The Difference Between Shorter-Versus Longer-Term Psychotherapy For Adult Mental Health Disorders: A Systematic Review With Meta-Analysis

Juul et al.
BMC Psychiatry (2023) 23:438 BMC Psychiatry

https://doi.org/10.1186/s12888-023-04895-6
RESEARCH Open Access
The difference between shorter‑

versus longer‑term psychotherapy for adult
mental health disorders: a systematic review
with meta‑analysis
Sophie Juul1,2*, Janus Christian Jakobsen2,3, Caroline Kamp Jørgensen2,3, Stig Poulsen4, Per Sørensen1 and
Sebastian Simonsen1
Abstract
Background The optimal psychotherapy duration for mental health disorders is unclear. Our aim was to assess the
beneficial and harmful effects of shorter- versus longer-term psychotherapy for adult mental health disorders.
Method We searched relevant databases and websites for published and unpublished randomised clinical trials
assessing different durations of the same psychotherapy type before June 27, 2022. Our methodology was based on
Cochrane and an eight-step procedure. Primary outcomes were quality of life, serious adverse events, and symptom
severity. Secondary outcomes were suicide or suicide-attempts, self-harm, and level of functioning.
Results We included 19 trials randomising 3,447 participants. All trials were at high risk of bias. Three single trials met
the required information size needed to confirm or reject realistic intervention effects. One single trial showed no
evidence of a difference between 6 versus 12 months dialectical behavioral therapy for borderline personality when
assessing quality of life, symptom severity, and level of functioning. One single trial showed evidence of a beneficial
effect of adding booster sessions to 8 and 12 weeks of internet-based cognitive behavioral therapy for depression
and anxiety when assessing symptom severity and level of functioning. One single trial showed no evidence of a
difference between 20 weeks versus 3 years of psychodynamic psychotherapy for mood- or anxiety disorders when
assessing symptom severity and level of functioning. It was only possible to conduct two pre-planned meta-analyses.
Meta-analysis showed no evidence of a difference between shorter- and longer-term cognitive behavioural therapy
for anxiety disorders on anxiety symptoms at end of treatment (SMD: 0.08; 95% CI: -0.47 to 0.63; p = 0.77; I2 = 73%; four
trials; very low certainty). Meta-analysis showed no evidence of a difference between shorter and longer-term psy-
chodynamic psychotherapy for mood- and anxiety disorders on level of functioning (SMD 0.16; 95% CI -0.08 to 0.40;
p = 0.20; I2 = 21%; two trials; very low certainty).
Conclusions The evidence for shorter versus longer-term psychotherapy for adult mental health disorders is cur-
rently unclear. We only identified 19 randomised clinical trials. More trials at low risk of bias and at low risk of random
errors assessing participants at different levels of psychopathological severity are urgently needed.
Systematic review registration PROSPERO CRD42019128535.
*Correspondence:
Sophie Juul
sophie.juul@regionh.dk
Full list of author information is available at the end of the article
© The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco
mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Juul et al. BMC Psychiatry (2023) 23:438 Page 2 of 23
Keywords Mental health disorders, Duration of psychotherapy, Systematic review, Meta-analysis
Background systematic review is described in detail in our protocol

The annual prevalence of mental health disorders is esti- [6], which was also registered in the PROSPERO database
mated to be 38.2% of the European population [1]. The (CRD42019128535) prior to the systematic literature
economic burden from mental health disorders is high, search.
both because of direct health care costs, but also because
of indirect costs like sick days, disability, and early retire-
Search strategy and selection criteria
ment [1–3]. Psychotherapy is among the recommended
Electronic searches
and widely used interventions for most disorders [4].
An experienced information specialist searched for eli-
Accordingly, it would be highly relevant to identify the
gible trials comparing a shorter with a longer-term ver-
optimal duration of psychotherapy for various mental
sion of the same psychotherapy type for one or more
health disorders and conditions. If short-term psycho-
adult mental health published before June 27, 2022 in the
therapy is the optimal treatment approach for a given dis-
following databases: Cochrane Central Register of Con-
order, this could result in a reduction of waitlists and thus
trolled Trials (CENTRAL), Medical Literature Analy-
a greater access to evidence-based care. On the contrary,
sis and Retrieval System Online (MEDLINE), Excerpta
if long-term psychotherapy is the most optimal treat-
Medica database (EMBASE), Latin American and Car-
ment, it would be sensible for mental health systems to
ibbean Health Sciences Literature (LILACS), PsycINFO,
invest in these treatments, as they would translate into
Science Citation Index Expanded (SCI-EXPANDED),
greater long-term health and occupational benefits [5, 6].
Social Sciences Citation Index (SSCI), Conference Pro-
The relationship between dose and effect in psycho-
ceedings Citation Index—Science (CPCI-S), and Con-
therapy has been studied with mixed results in non-
ference Proceedings Citation Index—Social Science &
controlled studies [5, 7]. While several non-controlled
Humanities (CPCI-SSH). The electronic search strategies
studies indicate that there is a linear or negatively accel-
can be found in Supplementary material 2. Additionally,
erating relationship between number of psychotherapy
we checked the reference lists of relevant publications for
sessions and outcome for most mental health disorders
any unidentified trials, and we hand searched conference
[8, 9], these findings have been criticized on methodo-
abstracts from psychiatry conferences for relevant trials.
logical grounds [10].
We also considered unpublished and gray literature trials
The inconclusiveness of the existing research and the
if these were identified.
general lack of internal validity of non-controlled studies
[11, 12] indicate the need for a systematic review of well-
Inclusion and exclusion criteria
designed randomised clinical trials directly comparing
We only included randomized clinical trials. Trials were
psychotherapies of different durations for clearly speci-
included irrespective of setting, publication status, publi-
fied populations, including patients treated for mental
cation year, language, and the reporting of our outcomes.
health disorders in secondary mental health care settings
We relied on the trialists defining their compared inter-
[11, 12]. However, such systematic review has not previ-
ventions as shorter and longer-term (or similar terminol-
ously been performed [6].
ogy). We did not include cluster randomized trials, quasi
The present systematic review aims at forming the
randomized trials, or observational studies.
basis for evidence-based guideline recommendations for
the optimal duration of psychotherapy for adult mental
health disorders taking both benefits and harms, bias risk Data extraction and risk of bias assessment
(systematic errors), play of chance (random errors), and Two review authors (SJ, CKJ) independently screened
certainty of the findings into consideration. relevant trials, extracted data using a standardised data
extraction sheet, and assessed the risk of bias accord-
ing to the Risk of Bias (ROB) assessment tool provided
Methods in Cochrane Handbook of Systematic Reviews of Inter-
We report this systematic review in accordance with ventions [14]. Any discrepancies were resolved through
the Preferred Reporting Items for Systematic Reviews discussion or, if required, through discussion with a third
and Meta-Analysis (PRISMA) guidelines [13] A author (JCJ, SS). We contacted trial authors by e-mail if
PRISMA 2020 checklist can be found in Supplemen- relevant data were unclear or missing. For more informa-
tary material 1. The Cochrane methodology used in this tion on our risk of bias assessments, see our protocol [6].
Outcomes and subgroup analyses information, we calculated the required information size
Our primary outcomes were quality of life, serious for each single trial result and assessed if there was ade-
adverse events (as defined by the ICH-GCP guidelines) quate power to confirm or reject realistic intervention
[15], and symptom severity. Our secondary outcomes effects of single trial results. For dichotomous outcomes,
were suicide or suicide attempts (dichotomous data), we used the proportion of participants with an event in
self-harm (dichotomous data), and level of functioning. the control group, a relative risk reduction of 20%, an
For all outcomes, we used the trial results reported at the alpha of 1.4%, and a beta of 20% as predefined in our pro-
time point closest to the end of treatment in the long- tocol [6]. For continuous outcomes, we used the observed
term treatment group. mean and standard deviation for the control group, the
We planned the following subgroup analyses on our observed mean in the control group plus or minus the
primary outcomes: observed standard deviation in the control group/2 for
the experimental group, an alpha of 1.4%, and a beta of
• High risk of bias trials compared to low risk of bias 20% as predefined in our protocol [6]. We assessed a total
trials of six primary and secondary outcome and, hence, con-
• Types of mental health disorders sidered a p-value of 0.014 as the threshold for statistical
• Types of psychotherapy comparisons significance [17, 30]. We performed independent samples
• Trials above and below the mean difference in inter- t-tests to calculate p-values for single trial results for con-
vention lengths tinuous outcomes, and Fisher’s exact test for single trial
results for dichotomous outcomes. We used The Grad-
ing of Recommendations Assessment, Development and
Assessment of statistical and clinical significance Evaluation (GRADE) to assess the certainty of evidence
We performed our meta-analyses according to the rec- [17, 31–33].
ommendations stated in the Cochrane Handbook for
Systematic Reviews of Interventions [14], Keus et al. [16], Results
and the eight-step procedure suggested by Jakobsen et al. Study characteristics
[17] for better validation of meta-analytic results in sys- On June 27, 2022 our literature search identified a total of
tematic reviews. Review Manager 5.4 and Stata 16 were 31,689 records after duplicates were removed (Fig. 1). We
used for all meta-analyses [18, 19]. We planned to use included 19 randomised clinical trials enrolling a total of
risk ratios (RR) for dichotomous outcomes, mean dif- 3,447 participants [34–52] (McMain S: The effectiveness
ferences (MD) for continuous outcomes assessed with of 6 versus 12-months of dialectical behaviour therapy
homogeneous measures, and standardised mean differ- for borderline personality disorder: the feasibility of a
ence (SMD) for continuous outcomes with heterogene- shorter treatment and evaluating responses (FASTER)
ous measures. We reported both the random-effects trial, Unpublished) (Supplementary material 3). A list of
and the fixed-effect meta-analysis results, but primar- excludes studies with reasons can be found in Supple-
ily emphasized the most conservative result (highest P mentary material 4.
value) of the two results, and considered the less con- Characteristics of included trials can be found in
servative results a sensitivity analysis [17]. We used the Table 1. All trials were assessed as at high risk of bias
best–worst/worst-best case scenarios to assess the poten- (Supplementary material 5). Five trials assessed the dif-
tial impact of missing outcome data [6, 17]. We planned ference between shorter- and longer-term cognitive
to use Trial Sequential Analysis to control for random behavioural therapy for anxiety disorders [36–38, 42,
errors and to report Trial Sequental Analysis-adjusted 48]. Four trials assessed the difference between shorter-
CIs if the cumulative Z-curves did not reach the futil- and longer-term cognitive behavioural therapy for major
ity area or passed the diversity-adjusted required infor- depressive disorder [39–41, 43]. Three trials assessed the
mation size (DARIS) [6, 17, 20–28]. Trial Sequential difference between shorter- and longer-term psycho-
Analysis estimates the DARIS (that is the number of par- dynamic psychotherapy for major depressive disorder
ticipants needed in a meta-analysis to detect or reject a [40, 41, 44]. Two trials assessed the difference between
certain intervention effect). When analysing continuous shorter- and longer-term psychodynamic psychotherapy
outcomes, we pragmatically anticipated an intervention for mood- and anxiety disorders [34, 35]. Three trials
effect equal to the MD of the observed SD/2 [29]. Het- assessed the difference between shorter- and longer-term
erogeneity was assessed by calculating inconsistency (I2) prolonged exposure for post-traumatic stress disorder
for traditional meta-analyses and diversity (D2) for Trial [46, 47, 49]. One trial assessed the difference between
Sequential Analysis. If it was not possible to perform shorter- and longer-term interpersonal therapy for major
Trial Sequential Analysis to estimate if there was enough depressive disorder [39]. One trial assessed the difference
Fig. 1 PRISMA flow diagram
between shorter- and longer-term cognitive behavioural 16 sessions delivered over 16 weeks) [34, 35, 40–42, 44,
therapy for post-traumatic stress disorder [45]. One trial 45, 48, 50, 51] (McMain S: The effectiveness of 6 versus
assessed the difference between shorter- and longer-term 12-months of dialectical behaviour therapy for borderline
internet-based cognitive behavioural therapy for post- personality disorder: the feasibility of a shorter treatment
traumatic stress disorder [51]. One factorial trial com- and evaluating responses (FASTER) trial, Unpublished).
pared internet-based cognitive behavioral therapy for Some trials compared different numbers of sessions
8 versus 12 weeks with or without booster sessions for delivered over the same duration (e.g. six sessions deliv-
depression and anxiety [50]. One trial assessed the differ- ered over 12 weeks compared with 12 sessions delivered
ence between shorter- and longer-term dialectical behav- over 12 weeks) [37, 38]. Some trials compared the same
ioural therapy for borderline personality disorder [52, 53] number of sessions over different durations (e.g. 10 ses-
(McMain S: The effectiveness of 6 versus 12-months of sions delivered over two weeks compared with 10 ses-
dialectical behaviour therapy for borderline personality sions delivered over 10 weeks) [36, 39, 46, 49]. Two trials
disorder: the feasibility of a shorter treatment and evalu- compared the same number of sessions, but with differ-
ating responses (FASTER) trial, Unpublished). ent sessions lengths in minutes (e.g. 10–15 sessions of
All trials compared different durations (weeks of treat- 60 min compared with 10–15 sessions of 90 min) [43, 47].
ment), dosages (number of sessions), and session lengths We planned to assess serious adverse events. However,
(minutes) (Table 1). Furthermore, trialists’ definitions of only one of the trials reported on this outcome (McMain
short-term and long-term psychotherapy were not con- S: The effectiveness of 6 versus 12-months of dialecti-
sistent across studies. Most trials compared different cal behaviour therapy for borderline personality disor-
numbers of sessions delivered over different durations der: the feasibility of a shorter treatment and evaluating
(e.g. 8 sessions delivered over 8 weeks compared with responses (FASTER) trial, Unpublished). For several of
Table 1 Characteristics of included trials
Trial Country Number of Mental health disorder Shorter-term Longer-term Overall Primary outcome
randomised intervention intervention risk of
participants bias
Barkham et al. 1996a [40] United Kingdom 54 Major depressive disorder 8 sessions CBT (8 weeks) 16 sessions CBT (18 weeks) High No primary outcome was
Juul et al. BMC Psychiatry
8 sessions psychody- 16 sessions psychody- reported

namic-interpersonal namic-interpersonal
therapy (8 weeks) therapy (18 weeks)
Bohni et al. 2009a [42] Denmark 48 Panic disorder 8 sessions CBT (3 weeks) 13 sessions CBT (13 weeks) High No primary outcome was
reported
(2023) 23:438
Böttche et al. 2021 [51] Egypt, Saudi Arabia, 224 PTSD 6 sessions internet-based 10 sessions internet-based High Primary outcome was PTSD
Algeria, Syria, CBT (3 weeks) CBT (5 weeks) symptoms as measured
Morocco, Palestine with the Posttraumatic
Stress Diagnostic Scale (PDS)
Bruijniks et al. 2020 [39] The Netherlands 200 Major depressive disorder 20 sessions CBT (24 weeks) 20 sessions CBT (16 weeks) High Primary outcome was
20 sessions interpersonal 20 sessions interpersonal depression severity as meas-
therapy (24 weeks) therapy (16 weeks) ured with the BDI-II
Christensen et al. 2006a Australia 931 Major depressive disorder Brief online CBT and Extended online CBT and High Primary outcome was
[43] problem solving (unclear problem solving (unclear depression severity as meas-
duration) duration) ured with the Goldberg
Depression Scale
Clark et al. 1999 [37] United Kingdom 29 Panic disorder 5 sessions CBT (12 weeks) 12 sessions CBT (12 weeks) High No primary outcome was
reported
Dekker et al. 2005 [44] The Netherlands 103 Major depressive disorder 8 sessions short psychody- 16 sessions short psy- High No primary outcome was
namic supportive psycho- chodynamic supportive reported
therapy (8 weeks) psychotherapy (24 weeks)
Dell et al. 2022 [49] Australia 138 PTSD 10 sessions prolonged 10 sessions prolonged High Primary outcome was
exposure (2 weeks) exposure (10 weeks) severity of PTSD symptoms
assessed with the Clinician-
Administered PTSD Scale
(CAPS).3
Ehlers et al. 2014 [45] United Kingdom 61 PTSD 14 sessions cognitive 12 sessions cognitive High Primary outcome was
therapy (5 weeks) therapy (12 weeks) severity of PTSD symptoms
assessed with the Clinician-
Administered PTSD Scale
(CAPS)
Foa et al. 2018 [46] USA 219 PTSD 10 sessions prolonged 10 sessions prolonged High Primary outcome was
exposure therapy exposure therapy severity of PTSD symptoms
(2 weeks) (8 weeks) assessed with the PTSD
Symptom Scale–Interview
(PSS-I)
Page 5 of 23
Table 1 (continued)
Trial Country Number of Mental health disorder Shorter-term Longer-term Overall Primary outcome
randomised intervention intervention risk of
participants bias
Hadjistavropoulos et al. Canada 434 Depression and/or anxiety 8 sessions internet-based 11 sessions internet-based High Primary outcomes were
2022 [50] CBT (8 weeks) CBT (8 weeks + 3 booster severity of depression and
sessions) anxiety assessed with the
Patient Health Question-
12 sessions internet-based 15 sessions internet-based naire (PHQ-9) and the Gen-
CBT (12 weeks) CBT (12 weeks + 3 booster eralized Anxiety Disorder-7
(2023) 23:438
sessions) (GAD-7)
Herbert et al. 2004 [36] USA 34 Social anxiety disorder 12 sessions CBT (12 weeks) 12 sessions CBT (18 weeks) High No primary outcome was
reported
Kenardy et al. 2003 [48] Australia and Scotland 81b Panic disorder 6 sessions CBT (6 weeks) 12 sessions CBT (12 weeks) High Primary measures included
a comprehensive battery of
panic and anxiety measures
Knekt et al. 2008 [34] Finland 229 Mood- and anxiety 20 sessions psychody- 468 sessions psychody- High Primary outcomes were
disorders namic therapy (20 weeks) namic therapy (156 weeks) depressive and anxiety
symptoms
Lorentzen et al. 2013 [35] Norway 167 Mood- anxiety- and per- 20 sessions psychody- 80 sessions psychody- High No primary outcome was
sonality disorders namic group therapy namic group therapy reported
(20 weeks) (80 weeks)
McMain et al Canada 240 BPD 26 sessions DBT (26 weeks) 52 sessions DBT (52 weeks) High Primary outcome was
frequency of suicidal or
non-suicidal self-injurious
episodes
Nacasch et al. 2015 [47] Israel 40 PTSD 10–15 sessions prolonged 10–15 sessions prolonged High Primary outcome was
exposure therapy (60 min) exposure therapy (90 min) severity of PTSD symptoms
assessed with the Post-
traumatic Symptom Scale-
Interview (PSS-I)
Roberge et al. 2008 [38] Canada 65 Panic disorder with agora- 7 sessions CBT (16 weeks) 14 sessions CBT (15 weeks) High No primary outcome was
phobia reported
Shapiro et al. 1994a [41] United Kingdom 150 Major depressive disorder 8 sessions CBT (8 weeks) 16 sessions CBT (18 weeks) High No primary outcome was
8 sessions psychody- 16 sessions of psychody- reported
namic-relationship-ori- namic-relationship-ori-
ented therapy (8 weeks) ented therapy (18 weeks)
a
The results of these trials were not reported in a usable way; i.e. the results were reported in a graph, and standard deviations were not provided for the point estimates
b
This trial randomised a total of 186 participants to four groups. The number of randomised participants for the two relevant groups were not sufficiently reported, as only the number of participants who commenced
treatment was reported
BPD Borderline personality disorder, CBT Cognitive behavior therapy, DBT Dialectical behavior therapy, PTSD Post-traumatic stress disorder
Page 6 of 23
our review outcomes it was not possible to conduct to incomplete outcome data (Supplementary material
meta-analysis due to insufficient data. Four trials did 5), and the certainty of evidence was assessed as “very
not report the results in a usable way [40–43], i.e. they low” for all outcomes (Supplementary material 7).
reported the results on a graph and/or did not include
standard deviations for each point estimate on a group Trials including participants with mood‑ and anxiety
level. We contacted trial authors to receive relevant data, disorders
but we have not received any responses. It was not possi- We identified three trials assessing the effects of shorter-
ble to perform Trial Sequential Analyses to assess the risk versus longer-term psychotherapy for mood- and anxiety
of random errors on any of our review outcomes because disorders [34, 35, 50].
of lack of relevant data. Only a few trials reported on our One trial randomising 229 participants with mood- and
dichotomous outcomes, and the continuous outcomes anxiety disorders to 20 weeks versus 156 weeks of psy-
were assessed with heterogeneous measures. We there- chodynamic psychotherapy [34] showed no evidence of
fore performed sample size calculations for all single trial a difference when assessing symptom severity (p = 0.037,
results to estimate the required information size needed required information size reached), considering our
to confirm or reject realistic intervention effects for all adjusted threshold for significance was pre-defined
outcomes. Results of these sample size calculations can at 0.014 in our protocol [6], or level of functioning
be found in Supplementary material 6. (p = 0.066, required information size reached). The trial
Due to the large heterogeneity in participants, inter- almost reached their sample size (230 participants) [34],
ventions, and lengths of trials included in this review, we but it was unclear whether this sample size was pre-
will present the single trial results first. Second, we will defined. One trial randomising 167 participants with
present the meta-analysis results. mood- and anxiety disorders to 20 weeks versus 80 weeks
of psychodynamic psychotherapy [35] showed no evi-
Single trial results dence of a difference when assessing the proportion of
Trials including participants with borderline personality participants with a suicide or a suicide attempts (zero
disorder events in both groups) or level of functioning (p = 0.889,
We identified one trial randomising 240 participants required information size not reached) (Table 2, Sup-
with borderline personality disorder to six months ver- plementary material 6). Both trials were assessed at high
sus 12 months dialectical behavioral therapy [52, 53] risk of bias (Supplementary material 5) and the certainty
(McMain S: The effectiveness of 6 versus 12-months of of evidence was assessed as “very low” for all outcomes
dialectical behaviour therapy for borderline personal- (Supplementary material 8). These two trials are included
ity disorder: the feasibility of a shorter treatment and in a meta-analysis (see below).
evaluating responses (FASTER) trial, Unpublished). We We also identified one factorial trial randomising 496
retrieved the data through the published trial report participants with major depressive disorder and anxiety
and personal communication with the trialists. This disorders to internet-based cognitive behavioral therapy
trial reported data on all our pre-defined review out- for 8 versus 12 weeks with or without 3 booster ses-
comes. It was not possible to include the trial in a pre- sions [50]. This trial showed no evidence of a difference
defined meta-analysis, as it was the only trial including when assessing quality of life for either of the two pair-
participants with borderline personality disorder. The wise comparisons (8 weeks versus 8 weeks plus boost-
trial reached their pre-calculated sample size of 240 ers p = 0.858; 12 weeks versus 12 weeks plus boosters
participants [52, 53]. The trial showed no evidence of p = 0.089; required information size reached). The trial
a difference between short-term and long-term dialec- showed evidence of a beneficial effect of adding booster
tical behavioral therapy when assessing quality of life sessions in both pairwise comparisons when assessing
(p = 0.831, required information size reached), serious symptom severity (8 weeks versus 8 weeks plus boosters
adverse events (p = 1, required information size not p = 0.01; 12 weeks versus 12 weeks plus boosters p = 0.01;
reached), symptom severity (p = 0.833, required infor- required information size reached) and level of func-
mation size reached), suicide or suicide attempts (p = 1, tioning (8 weeks versus 8 weeks plus boosters p = 0.01;
required information size not reached), self-harm 12 weeks versus 12 weeks plus boosters p = 0.01; required
(p = 0.28, required information size not reached), and information size reached) (Table 2, Supplementary mate-
level of functioning (p = 0.731, required information rial 6). Both trials were assessed at high risk of bias (Sup-
size reached) (Table 2, Supplementary material 6). This plementary material 5), and the certainty of evidence was
trial was assessed as at overall high risk of bias due to assessed as “very low” for all outcomes (Supplementary
lack of blinding of participants and personnel, and due materials 9 and 10).
Table 2 Single trial results
Trials including participants with major depressive disordera
Trial characteristics Primary review outcomes Secondary review outcomes Trialists’ own
conclusions
Trial ID Shorter-term Longer-term Quality of Life Serious Adverse Symptom sever- Suicide/Suicide Self-harm Level of func-
intervention intervention Events ity attempts tioning

Barkham et al. 8 sessions CBT 16 sessions CBT - - - - - - Clients given 16
1996 [40] (8 weeks) (18 weeks) sessions showed
8 sessions 16 sessions a statistically sig-
psychodynamic- psychodynamic- nificant advantage
interpersonal interpersonal over clients given
(2023) 23:438
therapy (8 weeks) therapy 8 sessions on some

(18 weeks) measures at some
assessments
Bruijniks et al. 20 sessions CBT 20 sessions CBT CBT: The mean - CBT: The mean - - - In clinical practice
2020 [39] (24 weeks) (16 weeks) RAND-36 scores (SD) BDI scores at settings, delivery
20 sessions inter- 20 sessions inter- at EoT were 50.13 EoT were 24.16 of twice weekly
personal therapy personal therapy (22.20) for the (15.09) for the sessions of CBT and
(24 weeks) (16 weeks) short-term group short-term group IPT for depres-
(n = 49) and 51.53 (n = 37) and 21.25 sion was superior
(22.36) for the (12.90) for the to once weekly
long-term group long-term group sessions when
(n = 39) (p = 0.77) (n = 35) (p = 0.38) assessing depres-
IPT: The mean IPT: The mean sion outcomes
RAND-36 scores (SD) BDI scores at
at EoT were 46.8 EoT were 22.91
(20.46) in the (14.75) for the
short-term group short-term group
(n = 36) and 53.46 (n = 34) and 20.02
(20.67) in the (16.05) for the
long-term group long-term group
(n = 47) (p = 0.14) (n = 39) (p = 0.42)
Christensen Brief online CBT Extended online - - - - - - Brief CBT-based
2006 [43] and problem CBT and problem interventions are
solving (unclear solving (unclear not as effective as
duration) duration) extended interven-
tions
Page 8 of 23
Table 2 (continued)
Dekker et al. 8 sessions short 16 sessions short The mean (SD) - The mean (SD) - - - Eight or 16 psycho-
2005 [44] psychodynamic psychodynamic QLDS scores at HDRS scores therapy sessions in
supportive supportive EoT were 22.6 at EoT were addition to 8 ses-
psychotherapy psychotherapy (8.6) for the 11.1 (6.8) for sions of pharma-
(8 weeks) (24 weeks) short-term group the short-term cotherapy over a
(n = 45) and group (n = 45) period of 6 months

22.8 (8.3) for the and 12.1 (7.6) for would appear to
long-term group the long-term be equally effective
(n = 45) (p = 0.911) group (n = 45) in terms of dealing
(p = 0.512) with symptoms
Shapiro et al. 8 sessions CBT 16 sessions CBT - - - - - - There is no added
(2023) 23:438
1994 [41] (8 weeks) (18 weeks) benefit from 16

treatment sessions
8 sessions 16 sessions of compared with 8
psychodynamic- psychodynamic-
relationship- relationship-
oriented therapy oriented therapy
(8 weeks) (18 weeks)
Trials including participants with anxiety disordersa
conclusions
Bohni et al. 2009 8 sessions CBT 13 sessions CBT - - - - - - Patients in massed
[42] (3 weeks) (13 weeks) CBT achieved their
results at a faster
rate than patients
in spaced CBT,
with outcomes
after 3 weeks
in massed CBT
comparable with
those achieved
after approximately
3 months in spaced
CBT
Page 9 of 23
Table 2 (continued)
Clark et al. 1999 5 sessions CBT 12 sessions CBT - - The mean (SD) - - - Brief CT did not
[37] (12 weeks) (12 weeks) BAI scores at EoT differ from full CT
were 9.8 (6.7) for at posttreatment
the short-term or at follow-up,
group (n = 14) and effect sizes
and 8.4 (8.0) for were essentially the

the long-term same
group (n = 15)
(p = 0.615). This
result is included
in a meta-analysis
(2023) 23:438
Herbert et al. 12 sessions CBT 12 sessions CBT - - The mean (SD) - - - The results revealed
2004 [36] (12 weeks) (18 weeks) SPAI-SP scores at that the standard
EoT were 76.71 treatment program
(47.18) for the in which therapy
short-term group was provided
(n = 15) and over 12 successive
113.77 (39.56) for weeks resulted
the long-term in more rapid
group (n = 19) symptom reduc-
(p = 0.018). This tion and lower
result is included dropout relative to
in a meta-analysis the extended treat-
ment delivered
over 18 weeks
Kenardy et al. 6 sessions CBT 12 sessions CBT - - The mean (SD) - - - A brief version per-
2003 [48] (6 weeks) (12 weeks) STAI-T scores at forms significantly
EoT were 47.86 worse than the
(12.31) for the standard duration
short-term group treatment at post-
(n = 39) and treatment
41.10 (13.14) for
the long-term
group (n = 42)
(p = 0.0195). This
result is included
in a meta-analysis
Page 10 of 23
Table 2 (continued)
Roberge et al. 7 sessions CBT 14 sessions CBT - - The mean (SD) - - - Brief CBT effective-
2008 [38] (16 weeks) (15 weeks) PAS scores at EoT ness appears
were 10.2 (8.8) for comparable to
the short-term standard CBT in the
group (n = 32) short term
and 9.5 (10.3) for

the long-term
group (n = 33)
(p = 0.77). This
result is included
in a meta-analysis
(2023) 23:438
Trials including participants with mood- and anxiety disordersa

conclusions
Knekt et al. 2008 Short-term Long-term - - The mean (SD) - - The mean (SD) Patients receiving
[34] psychodynamic psychodynamic HDRS scores SAS-work scores short-term psycho-
therapy therapy at EoT were at EoT were dynamic psycho-
10.8 (5.65) for 1.88 (0.55) for therapy recovered
the short-term the short-term faster from both
group (n = 83) group (n = 83) depressive and
and 9.0 (6.0) for and 1.72 (0.62) anxiety symptoms
the long-term for the long-term during the first year
group (n = 107) group (n = 107). of follow-up. Dur-
(p = 0.037) (p = 0.066). This ing the following
result is included 2 years, the symp-
in a meta-analysis toms persisted at
the level reached
in the brief therapy
group, whereas in
the long-term psy-
chodynamic psy-
chotherapy group
the improvement
continued during
the entire 3-year
period. In the long
run, long-term
psychodynamic
psychotherapy
thus gave greater
benefits than those
achieved by the
brief therapies
Page 11 of 23
Table 2 (continued)
Lorentzen et al. Short-term Long-term - - - There were - The mean (SD) We observed that
2013 [35] psychodynamic psychodynamic 0/77 suicides or GAF scores at EoT short- and long-
group therapy group therapy suicide attempts were 67.8 (11.7) term therapy were
in the short-term for the short-term equally effective
group compared group (n = 71) across 3 years,
to 0/90 in the and 68.1 (14.2) using IIP, GAF-S

long-term group for the long-term and GAF-F as the
(p = not appli- group (n = 79) outcome variables.
cable) (p = 0.889). This However, there
result is included was a trend in
in a meta-analysis favour of long-term
(2023) 23:438
therapy (P = 0.10)
using GAF-S as the
outcome variable
Hadjistavropou- Internet-based Internet-based The mean (SD) - The mean (SD) - - The mean (SD) No significant
los et al. 2022 CBT (8 weeks) CBT (8 weeks + 3 EQ-5D-5L scores PHQ-9 scores at SDS scores at EoT group differences
[50] booster sessions) at EoT were 71.92 EoT were 7.93 were 13.86 (7.86) were found in this
(18.93) for the (5.36) for the for the short-term study
short-term group short-term group group (n = 79)
(n = 79) and 71.36 (n = 79) and 5.84 and 10.66 (8.7)
(21.34) for the (5.07) for the for the long-term
long-term group long-term group group (n = 87)
(n = 87) (p = 0.858) (n = 87) (p = 0.01) (p = 0.01)
The mean (SD)
GAD-7 scores at
EoT were 7.56
(5.37) for the
short-term group
(n = 79) and 5.56
(4.60) for the
long-term group
(n = 87) (p = 0–01)
Page 12 of 23
Table 2 (continued)
Internet-based Internet-based The mean (SD) - The mean (SD) - - The mean (SD)
CBT (12 weeks) CBT (12 weeks + 3 EQ-5D-5L scores PHQ-9 scores SDS scores at EoT
booster sessions) at EoT were 74.06 at EoT were were 10.29 (8.01)
(15.94) for the 6.52 (5.23) for for the short-term
short-term group the short-term group (n = 87)
(n = 87) and 69.13 group (n = 87) and 11.39 (8.46)

(21.93) for the and 7.55 (6.24) for the long-term
long-term group for the long-term group (n = 91)
(n = 91) (p = 0.089) group (n = 91) (p = 0.374)
(p = 0.235)
The mean (SD)
(2023) 23:438
GAD-7 scores
at EoT were
6.33 (5.19) for
the short-term
group (n = 87)
and 6.96 (5.8) for
the long-term
group (n = 91)
(p = 0.446)
Trials including participants with post-traumatic stress disordera
Trial characteristics Primary review outcomes Secondary review outcomes Trialists’ own con-
clusions
Trial ID Shorter-term Longer-term Quality of Life Serious Adverse Symptom severity Suicide/Suicide Self-harm Level of function-
intervention intervention Events attempts ing
Böttche et al. 6 sessions 10 sessions The mean (SD) - The mean (SD) - - - The shorter condi-
2021 [51] internet-based internet-based EUROHIS-QOL-8 PDS change tion results in the
CBT (3 weeks) CBT (5 weeks) scores at EoT scores at EoT same symptom
were 5.53 (0.83) were -14.73 change and
for the short- (1.45) for the dropout rate as the
term group short-term group longer condition
(n = unclear) and (n = unclear)
5.11 (1.02) for the and -15.03 (1.64)
long-term group for the long-
(n = unclear) term group
(p = 0.75) (n = unclear)
(p = 0.89)
Page 13 of 23
Table 2 (continued)
Dell et al. 2022 Massed pro- Standard pro- - 0 events The mean (SD) 0 events - - Massed prolonged
[49] longed expoure longed exposure CAPS scores at exposure was non-
EoT were 27.69 inferior to standard
(18.42) for the prolonged expo-
short-term group sure in reducing
(n = 63) and symptoms of PTSD

25.68 (16.59) for
the long-term
group (n = 71)
(p = 0.664)
Ehlers et al. Intensive cogni- Standard cogni- The mean (SD) - The mean (SD) - - The mean (SD) A novel 7-day
(2023) 23:438
2014 [45] tive therapy tive therapy Q-LES-Q scores CAPS scores at SDS scores at EoT intensive version of
at EoT were 52.67 EoT were 32.22 were 9.30 (8.20) cognitive therapy
(20.21) for the (27.20) for the for the short-term for PTSD was well
short-term group short-term group group (n = 30) tolerated, achieved
(n = 30) and 62.93 (n = 30) and and 10.02 (9.76) faster symptom
(21.70) for the 26.97 (28.68) for for the long-term reduction, and led
long-term group the long-term group (n = 31) to comparable
(n = 31) (p = 0.061) group (n = 31) (p = 0.757) overall outcomes
(p = 0.466) as the standard
once-weekly
cognitive therapy
delivered over
3 months
Foa et al. 2018 Massed pro- Extended pro- - - The mean (SD) - - - Among active duty
[46] longed exposure longed exposure PSS-I scores at military personnel
EoT were 18.88 with PTSD, massed
(no SD reported) prolonged expo-
for the short-term sure therapy (10
group (n = 110) sessions delivered
and 18.34 (no SD over 2 weeks) was
reported) for the noninferior to
long-term group spaced prolonged
(n = 110) (p = not exposure therapy
applicable) (10 sessions deliv-
ered over 8 weeks)
Page 14 of 23
Table 2 (continued)
Nacasch et al. 60 min sessions 90 min sessions - - The mean (SD) - - - In sum, 20-min
2015 [47] of prolonged of prolonged PSS-I scores at imaginal exposure
exposure exposure EoT were 13.3 within 60-min ses-
(9.52) for the sions yielded non-
short-term group inferior outcomes
(n = 20) and in PTSD symptoms

12.24 (8.02) for and posttrau-
the long-term matic negative
group (n = 17) cognitions at
(p = 0.719) posttreatment and
follow-up to the
(2023) 23:438
40-min imaginal
exposures and
90-min sessions
Trials including participants with borderline personality disordera
Trial characteristics Primary review outcomes Secondary review outcomes Trialists’ own con-
clusions
Trial Shorter-term Longer-term Quality of Life Serious Adverse Symptom severity Suicide/Suicide Self-harm Level of function-
intervention intervention Events attempts ing
McMain 6 months of DBT 12 months of DBT The mean (SD) 2 / 90 participants The mean (SD) 2 / 90 par- 28 / 90 partici- The mean (SD) Half the dose of
et al. 2022 [52] overall EQ5DL had one or more BSL scores at EoT ticipants had pants had one or SAS scores at EoT the standard DBT
(McMain S: The scores at EoT serious adverse were 38.6 (22.4) one suicide or more deliber- were 2.51 (0.58) yielded noninferior
effectiveness of 6 were 60.7 (21.43) events in the for the short-term suicide-attempts ate self-harm for the short-term improvements
versus 12-months for the short-term short-term group group (n = 90) in the short-term incidents in the group (n = 90) across time points
of dialecti- group (n = 91) at EoT compared compared with group at EoT short-term group compared with for the primary
cal behaviour compared with with 2 / 93 in the 39.3 (22.2) in compared with at EoT compared 2.54 (0.59) in outcome, total self-
therapy for 61.41 (23.17) in long-term group the long-term 2 / 93 in the with 37/ 93 in the the long-term harm frequency,
borderline per- the long-term (p = 1) (based on group (n = 91) long-term group long-term group group (n = 91) as well as several
sonality disorder: group (n = 90) suicide/ suicide (p = 0.833) (p = 1) (p = 0.28) (p = 0.731) clinical outcomes
the feasibility of a (p = 0.831) attempt data
shorter treatment only)
and evaluat-
ing responses
(FASTER) trial,
Unpublished)
a
Data is presented for the primary time-point of assessment (end of treatment)
BAI Beck Anxiety Inventory, BSL Borderline Symptom List-23, DBT Dialectical Behavioural Therapy, CAPS Clinician Administered PTSD Scale, CBT Cognitive Behavioural Therapy, EoT End of treatment, EQ5DL Euroqol-5D-5,
HDRS Hamilton Depression Rating Scale, IPT Interpersonal Therapy, PAS Panic and Agoraphobia Scale, PSS-I PTSD Symptom Scale Interview, SAS Social Adjustment Scale, SD Standard deviation, SDS Sheehan Disability
Scale, SPAI-SP Social Phobia Anxiety Inventory – Social Phobia, STAI-T State Trait Anxiety Inventory-Trait, QLDS Quality of Life Depression Scale, Q-LES-Q Quality of Life Enjoyment and Satisfaction Questionnaire
Page 15 of 23
Trials including participants with major depressive disorder of internet-based cognitive behavioral therapy showed
We identified five trials including eight comparisons no evidence of a difference when assessing quality of
assessing the effects of shorter- versus longer-term life (p = 0.75, required information size not reached)
psychotherapy for participants with major depressive and symptom severity (p = 0.89, required informa-
disorder [39–41, 43, 44]. Four trials compared shorter- tion size not reached) [51]. One trial randomising
versus longer-term cognitive behavioural therapy for 138 participants with post-traumatic stress disorder
major depressive disorder [39–41, 43]. Three trials com- to massed prolonged exposure (10 sessions delivered
pared shorter- versus longer-term psychodynamic psy- over 2 weeks) versus standard prolonged exposure
chotherapy for major depressive disorder [40, 41, 44]. (10 sessions delivered over 10 weeks) showed no evi-
One trial compared shorter- versus longer-term inter- dence of a difference when assessing symptom sever-
personal therapy for major depressive disorder [39]. It ity (p = 0.664; required information size not reached)
was not possible to perform meta-analyses, as the trials [49]. One similar trial did not report standard devia-
differed in the assessed psychotherapy traditions, and tions, but the trialists concluded that massed pro-
only two trials reported on our pre-defined review out- longed exposure therapy (10 sessions delivered over
comes [39, 44]. 2 weeks) was noninferior to spaced prolonged exposure
One trial randomising 200 participants with major therapy (10 sessions delivered over 8 weeks) [46]. One
depressive disorder to once- versus twice weekly cogni- trial randomising 61 participants with post-traumatic
tive behavioral therapy or interpersonal therapy [39] stress disorder to intensive (5 weeks) versus standard
showed no evidence of a difference when assessing (12 weeks) cognitive therapy [45] showed no evidence
quality of life and symptom severity for either cogni- of a difference when assessing quality of life (p = 0.061,
tive behavioral therapy (p = 0.77 and p = 0.38, required required information size not reached), symptom sever-
information size not reached) or interpersonal therapy ity (p = 0.466, required information size not reached),
(p = 0.14 and p = 0.42, required information size not or level of functioning (p = 0.757, required informa-
reached). One trial randomising 103 participants with tion size not reached). One trial randomising 40 par-
major depressive disorder to eight versus 16 sessions of ticipants with post-traumatic stress disorder to 60 min
short-term psychodynamic supportive psychotherapy versus 90 min sessions of prolonged exposure therapy
[44] showed no evidence of a difference when assess- [47] showed no evidence of a difference when assessing
ing quality of life (p = 0.911, required information size symptom severity (p = 0.719, required information size
not reached) or symptom severity (p = 0.512, required not reached) (Table 2, Supplementary material 6). All
information size not reached) (Table 2, Supplementary trials were assessed at high risk of bias (Supplementary
material 6). Both trials were assessed at high risk of bias material 5) and the certainty of evidence was assessed
(Supplementary material 5) and the certainty of evidence as “very low” for all outcomes (Supplementary materi-
was assessed as “very low” for all outcomes (Supplemen- als 14, 15, and 16).
tary materials 11, 12, and 13).
Trials including participants with post‑traumatic stress Trials including participants with anxiety disorders
disorder We identified five trials assessing the effects of shorter-
We identified five trials assessing the effects of shorter- versus longer-term cognitive behavioral therapy for
versus longer-term psychotherapy for participants with anxiety disorders [36–38, 42, 48]. One trial did not
post-traumatic stress disorder [45–47, 49, 51]. Three report the results in a usable way; i.e. the results were
trials compared shorter- versus longer-term prolonged reported on a graph and standard deviations were not
exposure for post-traumatic stress disorder [46, 47, 49]. reported [42].
One trial compared shorter- versus longer-term cogni- One trial randomising 29 participants with panic disor-
tive behavioral therapy for post-traumatic stress disorder to five versus 12 sessions cognitive behavioral therapy
der [45]. One trial compared shorter- versus longer-term [37] showed no evidence of a difference when assessing
internet-based cognitive behavioral therapy for post- symptom severity (p = 0.615, required information size
traumatic stress disorder [51]. It was not possible to per- not reached). One trial randomising 34 participants with
form meta-analyses, as the trials differed in the assessed social anxiety disorder to 12 versus 18 weeks of cogni-
psychotherapy traditions, and one of them did not tive behavioral therapy [36] showed no evidence of a
report standard deviations [46]. The two remaining trials difference when assessing symptom severity (p = 0.018,
reported on some of our pre-defined review outcomes. required information size not reached), considering
One trial randomising 224 participants with post- our adjusted threshold for significance was pre-defined
traumatic stress disorder to 6 versus 10 assignments at 0.014 in our protocol [6]. One trial randomising 81
participants to six versus 12 weeks of cognitive behav- and agoraphobia [38]. We chose to analyse anxiety symp-
ioral therapy for participants with panic disorder [48] toms using SMD.
showed no evidence of a difference when assessing symp-
tom severity (p = 0.0195, required information size not Meta‑analysis of anxiety symptoms at end of treatment
reached), considering our adjusted threshold for signifi- Random-effects meta-analysis showed no evidence of a
cance was pre-defined at 0.014 in our protocol [6]. One difference between shorter (5, 6, 7, 12 weeks) and longer-
trial randomising 65 participants with panic disorder and term (12, 12, 14, 18 weeks) cognitive behavioural therapy
agoraphobia to 7 sessions versus 14 sessions cognitive for anxiety disorders (including social anxiety disorder,
behavioral therapy [38] showed no evidence of a differ- panic disorder, and panic disorder with agoraphobia) on
ence when assessing symptom severity (p = 0.77, required anxiety symptoms at end of treatment (SMD: 0.08; 95%
information size not reached). All trials were assessed at CI: -0.47 to 0.63; p = 0.77; I2 = 73%; four trials; very low
high risk of bias (Supplementary material 5) and the cer- certainty) (Fig. 2). Visual inspection of the forest plot
tainty of evidence was assessed as “very low” for all out- and measures to quantify heterogeneity indicated sub-
comes (Supplementary material 17). stantial heterogeneity (I2 = 73%). The end of treatment
It was only possible to perform two pre-planned meta- assessment time point was 12 weeks [37, 48], 15 weeks
analyses: one assessing the effects of shorter- versus [38], and 18 weeks [36]. It was not possible to assess the
longer-term cognitive behavioral therapy for anxiety dis- possible impact of missing outcome data, due to unclear
orders at end of treatment and at maximum follow-up, or lack of reporting of number of analysed participants
and another one assessing the effects of shorter- versus in some of the included trials. It was not possible to per-
longer-term psychodynamic psychotherapy for mood form Trial Sequential Analysis for this outcome, because
and anxiety disorders at end of treatment. the outcome was assessed using SMD [24]. This outcome
result was assessed as at high risk of bias. Certainty of
Shorter‑ versus longer‑term cognitive behavioural therapy the evidence was assessed as ‘very low’. See Supplemen-
for anxiety disorders tary material 17. The fixed-effect meta-analysis showed
We identified five trials assessing the effects of shorter- similar results (SMD 0.16; 95% CI: -0.11, 0.44; p = 0.25;
versus longer-term cognitive behavioural therapy for I2 = 73%; four trials; very low certainty) Supplementary
anxiety disorders [36–38, 42, 48]. All trials were assessed material 18.
as at high risk of bias (Supplementary material 5). One
trial was not eligible for meta-analysis, as the results Shorter‑ versus longer‑term psychodynamic therapy
were not reported in a usable way; i.e. the results were for mood and anxiety disorders
reported on a graph and standard deviations were not We identified two trials assessing the effects of shorter-
reported [42]. versus longer-term psychodynamic therapy for mood-
Four trials randomising a total of 209 participants and anxiety disorder [34, 35, 54]. Both trials were assessed
reported on anxiety symptoms [36–38, 48]. Four different as at high risk of bias (Supplementary material 4).
symptom scales were used: Beck Anxiety Inventory (BAI) Two trials randomising a total of 393 participants
[37], Social Phobia Anxiety Inventory – Social Phobia reported on level of functioning [34, 35]. Two different
[36], State Trait Anxiety Inventory-Trait (STAI-T) [48], assessment scales were used, including Global Assess-
and Panic and Agoraphobia Scale (PAS) [38]. One trial ment of Functioning – Function (GAF-F) [35] and the
included participants with social anxiety disorder [36]. work subscale (SAS-Work) of the Social Adjustment
Two trials included participants with panic disorder [37, Scale [34]. We chose to analyze level of functioning using
48]. One trial included participants with panic disorder standardised mean difference. In order to assure the
Fig. 2 Forest plot of shorter- versus longer-term cognitive behavioural therapy for anxiety disorders on severity of anxiety symptoms at end of
treatment
Fig. 3 Forest plot of shorter- versus longer-term psychodynamic therapy for mood- and anxiety disorders on level of functioning at end of
treatment
scales pointed in the right direction, we multiplied the publication bias by testing for funnel plot asymmetry due
mean in one of the trials with ‘-1’. to lack of trials. Last, it was not possible to perform Trial
Sequential Analyses because all included outcomes were
Meta‑analysis of level of functioning at end of treatment assessed using SMD.
Random effects meta-analysis showed no evidence of
a difference between shorter- (20 and 20 weeks) and The possible contribution of ongoing trials
longer-term (80 and 156 weeks) psychodynamic psy- We identified two ongoing trials [55, 56] that might con-
chotherapy for mood and anxiety disorders on level of tribute to the current evidence on shorter- versus longer-
functioning at end of treatment (SMD 0.16; 95% CI -0.08 term psychotherapy for adult mental health disorders.
to 0.40; p = 0.20; I2 = 21%; two trials; very low certainty) These ongoing trials will contribute to the evidence on
(Fig. 3). Visual inspection of forest plot and measures to quality of life, serious adverse events, symptom sever-
quantify heterogeneity (I2 = 21%) showed some hetero- ity, suicide and suicide attempts, self-harm, and level of
geneity. The end of treatment time point of assessment functioning.
was 36 months after randomisation for both trials. It was
not possible to perform Trial Sequential Analysis for this Discussion
outcome, because the outcome was assessed using SMD We conducted the first systematic review assessing the
[24]. This outcome result was assessed as at high risk of difference between shorter- and longer-term psychother-
bias. Certainty of the evidence was assessed as ‘very low’. apy for adult mental health disorders. We included 19 tri-
See Supplementary material 8. The fixed-effect meta- als randomising a total of 3,447 participants to a shorter
analysis showed similar results (SMD 0.16; 95% CI: -0.05, or a longer-term version of the same psychotherapy type.
0.37; p = 0.14; I2 = 21%; two trials; very low certainty) All trials and outcome results were at high risk of bias,
Supplementary material 19. and the certainty of the evidence according to GRADE
was `very low’ for all outcomes.
Incomplete outcome data One single trial showed no evidence of a difference
Random effects meta-analysis of the best–worst case between shorter- versus longer-term dialectical behav-
scenario adding 2 SD (SMD -0.16; 95% CI -8.13 to 7.81; ioral therapy for borderline personality disorder and
p = < 0.00001; I2 = 95%) and adding 1 SD (SMD -0.15; 95% reached the required information size needed to con-
CI -4.26 to 3.95; p = < 0.94; I2 = 100%) for missing data firm or reject realistic intervention effects when assessing
showed no evidence of a difference between shorter- and quality of life, symptom severity, and level of functioning
longer-term psychodynamic psychotherapy. Random [53] (McMain S: The effectiveness of 6 versus 12-months
effects meta-analysis of the worst-best case scenario add- of dialectical behaviour therapy for borderline person-
ing 2 SD (SMD -0.14; 95% CI -7.62 to 7.35; p = < 0.97; ality disorder: the feasibility of a shorter treatment and
I2 = 100%) and adding 1 SD (SMD -0.14; 95% CI -3.76 to evaluating responses (FASTER) trial, Unpublished). One
3.48; p = < 0.94; I2 = 100%) for missing values showed no single trial showed evidence of a beneficial effect of add-
evidence of a difference between shorter- and longer- ing booster sessions to 8 and 12 weeks of internet-based
term psychodynamic psychotherapy. cognitive-behavioral therapy when assessing symp-
Because of lack of relevant data, it was not possible to tom severity and level of functioning and reached the
conduct other pre-defined meta-analyses. It was only required information size needed to confirm or reject
possible to perform one sensitivity analysis (best–worst realistic intervention effects [50]. One single trial showed
worst-best scenarios) to assess the potential impact of no evidence of a difference between shorter- versus
incomplete outcome data. We also planned several sub- longer-term psychodynamic psychotherapy for mood- or
group analyses to test for heterogeneity [6], but it was anxiety disorders and reached the required information
not possible to conduct them because of lack of relevant size needed to confirm or reject realistic intervention
data. Further, it was not possible to assess the risk of effects when assessing symptom severity and level of
functioning [34]. The remaining single trials did not meet due to our pre-defined broad inclusion criteria, i.e. we
the required information size needed to confirm or reject used the trialists’ own definitions of short-term and long-
realistic intervention effects. It was only possible to per- term psychotherapy. However, we believe that this choice
form two pre-planned meta-analyses. Meta-analysis of methodology from a pragmatic point of view is the
showed no evidence of a difference between short-term best solution there is, as introducing specific thresholds
and long-term cognitive behavioural therapy for anxiety may have excluded important data from our review [6].
symptoms at end of treatment or at maximum follow- If we had used a specific threshold distinguishing short-
up. Meta-analysis showed no evidence of a difference term from long-term psychotherapy, e.g. by applying a
between short-term and long-term psychodynamic psy- definition of short-term psychotherapy as including up
chotherapy on level of functioning at end of treatment. to 24 sessions and long-term psychotherapy as including
All trials and outcomes were assessed as at high risk of at least 50 sessions or having a duration of at least one
bias, and the certainty of evidence was assessed as ‘very year as suggested by others [57, 58], we would have only
low’ for all outcomes. It was not possible to perform Trial been able to include three trials in the review, and the
Sequential Analysis or tests for publication bias. Fur- aim of presenting a complete overview would not be pos-
ther, due to poor reporting in the included trials, we only sible. Second, all trials were at high risk of bias. There-
performed one planned sensitivity analysis to assess the fore, there is a risk that our results overestimated the
potential impact of missing data. Only one trial reported beneficial effects and underestimated the harmful effects
on serious adverse events (McMain S: The effectiveness of the experimental interventions being studied [59–66].
of 6 versus 12-months of dialectical behaviour therapy Third, we only identified 19 trials, and it was not possi-
for borderline personality disorder: the feasibility of a ble to assess the risk of random errors in the meta-anal-
shorter treatment and evaluating responses (FASTER) yses with Trial Sequential Analysis due to the inclusion
trial, Unpublished). Two trials reported on suicide and of continuous outcomes assessed with heterogeneous
suicide attempts [35] (McMain S: The effectiveness of measures (i.e. we assessed the effects with standardised
6 versus 12-months of dialectical behaviour therapy mean difference). This is a major limitation, as we can-
for borderline personality disorder: the feasibility of a not assess if the shown lack of difference is an indication
shorter treatment and evaluating responses (FASTER) of a “true” lack of difference, or if it is an indication that
trial, Unpublished), and one trial reported on self-harm more trials are needed. We calculated the required infor-
(McMain S: The effectiveness of 6 versus 12-months of mation sizes for single trial results post-hoc, but these
dialectical behaviour therapy for borderline personality should primarily be considered exploratory, as they rely
disorder: the feasibility of a shorter treatment and evalu- on the observed means and standard deviations instead
ating responses (FASTER) trial, Unpublished). of pre-defined minimal clinically important differences
Our review has several strengths. We followed our on the assessed scales. Fourth, only few trials reported
protocol which was registered prior to the systematic lit- on serious adverse events, suicide, suicide attempts, and
erature search (PROSPERO ID: CRD42019128535). Data self-harm. It is of utmost importance to always assess
were double-extracted by independent authors minimiz- beneficial and harmful intervention effects on patient-
ing the risk of inaccurate data extraction, and we assessed important outcomes [14, 67].
the risk of bias in all trials according to Cochrane meth- We have identified one previous systematic review
odology [14]. We used GRADE to assess the certainty comparing short-term and long-term psychotherapy
of the evidence [31–33], and the eight-step assessment for schizophrenia [68]. However, the review did not
suggested by Jakobsen et al. to assess if the thresholds identify any trials. We have also identified a meta-
for significance were crossed [17]. Hence, this systematic regression study investigating the effects of psycho-
review considered both risks of random errors and risks therapy for major depressive disorder [5]. This study
of systematic errors which adds further robustness to our found no significant association between the duration
results and conclusions. Another strength of our review of psychotherapy and effect-size, which is similar to the
is that we pragmatically accepted any short-term psycho- conclusion of the present review. However, in the meta-
therapy type and any long-term psychotherapy type, thus regression study, there was a strong association between
results may therefore guide a clinician when choosing number of sessions per week and effect size. An increase
between different treatment durations. from one to two sessions per week increased the effect
Our review also has several limitations. First, due to size with g = 0.45, while keeping the total number of
large heterogeneity in participants, interventions, com- treatment sessions constant [5]. The results of the pre-
parisons, and outcomes, we decided to primarily report sent review could neither confirm nor reject that two
the results narratively and only perform two small pre- sessions per week were more efficacious than one ses-
planned meta-analyses. The observed heterogeneity is sion per week.
The included trials in this review typically assessed the preference with clinical experience which may include
effects of different durations of psychotherapy for anxiety knowledge of specific prognostic factors such as early
disorders, major depressive disorder, and post-traumatic onset or co-occurring disorders, while also considering
stress disorder. Our findings indicate that there may be the poor evidence regarding the optimal treatment dura-
no evidence of a difference between short-term and long- tion currently available.
term psychotherapy when assessing symptom severity
and level of functioning. There are, however, indications
from non-controlled studies that patients with complex Conclusions
and severe psychopathology, defined by the presence The evidence for shorter- versus longer-term psychother-
of, e.g., co-occurring mental health disorders, longer apy for adult mental health disorders is currently unclear.
duration and early onset of the disorder, and unemploy- We only identified 19 randomised clinical trials. More
ment, may have better outcomes in high-intensity than trials at low risk of bias and at low risk of random errors
in low-intensity treatments [69, 70]. We included one assessing participants at different levels of psychopatho-
trial including participants with borderline personality logical severity are urgently needed.
disorder. This trial did not find evidence of a difference
between six versus 12 months dialectical behavioral ther- Differences between the protocol and the review
apy, and the trial reached the required information size In addition to assessing all outcomes at end of treatment,
needed to confirm or reject realistic intervention effects we planned to assess all outcomes at maximum follow-up
for quality of life, symptom severity, and level of func- as a secondary analysis. However, only few trials reported
tioning. However, the trial was assessed as at high risk of data at maximum follow-up. Because of lack of relevant
bias and the certainty of evidence was “very low” for all data, we chose to only report data at end of treatment.
outcomes. Accordingly, future randomised clinical trials
comparing the outcomes of short- and long-term psycho- Abbreviations
therapy for patients with low and high problem complex- CENTRAL Cochrane Central Register of Controlled Trials
CI Confidence Interval
ity should be conducted. We are currently performing a CPCI-S Conference Proceedings Citation Index—Science
similar randomised clinical trial assessing the effects of CPCI-SSH Conference Proceedings Citation Index—Social Science &
five months versus 14 months of mentalization-based Humanities
DARIS Diversity Adjusted Required Information Size
therapy for borderline personality disorder [55, 71]. We EMBASE Excerpta Medica database
are planning a protocol for an individual patient data GRADE Grades of Recommendation, Assessment, Development,
meta-analysis of shorter- versus longer-term psycho- and Evaluation
LILACS Latin American and Caribbean Health Sciences Literature
therapy for borderline personality disorder, which will be MEDLINE Medical Literature Analysis and Retrieval System Online
conducted once data from the two trials become avail- MD Mean Difference
able. Results of the individual patient data meta-analysis PRISMA Preferred Reporting Items for Systematic Reviews and
Meta-Analysis
will increase the possibility of identifying subgroups of ROB Risk of Bias
participants with specific effects of the assessed inter- RR Risk Ratio
ventions. We identified no trials including participants SCI-EXPANDED Science Citation Index Expanded
SD Standard Deviation
with other severe personality pathology, schizophrenia, SMD Standardised Mean Difference
or other psychotic disorders. Hence, it is still unclear SSCI Social Sciences Citation Index
whether patients with severe psychopathology requires TSA Trial Sequential Analysis
short-term or long-term psychotherapy.
Evidence-based practice and decision-making should Supplementary Information
be based on the best available evidence, patient prefer- The online version contains supplementary material available at https://doi.
org/10.1186/s12888-023-04895-6.
ences, and the clinician’s expertise [72]. For severe and
complex cases there is evidence of beneficial effects of Additional file 1. PRISMA 2020 Checklist for “Short-term versus long-term
psychotherapy of specific treatment lengths (e.g. long- psychotherapy for adult psychiatric disorders: A systematic review with
term specialized treatment for borderline personality meta-analysis”
disorder [73]) but very low certainty evidence to guide cli- Additional file 2.
nicians in choosing the optimal treatment duration. Evi- Additional file 3.
dently, clinicians should by default offer psychotherapy Additional file 4.
in a duration supported by the best available evidence. Additional file 5:Supplementary material 5. Risk of bias table.
But when there is a question of treatment duration, e.g. Additional file 6.
a patient asking for a shorter treatment because of life Additional file 7.
circumstances, the clinician is advised to balance this
References
Additional file 8. 1. Wittchen HU, Jacobi F, Rehm J, Gustavsson A, Svensson M, Jonsson B,
Additional file 9. et al. The size and burden of mental disorders and other disorders of the
brain in Europe 2010. Eur Neuropsychopharmacol. 2011;21(9):655–79.
Additional file 10.
2. Olesen J, Gustavsson A, Svensson M, Wittchen HU, Jonsson B, group Cs, et al. The
Additional file 11. economic cost of brain disorders in Europe. Eur J Neurol. 2012;19(1):155–62.
Additional file 12. 3. Wittchen HU, Jacobi F. Size and burden of mental disorders in Europe-a
critical review and appraisal of 27 studies. Eur Neuropsychopharmacol.
Additional file 13. 2005;15(4):357–76.
Additional file 14. 4. Barlow DH. The Oxford Handbook of Clinical Psychology: Updated Edi-
tion. Oxford Library of Psychology. Oxford: Oxford University Press; 2014.
Additional file 15. 5. Cuijpers P, Huibers M, Ebert DD, Koole SL, Andersson G. How much
Additional file 16. psychotherapy is needed to treat depression? A metaregression analysis.
J Affect Disord. 2013;149(1–3):1–13.
Additional file 17. 6. Juul S, Poulsen S, Lunn S, Sorensen P, Jakobsen JC, Simonsen S. Short-
Additional file 18:Supplementary material 18. Fixed-effect meta-analy- term versus long-term psychotherapy for adult psychiatric disorders: a
sis of short-term versus long-term cognitivebehavioral therapy for anxiety protocol for a systematic review with meta-analysis and trial sequential
disorders on anxiety symptoms (sensitivity analysis). analysis. Syst Rev. 2019;8(1):169.
7. Castonguay LG, Barkham M, Lutz W, McAleavey A. Practice-oriented
Additional file 19:Supplementary material 19. Fixed-effect meta-
research: Approaches and applications. In: Lambert MJ, editors. Bergin
analysis of short-term versus long-term psychodynamic psychotherapy for
and Garfield’s handbook of psychotherapy and behavior change. 6th
mood- and anxiety disorders on level of functioning (sensitivity analysis).
edition. Hoboken: Wiley; 2013. p. 85-133.
8. Howard KI, Kopta SM, Krause MS, Orlinsky DE. The Dose-Effect Relation-
Acknowledgements ship in Psychotherapy. Am Psychol. 1986;41(2):159–64.
The expert help from Sarah Louise Klingenberg (Information Specialist, 9. Robinson L, Delgadillo J, Kellett S. The dose-response effect in routinely
The Cochrane Hepato-Biliary Group, Copenhagen Trial Unit, Copenhagen, delivered psychological therapies: A systematic review. Psychother Res.
Denmark) in making the search strategy and conducting electronic searches 2020;30(1):79–96.
is hugely appreciated. 10. Bone C, Delgadillo J, Barkham M. A systematic review and meta-analysis
of the good-enough level (GEL) literature. Journal of Counseling Psychol-
Authors’ contributions ogy. 2021;68(2):219.
SJ and CKJ independently assessed eligibility, extracted data, and performed 11. Jakobsen JC, Gluud C. The necessity of randomized clinical trials. Br J Med
risk of bias assessments with ongoing supervision from JCJ and SS. SJ con- Med Res. 2013;3(4):1453–68.
ducted data analyses and wrote up the manuscript draft. All authors read, 12. Deeks JJ, Dinnes J, D’Amico R, Sowden AJ, Sakarovitch C, Song F, et al.
commented on, and approved the final manuscript. Evaluating non-randomised intervention studies. Health Technol Assess.
2003;7(27):iii–x, 1–173.
Funding 13. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD,
Open access funding provided by Royal Danish Library. This work was et al. The PRISMA 2020 statement: an updated guideline for reporting
funded by Trygfonden (SJ, grant number: 123488), and by the Mental Health systematic reviews. Syst Rev. 2021;10(1):89.
Services in the Capital Region of Denmark Research Fund (SJ and SS, grant 14. Higgins J, Thomas J, Chandler J, Cumpston M, Li T, Page M, Welch V.
number: N/A). Cochrane Handbook for Systematic Reviews of Interventions (Version 5.1).
The Cochrane Collaboration. Wiley. 2011. Available at: https://handbook-5-
Availability of data and materials 1.cochrane.org/. Accessed 2 Mar 2023.
All data generated or analysed during this study are included in this published 15. International Conference on Harmonisation of Technical Requirements for
article (and its supplementary information files). Registration of Pharmaceuticals for Human Use). ICH Harmonised Guideline:
Integrated Addemdum to ICH E6(R1): Guideline for Good Clinical Practice
(ICH-GCP). 2015. Available at: https://ichgcp.net/da. Accessed 2 Mar 2023.
Declarations 16. Keus F, Wetterslev J, Gluud C, van Laarhoven CJ. Evidence at a glance:
error matrix approach for overviewing available evidence. BMC Med Res
Ethical approvement and consent to participate Methodol. 2010;10(1):90.
Not applicable. 17. Jakobsen JC, Wetterslev J, Winkel P, Lange T, Gluud C. Thresholds for statis-
tical and clinical significance in systematic reviews with meta-analytic
Consent for publication methods. BMC Med Res Methodol. 2014;14(1):120.
Not applicable. 18. Stata Statistical Software: Release 16. College Station: StataCorp LLC;
2019. Available at: http://www.stata.com. Accessed May 26 2023.
Competing interests 19. The Cochrane Collaboration. Review manager (Version 5.4) [Computer
The authors declare no competing interests. Software]. The Nordic Cochrane Centre; 2020. Available at: https://rev-
man.cochrane.org. Accessed 2 Mar 2023.
Author details 20. Wetterslev J, Thorlund K, Brok J, Gluud C. Trial sequential analysis may
1
Stolpegaard Psychotherapy Centre, Mental Health Services in the Capital establish when firm evidence is reached in cumulative meta-analysis. J
Region of Denmark, Gentofte, Denmark. 2 Copenhagen Trial Unit, Centre Clin Epidemiol. 2008;61(1):64–75.
for Clinical Intervention Research, Copenhagen University Hospital – Rigshos- 21. Copenhagen Trial Unit. TSA - Trial Sequential Analysis. [Computer Soft-
pitalet, The Capital Region of Denmark, Copenhagen, Denmark. 3 Department ware] Available at: http://www.ctu.dk/tsa. Accessed 2 Mar 2023.
of Regional Health Research, Faculty of Health Sciences, University of South- 22. Brok J, Thorlund K, Gluud C, Wetterslev J. Trial sequential analysis reveals
ern Denmark, Odense, Denmark. 4 Department of Psychology, University insufficient information size and potentially false positive results in many
of Copenhagen, Copenhagen, Denmark. meta-analyses. J Clin Epidemiol. 2008;61(8):763–9.
23. Brok J, Thorlund K, Wetterslev J, Gluud C. Apparently conclusive meta-
Received: 6 September 2022 Accepted: 22 May 2023 analyses may be inconclusive—trial sequential analysis adjustment of ran-
dom error risk due to repetitive testing of accumulating data in apparently
conclusive neonatal meta-analyses. Int J Epidemiol. 2008;38(1):287–98.
24. Thorlund K, Engstrøm J, Wetterslev J, Brok J, Imberger G, Gluud C. User therapy for PTSD and emotion-focused supportive therapy. Am J Psychia-
manual for trial sequential analysis (TSA). 2011. Available at: https://ctu. try. 2014;171(3):294–304.
dk/wp-content/uploads/2021/03/2017-10-10-TSA-Manual-ENG_ER.pdf. 46. Foa EB, McLean CP, Zang Y, Rosenfield D, Yadin E, Yarvis JS, et al. Effect
Accessed 2 Mar 2023. of prolonged exposure therapy delivered over 2 weeks vs 8 weeks vs
25. Wetterslev J, Thorlund K, Brok J, Gluud C. Estimating required information present-centered therapy on PTSD symptom severity in military person-
size by quantifying diversity in random-effects model meta-analyses. nel: A randomized clinical trial. JAMA. 2018;319(4):354–64.
BMC Med Res Methodol. 2009;9(1):86. 47. Nacasch N, Huppert JD, Su Y-J, Kivity Y, Dinshtein Y, Yeh R, et al. Are
26. Thorlund K, Anema A, Mills E. Interpreting meta-analysis according to 60-Minute Prolonged Exposure Sessions With 20-Minute Imaginal
the adequacy of sample size. An example using isoniazid chemoprophy- Exposure to Traumatic Memories Sufficient to Successfully Treat PTSD? A
laxis for tuberculosis in purified protein derivative negative HIV-infected Randomized Noninferiority Clinical Trial. Behav Ther. 2015;46(3):328–41.
individuals. Clin Epidemiol. 2010;2:57. 48. Kenardy JA, Dow MG, Johnston DW, Newman MG, Thomson A, Taylor CB.
27. Thorlund K, Devereaux P, Wetterslev J, Guyatt G, Ioannidis JP, Thabane L, A comparison of delivery methods of cognitive-behavioral therapy for
et al. Can trial sequential monitoring boundaries reduce spurious infer- panic disorder: an international multicenter trial. J Consult Clin Psychol.
ences from meta-analyses? Int J Epidemiol. 2008;38(1):276–86. 2003;71(6):1068.
28. Imberger G, Thorlund K, Gluud C, Wetterslev J. False-positive findings in 49. Dell L, Sbisa AM, Forbes A, O’Donnell M, Bryant R, Hodson S, et al. Effect
Cochrane meta-analyses with and without application of trial sequential of massed v. standard prolonged exposure therapy on PTSD in military
analysis: an empirical review. BMJ Open. 2016;6(8):e011890. personnel and veterans: a non-inferiority randomised controlled trial.
29. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health- Psychol Med. 2022:1–8. https://doi.org/10.1017/S0033291722000927.
related quality of life: the remarkable universality of half a standard devia- 50. Hadjistavropoulos H, Peynenburg V, Thiessen D, Nugent M, Karin E, Dear
tion. Med Care. 2003;41:582–92. B, et al. A randomized factorial trial of internet-delivered cognitive behav-
30. Jakobsen JC, Wetterslev J, Lange T, Gluud C. Viewpoint: taking into ioural therapy: An 8-week program with or without extended support
account risks of random errors when analysing multiple outcomes in and booster lesson. Internet Int. 2022;27:100499.
systematic reviews. Cochrane Database of Syst Rev. 2016(3)ED000111. 51. Böttche M, Wagner B, Vöhringer M, Heinrich M, Stein J, Selmo P, et al. Is
https://doi.org/10.1002/14651858.ED000111. only one cognitive technique also effective? Results from a randomized
31. Guyatt GH, Oxman AD, Schünemann HJ, Tugwell P, Knottnerus A. GRADE controlled trial of two different versions of an internet-based cognitive
guidelines: a new series of articles in the Journal of Clinical Epidemiology. behavioural intervention for post-traumatic stress disorder in Arabic-
J Clin Epidemiol. 2011;64(4):380–2. speaking countries. Eur J Psychotraumatol. 2021;12(1):1943870.
32. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, 52. McMain SF, Chapman AL, Kuo JR, Dixon-Gordon KL, Guimond TH, Labrish
et al. GRADE: an emerging consensus on rating quality of evidence and C, et al. The effectiveness of 6 versus 12 months of dialectical behavior
strength of recommendations. BMJ. 2008;336(7650):924–6. therapy for borderline personality disorder: a noninferiority randomized
33. Schünemann HJ, Best D, Vist G, Oxman AD, Group GW. Letters, numbers, clinical trial. Psychother Psychosom. 2022;91(6):382-97.
symbols and words: how to communicate grades of evidence and 53. McMain SF, Chapman AL, Kuo JR, Guimond T, Streiner DL, Dixon-Gordon
recommendations. CMAJ. 2003;169(7):677–80. KL, et al. The effectiveness of 6 versus 12-months of dialectical behaviour
34. Knekt P, Lindfors O, Harkanen T, Valikoski M, Virtala E, Laaksonen MA, et al. therapy for borderline personality disorder: the feasibility of a shorter
Randomized trial on the effectiveness of long-and short-term psycho- treatment and evaluating responses (FASTER) trial protocol. BMC Psychia-
dynamic psychotherapy and solution-focused therapy on psychiatric try. 2018;18(1):230.
symptoms during a 3-year follow-up. Psychol Med. 2008;38(5):689–703. 54. Knekt P, Lindfors O, Laaksonen MA, Raitasalo R, Haaramo P, Jarvikoski A,
35. Lorentzen S, Ruud T, Fjeldstad A, Hoglend P. Comparison of short- and et al. Effectiveness of short-term and long-term psychotherapy on work
long-term dynamic group psychotherapy: randomised clinical trial. Br J ability and functional capacity–a randomized clinical trial on depressive
Psychiatry. 2013;203(3):280–7. and anxiety disorders. J Affect Disord. 2008;107(1–3):95–106.
36. Herbert JD, Rheingold AA, Gaudiano BA, Myers VH. Standard Versus 55. Juul S, Lunn S, Poulsen S, Sorensen P, Salimi M, Jakobsen JC, et al. Short-
Extended Cognitive Behavior Therapy for Social Anxiety Disorder: A Rand- term versus long-term mentalization-based therapy for outpatients with
omized-Controlled Trial. Behav Cognitive Psychother. 2004;32(2):131–47. subthreshold or diagnosed borderline personality disorder: a protocol for
37. Clark DM, Salkovskis PM, Hackmann A, Wells A, Ludgate J, Gelder M. Brief a randomized clinical trial. Trials. 2019;20(1):196.
Cognitive Therapy for Panic Disorder: A Randomized Controlled Trial. J 56. Kool M, Van HL, Bartak A, de Maat SCM, Arntz A, van den Eshof JW, et al.
Consult Clin Psychol. 1999;67(4):583–9. Optimizing psychotherapy dosage for comorbid depression and per-
38. Roberge P, Marchand A, Reinharz D, Savard P. Cognitive-Behavioral Treat- sonality disorders (PsyDos): a pragmatic randomized factorial trial using
ment for Panic Disorder With Agoraphobia A Randomized, Controlled schema therapy and short-term psychodynamic psychotherapy. BMC
Trial and Cost-Effectiveness Analysis. J Behavior Mod. 2008;32(3):333–51. Psychiatry. 2018;18(1):252.
39. Bruijniks SJE, Lemmens L, Hollon SD, Peeters F, Cuijpers P, Arntz A, et al. 57. Gabbard, G. Long-term Psychodynamic Psychotherapy: A Basic Text.
The effects of once- versus twice-weekly sessions on psychotherapy Arlington: American Psychiatric Publishing Inc; 2004.
outcomes in depressed patients. Br J Psychiatry. 2020;216(4):222–30. 58. Crits-Christoph P, Barber JP. Long-term psychotherapy. In: Ingram RE, Snyder
40. Barkham M, Rees A, Shapiro DA, Stiles WB, Agnew RM, Halstead J, et al. CR, editors. Handbook of Psychological Change: Psychotherapy Processes &
Outcomes of time-limited psychotherapy in applied settings: replicat- Practices for the 21st Century. Hoboken: John Wiley & Sons; 2000. p. 455–73.
ing the Second Sheffield Psychotherapy Project. J Consult Clin Psychol. 59. Gluud LL. Bias in clinical intervention research. Am J Epidemiol.
1996;64(5):1079. 2006;163(6):493–501.
41. Shapiro DA, Barkham M, Rees A, Hardy GE, Reynolds S, Startup M. Effects 60. Kjaergard LL, Villumsen J, Gluud C. Reported methodologic quality and
of treatment duration and severity of depression on the effectiveness of discrepancies between large and small randomized trials in meta-analy-
cognitive-behavioral and psychodynamic-interpersonal psychotherapy. J ses. Ann Intern Med. 2001;135(11):982–9.
Consult Clin Psychol. 1994;62(3):522. 61. Jüni P, Altman DG, Egger M. Assessing the quality of controlled clinical
42. Bohni M, Spindler H, Arendt M, Hougaard E, Rosenberg N. A randomized trials. BMJ. 2001;323(7303):42–6.
study of massed three-week cognitive behavioural therapy schedule for 62. Moher D, Jones A, Cook DJ, Jadad AR, Moher M, Tugwell P, et al. Does
panic disorder. Acta Psychiatr Scand. 2009;120(3):187–95. quality of reports of randomised trials affect estimates of intervention
43. Christensen H, Griffiths K, Mackinnon A, Brittliffe K. Online randomized efficacy reported in meta-analyses? Lancet. 1998;352(9128):609–13.
controlled trial of brief and full cognitive behaviour therapy for depres- 63. Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias:
sion. Psychol Med. 2006;36(12):1737. dimensions of methodological quality associated with estimates of treat-
44. Dekker J, Molenaar PJ, Kool S, Van Aalst G, Peen J, de Jonghe F. Dose– ment effects in controlled trials. JAMA. 1995;273(5):408–12.
effect relations in time-limited combined psycho-pharmacological treat- 64. Hrobjartsson A, Emanuelsson F, Skou Thomsen AS, Hilden J, Brorson S.
ment for depression. Psychol Med. 2005;35(1):47–58. Bias due to lack of patient blinding in clinical trials. A systematic review
45. Ehlers A, Hackmann A, Grey N, Wild J, Liness S, Albert I, et al. A rand- of trials randomizing patients to blind and nonblind sub-studies. Int J
omized controlled trial of 7-day intensive and standard weekly cognitive Epidemiol. 2014;43(4):1272–83.
65. Hrobjartsson A, Skou Thomsen AS, Emanuelsson F, Tendal B, Hilden J,

Boutron I, et al. Observer bias in randomized clinical trials with measure-
ment scale outcomes: a systematic review of trials with both blinded and
nonblinded assessors. CMAJ. 2013;185(4):E201-11.
66. Hrobjartsson A, Thomsen AS, Emanuelsson F, Tendal B, Hilden J, Boutron
I, et al. Observer bias in randomised clinical trials with binary outcomes:
systematic review of trials with both blinded and non-blinded outcome
assessors. BMJ. 2012;344:e1119.
67. Higgins J, Thomas J, Chandler J, Cumpston M, Li T, Page M, Welch V.
Cochrane Handbook for Systematic Reviews of Interventions (Version
6.3). The Cochrane Collaboration. Wiley; 2019. Available at: https://train
ing.cochrane.org/handbook. Accessed 2 Mar 2023.
68. Naeem F, Farooq S, Kingdon D. Cognitive behavioural therapy (brief
versus standard duration) for schizophrenia. Cochrane Database Syst Rev.
2015;10:CD010646.
69. Delgadillo J, Moreea O, Lutz W. Different people respond differently to
therapy: A demonstration using patient profiling and risk stratification.
Behav Res Ther. 2016;79:15–22.
70. Delgadillo J, Huey D, Bennett H, McMillan D. Case complexity as a
guide for psychological treatment selection. J Consult Clin Psychol.
2017;85(9):835.
71. Juul S, Simonsen S, Poulsen S, Lunn S, Sørensen P, Bateman A, et al.
Detailed statistical analysis plan for the short-term versus long-term men-
talization-based therapy for outpatients with subthreshold or diagnosed
borderline personality disorder randomized clinical trial (MBT-RCT). Trials.
2020;22:1–8.
72. Sackett DL, Rosenberg WM, Gray JM, Haynes RB, Richardson WS. Evidence
based medicine: what it is and what it isn’t. BMJ. 1996;321:71–2.
73. Storebø OJ, Stoffers-Winterling JM, Völlm BA, Kongerslev MT, Mattivi JT,
Jørgensen MS, et al. Psychological therapies for people with borderline
personality disorder. Cochrane Database Syst Rev. 2020(5). https://doi.
org/10.1002/14651858.CD012955.pub2.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in pub-
lished maps and institutional affiliations.
Ready to submit your research ? Choose BMC and benefit from:
• fast, convenient online submission

• thorough peer review by experienced researchers in your field
• rapid publication on acceptance
• support for research data, including large and complex data types
• gold Open Access which fosters wider collaboration and increased citations
• maximum visibility for your research: over 100M website views per year
At BMC, research is always in progress.
Learn more biomedcentral.com/submissions

The Difference Between Shorter-Versus Longer-Term Psychotherapy For Adult Mental Health Disorders: A Systematic Review With Meta-Analysis

Uploaded by

Copyright:

Available Formats

The Difference Between Shorter-Versus Longer-Term Psychotherapy For Adult Mental Health Disorders: A Systematic Review With Meta-Analysis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Difference Between Shorter-Versus Longer-Term Psychotherapy For Adult Mental Health Disorders: A Systematic Review With Meta-Analysis

Uploaded by

Copyright:

Available Formats

Juul et al.

BMC Psychiatry (2023) 23:438 BMC Psychiatry

RESEARCH Open Access

The difference between shorter‑

Keywords Mental health disorders, Duration of psychotherapy, Systematic review, Meta-analysis

Background systematic review is described in detail in our protocol

Fig. 1 PRISMA flow diagram

8 sessions psychody- 16 sessions psychody- reported

intervention intervention Events ity attempts tioning

therapy (8 weeks) therapy 8 sessions on some

(n = 45) and group (n = 45) period of 6 months

1994 [41] (8 weeks) (18 weeks) benefit from 16

and 8.4 (8.0) for were essentially the

and 9.5 (10.3) for

Trials including participants with mood- and anxiety disordersa

to 0/90 in the and 68.1 (14.2) using IIP, GAF-S

(n = 87) and 69.13 group (n = 87) and 11.39 (8.46)

(n = 63) and symptoms of PTSD

(n = 20) and in PTSD symptoms

65. Hrobjartsson A, Skou Thomsen AS, Emanuelsson F, Tendal B, Hilden J,

Ready to submit your research ? Choose BMC and benefit from:

• fast, convenient online submission

At BMC, research is always in progress.

Learn more biomedcentral.com/submissions

You might also like