all, participants wanted opportunities to discuss the problem, and a plan for resolving it. One participant reported discontinuing therapy when the therapist failed to address the presenting problem or offer a plan for resolving it. She said,

[the therapist] was a very nice lady, probably too nice, she was very passive. She didn’t talk much and I needed direction. There were times where I asked her, ‘what was my point again’ and she said, ‘I’m not sure’. She wasn’t there in my conversation and I needed direction.

Ten participants (52%) identified therapy pacing as a critical consideration in their decision to persist in therapy or to discontinue therapy. Participants preferred therapy to progress at a gradual and logical, yet steady and progressive pace. Participants were particularly sensitive to therapeutic processes or activities that felt more advanced than their current stage of therapy. One couple therapy participant who had previously dropped out described a time when he felt that therapy moved too quickly. He said, “The second session escalated quickly. We weren’t having sex and our homework assignment was to go home and lay in bed naked together and explore each other’s bodies. That wasn’t our problem, for us it was awkward and stupid.” Four participants (21%) reported that feeling productively challenged by the therapist lead them to decide to persist in therapy. In other words, participants appreciated a therapist who was willing to deliver a tough, but important message. For example, one participant said, “It was important that [the therapist] is not afraid to call me out when I’m not facing [the problem].”

Of the twelve participants in relational therapy, three participants (25%) reported that therapist neutrality lead them to persist in therapy. When participants perceived the therapist as aligned with another member of the client system, they concluded that their own perspective of the problem would not be honored. One participant described his decision to discontinue therapy. He said, “I felt the therapist sided with my ex-fiancé. Her body language, the way she looked at me and tilted her head. She was very cold to me, I picked up on that stuff.” Another participant who ultimately persisted in therapy described a time she considered discontinuing. She said, “There were times where I felt [the therapist] made excuses for [my husband’s] behavior and my issues weren’t addressed.” Participants recognized therapist neutrality, and reported that it lead them to persist in therapy. For example, one said, “[The therapist] looks at me, and it’s just her and I for a minute. Then I wait my turn, then I feel good because [my husband] gets to say what he needs to say.” Another participant said, “[The therapist] seems fair, I want her to listen to [my husband], I want her to help us. So that’s our goal, I know that she’s not against me.”

Therapy Outcomes

Twelve participants (63%) reported that when they noticed changes in the presenting problem or reduction in their symptoms, they persisted in therapy. One participant said, “I have not considered [discontinuing my current therapy] because it seems to be helping and moving forward. We’re actually looking to bring our kids into family therapy rather than just my wife and I.” Another participant said, “[I notice] progress in general. When I look back at the person I was like a year ago I’m such a different person now. I’ve resolved a lot of issues with my family.”

Fourteen participants (74%) reported leaving therapy because they did not feel it was effective in resolving the presenting problem. Many participants felt it would be awkward or uncomfortable to tell the therapist directly that therapy did not feel productive. For example, one participant said, “I just emailed [the therapist] and told her I didn’t think it was working out and that I wouldn’t be coming back. I didn’t want to go through a confrontation with her and hurt her feelings. I was kind of afraid to tell her to her face, so that’s why I went through email.” Another participant described his experience of being unsure of what to do when he felt that his first therapist was not a strong fit. He said,

[therapy] was new to us, neither of us had been before. We didn’t even know if we could ask for a different therapist. Now that I look back on it, I think, why didn’t I just switch therapists? But you’re going through the embarrassment of having to go to therapy, so you have to break that wall down, and I didn’t want to go through that again.

This excerpt exemplifies a participant’s experience of continuing therapy that did not feel helpful because they did not know how to find another therapist.

After a negative therapy experience, all participants described feeling reluctant to return to therapy. One participant described his experience of returning to therapy. He said, “I didn’t want to because I had a bad experience before, but my girlfriend wanted me to go. I was like fine I’ll go once. Now I’m on my fifth session and I’ve had an amazing experience.” Another participant explained, “I was very off-put by coming to the realization that I should go to therapy again. The first time wasn’t that good.”

Discussion

Results of this study suggest that participants’ decisions to persist or discontinue therapy is based on the intersection of three factors. Clients continually evaluate their motivation, their relationship with their therapist, therapy productiveness, which included the therapists’ frame of the problem, plan for resolving the problem, therapist neutrality and pacing when making decision about whether they will persist in therapy or discontinue. The factors that participants identified as important in their decision about therapy persistence are consistent with common factors. The common factors perspective of therapy effectiveness supports the position that common factors, rather than model specific elements, are largely responsible for therapeutic change (Sprenkle and Blow 2004; Sprenkle et al. 2009). Emergent results of the present study that align with common factors are: motivation, hopefulness that therapy will result in change, a strong therapeutic relationship, therapist neutrality in relational therapy, and a comfortable therapeutic pace (Sprenkle and Blow 2004; Sprenkle et al. 2009). While previous common factors research (Sprenkle and Blow 2004; Sprenkle et al. 2009) and meta analytic research (Shadish and Baldwin 2003; Shadish et al. 1995) discussed the role of common factors in contributing to successful therapeutic outcome; the present study supports the important role of common factors in leading clients to persist in therapy. Specifically, these results suggest that common factors may support therapy persistence.

The therapeutic alliance is perhaps the most widely researched common factor, (Sprenkle et al. 2009) and emerged as an important factor in clients’ decision to persist, or discontinue therapy, which is consistent with previous study findings (Beutler and Harwood 2000; Horvath and Symonds 1991; Knerr et al. 2011). Bordin (1979) defines the therapeutic alliance as a conscious, collaborative relationship between the therapist and the client characterized by a strong emotional bond, agreement on therapeutic tasks and goals. Going beyond literature that calls therapists to be attentive to the relationship they have with their clients, (Chenail et al. 2012; Sprenkle et al. 2009) participants in this study identified the importance of devoting time to listen closely to the presenting problem. Participants were sensitive to the amount of time the therapist spent listening to the presenting problem because they believed that the therapist needed to possess a thorough understanding of the presenting problem to provide a revolutionary frame of the problem. Participants explained that when the therapist provided a revolutionary frame of the problem that they could buy into, they felt therapy was productive and intended to persist in treatment. This finding relates to the technique of reframing, in which therapists provide a cognitive frame, or perception of the problem that modifies the client’s understanding of the problem (Watzlawick et al. 1974; Weeks and D’Aniello 2017).

Related to the therapeutic alliance, an unexpected finding that emerged was the importance of therapist neutrality in relational therapy. Both male and female clients explained that they considered discontinuing therapy, or actually discontinued therapy when they felt that the therapist had taken their partner’s side. When clients perceived the therapist as bias toward their partner, they concluded that their own concerns would not be addressed. Knudson-Martin (2013) writes that members of a couple often possess unequal power, and unequal power affirms the dominant power at the expense of the other. Therapists who are attuned to power dynamics within couples can provide emotional support to the dominant partner without inadvertently organizing therapy around their definition of the problem (Knudson-Martin 2013; Ward and Knudson-Martin 2012).

An additional common factor that emerged as critically important in participants’ intentions to continue therapy is motivation to resolve the presenting problem. Participants in this study described beginning therapy with varying levels of motivation. Many participants explained that during their initial course of therapy, they expected therapy to be ineffective in resolving the presenting problem, and therefore, experienced low motivation. It is possible that holding the expectation that therapy would not be effective contributed to low motivation.

At the time of this interview, most participants had ended therapy with a previous provider, and found a new one. Previous research that asserts that clients who have an initial negative therapy experience are less likely to seek therapy with another provider (Garfield 1986; Hamilton et al. 2011). It is possible that the present sample is comprised of highly motivated clients who sought further therapy after an initial negative experience. This finding may reflect that in general, people feel that therapy is an effective way to resolve mental health problems, and therefore, are more likely to search for a therapist who represents a strong fit with their personality.

Other participants in the present sample described beginning therapy with the mindset that they were trying out one session. When these participants found a session helpful, they returned for subsequent sessions. These participants did not inform their therapist that they were unsure about how long they would persist in therapy. This finding aligns with Prochaska et al. (1992) position that clients are motivated differently depending on which stage of change they are in. Common factors theorists, Sprenkle et al. (2009) assert that therapists can do a great deal to enhance or diminish client motivation. Results of this study support the idea that therapists must understand client’s level of motivation, and find ways to become in sync with the client’s motivation level (Miller and Rollnick 2002; Sprenkle et al. 2009).

An emergent finding of this grounded theory is the concept of therapy productiveness, which may be a unique common factor. Participants assess therapy sessions and think about what has happened over the course of several sessions and see if it feels like it is working and if they have noticed some initial changes or symptom reduction. This conceptualization of productive therapy indicates that the therapeutic event, behavior, cognition, or action is identified as giving rise to the significant event of making progress toward clients’ therapeutic goals. Because research indicates that discontinuation may be impacted by clients’ sense of a lack of productivity in therapy, identifying the elements that contribute to therapy productivity may help therapists target dropout triggers early in treatment, increasing the likelihood that clients will remain in therapy until meeting their goals (Marchionda and Slesnick 2013). Further, research suggests that collecting feedback from therapy clients about their satisfaction with their therapy treatment leads to improved therapy outcomes, and reduces risk for therapy dropout (Lambert and Shimokawa 2011; Lambert and Ogles 2004; Simon et al. 2012). It is possible that the process of soliciting client feedback, and using such feedback to initiate changes in the therapeutic processes that are based on client feedback would serve to enhance the therapeutic relationship and contribute to client intentions to persist in therapy.

Participants who discontinued therapy reported that they did so because it was not helpful in resolving the presenting problem. Participants also reported reluctance to tell their therapist when they felt therapy was not productive. Some reported that they were concerned about having an uncomfortable or awkward conversation, or hurting the therapist’s feelings. No participants indicated that they felt comfortable telling the therapist that therapy did not feel productive. While many therapists reject the image of a hierarchical therapist (Anderson and Goolishian 1988; Simon 1993) and aspire to conduct therapy that is nonhierarchical, neutral and conversational (Anderson and Goolishian 1988; Simon 1993) it is possible that clients perceive the therapist as in a hierarchical position.

Researchers of therapy retention have largely aimed to identify demographic factors that predispose clients to dropping out of therapy. Specifically, MFT research of dropout has focused on therapy constellation (Masi et al. 2003), demographic factors (Bischoff and Sprenkle 1993) and diagnosis or presenting problems (Hamilton et al. 2011) associated with dropout. Results indicate that premature discontinuation is a process rather than a static event. Results of this study support the idea that clients go through a decision-making process prior to dropping out of therapy in which they evaluate the usefulness or helpfulness of therapy services, and weigh the cost and benefits of staying in therapy versus living with the problem (Walitzer et al. 1999). It is important for therapists to be aware that clients evaluate their own motivation for change, the therapeutic alliance, and therapy productiveness and use this evaluation to make decisions about therapy persistence. Conceptualizing therapy persistence as a process rather than a static event provides therapists with opportunities to intervene in this process in effort to retain clients who are vulnerable to premature discontinuation. Collecting feedback from clients and using this feedback to initiate changes in the therapeutic process is a key way that therapists can prevent premature discontinuation.

Limitations and Future Directions

A notable limitation of the present study is that we did not collect information about participants’ current or past therapists. Participations disclosed sensitive information about their therapeutic relationship, and many potential participants asked the research team if their therapist would know they participated or have access to their responses. Client participants expressed relief when we informed them that we would not collect information that could link them to their therapist. We were also cognizant that we were guests in the therapy centers where we were recruiting, and our aim was not to disrupt clinic process or flow. It is likely that therapist characteristics influence client persistence. For example, the skill level of the therapist, years of experience, training orientation, and practice approach likely influence client persistence. Therefore, research that triangulates client data with therapist data would be useful in continuing this research and understanding the therapist factors that contribute to client persistence. This research may also identify therapist skill areas that are critical to client persistence.

Though not a focus of this study, participants’ therapy history was noteworthy. Previous research found that clients who dropped out of therapy were unlikely to seek therapy with another provider (Garfield 1986; Hamilton et al. 2011); however, 14 of the 19 participants had left a previous therapist. It is likely that other clients who aren’t as motivated would discontinue therapy after a negative experience. Future research that investigates the differences between clients who return to therapy after an initial negative experience and clients who do not would be useful.

Though all participants in the present study were in therapy with an MFT, this sample included participants of varied therapy constellation (individual, couple and family), stage of therapy and presenting problem. It is likely that there are differences in client persistence patterns that are specific to these client variables. Specifically, the elements that contribute to productiveness of relational therapy could be different from the productive elements of individual therapy. Therefore, future studies could examine the findings by therapy constellation to determine if there are differences. Further, future research investigating the interaction between client and therapist would be useful to identify relational components of the therapeutic alliance that contribute to therapy persistence. Perhaps video analysis would be a useful research method that would address the interaction between the client and therapist. Video analysis would also allow trained coders to identify the specific techniques that therapists use, which clients experience as helpful.