Franklin 2016
Franklin 2016
Franklin 2016
doi: 10.1111/jmft.12193
© 2016 American Association for Marriage and Family Therapy
Adam Froerer
Mercer University, Atlanta
Shannon Johnson
The Catholic University of America
This article presents a systematic review of the process research on solution-focused brief
therapy (SFBT). We searched published and unpublished studies in English across five data-
bases, five major journals, two book chapters, and four websites to locate studies that investi-
gate why and how SFBT works. Thirty-three studies that used various research methods
were located and included for further analysis using a meta-summary approach. The findings
supported the significance of the co-construction process within SFBT and the effects of
specific types of SFBT techniques. The most empirical support was found for the strength-
oriented techniques in comparison to the other techniques and for the co-construction of
meaning. Current studies require replications with larger samples and experimental designs
that study SFBT process in relationship to outcomes.
Process change studies on therapies use a pluralistic approach to research including quantita-
tive, qualitative, and mixed-methods designs, and are rich in perspectives for helping therapists to
identify, describe, explain, and predict the effects of the processes that bring about therapeutic
change, thus the “how and why” a therapy works (Elliott, 2010; Nock, 2007). Elliott (2010) identi-
fied four major approaches that are used to study therapy process (a) the Quantitative Process-
Outcome Design (QPOD), (b) the Qualitative Helpful Factors Design (QHFD), (c) the Microana-
lytic Sequential Process Design (MSPD), and (d) the Significant Events Approach (SEA). The
QPOD samples key processes from one or more therapy sessions and uses these to predict post-
therapy outcome. Researchers who use experimental designs consider QPOD to be the most rigor-
ous research method for studying mechanisms of change because this approach aims to establish a
causal relationship between specific therapeutic process(es) and therapeutic change (Kazdin, 2007;
Nock, 2007).
In contrast, the QHFD emerged over the past 20 years as a result of the popularization of
qualitative research often using interview methods to understand the client experiences during the
course of their therapy sessions. The MSPD also known as microanalysis focuses on the turn-to-
turn in-session interaction between client and therapist. Sequential process studies typically focus
on a small number of process variables, which means that they lend themselves to testing theories
about fundamental processes of influence in therapy session. In contrast, Change Process Research
(CPR) describes the entire course of therapy to understand the process of therapeutic change
Cynthia Franklin, PhD, Associate Dean for Doctoral Education and Stiernberg/Spencer Family Professor in
Mental Health, School of Social Work, The University of Texas at Austin. Anao Zhang, MSW, Doctoral Student,
School of Social Work, The University of Texas at Austin. Adam Froerer, PhD, Associate Professor, School of Med-
icine, Mercer University. Shannon Johnson, PhD, Assistant Professor, National Catholic School of Social Service,
The Catholic University of America.
Address correspondence to Cynthia Franklin, The University of Texas, School of Social Work, 1925 San
Jacinto Blvd D3500, Austin, Texas 78712-0358; E-mail: [email protected]
Solution-focused brief therapy (SFBT) change processes were originally grounded in the con-
structivist approaches to communication and social interactional theories (de Shazer, 1991) and
over time SFBT also became associated with social constructionism and the philosophical, post-
structural views of language such as Wittgenstein’s language games (Bavelas, De Jong, Jordan &
Korman; 2014; de Shazer, 1994). Researchers have noted that the specific questioning techniques
(e.g., miracle questions, scaling, etc.) are an important means of facilitating changes with clients
(e.g., Beyebach, 2014), and that increasing positive expectancies, and positive emotion, such as
hope and optimism, may be associated with positive outcomes within SFBT (e.g., Kim & Franklin,
2015; Kiser, Piercy, & Lipchik, 1993; Lipchik, 2002b).
A treatment manual on SFBT was first developed in 2008, and updated in 2013 (Bavelas et al.,
2013; Trepper et al., 2012) by the Solution-focused Brief Therapy Association (SFBTA). The
research committee identifies active ingredients and the core processes of conversations that are
important in SFBT. These ingredients include conversations that involve a therapeutic process of
co-constructing, by altering and/or creating new meanings with clients. Co-construction is a col-
laborative process in communication where speaker and listener collaborate to negotiate mean-
ings, and this jointly produced information in turn acts to shift meanings and social interactions
(Bavelas et al., 2013, p. 5). According to the SFBTA treatment manual, clients are specifically
asked to co-construct a vision of a preferred future and draw on their past successes, strengths,
and resources to make that vision a part of their everyday lives.
No systematic reviews have been completed on the SFBT process research. This article will
systematically review process studies in English that have been conducted on SFBT including those
that use the four major approaches to change processes research that have been identified by
Elliott (2010) and all other research designs that can be identified and are relevant to understand-
ing the change processes of SFBT. The purpose of this current review is to summarize the empirical
change process literature on SFBT and to review the use of practices and processes in SFBT as
reported by the literature.
We will examine studies that look at the active ingredients of therapy, for example, co-con-
struction of meaning and relationship factors, and therapeutic techniques, such as the miracle
question and other techniques. It is the intention of the authors to be inclusive of all the change
process research and to increase the rigor of this review by following a systematic approach to the
review resulting in a meta-summary (Sandelowski & Barroso, 2007) of the results for both quanti-
tative and qualitative process studies on SFBT.
METHOD
Selection Criteria
Our objective was to review (a) all available, (b) process studies of SFBT that were (c) deliv-
ered in clinical settings (direct client contact). We reviewed all studies in English, published or
unpublished, that purported to study the processes of SFBT. Unpublished studies, dissertations,
and conference reports, including published conference proceedings and conference presentations,
were reviewed just as published study (Higgins & Green, 2011).
In defining process studies, we started with a four-category conceptual framework based
on Elliott’s (2010) model: (a) the Quantitative Process-Outcome Design, (b) the Qualitative
Helpful Factors Design, (c) the Microanalytic Sequential Process Design, and (d) the Signifi-
cant Events Approach. Across each category, we assigned 11 analytical techniques extracted
from three methodology articles on change process research (Elliott, 2010; Greenberg, 1986;
Woolley, Butler, & Wampler, 2000) that fit a certain category. This conceptual framework
was important for this review because change process research is a broad concept that houses
a plurality of research designs. Our approach allowed this article to maximize the scope of
our search while keeping the focus on all analytical techniques and research designs that are
associated with process studies. During review of the abstracts of articles, we found one
research design that was not identified in Elliott’s review, recursive frame analysis, and we
added this term to our search criteria.
Search Strategy
We used two strategies to create the initial pool of potential studies for review. First, we con-
ducted five electronic databases searches including PsychINFO, Academic Search Complete, Med-
line, PUBMED, and Education Resource Information Center (ERIC). Within each database, we
used *Solution-Focused* or *Solution Focused* or *SFBT* or *Solution-Focused Brief Therapy*
AND four different types of Psychotherapy Change Process Research. For example, for type one
studies, search within PsychINFO, we used the key words: *Solution-Focused* or *Solution
Focused* or *SFBT* or *Solution-Focused Brief Therapy* AND *Mechanism of Change* or
*Mediator* or *Mediation*. As a result, we conducted four different sets of key word search
within each database and repeated five times for the period 1980 to January 2015.
Second, we conducted a comprehensive hand search including all possible sources of studies
like (a) major journal search with key words: *Solution-Focused* or *Solution Focused* or
*SFBT* or *Solution-Focused Brief Therapy*. We included: Journal of Systemic Therapy; Jour-
nal of Marital and Family Therapy; American Journal of Family Therapy; Journal of Family
Therapy; and “Family Process”, (b) all additional studies from McKeel’s (2012) book chapter ref-
erences, and (c) manually searched major Internet resources for additional candidate studies
including Solution-Focused Brief Therapy Association at http://www.sfbta.org/, European Brief
(n = 272)
(n = 272) (n = 172)
(n = 100) (n = 0)
Full-text articles assessed for eligibility Full-text articles excluded, with reason
(n = 100) (n = 67 + 1)
(n = 32 + 1)
Figure 1. Flow diagram showing the number of studies at each step in the selection process.
Data Analysis
This study used meta-summary, one of the mixed-method research synthesis approaches (San-
delowski & Barroso, 2007), to analyze data. Meta-summary has been suggested to be an appropri-
ate method for reviewing active ingredients of interventions (Heyvaert, Hannes, Maes, &
Onghena, 2013) and provides analytic methods for summarizing primary studies that include
quantitative, qualitative, and mixed-methods designs (Sandelowski, Voils, Leeman, & Crandell,
2012). Meta-summary follows the logic of meta-analysis and uses a comparability approach where
“the differences [between quantitative and qualitative results] are reconciled by converting the one
into the other” in order to provide an aggregate summary of primary studies (Sandelowski, Voils,
& Barroso, 2007, p. 239). For this study, both quantitative and qualitative results were retrieved
from primary studies and were aggregated using a data transformation approach. Using this
approach, we converted quantitative results into qualitative themes so that the results would be
comparable. Figure 2 summarizes the steps of the analysis.
In order to conduct a meta-summary, quantitative results and qualitative results were first sep-
arately retrieved from primary studies. Specifically, [STEP 1], we assigned results from primary
studies’ into either quantitative results based on the statistical data or qualitative results based on
the reported narrative results and themes. For mixed-method studies, quantitative and qualitative
results from the studies were separated and the results were assigned into quantitative or qualita-
tive categories. For quantitative results [STEP 2], we used a vote counting (Voils, Sandelowski,
Barroso, & Hasselblad, 2008) method and if p < 0.05, then a result was assigned as a positive
quantitative result. If p ≥ 0.05, then a result was assigned as a non-positive quantitative result. To
reiterate, the term “positive” was used when there was statistical support for the concept in pri-
mary studies, and the term “non-positive” was used when there was no statistical support for the
concept. For qualitative results, if a qualitative theme supported the research question, or if a qual-
itative theme emerged that was reported as a positive finding in the study, then the theme was
assigned as a positive qualitative result. If a qualitative theme did not support the research
If p < 0.05, then If p ≥ 0.05, then If a qualitative theme If a qualitative theme did not
supported research question, support research question, or, if
or, if a theme emerged, then a theme did not emerge, then
** The terminology of “relational result” in the context of this study is not about relationship of the client-therapist dyad. It specifically refers to those results that related the
techniques or practices of SFBT to therapeutic outcomes.
** For a relational, positive [quantitative] result, it was transformed into “SFBT techniques significantly improve therapeutic outcomes”. For a process, positive
[quantitative] result, it was transformed into “SFBT therapists used significantly more process (e.g., client’s own words) than other therapies (e.g., CBT)”. For non-
positive quantitative results, they were transformed into either “SFBT techniques did not significantly improve therapeutic outcomes” or “SFBT therapists did not use
more of clients’ words than CBT therapists”.
question, or if a hypothesized theme failed to emerged, then the theme was assigned as a non-posi-
tive qualitative theme. [STEP 3] If the reported positive [quantitative] results specified SFBT tech-
niques in relation to therapeutic change, then the result was assigned as a relational positive result
(e.g., “clients in the solution-talk group reported a significantly higher (p < 0.001) proportion of
treatment continuation” is a quantitative relational positive result). However, if the positive [quan-
titative] results described the process of SFBT in comparison to other therapies, then the result was
assigned as a process positive result (e.g., “positive utterance among SFBT therapists was signifi-
cantly higher that CBT therapists, p < 0.001” is a positive quantitative results that indicated pro-
cess differences between solution-talk and CBT). For clarity purpose, it should be emphasized that
the terminology of “relational result” in the context of this study is not about relationship of the
client–therapist dyad. It specifically refers to those results that related the techniques or practices
We identified two studies (6%) published before 1991, eleven studies (34%) published between
1991 and 2000, and twenty studies (60%) published after 2001. Twenty-seven studies (81%) are
journal articles and the six are dissertations (19%). Regarding research designs used in the studies,
42.4% of the studies (n = 14) used MSPD and 39.4% of the studies (n = 13) used process-outcome
or mechanism of change method of analysis. Four articles (12.1%) used qualitative method of
analysis and only two articles (6.1%) used comprehensive or task analysis. Twenty-five studies
(75.8%) examined SFBT as an individual intervention, six studies (18.1%) examined SFBT for
couples and families, and two studies (6.1%) used SFBT as an intervention for both individuals
and family/couple.
Quality of Studies
We assessed the studies on quality indicators that have been recommended in systematic
reviews including study designs, measures, sample and population, and intervention fidelity. The
study designs were diverse and only 12 studies (36.36%) used an experimental design that exam-
ined process along with outcomes. Eighteen studies examined processes only and did not link pro-
cesses to outcomes. One study (Lloyd & Dallos, 2008) used semistructured qualitative interview to
examine the relationship between SFBT components and self-efficacy and self-locus of control,
and two studies (Monro, 1998; Simon & Nelson, 2004) used qualitative interview on clients’ per-
ceived helpful factors of SFBT. Only nine studies compared SFBT against another intervention,
and eighteen studies did not compare the process of SFBT with another intervention. Six addi-
tional studies added SFBT component(s) to an existing treatment or intervention. There were no
studies that would meet the gold standard for mechanism of change research as described by Kaz-
din and Nock (2003), including (a) strong association, (b) specificity, (c) gradient, (d) experiment,
(e) temporal relation, (f) consistency, and (g) plausiblity and coherence. This is a criticism of the
quality of the studies in relationship to the overall confidence of the findings but it is additionally
noted that few psychotherapy process studies may meet this gold standard.
There were also strengths and limitations within measures used. Out of the sixteen studies that
included some types of measures, only seven studies used standardized rating scales, and five stud-
ies used either a therapeutic coding systems or client self-reported survey. The rest of the studies
used direct behavioral observations or observational measures, such as the Topic Initiation/Topic
Following coding scheme (Beyebach & Carranza, 1997a,b). Studies also used clinical data includ-
ing progress note (e.g., Franklin, 1996), transcribed videotapes (e.g., Gale & Newfield, 1992), and
audiotapes (e.g., Shields, Sprenkle, & Constantine, 1991).
Approximately half of the studies had small, clinical, convenience samples (n = 16) limiting
the overall generalizability of the findings. Of these studies, small sample sizes that used part or full
sessions of therapy were available and these studies lacked replications with larger samples. Most
of the studies (n = 25) reported 20 or less participants in the treatment condition. Only four studies
had more than 45 participants for the SFBT group, leaving the other four studies with a treatment
group sample size ranging from 21 to 40. Although smaller sample sizes are not uncommon in pro-
cess research, there were other characteristics of the samples that may limit the conclusions that
can be drawn from a synthesis of these studies. There was a lack of diverse populations and several
studies did not report about the client characteristics within the studies. Only 14 studies (42%)
reported clients’ racial background and they were predominantly Caucasian.
The experience of the therapists delivering the SFBT interventions also varied with 13 studies
using experienced therapists and 4 studies using intermediate therapists with at least 2 years of
experience. Five studies used both experienced and intermediate therapists and two studies used
both experienced therapists and master level interns, while an additional two studies used only the
interns. It should be noted that information regarding the therapists’ training, competencies and
experiences in using SFBT were generally not available in the primary studies. Limited information
could be drawn from the 13 studies that used experienced therapists. For a study’s therapist(s) to be
coded as experienced, he/she must meet the following criteria (a) well-known experts of SFBT like
Insoo Kim Berg, or (b) a doctoral level therapist with a marriage and family therapy background,
Notes. aTypes of techniques and practices of SFBT that were examined in primary studies.
LCL = Linguistic and collaborative language, included co-construction and solution-talk;
TRTS = therapeutic relationship and therapist style, included positive, client-focused stance,
supportive and flexible approach; SRT = strengths and resources techniques, included pre-
session change, formulation first session task, scaling question, exception question;
FOT = future-oriented techniques, included miracle question, goal setting, and end of ses-
sion homework; MTI = multiple SFBT techniques and interviews. bNumber of themes
reported and by how many studies. cHow many themes were derived from quantitative data
and how many were from qualitative data in primary studies. dThe percentage of number of
themes out of all themes from primary studies.
or (c) a licensed clinician with several years of practice experience including SFBT. With this said,
however, except those who are well-known experts of SFBT, most studies were not clear about how
much actual training that an experienced therapist had in relationship to the practice of SFBT.
Study Findings
Descriptive statistics in Table 1 presented the frequency of the positive and non-positive
themes reported across studies. The qualitative themes are first grouped based on the nature of the
themes (positive relational themes, positive process themes, and non-positive themes). Then,
within each group, themes were further assigned into a configuration of the types of SFBT thera-
peutic techniques and processes that were examined in primary studies described previously.
Overall, 87.9% (n = 102) of the themes were positive themes from primary studies, meaning
87.9% of the results from primary studies retrieved either received statistical support or qualitative
support for the techniques and practices of SFBT. Out of the 102 themes, 68 themes (66.67%) were
DISCUSSION
The findings from this study have important implications for clinical practice/supervision,
education, and future research. The results of this current review suggests that both SFBT tech-
niques and linguistic/style-oriented methods show positive support and that both may be co-func-
tioning in the change processes of SFBT. Even though, SFBT techniques and linguistic processes
are presumed to interact together to create change, the studies on techniques are different than the
studies that focus on linguistic processes resulting in the two being discussed separately within this
review. Only a few studies, for example, discussed the use of techniques in relationship to linguistic
change (e.g., Strong & Pyle, 2009; Strong, Pyle, & Sutherland, 2009). Future researchers may want
to pay more attention to how both linguistic processes and techniques co-function together in the
change process because a more thorough sequential or systemic analysis would be useful for educa-
tion and supervision.
Overall findings from this review add to the scientific foundation for SFBT practice and pro-
vide support for the active ingredients that have been identified in the SFBTA treatment manual
and in other training and clinical literature on SFBT (Berg & De Jong, 1996; Trepper et al., 2012;
Bavelas et al., 2013; Taylor & Simon, 2014). This review indicated that the strongest overall find-
ings based on the most positive themes for effectiveness (0.45%) were for all the SFBT techniques
combined (strength and resources, future-oriented, and multiple techniques. In comparison to
other techniques, the strengths and resources techniques received most positive findings from pri-
mary studies (i.e., eight studies support the use of these techniques). These findings support the
continued need for clinicians to be thoughtful and purposefully integrate the specific strength/
Limitations
First, there is always a possibility that we did not find all the process studies within the
proposed timeframe, although, we followed a rigorous search criteria and process making use
of methods that have been recommended for systematic reviews. This study is also limited to
studies in English and no doubt a broader search in different languages may enhance or even
possibly change the findings of this study. The mixed-method research approach, meta-sum-
mary that was used to aggregate primary studies while adding strengths to this study by allow-
ing a summarization of findings from diverse study designs also has limitations in that some
of the information from different studies may be lost in the recoding and analysis. The config-
uration of study variables is also based on an analytical qualitative technique, and it may not
represent all the quantitative findings within each study or the thick descriptions of the sum-
maries within the qualitative studies. Even with these limitations, we believe that this study
provides a contribution to literature by providing a systematic review and analysis of the cur-
rent process studies on SFBT in English.
CONCLUSION
This review show that SFBT techniques have considerable support across the process studies
and that the techniques directed toward the strengths and resources of the clients show the most
positive results. The linguistic and collaborative language approaches also have relatively strong
support in the SFBT and this finding adds to the empirical support for the co-construction of
meaning. The current findings have limitations in their evidence due to the quality and numbers of
the current studies. The next steps for SFBT process research will be to replicate current findings
and improve upon research designs regardless of what methods are used. One way to improve the
studies is to design prospective studies that can show that the co-constructive processes in SFBT
are directly linked to client outcomes. It may be useful to study outcomes that have been found to
be significant in effectiveness studies because a normal progression for evidence based practice is to
first show a therapy works and then proceed to study how and why it works. Two areas that have
shown promise for example, are with internalizing disorders such as depression, (Gingerich &
Peterson, 2013) and with problems that occur in children and families (Bond et al., 2013). Samples
would need to be large enough to meet the recommended standards for the research designs being
used. In particular, the development of larger randomized controlled trials that study both process
and outcomes are indicated.
REFERENCES
Note: References marked with an asterisk indicate studies included in this systematic review and
meta-summary.
*Adams, J. F., Piercy, F. P., & Jurich, J. A. (1991). Effects of solution focused therapy’s “formula first session task”
on compliance and outcome in family therapy. Journal of Marital and Family Therapy, 17, 277–290.
Bavelas, J., De Jong, P., Franklin, C., Froerer, A., Gingerick, W., & Kim, J. (2013). Solution focused therapy treat-
ment manual for working with individuals (2nd version). Solution Focused Brief Therapy Association, Retrieved
December 9, 2015, from http://www.sfbta.org/PDFs/researchDownloads/fileDownloader.asp?fname=SFBT_
Revised_Treatment_Manual_2013.pdf
Bavelas, J. B., De Jong, P., Jordan, S., & Korman, H. (2014). The theoretical and research basis of co-constructing
meaning in dialogue. Journal of Solution-focused Brief Therapy, 2(2), 1–24.