Single-Case Design Review and Meta-Analysis for Supporting the Method of Transactional Analysis towards Recognition as an Empirically Supported Treatment for Depression

Abstract

Common Mental Disorders represent a severe burden for a country's health, society and economy. Several psychotherapies have shown their efficacy in treating such common mental disorders using randomized clinical trials (RCT). Psycho-therapies that are not supported by this sort of research evidence are disenfran-chised and marginalized. A way to obtain recognition as an 'Empirically Supported Treatment' relies on systematic replication of single-case designs and on the aggregation of results through a meta-analysis. The purpose of this meta-analytic review was to synthesize the single-case research on Transactional Analysis (TA) treatment for depression. Specifically, the effect of TA treatment for depression was examined in 11 studies, published between 2012 and 2017. Results indicated that, on average, TA psychotherapy for depression had a large effect on depressive symptoms: g = 0.89, 95% confidence interval (CI) [0.29-1.50]. Implication for future research on such TA manualized treatments for specific Common Mental Disorders is discussed.

Full text

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
93
Enrico Benelli & Mariavittoria Zanchetta
Single-Case Design Review
and Meta-Analysis for Supporting
the Method of Transactional
Analysis towards Recognition
as an Empirically Supported
Treatment for Depression
Abstract: Common Mental Disorders represent a severe burden for a country’s health, so-
ciety and economy. Several psychotherapies have shown their ecacy in treating
such common mental disorders using randomized clinical trials (RCT). Psycho-
therapies that are not supported by this sort of research evidence are disenfran-
chised and marginalized. A way to obtain recognition as an ‘Empirically Support-
ed Treatment’ relies on systematic replication of single-case designs and on the
aggregation of results through a meta-analysis. The purpose of this meta-ana-
lytic review was to synthesize the single-case research on Transactional Anal-
ysis (TA) treatment for depression. Specifically, the eect of TA treatment for
depression was examined in 11 studies, published between 2012 and 2017. Results
indicated that, on average, TA psychotherapy for depression had a large eect on
depressive symptoms: g = 0.89, 95% confidence interval (CI) [0.29-1.50]. Impli-
cation for future research on such TA manualized treatments for specific Com-
mon Mental Disorders is discussed.
Key Words:
Transactional Analysis, Depressive disorders, Marginalized and Emerging Psycho-
therapy, Single-Case Meta-Analysis, Hermeneutic Single-Case Ecacy Design.
Introduction
In the last decade, epidemiological studies
have drawn attention to the high prevalence of
Common Mental Disorders (CMD; e.g., depres-
sion, anxiety, personality disorders) and their
impact on health, society and economy (Steel
et al., 2014; Trautmann, Rehm & Wittchen,
2016; WHO, 2003). Research in psychothera-
py grew exponentially to establish the ecacy
International Journal of Psychotherapy
Nov. 2019, Vol. 23, No. 3, pp. 93-108; ISSN 1356-9082 (Print); ISSN 1469-8498 (Online)
© Author and European Association of Psychotherapy (IJP): Reprints and permissions: www.ijp.org.uk
Published Online: 31-Oct 2019; Print publication: 31-Oct 2019; DOI: 10.36075/IJP.2019.23.3.10/Benelli/Zanchetta

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
94
and the eectiveness of psychological ther-
apies for these CMD, progressively adopting
the methodologies of Evidence-Based Practice
(EBP), as used in medicine. EBP in psycholo-
gy promotes eective psychological practice
and enhances public health by applying em-
pirically supported principles of psychologi-
cal assessment, case formulation, therapeutic
relationship, and intervention, articulating a
decision-making process for integrating mul-
tiple streams of information, encompassing:
(a) research evidence on a treatment’s ecacy
and eectiveness; (b) clinical expertise; and
(c) patient characteristics (APA Presidential
Task Force on Evidence-Based Practice, 2006).
Research evidence on ecacy is generally
supported by experimental designs such as
Randomized Clinical Trials (RCTs), whereas
eectiveness is generally supported by ob-
servational studies, such as cohort studies
(Grimes & Schulz, 2002). However, Chamb-
less & Hollon (1998) argued that ecacy may
be supported, not only by RCTs, but also by a
series of Single-Case Experimental Designs
(SCED) with systematic replication by inde-
pendent research groups.
Despite this, the last 25 years has seen the ac-
cumulation of ever-larger and more complex
RCTs, with the widespread diusion of the
ideology of Empirically Supported Treatments
(EST), that implicitly equates EBP as needing
to be supported mostly by RCTs; this trend
often discounts any (all) clinical expertise,
dierent patient characteristics, and any ev-
idence based on SCED, or equivalent designs.
As a result, in several countries, the mental
health policy-makers only included methods
of psychotherapy that had gained the status
of EST in their national guidelines: thereby,
implicitly delegitimizing any methods that
cannot aord the costs of conducting RCTs.
An example of this happened in the United
Kingdom with the NICE clinical guideline for
depression (National Collaborating Centre for
Mental Health, 2009), that explicitly recom-
mended doctors and mental health profes-
sionals discussing with the patient the uncer-
tainty of the eectiveness of treatments, such
as counselling or psychodynamic psychother-
apy, and where a widespread, world-wide
method, such as Transactional Analysis, was
not even considered.
Methods of psychotherapy without EST sta-
tus, because they were relatively new or were
lacking in research support, were grouped
under the label of Marginalized and Emerg-
ing Psychotherapies (MEPs). Stiles, Hill and
Elliott (2015) proposed a four-step pathway
that might enhance recognition: (1) publish-
ing systematic single case studies by com-
mitted practitioners; (2) forming Practitioner
Research Networks (PRN), with online data
collection facilities, collecting substantial
amounts of practice-based evidence that can
be aggregated, analysed and compared with
larger population benchmarks; (3) conduct-
ing small RCTs and practice-based random-
ized trials or pragmatic trial,s and high profile
studies; and (4) developing political network-
ing and pressure groups.
Supporting Marginalized and
Emerging Psychotherapy
with Single-Case
Experimental Design
RCTs have become the ‘gold standard’ for sup-
porting EST, thanks to their sound methodol-
ogy based on: (a) registration of the trial in
advance, to prevent publication biases related
to publication of only good outcome trials; (b)
randomization of population into cohorts un-
der dierent conditions; (c) manualized treat-
ments ensuring a common methodology that
can be undertaken by dierent practitioners;
(d) intention-to-treat (ITT) to avoid publica-
ENRICO BENELLI & MARIAVITTORIA ZANCHETTA

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
95
tion biases on dropouts; and (e) the possibility
to compute meta-analyses and calculate valid
eect sizes.
RCTs’ rationale is similar to that of SCEDs. The
‘gold standard’ method evaluates eects of an
intervention by comparing performances of a
treated group with a control group, which does
not receive the treatment, or receives a dier-
ent one. On the other hand, SCEDs compares
changes in the same person in diverse mo-
ments: i.e. during a baseline phase, where data
is used to predict a level of performance for the
immediate future if treatment is not provided;
and then a therapy phase with the treatment
(Kazdin, 1978).
Generally, case studies are considered biased
by the author’s assumptions, ethically prob-
lematic, dicult to summarize, unable to test
causality, not generalizable, and their results
are not relevant for policy-making (McLeod,
2010). This is not true for outcome-oriented
case studies, able to support ecacy and ef-
fectiveness with SCED and the Hermeneutic
Single-Case Ecacy Design (HSCED) (Mc-
Leod, 2010; Benelli et al., 2015).
Results from well-conducted single-case
studies can approximate the results of RCTs
(Kazdin, 1981) when implying: (a) reliable and
valid outcome measures; (b) continuous as-
sessment of key outcome variables; (c) stabili-
ty of the baseline before treatment; (d) marked
eect of intervention on outcome variables
supported by time-series analysis; and (e)
replication of the same pattern over multiple
cases.
Causality and generalizability may be further-
ly enhanced by following some of the criteria
used by an EST approach, and built into the
guide for RCT (CONSORT, Schulz, Altman &
Moher, 2010); meta-analysis (PRISMA, Mo-
her, Liberati, Tetzla & Altman, 2009); and
guideline production (GRADE, Guyatt et al.,
2011).
In particular, it is necessary to specify the di-
mensions grouped in the acronym PICO: Par-
ticipant (eligibility criteria defining the target
population, e.g., people with depressive disor-
ders); Intervention (e.g., a manualized treat-
ment); Comparison (e.g., the pre-post treat-
ment, or time-series analysis); and Outcome
(primary and secondary outcome measures
and methods of assessment, e.g., a weekly
questionnaire on depression).
Single Case Meta-Analysis
Thanks to methodological rigour, systematic
outcome-oriented single case studies may be
aggregated through meta-analysis (Shadish,
Rindskopf & Hedges, 2008).
A meta-analysis is the statistical synthesis of
results from a systematic review of original re-
search related to a particular topic (Borenstein
et al., 2009), and research syntheses whose
aim is to integrate empirical research so as to
generalize the data (Cooper & Hedges, 2009);
evaluating the outcome through the eect size
(Cohen, 1988); which measures the dier-
ence between control and treatment groups,
in terms of standard deviation. The validity of
conclusions drawn from such meta-analytic
research is only as strong as the methods used
within each original study (Burns, 2012).
SCEDs may be a valid and important contri-
bution to evidence-based literature, but they
are not largely used in reviews about evi-
dence-practice because they lack of a widely
accepted and formally-developed statistical
method for analysis and meta-analysis (Shad-
ish, 2014). There are many possible methods
to conduct a meta-analysis with single-cases
(e.g., see review by Allison & Gorman, 1993;
Beretvas & Chung, 2008; Shadish, 2014).
For our meta-analysis, we selected a meth-
od that allows to use an eect size for SCEDs,
with the same metrics as the one used in RCTs,
SINGLE-CASE DESIGN REVIEW AND META-ANALYSIS

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
96
making possible the comparison between re-
sults of dierent designs and outcome mea-
sures. We adopted a standardized mean dif-
ference statistic (d), as suggested by Shadish,
Hedges & Putejovsky (2014), because the d
statistic allows us able: (a) to compare SCEDs
and RCTs’ eect sizes because they are based
on the same metric; (b) to aggregate data from
studies that used dierent quantitative out-
come measures; (c) to use conventional sta-
tistical methods (like forest plots, diagnostic
plots (such as radial plots and residual plots),
cumulative meta-analysis, regression tests,
or publication bias analysis; and (d) that the d
statistic can be corrected for small sample bias
with Hedges’ ‘g’ (Hedges, 1981).
The Case of Transactional
Analysis
Transactional Analysis (TA) psychotherapy is
considered a MEP and is developing research
strategies to be recognized as an EST, follow-
ing the four-step pathway proposed by Stiles,
Hill & Elliott (2015).
Ecacy of TA psychotherapy for the treatment
of depression has been investigated through
a particular SCED, the HSCED (Elliott, 2002,
2009) in a direct replication (Widdowson,
2012a, 2012b, 2012c, 2013, 2014), followed by
systematic replications (Benelli et al., 2016a,
2016b, 2016c, 2017a, 2017b, 2017c).
Nowadays, nine single cases demonstrated the
ecacy of TA for mood disorders (Major De-
pressive Disorder, Persistent Depressive Dis-
order, Subthreshold Depression), all of them
fulfilling the Chambless and Hollon criteria
for claiming recognition as a well-established
treatment.
The aim of this article is to conduct a review
of SCEDs supporting Transactional Analysis
treatment for depression and to conduct a me-
ta-analysis.
Method
Review
A literature search was conducted using the
following keywords: transactional analysis,
psychotherapy, counselling, mood disorders,
major depressive disorder, persistent depres-
sive disorder (dysthymia), subthreshold de-
pression. We identified 11 articles that focused
on supporting ecacy of TA treatment for
mood disorders (Major Depressive Disorder,
Persistent Depressive Disorder, Subthreshold
Depression). Each document was evaluated
according to whether or not it met all of the
following criteria according to the PICO strat-
egy (see Table 1, below).
Table : The list of inclusion criteria for this meta-analysis
Population Patient received mainly a DSM diagnosis of mood disorder
Intervention Patient received a manualized treatment
Comparison – The presence of a sound research design (e.g., SCED, HSCED);
– The presence of a systematic case study research with at least AB phases
(baseline, treatment)
Outcome The presence of highly validated outcome measures (e.g., BDI-II, PHQ-9)
ENRICO BENELLI & MARIAVITTORIA ZANCHETTA

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
97
(2)
 jid (journal identification number):  (International Journal of Transactional Analysis,
IJTAR);
 SID (study identification number):  (Widdowson case series), 2 (Benelli et al., 2016 case
series), and 3 (Benelli et al., 2017 case series);
 DVID (dependent variable identification number):  (for PHQ-9), and 2 (for BDI-II);
 DesType (0 = multiple baseline design, 1 = ABk design, and 9 = mixture of both designs): 1;
 DesVar (design variable, required only when DesType = 9): not considered;
 PID (case identification number). We chronologically ordered the case series for each study,
and numbered each single-case study from  to  for the cases of Widdowson and from  to
 for each Benelli et al. case series;
 SessIDX (session number): as for AB phase, we considered  the first assessment score and
16+n the last treatment score (according to n-point baseline); for BC phase, we considered 
the first treatment score, and  the last follow-up score;
 PhaseBTM (for AB comparison:  = baseline,  = treatment; for BC comparison:  = treat-
ment,  = follow-up);
 NumPh (phase number): with AB:  = baseline,  = treatment; with BC:  = treatment,  =
follow-up1;
 DVDir ( = outcome increases if treatment works, or  = outcome decreases if treatment
works): ;
 DVY (the outcome variable on the y-axis): the score of the BDI-II or PHQ-9;
 Detrend (optional detrending variable): not considered.
1
PhaseBTM and NumPh overlap because this study was an AB
1
design.
Meta-Analysis
In BSDs, Glass (1976) suggested to use Cohen’s
d statistic, which expresses treatments eects
in terms of outcome standard deviations:
where Mt is the mean of the treatment group,
Mc is the mean of the comparison group, and
S is an estimate of the standard deviation as-
sumed to be common across treatment and
control conditions.
The following equation (Shadish et al., 2014)
represents the eect size parameter as the
standardized mean dierence from BSDs:
(1)
where μT is the mean outcome treatment, μC
denotes the mean outcome baseline, is the
variance of observations within cases, is
the variance of observations between cases,
and is the total variance. Cohen’s d
in BSDs is based on one observation per case,
and for this reason and are not sepa-
rately identified, so it is not possible to dis-
tinctly estimated them. Therefore, equations
(1) and (2) overlap, leading to an eective
comparison between BSDs and SCEDs. More-
over, as previously mentioned, this equation
can be corrected for small sample bias using
Hedges’ g (Hedges, 1981).
To analyse these considered cases, we followed
the indications of Shadish (2014) and used the
SPSS macro with the software for meta-anal-
ysis of SCED (Shadish, DHPS version March 7,
2015). The variables that we considered for the
analysis are the following:
SINGLE-CASE DESIGN REVIEW AND META-ANALYSIS

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
98
The macro used in DHPS is illustrated in Ap-
pendix A. SPSS calculated for each study
Hedges’ g, the variance of g (VarG), the auto-
correlation (PhiHat), and the intraclass cor-
relation (Rho).
Finally, we used R (R Development Core Team,
2016) and the metafore package (Viechtbauer,
2010) to work on the analysis (the R metafore
commands are illustrated in Appendix B), and
we produced a forest plot to display the eect
size and confidence interval for each study and
the overall meta-analytic average.
Results
Review: This meta-analysis includes the anal-
ysis of 11 single-case studies. The cases that
follow the PICO strategy above mentioned are
described in the following tables (Table 2).
Population: The research recorded the ther-
apist’s first new patient that had received
principally a DSM diagnosis of mood disorder
(Major Depressive Disorder, MDD; and Per-
sistent Depressive Disorder, PDD) and who
also agreed to participate in the research proj-
ect (see Table 3).
Intervention: All therapies followed (at first)
the draft, lately, the manualised therapy pro-
tocol of Widdowson (2016). The therapists
received supervision on a weekly to monthly
basis from a Provisional Teaching and Super-
vising Transactional Analyst (Psychotherapy)
(PTSTA-P), or from a Teaching and Super-
vising Transactional Analyst (Psychotherapy)
(TSTA-P). and follow-ups outcome (see Table
4).
Outcome: In the Widdowson case series, the
Beck Depression Inventory (BDI-II) (Beck,
Steer & Brown, 1996; see Table 2) was ad-
ministrated: a widely used 21-items self-re-
port inventory. Scores for each item are rep-
resentative of the gravity range and go from
0 (absent) to 3 (severe). Total scores up to 13
are considered healthy, scores of 14, 20, 29 are
respectively cut-o points for mild, moderate
and severe depression, scores above 63 con-
sider an extremely severe depression, and its
caseness cut-o is of 16 points. A change of
at least 5.78 points on BDI-II is considered to
assess a reliable improvement or deteriora-
tion (Reliable Change Index, RCI). Instead, in
both the Benelli et al. (2016 & 2017) case se-
ries, the measure used to evaluate depression
was the Patient Health Questionnaire 9-item
for depression (PHQ-9) (Spitzer, Kroenke &
Williams, 1999; see Table 2), which is a quan-
titative measure that scores each of the nine
DSM-5 criteria from 0 (not at all) to 3 (nearly
every day). Total scores up to 4 are considered
healthy, scores of 5, 10, 15 and 20 are taken re-
spectively as the cut-o points for mild, mod-
erate, moderately severe and severe depres-
sion, and scores lower than 10 are considered
sub-clinical. To reach RCI, there must be a
change of at least 6 points on the PHQ-9 score.
The trial registration of the considered cases is
reported in the table below (see Table 3).
The considered HSCEDs and the pragmatic
case study used an ABA design, with baseline,
treatment and follow-up. We considered only
an AB comparison, between the baseline phase
(A) and treatment (B). Table 4 summarizes the
considered cases.
The three case series studies all have a large
eect size, respectively: Widdowson series: g
= 0.74 (95% CI [-0.06 to 1.53]); Benelli 2016
series: g = 0.77 (95% CI [-0.30 to 1.83]); and
Benelli 2017 series: g = 1.16 (95% CI [0.31
to 2.02]). However, Widdowson (2012) and
Benelli (2016) cannot be considered as repre-
sentative because their lower CI is negative.
Nevertheless, the average for all studies is
0.89 (95% CI [0.29 to 1.50]), which represents
a significant and large eect-size. We can
therefore observe the ecacy of Transaction-
al Analysis on depression, with a large overall
dierence between assessment and treatment.
ENRICO BENELLI & MARIAVITTORIA ZANCHETTA

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
99
Table 2: The list of the considered cases for this meta-analysis on published data.
Case series Case
Population
(main diagno-
sis)
Intervention
(treatment)
Com-
parison
(design)
Outcome
(mea-
sure)
Widdowson
“Peter”
(Widdowson,
2012a)
MDD
Manualized TA
treatment for
depression
ABA2 study BDI-II
“Denise”
(Widdowson,
2012b)
MDD
Manualized TA
treatment for
depression
ABA study BDI-II
“Tom”
(Widdowson,
2012c)
MDD
Manualized TA
treatment for
depression
ABA study BDI-II
“Linda” (Wid-
dowson, 2013) MDD
Manualized TA
treatment for
depression
ABA study BDI-II
“Alastair”
(Widdowson,
2014)
PDD
Manualized TA
treatment for
depression
ABA study PHQ-9
Benelli
(2016)
“Sara”
(Benelli et al.,
2016a)
MDD
Manualized TA
treatment for
depression
ABA study PHQ-9
“Penelope”
(Benelli et al.,
2016b)
MDD
Manualized TA
treatment for
depression
ABA study PHQ-9
“Luisa”
(Benelli et al.,
2016c)
PDD
Manualized TA
treatment for
depression
ABA study PHQ-9
Benelli
(2017)
“Anna”
(Benelli et al.,
2017a)
PDD
Manualized TA
treatment for
depression
ABA study PHQ-9
“Caterina”
(Benelli et al.,
2017b)
MDD
Manualized TA
treatment for
depression
ABA study PHQ-9
“Deborah”
(Benelli et al.,
2017c)
MDD
Manualized TA
treatment for
depression
ABA study PHQ-9
2
ABA study: A: baseline, B: treatment; A: follow-up
SINGLE-CASE DESIGN REVIEW AND META-ANALYSIS

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
100
Table 3: Trial registration of the considered cases for this meta-analysis.
Case series Case
Started (month/year) Ended (month/year)
Assessment Treatment Last follow
up End
Widdowson3
“Peter” (Widdow-
son, 2012a)
“Denise” (Widdow-
son, 2012b)
“Tom” (Widdowson,
2012c)
“Linda” (Widdow-
son, 2013)
“Alastair” (Widdow-
son, 2014)
Benelli
(2016)
“Sara” (Benelli et al.,
2016a) 11/2013 11/2013 04/2014 12/2014
“Penelope” (Benelli
et al., 2016b) 10/2013 10/2013 03/2014 09/2014
“Luisa” (Benelli et
al., 2016c) 10/2013 11/2013 06/2014 12/2014
Benelli
(2017)
“Anna” (Benelli et
al., 2017a) 08/2014 10/2014 02/2015 09/2015
“Caterina” (Benelli
et al., 2017b) 05/2014 06/2014 12/2014 06/2015
“Deborah” (Benelli
et al., 2017c) 07/2014 10/2014 03/2015 11/2015
3
Missing data: pending request
Discussions
The cases of Widdowson have a -point base-
line. A baseline is conventionally defined as a
minimum of three-data-points recorded be-
fore the beginning of a treatment, in order to
compare absence of therapy with its eects
(Kazdin, 2010), therefore a real baseline is
missing. This is justifiable because research
covered only a period of sixteen sessions (Wid-
dowson, 2012a). Moreover, only three sin-
gle-cases have 16 treatment session, where-
as “Linda” dropped out attending only 9 and
“Alastair” therapy considered only  sessions.
All patients took part in  follow-ups, at -, -
and -month after the end of therapy (Table ).
The first two cases of Benelli (2016a, 2016b)
have a -points baseline, however both pa-
ENRICO BENELLI & MARIAVITTORIA ZANCHETTA

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
101
tients attended a prior session, after which the
therapist proposed to participate to the re-
search, but quantitative data from that session
is missing. The third case (2016c) has -point
baseline (see Table ).
The cases of “Anna” and “Caterina” (Benel-
li, 2017a, 2017b) extended the baseline to
4-points, whereas the first case with adoles-
cents (Benelli, 2017c) has a -points baseline.
All cases in both Benelli case series had  ses-
sions and  follow-ups, at -, - and -months
after the end of therapy (see Table 4).
Overall, this meta-analysis indicates a sig-
nificative global eect size for the ecacy of
change in the comparison between assessment
and treatment (AB), in particular for those
case series with a more structured form and
more data available. In fact, not all these con-
sidered cases were projected to conduct this
meta-analysis, so some statistical artefacts
have been made to calculate such eect sizes.
Shadish et al. (2014) explained that both the
number of observations (n) and the number
of cases (m) are proportional to the estimat-
ed power, and higher these are, stronger will
be the estimated power. Therefore, even if the
number of observations (1 ≤ n ≤ 4) in the as-
sessment phase is very poor, the high number
of cases (m = 9) fills this gap. So, we can arm
that the more the cases were structured, thus
fulfilling the criteria to compute a strong es-
timated power, the more significant the eect
size was. This explains why the Widdowson
series’ and the Benelli 2016 series’ eect-sizes
were not significative in the first comparison.
As previously mentioned, SCEDs have gener-
alization issues, but there are several ways to
overcome this critical aspect: (a) creating a
trial register, like the CONSORT, the Interna-
tional Standard Randomised Controlled Trial
Number (ISRCTN) set out by the WHO Interna-
tional Clinical Trials Registry Platform (ICTRP)
and the International Committee of Medical
Journal Editors (ICMJE) guidelines to prevent
selective reporting and publication bias; (b)
Table 4: The contingency table for the considered cases.
Case series Case Baseline
(A)
Treatment
(B)
Follow Up
(A)
Widdowson
(2012)
“Peter” (Widdowson, 2012a) 116 3
“Denise” (Widdowson, 2012b) 116 3
“Tom” (Widdowson, 2012c) 116 3
“Linda” (Widdowson, 2013) 1 9 3
“Alastair” (Widdowson, 2014) 114 3
Benelli (2016)
“Sara” (Benelli et al., 2016a) 216 3
“Penelope” (Benelli et al., 2016b) 216 3
“Luisa” (Benelli et al., 2016c) 316 3
Benelli (2017)
“Anna” (Benelli et al., 2017a) 416 3
“Caterina” (Benelli et al., 2017b) 416 3
“Deborah” (Benelli et al., 2017c) 316 3
SINGLE-CASE DESIGN REVIEW AND META-ANALYSIS

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
102
“randomly select” the first patient that askes
therapy with the diagnosis that research aims
to study; (c) avoid publication bias including
drop outs (following the ITT criteria); (d) man-
ualizing treatments; (d) making systematic
replications of single-cases; and (e) calculat-
ing eect sizes with meta-analyses.
Limitations
This analysis is not free from limitations.
Number of observations (n) in the assessment
phase is: one in the Widdowson series and two
in the Benelli (2016a & 2016b), leading to a
poorly calculated eect size. It is an important
limitation to this meta-analysis, but the case
series that we considered were not structured
to satisfy this requirement. We hope that fur-
ther researchers that aim to support Transac-
tional Analysis through SCEDs will make more
assessment observations in order to make fu-
ture meta-analysis statistically more powerful.
Another limitation to this study is due to miss-
ing data from the cases from the Widdowson
series, and doing a linear interpolation cre-
ated a straight trend, which not only lead to a
non-significant eect size, but it’s also poorly
representative of real clinical change.
Finally, the DHPS program needed at least two
baseline points to calculate ‘g’, VarG, PhiHat
and Rho, therefore we had to duplicate the
baseline score for Widdowson series, creating
a 2-point stable baseline.
Future Directions
Future research should be considered – spe-
cially in order to obtain a stronger and more
significative eect-size. Therefore, inspiring to
Primary Registries in the World Health Orga-
nization (WHO) Registry Network, we suggest:
(a) that future cases respected the definition
of interventional clinical trial, which is “any
research study that prospectively assigns hu-
man participants or groups of humans to one or
more health-related interventions to evaluate
the eects on health outcomes” (WHO, 2012, pp
8), therefore researchers should follow the in-
dications proposed by the WHO and CONSORT
guidelines, and register studies in databases
(such as clinicaltrials.gov for RCTs), as soon as
any patients agree to participate to research; (b)
the creation of a more structured therapy with at
least three assessment sessions in order to ob-
tain a stronger power analysis; and (c) to possess
all sessions’ scores and avoid linear interpola-
tions that level trends. As to the first point, the
inclusion in a register of the considered SCEDs
was not planned, but indirectly happened thanks
to the audio files, which have been archived.
We hope that future researchers will chose to
conduct meta-analysis of SCEDs to support
the acknowledgement of their therapy as EST,
and specifically the ecacy and eectiveness
of TA for Common Mental Disorders.
Conclusions
In the last decades, there have been many me-
ta-analyses on SCEDs, especially in education,
and these have been considered as a flexible
and “useful alternative to RCTs […] for the goal of
empirically demonstrating that an intervention is
eective […]. SCEDs are ideal for both research-
ers and clinicians working with small or very
heterogeneous populations in the development
and implementation of evidence-based prac-
tice” (Byiers, Reichle & Symons, 2012, pp. 412),
therefore, since Chambless et al. (1998) had
defined that both RCTs and SCEDs can be used
to demonstrate the ecacy of a treatment, so
there is no reason why SCEDs shouldn’t also
become a valid substitute to traditional group
designs in psychotherapy.
In fact, the considered SCEDs and this me-
ta-analysis reflect some characteristics of
RCTs’, such as: (a) a prior registration of the
cases; (b) a “randomization”, meaning in-
ENRICO BENELLI & MARIAVITTORIA ZANCHETTA

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
103
cluding in the research the first patient that
presented to therapy with a DSM diagnosis of
mood disorder and agreed to participate; (c)
the use of a manualized treatment; (d) the in-
clusion of cases of patients that dropped out to
prevent biases (ITT); (e) the use of a d statistic
(Shadish et al., 2014) that uses the same met-
rics as the one used in BSD and that allows to
compare dierent quantitative instruments
(BDI-II and PHQ-9), that can also be corrected
for small samples (Hedges’ g; Hedges, 1981);
and (f) follows the PICO strategy, where there
is a Population with a DSM diagnosis (e.g. mood
disorders), a manualized Intervention (e.g., AT
for mood disorders), a Comparison between
pre-therapy and post-therapy (instead of a
control group like in RCTs), and a measured
Outcome with highly validated quantitative
instruments (e.g., BDI-II and PHQ-9).
This form of meta-analysis is the first step in
supporting TA as an eective treatment for
mood disorders, as an EST, and it showed how
this form of psychotherapy is ecacious in
treating depression in short-term therapies.
Since the first case of Widdowson (2012a), the
methodology of these HSCEDs has improved
throughout the years and via various publica-
tions, especially in extending the baseline from
the 1-point baseline (Widdowson case series)
to the 4-point baseline (Benelli, 2017a, 2017b).
Moreover, the last case had included an ado-
lescent in the research (Benelli, 2017c), which
evaluated the ecacy of TA with mood dis-
orders, also in adolescence. With the Italian’s
systematic replications, this type of herme-
neutic analysis evolved, even focusing on the
incongruences between quantitative and qual-
itative data. Furthermore, in order to keep pace
with growing diagnosis, Widdowson (2015)
manualized TA for depression and it has also
been translated into Italian, published in 2018.
To conclude, systematic replications are an
extremely important instrument in establish-
ing external validity (Sidman, 1960), and con-
ducting a meta-analysis is the via regia (Royal
Road) for a systematic review of RCTs, where-
as single-cases are left aside assuming they
possessed a poor generality. Therefore, since
the considered SCEDs for this meta-analysis
are part of these systematic replications, they
cannot lack in generalizability. Nevertheless,
future research will be necessary to enhance
the generality of TA psychotherapy’s ecacy
for mood disorders.
SINGLE-CASE DESIGN REVIEW AND META-ANALYSIS

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
104
The following table is the obtained output from the macros used in DHPS for the AB comparison
(see Table 5).
Table 5: DHPS output for AB comparison
(for space issues only four decimals have been included)
jid SID DVID St.
Type
Err.
Code PhiHat Rho GVarG yi vi
1 1 2 1 0 ,8795 ,4248 ,7351 ,1650 ,7351 ,1650
1 2 1 1 0 ,7052 ,1885 ,7658 ,2971 ,7658 ,2971
1 3 1 1 0 ,2811 01,1627 ,1912 1,1627 ,1912
The following figure (Fig. 1) represents the AB (baseline-treatment) comparison of the three
considered case series for the meta-analysis, with respective eect sizes and CIs 95%.
17
The following table is the obtained output from the macros used in DHPS for the AB
comparison (see Table 5).
Table 5: DHPS output for AB comparison (for space issues only four decimals
have been included).
jid
SID
DVID
St.
Type
Err.
Code
PhiHat
Rho
G
VarG
yi
vi
1
1
2
1
0
,8795
,4248
,7351
,1650
,7351
,1650
1
2
1
1
0
,7052
,1885
,7658
,2971
,7658
,2971
1
3
1
1
0
,2811
0
1,1627
,1912
1,1627
,1912
The following figure (Fig. 1) represents the AB (baseline-treatment) comparison of the three considered
case series for the meta-ana lysis, with respective effect sizes and CIs 95%.
Figure 1: The forest plot of effect sizes and 95% confidence intervals of the three case series in the AB
comparison and the average of all studies.
Fig. 1: The forest plot of eect sizes and 95% confidence intervals of the
three case series in the AB comparison and the average of all studies.
Appendix A
Appendix B
ENRICO BENELLI & MARIAVITTORIA ZANCHETTA

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
105
References
ALLISON, D. B. & GORMAN, B. S. (1993). Calculating Eect Sizes for Meta-Analysis: The Case of the
Single Case. Behaviour Research and Therapy, 31(6), pp. 621-631.
APA PRESIDENTIAL TASK FORCE ON EVIDENCE-BASED PRACTICE. (2006). Evidence-Based Practice
in Psychology. The American Psychologist, 61 (4), pp. 271-285.
BECK, A. T., STEER, R. A. & BROWN, G. K. (1996). Manual for the Beck Depression Inventory-II. San Anto-
nio, TX: Psychological Corporation.
BENELLI, E., DE CARLO, A., BIFFI, D. & MCLEOD, J. (2015). Hermeneutic Single Case Ecacy Design: A
Systematic Review of Published Research and Current Standards. Testing, Psychometrics, Methodol-
ogy in Applied Psychology, 22, pp. 97-133.
* BENELLI, E., REVELLO, B., PICCIRILLO, C., MAZZETTI, M., CALVO, V., PALMIERI, A., SAMBIN, M.
& WIDDOWSON, M. (2016a). TA Treatment of Depression: A Hermeneutic Single-Case Ecacy De-
sign Study – ‘Sara’. International Journal of Transactional Analysis Research, 7 (1), pp. 3-18.
* BENELLI, E., SCOTTA, F., BARRECA, S., PALMIERI, A., CALVO, C., DE RENOCHE, G., COLUSSI, S.,
SAMBIN, M. & WIDDOWSON, M. (2016b). TA treatment of depression: a hermeneutic single-case
ecacy design study – ‘Penelope’. International Journal of Transactional Analysis Research, 7 (1), pp.
19-34.
* BENELLI, E., BOSCHETTI, D., PICCIRILLO, C., QUAGLIOTTI, L., CALVO, V., PALMIERI, A., SAMBIN,
M. & WIDDOWSON, M. (2016c). TA Treatment of Depression: A Hermeneutic Single-Case Ecacy
Design Study – ‘Luisa’, International Journal of Transactional Analysis Research, 7 (1), pp. 35-50.
* BENELLI, E., MORETTI, E., CAVALLERO, G. C., GRECO, G., CALVO, V., MANNARINI, S., PALMIERI,
A. & WIDDOWSON, M. (2017a). TA Treatment of Depression: A Hermeneutic Single-Case Ecacy
Design Study – ‘Anna’. International Journal of Transactional Analysis Research, 8 (1), pp. 3-20.
* BENELLI, E., FILANTI, S., MUSSO, R., CALVO, V., MANNARINI, S., PALMIERI, A. & WIDDOWSON, M.
(2017b). TA Treatment of Depression: A Hermeneutic Single-Case Ecacy Design Study – ‘Cateri-
na’. International Journal of Transactional Analysis Research
ENRICO BENELLI, PhD, Adjunct Professor of Dynamic Psychology, University of Padua, De-
partment of Philosophy, Sociology, Pedagogy, & Applied Psychology – FISPPA, Psychologist,
Psychotherapist, PTSTA-P.
E-mail: enrico.benelli@unipd.it
MARIAVITTORIA ZANCHETTA, Psychologist, University of Padua, Department of Philoso-
phy, Sociology, Pedagogy, and Applied Psychology – FISPPA. CTA Trainee at Centro Psicolo-
gia Dinamica (CPD) in Padua. Honorary fellowship in Dynamic Psychology at the University of
Padua.
E-mail: zanchettamv@gmail.com
Authors
SINGLE-CASE DESIGN REVIEW AND META-ANALYSIS

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
106
*BENELLI, E., BERGAMASCHI, M., CAPOFERRI, C., MORENA, S., CALVO, V., MANNARINI, S., PALM-
IERI, A., ZANCHETTA, M., SPINELLI, M. & WIDDOWSON, M. (2017c). TA Treatment of Depression:
A Hermeneutic Single-Case Ecacy Design Study – ‘Deborah’. International Journal of Transactional
Analysis Research, 8 (1), pp. 39-58.
BERETVAS, S. N. & CHUNG, H. (2008). A Review of Meta-Analyses of Single-Subject Experimental De-
signs: Methodological Issues and Practice. Evidence-Based Communication Assessment and Interven-
tion, 2 (3), pp. 129-141.
BORENSTEIN, M., COOPER, H., HEDGES, L. V. & VALENTINE, J. (2009). Eect Sizes for Continuous
Data. The Handbook of Research Synthesis and Meta-Analysis, 2, pp. 221-235.
BURNS, M. K. (2012). Meta-Analysis of Single-Case Design Research: Introduction to the Special Issue.
Journal of Behavioral Education, 21 (3), pp. 175-184.
BYIERS, B. J., REICHLE, J. & SYMONS, F. J. (2012). Single-Subject Experimental Design for Evi-
dence-Based Practice. American Journal of Speech-Language Pathology, 21 (4), pp. 397-414.
CHAMBLESS, D. L., BAKER, M. J., BAUCOM, D. H., BEUTLER, L. E., CALHOUN, K. S., CRITS-CHRIS-
TOPH, P., ... & WOODY, S. (1998). Update on Empirically Validated Therapies, II. The Clinical Psy-
chologist, 51 (1), pp. 3-16.
CHAMBLESS, D. L. & HOLLON, S. D. (1998). Defining Empirically Supported Therapies. Journal of con-
sulting and clinical psychology, 66 (1), pp. 7-18.
COHEN, J. (1988). Statistical Power Analysis for the Behavioral Sciences. London: Rutledge.
COOPER, H. & HEDGES, L. V. (2009). Potentials and Limitations. In: The Hand. of Res. Synthesis and Me-
ta-Analysis, 2nd Ed.. Russell Sage Foundation.
ELLIOTT, R. (2002). Hermeneutic Single-Case Ecacy Design. Psychotherapy Research, 12 (1), pp. 1-21.
ELLIOTT, R., PARTYKA, R., ALPERIN, R., DOBRENSKI, R., WAGNER, J., MESSER, S. B., WATSON, J. C.
& CASTONGUAY, L. G. (2009). An Adjudicated Hermeneutic Single-Case Ecacy Design Study of
Experiential Therapy for Panic/Phobia. Psychotherapy Research, 19 (4-5), pp. 543-557.
GLASS, G. V. (1976). Primary, Secondary, and Meta-Analysis of Research. Educational Researcher, 5 (10),
pp. 3-8.
GRIMES, D. A. & SCHULZ, K. F. (2002). An Overview of Clinical Research: The Lay of the Land. The Lan-
cet, 359 (9300), pp. 57-61.
GUYATT, G., OXMAN, A. D., AKL, E. A., KUNZ, R., VIST, G., BROZEK, J., ... & RIND, D. (2011). GRADE
Guidelines: 1. Introduction – GRADE Evidence Profiles and Summary of Findings Tables. Journal of
Clinical Epidemiology, 64 (4), pp. 383-394.
HEDGES, L. V. (1981). Distribution Theory for Glass’s Estimator of Eect Size and Related Estimators.
Journal of Educational Statistics, 6 (2), pp. 107-128.
KAZDIN, A. E. (1978). Methodological and Interpretive Problems of Single-Case Experimental Designs.
Journal of Consulting and Clinical Psychology, 46 (4), pp. 629-642.
KAZDIN, A. E. (1981). Drawing Valid Inferences from Case Studies. Journal of Consulting and Clinical Psy-
chology, 49 (2), pp. 183-192.
KAZDIN, A. E. (2010). Single-Case Research Designs: Methods for Clinical and Applied Settings (nd ed.). New
ENRICO BENELLI & MARIAVITTORIA ZANCHETTA

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
107
York. Oxford University Press
MCLEOD, J. (2010). Case Study Research in Counselling and Psychotherapy. London: Sage Publications.
MOHER, D., LIBERATI, A., TETZLAFF, J. & ALTMAN, D. G.: THE PRISMA GROUP. (2009). Preferred
Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med, 6
(6): e1000097, pp. 1-6.
NATIONAL COLLABORATING CENTRE FOR MENTAL HEALTH. (2009). Depression: The Treatment and
Management of Depression in Adults (Update) (NICE clinical guideline 90). London: National Institute
for Clinical Excellence. Retrieved from: www.nice.org.uk/CG90
SCHULZ, K. F., ALTMAN, D. G. & MOHER, D.: THE CONSORT GROUP. (2010). CONSORT 2010 Statement:
Updated Guidelines for Reporting Parallel Group Randomised Trials. PLoS Med, 7 (3): e1000251, pp.
1-7.
SHADISH, W. R., RINDSKOPF, D. M. & HEDGES, L. V. (2008). The State of the Science in the Me-
ta-Analysis of Single-Case Experimental Designs. Evidence-Based Communication Assessment and
Intervention, 2 (3), pp. 188-196.
SHADISH, W. R. (2014). Analysis and Meta-Analysis of Single-Case Designs: An Introduction. Journal of
School Psychology, 52 (2), pp. 109-122.
SHADISH, W. R., HEDGES, L. V. & PUSTEJOVSKY, J. E. (2014). Analysis and Meta-Analysis of Sin-
gle-Case Designs with a Standardized Mean Dierence Statistic: A Primer and Applications. Journal
of School Psychology, 52 (2), pp. 123-147.
SIDMAN, M. (1960). Tactics of Scientific Research: Evaluating Experimental Data in Psychology (Vol. 5). New
York: Basic Books.
SPITZER, R. L., KROENKE, K. & WILLIAMS, J. B. (1999). Validation and Utility of a Self-Report Version
of PRIME-MD: The PHQ Primary Care Study. Journal of the American Medical Association, Nov 10; 282
(18), pp. 1737-44.
STEEL, Z., MARNANE, C., IRANPOUR, C., CHEY, T., JACKSON, J. W., PATEL, V. & SILOVE, D. (2014). The
Global Prevalence of Common Mental Disorders: A Systematic Review and Meta-Analysis 1980–
2013. International Journal of Epidemiology, 43 (2), pp. 476-493.
STILES, W. B., HILL, C. E. & ELLIOTT, R. (2015). Looking Both Ways. Psychotherapy Research, 25 (3), pp.
282-293.
TRAUTMANN, S., REHM, J. & WITTCHEN, H. (2016). The Economic Costs of Mental Disorders: Do Our
Societies React Appropriately to the Burden of Mental Disorders? EMBO Reports, 17 (9), pp. 1245-
1249.
VIECHTBAUER, W. (2010). Conducting Meta-Analyses in R with the Metafore Package. J. of Statistical
Software, 36 (3), pp. 1-48.
WORLD HEALTH ORGANIZATION. (2003). The Mental Health Context. Geneva: Mental Health Policy &
Service Guidance Package.
WORLD HEALTH ORGANIZATION. (2012). International Standards for Clinical Trial Registries.
* WIDDOWSON, M. (2012a). TA Treatment of Depression – A Hermeneutic Single-Case Ecacy Design
Study – ‘Peter’. International Journal of Transactional Analysis Research, 3(1), pp. 3-13.
SINGLE-CASE DESIGN REVIEW AND META-ANALYSIS

INTERNATIONAL JOURNAL OF PSYCHOTHERAPY | Nov. 2019, Vol. 23, No. 3
108
* WIDDOWSON, M. (2012b). TA Treatment of Depression – A Hermeneutic Single-Case Ecacy Design
Study – ‘Denise’. International Journal of Transactional Analysis Research, 3(2), pp. 3-14.
* WIDDOWSON, M. (2012c). TA Treatment of Depression – A Hermeneutic Single-Case Ecacy Design
Study – ‘Tom’. International Journal of Transactional Analysis Research, 3(2), pp. 15-27.
* WIDDOWSON, M. (2013). TA Treatment of Depression – A Hermeneutic Single-Case Ecacy Design
Study – ‘Linda’: A Mixed Outcome Case. International Journal of Transactional Analysis Research, 4:2,
pp. 3-15.
* WIDDOWSON, M. (2014). Transactional Analysis Psychotherapy for a Case of Mixed Anxiety & De-
pression: A Pragmatic Adjudicated Case Study – ‘Alastair’. International Journal of Transactional
Analysis Research, 5:2, pp. 66-76.
WIDDOWSON, M. (2015). Transactional Analysis for Depression: A Step-by-Step Treatment Manual. Rout-
ledge.
WIDDOWSON, M, (2018). Analisi transazionale per i disturbi depressivi. Manuale per il trattamento [Tran-
sactional Analysis for Depressive Disorders. A Manual for Treatment]. Milano. Franco Angeli.
* Indicates articles included in meta-analysis (N=11).
ENRICO BENELLI & MARIAVITTORIA ZANCHETTA