See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/374935477

Trajectories of Change in Weekly and Biweekly Therapy

Article in Journal of Counseling Psychology · October 2023

DOI: 10.1037/cou0000711

CITATIONS READS

0 409

4 authors:

Tao Lin Timothy Anderson

Ohio University Ohio University
17 PUBLICATIONS 157 CITATIONS 117 PUBLICATIONS 3,507 CITATIONS

SEE PROFILE SEE PROFILE

David McConkie Erekson Benjamin M. Ogles

Brigham Young University - Provo Main Campus Brigham Young University - Provo Main Campus
62 PUBLICATIONS 375 CITATIONS 113 PUBLICATIONS 5,484 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Tao Lin on 25 October 2023.

The user has requested enhancement of the downloaded file.

Journal of Counseling Psychology
Trajectories of Change in Weekly and Biweekly Therapy
Tao Lin, Timothy Anderson, David M. Erekson, and Benjamin M. Ogles
Online First Publication, October 23, 2023. https://dx.doi.org/10.1037/cou0000711

CITATION
Lin, T., Anderson, T., Erekson, D. M., & Ogles, B. M. (2023, October 23). Trajectories of Change in Weekly and Biweekly
Therapy. Journal of Counseling Psychology. Advance online publication. https://dx.doi.org/10.1037/cou0000711
 Journal of Counseling Psychology
© 2023 American Psychological Association
ISSN: 0022-0167 https://doi.org/10.1037/cou0000711

Trajectories of Change in Weekly and Biweekly Therapy

Tao Lin1, Timothy Anderson1, David M. Erekson2, and Benjamin M. Ogles3
1
Department of Psychology, Ohio University
2
Counseling and Psychological Services, Brigham Young University
3
Department of Psychology, Brigham Young University

Higher therapy session frequency has been found to result in faster recovery, but few studies have considered
that clients follow diverse trajectories of change in psychotherapy. It is unknown how session frequency may
affect the sizes and shapes of change trajectories. The present study examined clients’ change trajectories in
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

weekly and biweekly therapy in a naturalistic setting, as well as predictors of these trajectories. Using a
This document is copyrighted by the American Psychological Association or one of its allied publishers.

sample of 5,102 clients receiving 3–10 therapy sessions at a university counseling center, we identified
886 clients attending approximately weekly therapy and 1,753 clients attending approximately biweekly
therapy. We examined the change trajectories of the weekly and biweekly samples using latent growth
mixture modeling (LGMM). Three trajectories were identified in weekly therapy: slow change (78.33%),
early improvement (17.61%), and worse before better (4.06%), and in biweekly therapy: slow change
(80.38%), early improvement (13.52%), and worse before better (6.1%). The worse before better subgroup in
weekly therapy experienced greater deterioration than those in biweekly therapy. The slow change and early
improvement subgroups in weekly therapy showed treatment outcome comparable to those of their
respective counterparts in biweekly therapy. Clients’ intake symptoms, including eating concerns,
frustration/anger, depression, and academic concerns, significantly predicted change trajectories. Compared
to biweekly therapy, weekly therapy leads to higher chances of early improvement and shortens the duration
of suffering but results in greater deterioration for individuals who deteriorate in therapy. The impact of
session frequency on treatment outcome varies across clients, and session frequency should be adjusted
individually.

Public Significance Statement

The study suggests that clients fall into three trajectories of change in both weekly and biweekly therapy
and that increasing session frequency can increase the chances of achieving early improvement and
shortens the duration of suffering but may intensify the negative effects of therapy. Individualizing
session frequency can increase treatment efficacy and save clinical resources.

Keywords: session frequency, trajectory, psychotherapy, outcome, growth mixture model

Supplemental materials: https://doi.org/10.1037/cou0000711.supp

Ensuring care access while minimizing treatment cost is a key has been increasing and even exceeded the availability of services
challenge for mental health systems all over the world (Drake & (Xiao et al., 2017). A common solution has been to reduce
Latimer, 2012). This task has become more pressing for university psychotherapy dosage and intensity (Bailey et al., 2022). According
counseling centers over the recent decade as demand for psychotherapy to the 2014 National Survey of College Counseling Centers, nearly

Tao Lin played a lead role in conceptualization, formal analysis,

Tao Lin https://orcid.org/0000-0002-8883-870X investigation, methodology, project administration, software, visualization,
This study was funded by Society for Psychotherapy Research. The writing–original draft, and writing–review and editing. Timothy Anderson
authors have no known conflicts of interest to disclose. played a supporting role in conceptualization, data curation, project
The data that support the findings of this study are available from the administration, writing–original draft, and writing–review and editing and
corresponding author upon reasonable request. an equal role in resources and supervision. David M. Erekson played a
Part of the research was previously presented at the 95th Annual Meeting of supporting role in conceptualization, supervision, writing–original draft,
Midwestern Psychological Association. Findings from the data collection and writing–review and editing and an equal role in data curation, project
have been reported in separate manuscripts. Manuscript 1 (Erekson et al., administration, and resources. Benjamin M. Ogles played a supporting role
2022) utilized a quasirandomized controlled design and included 1,322 clients in conceptualization, project administration, writing–original draft, and
who were randomly assigned to a treatment-as-usual group (attenuated writing–review and editing and an equal role in data curation and
session frequency) or an experimental group (weekly session frequency). supervision.
Manuscript 2 (Bailey et al., 2022) examined the impact of an adapting stepped Correspondence concerning this article should be addressed to Tao Lin,
care model on clinical capacity and treatment outcomes using data collected Department of Psychology, Ohio University, 22 Richland Avenue,
between 2017 and 2020. Athens, OH 45701, United States. Email: [email protected]

1
 2 LIN, ANDERSON, EREKSON, AND OGLES

30% of the counseling centers limit the number of therapy sessions recovery, and a greater likelihood of achieving recovery sooner.
(e.g., 10 sessions) and over 40% strive to offer short-term counseling Another RCT also provided evidence for the association between
services (Gallagher, 2014). Beyond session limit policies, more session frequency and treatment outcome (Bruijniks et al., 2020)
practitioners and clinics have begun to attenuate session frequency whereby 200 depressed adults were randomized to once or
from the traditional weekly schedule to biweekly or even less frequent twice weekly sessions of cognitive behavioral therapy (CBT) or
schedules to carry larger caseloads (Bailey et al., 2022). The majority interpersonal therapy (IPT) over 16–24 weeks. For both groups,
of counseling centers mandate policies that necessitate scheduling patients who received twice weekly sessions reported a greater
appointments biweekly, especially for large universities (Gorman reduction in depressive symptoms, lower attrition, and an increased
et al., 2021). rate of response compared with those who received weekly sessions,
Although this approach allows practitioners to schedule a larger suggesting that both CBT and IPT are more effective if provided on
number of clients on an infrequent basis, it raises concerns that a more frequent basis (Bruijniks et al., 2020).
decreasing session frequency may reduce the efficacy and efficiency Similarly, according to a meta-analysis of 70 RCTs on the efficacy
of the treatment. Indeed, previous research found that higher session of therapy for adult depression (Cuijpers et al., 2013), session
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

frequency can lead to superior treatment outcomes (Bruijniks et al., frequency was moderately associated with treatment outcome (g =
This document is copyrighted by the American Psychological Association or one of its allied publishers.

2020; Erekson et al., 2015). Researchers have suggested that session 0.45) whereas the number of therapy sessions, total contact time
frequency should be increased, at least for the early stages of with the therapist, and duration of treatment had minimal effects.
treatment (Cuijpers et al., 2013; Tiemens et al., 2019). Erekson et al. Specifically, increasing session frequency from once to twice per
(2022) highlighted the importance of prioritizing weekly therapy for week can significantly improve treatment effectiveness (Cuijpers et
increasing treatment efficiency and effectiveness. Given the limited al., 2013). Overall, these studies provided evidence that treatment
resources and institutional obstacles, however, such suggestions may outcome was associated with a higher level of session frequency,
not be feasible. As a result, clinics and practitioners face a challenging highlighting the importance of increasing session frequency.
dilemma: should they serve more clients at the cost of partial While these findings consistently demonstrate that session
effectiveness, or prioritize session frequency for a limited number of frequency is associated with treatment outcome based on the average
clients while keeping others on the waitlist? To address this issue, the level of change for the entire sample, it should not be assumed that all
present study aims to compare the treatment outcome of weekly and clients respond to psychotherapy of a certain frequency in a similar
biweekly therapy by examining the trajectories of change in both manner. Since clients fall into heterogeneous trajectories of change
therapies and to explore the predictors of change trajectories. even when receiving the same treatment (Lin & Farber, 2021; Owen
et al., 2015; Palmer et al., 2021), the average outcome may fail to
capture individual differences in treatment responses. Although
Session Frequency
weekly therapy results in a faster improvement trajectory than
To effectively balance the growing demand for psychotherapy biweekly therapy on average (Erekson et al., 2015), the average
with the limited availability of clinical services, it is crucial to trajectories may not apply to each individual. For example, some
understand how clients respond differently to therapies with lower subgroups of clients may improve rapidly in biweekly therapy,
and higher frequencies. Researchers have attempted to examine how whereas other subgroups of clients may stay symptomatic even when
session frequency may affect psychotherapy process and outcome receiving therapy on a more frequent basis. Some clients may equally
via naturalistic studies, randomized controlled trials (RCTs), and benefit from attenuated therapy frequency, whereas other clients may
quasi-experiments (e.g., Erekson et al., 2022; Reese et al., 2011). For need more frequent therapy. Therefore, it is worthwhile to examine
example, using an archival database of 21,488 clients receiving how biweekly and weekly therapy result in different change
individual therapy at a university counseling center, Erekson et al. trajectories and investigate which session frequency works best
(2015) found that more frequent therapy was associated with faster for whom.
recovery. Furthermore, compared to individuals who received
therapy biweekly, those who attended weekly sessions achieved
Trajectories of Change
clinically significant improvement faster and were less likely to
experience deterioration over the course of treatment. More recently, One way to address how clients might respond differentially to
Tiemens et al. (2019) explored the impact of the initial session different session frequencies is to include the trajectories of change by
frequency on treatment outcome among a naturalistic sample of using latent trajectory modeling methods such as latent growth
2,634 clients diagnosed with anxiety disorders, depressive disorders, mixture modeling (LGMM) and latent class growth analysis (LCGA).
or personality disorders. The authors monitored clients’ treatment These advanced statistical methods allow researchers to examine
outcome every 12 weeks and assessed the frequency of sessions in unique groupings of change trajectories within the context of other
the first 3 months of treatment, which ranged from 1 to 64 sessions. treatment factors. For example, in a study of a large naturalistic
In all diagnostic groups, clients’ improvement and recovery were database from university counseling centers (N = 10,854), Owen et al.
found to be associated with a higher frequency of sessions during the (2015) identified three trajectories of change in psychotherapy. The
first 3 months of treatment, and the impact of initial frequency on first trajectory (75.3% of the sample), termed early and late change,
recovery lasted for more than 3 years (Tiemens et al., 2019). was characterized by improvements in early treatment, followed by
Utilizing a quasirandomized controlled design within a university a plateau, and another improvement in later treatment. Another
counseling center, Erekson et al. (2022) extended previous findings trajectory (slow and steady; 19%) showed a slow but steady
by randomly assigning therapists to provide therapy weekly or less improvement in therapy. Finally, 5.4% of the sample (worse before
frequently. The findings showed that weekly therapy led to faster better) demonstrated an initial deterioration and then improved
trajectories of change over time, a greater likelihood of achieving rapidly. Later, Palmer et al. (2021) replicated these findings (Owen
 TRAJECTORIES IN WEEKLY AND BIWEEKLY THERPAY 3

et al., 2015) by using a similar sample of college counseling center The present study had two aims. First, we aimed to examine the
clients (N = 2,538) who completed the same outcome measure at the trajectories of change in biweekly and weekly therapy, respectively,
same time points and identified a two-class model. One class using naturalistic data collected from a university counseling center.
improved slowly and steadily and the other improved quickly in early Consistent with previous trajectory analyses (Allan et al., 2017;
psychotherapy, both of which were roughly similar to those in Owen et al., 2015), this aim was exploratory in nature. Second, we
previous studies, although the worse before better trajectory was not aimed to explore the predictors of the change trajectories in weekly
replicated. therapy and biweekly therapy, respectively.
Additionally, researchers have explored trajectories of change
using different methods and samples (Palmer et al., 2021). For Method
instance, trajectory analyses have been conducted among different
diagnostic categories, including depression (Cuijpers et al., 2005), Participants
anxiety (Joesch et al., 2013), and trauma (Allan et al., 2017) and in Participants were 18,223 students at a large Western University who
different treatment modalities and formats (e.g., Heckman et al., sought individual therapy at the university counseling center between
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

2017; Lindebø Knutsen et al., 2020). Although the identified change January 2014 and September 2021. We limited the analyses to the first
This document is copyrighted by the American Psychological Association or one of its allied publishers.

trajectories varied in the number and shapes of trajectories across course of treatment for each client to isolate the effects of initial
studies, these findings generally support that psychotherapy leads to exposure to psychotherapy, leading to 15,509 clients. Next, as only a
multiple trajectories for different subgroups of clients and that these small proportion of clients (4%) attended more than 10 sessions, the
trajectories start to emerge in the early stages of treatment (Lin et al., trajectories plotted for Session 11 or beyond may be less reliable.
2023; Lutz et al., 2014). Considering that most university counseling centers adopt short-term
Importantly, most studies that examined change trajectories have therapy and some even have session-limit policies, we only include
not taken session frequency into consideration. Specifically, most of clients receiving short-term therapy (i.e., ≤10 sessions) to ensure that
these studies were RCTs in which the treatment was delivered at a the findings can generalize to similar settings. We also excluded clients
fixed session frequency (e.g., weekly), thereby providing little who attended less than three sessions because LGMM requires at least
information on how session frequency may affect change trajectories. three time points to depict nonlinear change patterns. Therefore, clients
In contrast, studies based on naturalistic data have combined clients who received 3–10 sessions of individual therapy were included in this
receiving therapy at different frequencies (Owen et al., 2015), which study, which resulted in an initial sample of 5,102 clients (see
raise concerns that the diverse trajectories may be due to the varied Supplemental Figure 1).
intensity of treatments. Considering that more frequent therapy can As part of routine practice, student clients completed the Outcome
lead to more positive treatment outcome (Erekson et al., 2015), it is Questionnaire–45.2 (OQ-45.2), an outcome measure, before each
possible that trajectories with steeper slopes may be due to higher session. Therapy modalities included cognitive behavioral approaches,
session frequency. Jankowski et al. (2019) attempted to address this humanistic/existential approaches, and psychodynamic approaches.
issue by controlling for the time intervals between data points using Clients were assigned to therapists based on availability at the time
advanced statistical methods when examining trajectories of change. the client was seeking services, though if clients requested a
Nevertheless, this method adds considerable computational burden certain therapist or therapist characteristic they were accommodated.
and may not be suitable for samples who receive therapy at different Individual therapy sessions were typically 50 min.
lengths. Additionally, Palmer et al. (2021) examined session
frequency and consistency as a predictor of trajectories. Their Identifying Weekly and Biweekly Therapy Groups
findings showed that clients attending therapy more consistently are
more likely to show an early rapid improvement, whereas clients In order to identify clients receiving approximately weekly therapy
attending therapy more frequently are more likely to improve slowly and biweekly therapy, we further calculated the average frequency over
and steadily. the entire course of therapy following Erekson et al. (2015).
Specifically, we selected those whose mean levels of session frequency
ranged from 0.75 to 1.25 weeks (weekly group) and those whose mean
The Present Study levels of session frequency ranged from 1.5 to 2.5 weeks (biweekly
group). The weekly group consists of 886 clients, which were primarily
To date, no studies have directly examined and compared clients’ female (65.35%), White (83.41%), and single (70.27%) and had a
change trajectories in biweekly and weekly therapy, which are the mean age of 21.57 years old (SD = 4.35; Mdn = 21; range = 17–59).
most commonly adopted session frequency schedules. Such studies The biweekly group, consisting of 1,753 clients, was primarily female
may help us better understand how attenuating session frequency may (63.15%), White (83%), and single (71.97%) and had a mean age of
affect the sizes and shapes of change trajectories. For example, 21.81 years old (SD = 4.58; Mdn = 22; range = 17–62). The final
comparing the trajectories in biweekly and weekly therapy can inform sample (biweekly and weekly groups) was treated by 150 therapists,
whether weekly therapy can result in more favorable trajectories who had seen a mean of 17.59 clients (Mdn = 5, SD = 19.57; range =
(e.g., larger improvements) and/or lead to a larger proportion of 1–121). These therapists were full-time faculty, as well as trainees who
positive trajectories compared to biweekly therapy. Furthermore, receive clinical supervision from the counseling center faculty.
predicting the multiple trajectories in treatment of different intensities
can help us allocate mental health resources more efficiently. For
Transparency and Openness
example, clients who are predicted to be more likely to fall into a
negative trajectory of change in biweekly therapy may benefit from We used archival data and reported data inclusion and exclusion,
more frequent treatments. all manipulations, and all measures in the study, and we followed
 4 LIN, ANDERSON, EREKSON, AND OGLES

Journal Article Reporting Standards (Kazak, 2018). All data and the 523 clients who received psychotherapy remotely during the
analysis codes are available from the authors upon reasonable pandemic (i.e., March 2020 to September 2021), as well as the
request. Data were analyzed using R Version 4.2.2 (R Core Team, 2,116 clients who were seen prior to the pandemic. Independent
2022) and Mplus Version 8.3 (Muthén & Muthén, 2020). This t tests and chi-square tests were performed to compare the two
study’s design and its analysis were not preregistered. samples on demographic variables (i.e., age, gender, and ethnicity)
and treatment outcomes (i.e., baseline symptom severity, final
Measures symptom severity, and overall improvement). No significant
differences were found between the two samples regarding age,
Outcome Questionnaire–45.2 (OQ-45.2) ethnicity, baseline symptom severity, final symptom severity, and
overall improvement (p > .05). The only exception is that a larger
The OQ-45.2 is a 45-item self-report instrument that was designed to
proportion of female clients were found during the pandemic than
evaluate clients’ psychological distress and functioning and has been
prepandemic (p < .01). These results fail to support that the COVID-
widely used to assess treatment progress and outcome in routine practice
19 had directly impacted the treatment outcome for the current
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

(Lambert et al., 1996). Each item asks individuals to rate their feeling
sample. Thus, we included clients who were treated remotely during
This document is copyrighted by the American Psychological Association or one of its allied publishers.

over the past 2 weeks on a 5-point Likert-style scale from 0 (never) to 4

the pandemic in the following analyses. Additionally, we compared
(almost always). The total score was calculated by summing all items,
the weekly and biweekly groups on age, gender, ethnicity, and
with higher scores reflecting higher levels of distress. The OQ-45.2 has a
baseline symptom severity using independent t tests and chi-square
cutoff score of 63 that can distinguish clinical populations from
tests. The weekly group had a higher level of OQ at baseline compared
nonclinical populations. Further, the OQ-45.2 includes a reliable change
to the biweekly group ( p < .001). No significant differences in age,
index (RCI) and clinically significant improvement (CSI). Specifically,
gender, and ethnicity were found between the two groups.
reliable change is defined as a change in OQ scores that exceed
the smallest significant difference (i.e., at least 14 points). Accordingly,
an increase of 14 points or more results in reliable deterioration. Trajectory Analysis
Furthermore, CSI requires (a) a reliable change and (b) that the initial
LGMM was performed and described following the guidelines for
OQ score decreased from the clinical range (64 or above) to the
Reporting on Latent Trajectory Studies (GRoLTS), a tool for
nonclinical range (63 or below). Previous studies have shown that the
reporting on latent trajectory studies (van de Schoot et al., 2017). To
OQ has excellent internal consistency (α = .93) and the ability to
identify the trajectories of symptom change in weekly therapy, we
discriminate clinical and nonclinical samples (sensitivity = 0.85;
performed LGMM on the OQ scores of the weekly therapy group
specificity = 0.74). Furthermore, the OQ-45.2 has demonstrated good
(N = 886) using Mplus. The time was measured by session number
test–retest reliability on a time interval of 21 days (r = .84) and was
(approximately weekly). A multilevel model with clients nested
sensitive to short-term change following treatment (Lambert et al.,
within therapists was adopted. The random start values and iterations
1996). The Cronbach’s α based on the current sample was .92.
were set as 200 and 20, respectively. Consistent with previous studies
(Owen et al., 2015; Palmer et al., 2021), clients who received fewer
Counseling Center Assessment of Psychological than 10 therapy sessions were not considered to have missing data.
Symptoms–62 (CCAPS-62) For example, if a client attended five sessions, the last five time points
were not considered to be missing and therefore, were not imputed.
The CCAPS-62 is a self-report, multidimensional instrument that
To determine whether adding higher order terms improve model
was designed to assess psychological symptoms in multiple areas
fit, we compared the one-class LGMM models with linear term,
for college students and to facilitate initial assessment, treatment
linear and quadratic terms, and linear, quadratic, and cubic terms.
planning, and outcome monitoring in university counseling centers
After the model terms were determined, we further compared the
(Locke et al., 2011). The CCAPS-62 consists of 62 items, which are
LGMM models with different variance and covariance estimates.
clustered into eight subscales: Depression, Generalized Anxiety,
Specifically, the variance and covariance estimates for each growth
Social Anxiety, Eating Concerns, Substance Use, Family Distress,
factor within each class were first fixed to zero (i.e., LCGA), then
Academic Distress, and Frustration/Anger. Each subscale includes
fixed to be equal across classes, and finally set to be freely estimated
two cutoff scores that classify the symptom severity into three
across the classes. The model fit indices of the three models were
ranges: low, mild, and elevated (Youn et al., 2015). Clients were
compared to determine the best model specification.
asked to rate each item on a 5-point scale from 0 (not at all like me)
Next, we added one class each time until both adjusted Lo–
to 4 (extremely like me) based on their feelings and behaviors in the
Mendell–Rubin (LMR) and Vuong–Lo–Mendell–Rubin (VLMR)
past 2 weeks. Previous studies have demonstrated acceptable to
likelihood ratio tests, which are commonly used tests for LGMM
excellent internal consistency (α = .781–.913) and validity for the
model selection, were no longer significant, suggesting that adding
subscales of the CCAPS-62 (Locke et al., 2011). The Cronbach’s α
additional classes did not significantly improve the model fit
based on the current sample was .93.
(Nylund et al., 2007). The optimal model was determined by a
variety of statistical fit indices, including Akaike’s information
Data Analysis criterion (AIC), Bayesian information criterion (BIC), and sample
Sensitivity Analysis size adjusted BIC (SSA-BIC). We also considered each model’s
clinical utility and conceptual rationale when selecting models to
Sensitivity analyses were performed to examine whether the avoid the overextraction of classes and ensure parsimony (Nagin &
COVID-19 pandemic had affected the treatment outcome for the Odgers, 2010). Specifically, if the additional classes fail to add
current sample. Specifically, among the whole sample, we identified distinctive features (e.g., similar shape to existing trajectories) or
 TRAJECTORIES IN WEEKLY AND BIWEEKLY THERPAY 5

constitute a very small proportion (e.g., fewer than 3% of the suggest that the three-class model was superior to the one-class
sample), the more parsimonious model (i.e., fewer classes) would be model and the two-class model. Although the AIC and SSA-BIC
selected. favor the four-class model over the three-class model, the difference
We then performed the same analyses on the OQ scores of the in BIC between both models was minimal (BIC in four class:
biweekly therapy group (N = 1,753) to identify the change trajectories 34,061.57; BIC in three class: 34,062.31). Additionally, the fourth
in biweekly therapy. The time was measured by session number class only accounted for about 1% of the sample, thereby adding
(approximately biweekly). After the final models were determined, little clinical utility to the three-class model. Consequently, the
two multinominal logistic regression models were performed to three-class model was used for further analyses.
explore the predictors of the change trajectories for biweekly and Figure 1 shows the three trajectories of change in weekly therapy.
weekly therapy, respectively. Predictor variables entered into the The first and largest class, slow change, accounted for 78.33% of the
model were the eight CCAPS subscores. sample (n = 694) who received an average of 4.75 (SD = 1.63)
sessions. These clients had the lowest levels of initial distress
and demonstrated slow and steady improvements throughout the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Results
treatment (5.3 OQ points). The average OQ score (OQ = 65.05) of
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Overall Effectiveness of Weekly and Biweekly Therapy this trajectory at treatment termination remained above the clinical
The 886 clients attending weekly therapy received an average of cutoff of 63. Of note, the average OQ reached the subclinical range at
4.84 (SD = 1.64; Mdn = 5) sessions (see Supplemental Table 1, for the ninth session for those who continued to receive therapy. Among
details). The average treatment length for the weekly sample was this subgroup, 23.92% achieved reliable change and 18.01%
4.88 weeks (SD = 1.82, Mdn = 4.57). The average OQ of the weekly achieved clinically significant change and 4.47% deteriorated.
sample was 72.28 (SD = 19.61; range = 9–144) at the first session The second trajectory, early improvement, is comprised of 17.6% of
and 63.28 (SD = 21.56; range = 0–134) at the last session, achieving the sample (n = 156) attending an average of 4.92 (SD = 1.41) sessions.
an average decrease of 9 (SD = 16.2; range = −69–77) OQ points. This trajectory demonstrated rapid improvements in early treatment
Among these 886 clients, 323 (36.4%) achieved RCI, and 257 followed by slight deterioration starting at around Session 7. Clients in
(29.01%) achieved CSI whereas 57 (6.43%) deteriorated (Table 1). the early improvement trajectory, though having the highest levels of
As for biweekly therapy, the average treatment length was 5.02 initial distress, achieved substantial improvements (32.13 OQ points) in
(SD = 1.95; Mdn = 5) sessions. The average treatment length for the the first six sessions and became asymptomatic at the end of treatment,
biweekly sample was 9.88 weeks (SD = 4.27, Mdn = 9). The average followed by minimal deterioration at Session 6. This trajectory was only
OQ of the 1,753 clients attending biweekly therapy decreased from plotted to Session 7 because fewer than five clients attended more than
68.93 (SD = 22.19; range = 10–133) at the first session to 60.98 seven sessions. All clients in this trajectory achieved reliable change and
(SD = 34.34; range = 5–146) at the last session, achieving an average 83.97% achieved clinically significant change; none deteriorated.
change of 7.95 (SD = 17.18; range = −73–81) OQ points. Among The last and smallest trajectory, worse before better, accounted for
the 1,753 clients, 602 (34.34%) achieved RCI, and 456 (26.01%) 4.06% of the sample (n = 36). Attending an average of 6.36 (SD =
achieved CSI; 154 (8.78%) deteriorated (Table 2). 1.96) sessions, clients in the worse before better trajectory received
longer treatment compared to those in the other two trajectories.
However, these clients greatly deteriorated in the first five sessions
Trajectories of Change in Weekly Therapy and then slightly improved after around Session 6. On average, the
After comparing one-class linear, quadratic, and cubic LGMM OQ scores of this subgroup increased by 19.81 (SD = 15.77) points.
models, we found that the model with the cubic term showed better The majority of clients in this trajectory deteriorated (72.22%) and
model fit than the models with linear-only and quadratic terms only 2.78% of them achieved RCI and CSI. Similarly, this trajectory
(Supplemental Table 2). The within-class variances for intercept was only plotted to Session 9 because fewer than five clients attended
were fixed to be freely estimated for all classes, whereas the within- more than nine sessions.
class variances for linear, quadratic, and cubic terms were fixed to be
zero. We tested one-class, two-class, three-class, and four-class
Trajectories of Change in Biweekly Therapy
LGMM models of trajectories in weekly therapy as the adjusted
LMR and VLMR likelihood ratio tests were insignificant when Consistent with the models of weekly therapy, the model with the
adding the fourth class (Supplemental Table 3). All model fit indices cubic term showed a better model fit than the models with linear-only

Table 1
Three Trajectories of Change in Weekly Therapy

Class n Session length Total OQ change Initial OQ Final OQ RCI (%) CSI (%) Deterioration (%)

Worse before better 36 6.36 (1.96) 19.81 (15.77) 75 (19.96) 94.81 (19.02) 2.78 2.78 72.22
Slow change 694 4.75 (1.63) −5.3 (11.14) 70.35 (19.53) 65.05 (20.31) 23.92 18.01 4.47
Early improvement 156 4.92 (1.41) −32.13 (9.72) 80.24 (17.88) 48.11 (15.88) 100 83.97 0
Total 886 4.84 (1.64) −9 (16.2) 72.28 (19.61) 63.28 (21.56) 36.46 29.01 6.43
Note. RCI = reliable change index; CSI = clinically significant improvement; OQ = Outcome Questionnaire. Negative values of total OQ change
indicate symptom relief, and positive values of total OQ change indicate symptom deterioration. Data refer to mean (standard deviation) unless otherwise
specified.
 6 LIN, ANDERSON, EREKSON, AND OGLES

Table 2
Three Trajectories of Change in Biweekly Therapy

Class n Session length Total OQ change Initial OQ Final OQ RCI (%) CSI (%) Deterioration (%)

Worse before better 107 5.34 (1.88) 13.74 (24.1) 73.72 (23.24) 87.46 (25.72) 13.08 11.21 62.62
Slow change 1,409 5.01 (1.99) −5.22 (12.03) 66.71 (19.87) 61.49 (20.85) 25.34 17.46 6.17
Early improvement 237 4.93 (1.74) −33.97 (11.39) 79.97 (15.03) 46 (14.86) 97.47 83.54 0
Total 1,753 5.02 (1.95) −7.95 (17.18) 68.93 (20.06) 60.98 (22.19) 34.34 26.01 8.78
Note. RCI = reliable change index; CSI = clinically significant improvement; OQ = Outcome Questionnaire. Negative values of total OQ change
indicate symptom relief, and positive values of total OQ change indicate symptom deterioration. Data refer to mean (standard deviation) unless otherwise
specified.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

and quadratic terms for biweekly therapy (Supplemental Table 4). Figure 2 depicts the three trajectories of change in biweekly
This document is copyrighted by the American Psychological Association or one of its allied publishers.

The within-class variances for intercept and cubic terms were fixed to therapy. The first trajectory, slow change, comprised 80.3% of the
be freely estimated for all classes, whereas the within-class variances sample (n = 1,409) attending an average of 5.01 sessions. Compared
for linear and quadratic terms were fixed to be zero. We tested one- to those in other trajectories in biweekly therapy, clients in the slow
class, two-class, and three-class LGMM models in biweekly therapy change trajectory had lower levels of initial distress (OQ = 66.71),
as the adjusted LMR and VLMR likelihood ratio tests became which is slightly higher than the clinical cutoff of 63. They
insignificant when adding the third class (Supplemental Table 5). All demonstrated slow and steady improvements throughout the treatment
model fit indices suggest that the three-class model was superior to (5.22 OQ points) and reached the subclinical range (OQ = 61.49) at
the one-class model and the two-class model. In terms of clinical the end of treatment. Among these clients, 25.34% achieved reliable
utility, the three-class model extracted a distinct pattern that differs change and 17.46% achieved clinically significant change whereas
from the two patterns in the two-class model. Consequently, the 6.17% deteriorated. After controlling for initial OQ and the number
three-class model was used for further analyses. of sessions, the slow change group in biweekly therapy did not

Figure 1
Trajectories of Change in Weekly Therapy

Note. Dashed line indicates the clinical cutoff of OQ (>63). Shaded areas = ±1 SD of the mean OQ of each
trajectory. Scores for Session 8 and beyond of early improvement trajectory and for Session 10 of worse before
better trajectory were not plotted due to a lack of sufficient data. OQ = Outcome Questionnaire. See the online
article for the color version of this figure.
 TRAJECTORIES IN WEEKLY AND BIWEEKLY THERPAY 7

Figure 2
Trajectories of Change in Biweekly Therapy
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Note. Dashed line indicates the clinical cutoff of OQ (>63). Shaded areas = ±1 SD of the mean OQ of each trajectory. Scores for Session 10 of the worse
before better trajectory were not plotted due to a lack of sufficient data. OQ = Outcome Questionnaire. See the online article for the color version of this figure.

significantly differ from the slow change group in weekly therapy weekly therapy (Table 3). Clients with higher levels of depression at
regarding final OQ ( p > .05). intake were less likely to fall into the slow change trajectory compared
The second trajectory, early improvement, is comprised of 237 to the worse before better trajectory (p < .05) in weekly therapy. In
clients (13.5% of the sample). On average, clients in the early contrast, clients with higher levels of eating concerns at intake were
improvement trajectory had higher levels of initial distress (OQ = more likely to fall into the slow change ( p < .05) and early
79.97) and attended fewer sessions 4.93 (SD = 1.74) than those in improvement ( p < .05) trajectory than the worse before better
other trajectories in biweekly therapy. Nevertheless, the psychologi- trajectory in weekly therapy. None of the CCAPS domains
cal distress of these clients demonstrated a rapid and great decrease in significantly predicted whether clients fall into slow change versus
early treatment (i.e., first six sessions), followed by a mild increase early improvement trajectory in weekly therapy.
until Session 8 when another improvement emerged. Among these
clients, 97.47% achieved reliable change and 83.54% achieved
clinically significant change; none deteriorated. After controlling for
Predictors of Trajectories of Change in
initial OQ and the number of sessions, the early improvement group Biweekly Therapy
in biweekly therapy did not significantly differ from its counterpart in Table 4 shows the associations between CCAPS subscores and
weekly therapy regarding final OQ ( p > .05). trajectories of change in weekly therapy. Clients with higher levels of
The last trajectory, worse before better, accounted for 6.1% of the eating concerns ( p < .05) and frustration/anger (p < .05) were less
sample (n = 107) who received an average of 5.34 (SD = 1.88) therapy likely to be in the slow change trajectory compared to the worse before
sessions. These clients demonstrated an early deterioration in the first better trajectory in biweekly therapy. Clients with higher levels of
three sessions, followed by a plateau and then an improvement at Session eating concerns at intake were also less likely to be in the early
7. Throughout the treatment, the OQ scores of this trajectory decreased improvement trajectory compared to the worse before better trajectory
by 13.74 (SD = 24.1) points. Most clients in this trajectory deteriorated (p < .01) in biweekly therapy. Additionally, higher levels of
(62.62%); 12.08% achieved RCI and 11.21% achieved CSI. After depression (p < .01) and academic concerns ( p < .05) at intake were
controlling for initial OQ and the number of sessions, the worse before associated with a higher chance of being in the early improvement
better group in biweekly therapy showed lower levels of final OQ trajectory compared to the slow change trajectory in biweekly therapy.
compared to the worse before better in weekly therapy (p < .01).

Discussion
Predictors of Trajectories of Change in Weekly Therapy
This study was the first to examine the trajectories of change in
Multinominal logistic regression was performed to examine the both weekly and biweekly therapy and explore the predictors of
associations between CCAPS subscores and trajectories of change in trajectories, respectively. Using a naturalistic sample of 5,102 clients
 8 LIN, ANDERSON, EREKSON, AND OGLES

Table 3
Multinomial Logistic Regression Model of Trajectories in Weekly Therapy

Worse before better (ref.) versus Slow change (ref.) versus

Slow change Early improvement Early improvement
Outcome measure OR 95% CI OR 95% CI OR 95% CI

Depression 0.52* [0.27, 0.99] 0.53 [0.26, 1.05] 1.01 [0.72, 1.39]
Anxiety 0.85 [0.54, 1.34] 0.70 [0.42, 1.14] 0.82 [0.63, 1.04]
Social anxiety 1.36 [0.87, 2.10] 1.28 [0.79, 2.06] 0.94 [0.74, 1.19]
Academic concerns 1.08 [0.69, 1.66] 1.30 [0.80, 2.07] 1.20 [0.96, 1.49]
Eating concerns 1.68* [1.03, 2.73] 1.92* [1.15, 3.20] 1.14 [0.93, 1.39]
Frustration/anger 1.27 [0.74, 2.18] 1.48 [0.82, 2.63] 1.16 [0.88, 1.51]
Substance use 1.29 [0.22, 7.47] 1.38 [0.22, 8.55] 1.07 [0.54, 2.09]
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Family distress 0.78 [0.54, 1.11] 0.89 [0.60, 1.31] 1.14 [0.93, 1.39]
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Note. ref. = reference; OR = odds ratio; CI = confidence interval.

* p < .05.

receiving 3–10 therapy sessions at a university counseling center, we that, for most clients, reducing session frequency does not reduce
identified 886 clients attending approximately weekly therapy and treatment effectiveness when enough sessions were offered, thereby
1,753 clients attending approximately biweekly therapy. Consistent being a feasible solution to limited resources. Consistently, previous
with previous trajectory analyses of psychotherapy outcome (Lin studies have shown that biweekly therapy can lead to equal levels of
et al., 2023; Owen et al., 2015; Palmer et al., 2021), the present study RCI and CSI with enough dosage (Erekson et al., 2015). Of note, this
identified three trajectories in both weekly and biweekly therapy. does not necessarily suggest that weekly and biweekly therapy were
This finding provided further evidence for multiple trajectories even equally effective. Instead, for individuals following the slow change
when the session frequency was controlled. We further examined the trajectory, weekly therapy can lead to faster improvement than
roles of the client’s CCAPS profile at intake in predicting their biweekly therapy, thereby shortening the duration of suffering. Thus,
trajectories of symptom change in weekly and biweekly therapy, weekly therapy should be prioritized for these clients to enhance
respectively. treatment efficiency when such resources are available (Bailey et al.,
Notably, the trajectories in weekly therapy were more or less similar 2022; Erekson et al., 2022).
to those in biweekly therapy but slightly differed in size and shape (see The second trajectory, early improvement, the most favoring
Supplemental Table 6). Such findings enhanced our understanding of outcome pattern, was characterized by rapid improvement in early
how session frequency may affect change trajectories. The majority of treatment. The minimal deterioration at around Session 6 may be
clients in weekly (78.33%) and biweekly therapy (80.38%) entered because clients who had achieved recovery had terminated therapy and
therapy with clinical levels of symptoms and improved slowly and the remainder were more severe cases; nevertheless, they continued to
steadily during the treatment (slow change). This trajectory has been improve if attending more sessions. Clients in this trajectory started
consistently found in previous studies (Owen et al., 2015; Palmer therapy with the highest level of initial symptoms and demonstrated
et al., 2021). Although the slow change trajectory in the weekly substantial improvements. Their average OQ at the end of treatment
sample had substantially higher levels of initial distress and received was close to that of community populations (Lambert et al., 1996). In
shorter treatment than its equivalent in the biweekly sample, they both weekly and biweekly therapy, most clients in early improvement
achieved comparable outcome, including the overall OQ change and trajectories achieved RCI and CSI whereas no one experienced
the likelihood of RCI, CSI, and deterioration. This finding suggests deterioration. The early improvement trajectory in weekly therapy

Table 4
Multinomial Logistic Regression Model of Trajectories in Biweekly Therapy Using CCAPS

Worse before better (ref.) versus Slow change (ref.) versus

Slow change Early improvement Early improvement
Outcome measure OR 95% CI OR 95% CI OR 95% CI

Depression 0.83 [0.57, 1.18] 1.18 [0.78, 1.78] 1.43** [1.12, 1.82]
Anxiety 0.89 [0.66, 1.17] 1.04 [0.75, 1.44] 1.18 [0.96, 1.42]
Social anxiety 0.95 [0.72, 1.25] 0.93 [0.67, 1.27] 0.98 [0.81, 1.17]
Academic concerns 0.84 [0.65, 1.07] 1.00 [0.75, 1.33] 1.20* [1.01, 1.42]
Eating concerns 0.80* [0.64, 0.99] 0.70** [0.54, 0.91] 0.87 [0.73, 1.04]
Frustration/anger 0.75* [0.56, 0.98] 0.78 [0.56, 1.08] 1.05 [0.84, 1.30]
Substance use 1.14 [0.51, 2.54] 0.96 [0.36, 2.53] 0.84 [0.42, 1.68]
Family distress 0.86 [0.68, 1.06] 0.79 [0.61, 1.01] 0.92 [0.78, 1.08]
Note. ref. = reference; OR = odds ratio; CI = confidence interval; CCAPS= counseling center assessment of psychological symptoms.
* p < .05. ** p < .01.
 TRAJECTORIES IN WEEKLY AND BIWEEKLY THERPAY 9

highly resembles that in biweekly therapy. However, weekly therapy offering additional weekly sessions, therapists may consider offering
results in a larger proportion of early improvement than biweekly the additional sessions on a less frequent basis (e.g., biweekly),
therapy (17.61% in weekly therapy vs. 13.52% in biweekly therapy), which saves resources and extends treatment length.
suggesting that increasing therapy frequency may increase the The present study extended the findings of previous studies of the
likelihood of early and rapid improvement. This is consistent with impact of session frequency on treatment outcomes by identifying
previous findings that more frequent therapy is associated with faster how weekly and biweekly therapy result in different trajectories of
recovery (e.g., Erekson et al., 2022). change. While previous studies consistently suggest that weekly
The last trajectory, worse before better, accounted for 4.06% of therapy produces superior outcomes compared to biweekly therapy
the weekly sample and 6.1% of the biweekly sample. In both weekly on average, we found that the effects of adjusting session frequency
and biweekly therapy, clients in this trajectory deteriorated in the may be more complicated. Overall, weekly therapy leads to a larger
first several sessions and, if not discontinuing therapy, showed proportion of clients who showed rapid and dramatic improvements
improvement. Despite the later improvement, however, these clients and enables clients who change slowly and steadily to improve faster
experienced substantial deterioration throughout therapy. Indeed, compared to biweekly therapy. For individuals who deteriorated in
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

the later improvement may be partially because clients who early therapy, however, weekly therapy was found to result in greater
This document is copyrighted by the American Psychological Association or one of its allied publishers.

responded negatively to therapy have dropped out. Of note, the deterioration for them compared to biweekly therapy. Increasing
worse before better trajectory in weekly therapy appears to be less session frequency likely intensifies the positive effects, as well as the
favoring than its equivalent in biweekly therapy. Specifically, only negative effects of therapy (i.e., greater symptom deterioration, lower
2.78% of the worse before better subgroup in weekly therapy rates of RCI and CSI). Therefore, it is essential to identify the change
achieved RCI and CSI, whereas over 10% of their counterparts in trajectories and adjust session frequency accordingly.
biweekly therapy achieved RCI (13.08%) and CSI (11.21%). Clients’ CCAPS profiles were found to be predictors of change
Furthermore, the worse before better subgroup in weekly therapy trajectories in weekly and biweekly therapy. Interestingly, clients
had a higher likelihood and magnitude of deterioration than those in with higher levels of eating concerns at intake were more likely to be
biweekly therapy (deterioration: 72.22% vs. 62.62%; OQ increase: in the worse before better trajectory versus the other two trajectories
19.81 points vs. 13.74 points). in biweekly therapy, whereas they were less likely to be in the worse
This worse before better trajectory in the present study is in line before better trajectory in weekly therapy. This suggests that clients
with Owen et al.’s (2015) findings that approximately 5.5% with higher levels of eating concerns are more likely to benefit from
followed a worse before better trajectory. Other researchers have weekly therapy instead of biweekly therapy. This may be because
also found negative outcomes in psychotherapy (Cuijpers et al., eating concerns require close monitoring and intensive treatments.
2018; Rozental et al., 2018). Of note, this trajectory does not Likewise, clients with higher levels of frustration/anger ( p < .05)
necessarily indicate psychotherapy is ineffective or harmful. The were less likely to be in the slow change trajectory compared to the
initial deterioration may reflect several reasons. For example, worse before better trajectory in biweekly therapy. Therefore, high-
the initial deterioration may be because clients experience alliance intensity treatment should be prioritized for such clients.
ruptures with the therapists and the later improvement may be the In contrast, clients with higher levels of depression at intake were
evidence of rupture repair (Zilcha-Mano & Errázuriz, 2017). more likely to fall into the worse before better trajectory versus the
Additionally, clients may experience increasing distress initially slow change trajectory in weekly therapy, whereas such clients were
when they start overtly exposing themselves to presenting problems more likely to be in the early improvement trajectory versus slow
(e.g., trauma, crisis; Littleton et al., 2007). More frequent therapy change trajectory in biweekly therapy. This suggests clients with
likely makes the process more intense and struggling, thereby higher levels of depression may benefit from more attenuated
leading to greater deterioration. This may also explain why the session frequency schedules. Likewise, higher levels of academic
initial deterioration of the worse before better subgroup in weekly concerns at intake were associated with a higher likelihood of being
therapy lasted for more sessions than that in biweekly therapy. in the early improvement trajectory compared to the slow change
Indeed, the initial deterioration lasted for 5 weeks approximately for trajectory in biweekly therapy. This is likely because the additional
both groups. Compared to the weekly group, the biweekly groups time between sessions in biweekly therapy allows these college
may have benefited from the longer time intervals between sessions. students to resolve their academic concerns. Therefore, therapists
Therefore, it may be practical and even beneficial to attenuate may consider attenuating session frequency for those clients to save
session frequency for such clients in order to provide them with health care resources.
more time between sessions to process.
Regardless of the underlying mechanism, these findings suggest
Clinical Implications
that clients who experienced initial deterioration need to remain in
therapy for longer periods of time to engender later improvement. Our findings have several implications in addressing the dilemma
However, this may not be feasible for all cases considering the of prioritizing session frequency versus serving more clients. First,
limited session policy in many university counseling centers and consistent with previous studies (Bruijniks et al., 2020; Erekson
clinics (Gallagher, 2014). Many clients and therapists have to et al., 2015), weekly therapy should be prioritized as it increases the
discontinue therapy before they find an effective way to manage the chances of achieving early improvement and reduces the duration of
distress. Beyond offering additional sessions for certain clients, suffering when such resources are available. However, it may be
which has been adopted in some clinical settings (Jacques & Abel, more efficient and beneficial to attenuate session frequency and
2020), the current findings suggest that decreasing session frequency extend treatment duration for clients who showed initial deteriora-
may be an efficient and feasible solution. For example, for clients tion. This finding also highlights the importance of routine outcome
who have not improved when reaching the session limit, instead of monitoring.
 10 LIN, ANDERSON, EREKSON, AND OGLES

Second, given that individuals fall into different trajectories in completion status of the treatment, as well as the reasons for
biweekly and weekly therapy, one-size-fits-all policies may not termination. It is unknown whether the clients had discontinued or
address the dilemma when clinical demand exceeds clinical capacity. completed treatment.
Instead, our study suggests that individualizing session frequency Additionally, this study only compared biweekly therapy and
may help therapists allocate clinical resources and increase treatment weekly therapy. Although treating session frequency as a
efficacy. We found that individuals with higher levels of eating categorical variable can inform clinical decisions, it remains unclear
concerns and frustration/anger at intake were more likely to fall into how session frequency as a continuous variable affects trajectories
negative trajectories in biweekly therapy compared to weekly of change in therapy. Likewise, the trajectories of change in therapy
therapy. In contrast, clients who suffered from depression and of other session frequencies (e.g., twice per week) are unknown.
academic concerns were more likely to fall into positive trajectories Finally, our sample was relatively small considering the computa-
in biweekly therapy. Therefore, therapists can prioritize session tional burden of LGMM and was from one counseling center. Future
frequency for individuals with frustration/anger and eating concerns studies may benefit from a larger sample from multiple sites.
and attenuate session frequency for clients with depression and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

academic concerns. Such findings also contribute to the development

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Conclusion
of stepped care models in university counseling centers.
Additionally, while only a small proportion of clients showed the Our study shed light on the heterogeneity of treatment responses
worse before better trajectory, it is clinically important to offer and how adjusting session frequency may affect the shapes and sizes
enough therapy sessions for these clients to reach the “better” stage. of the trajectories. We found that both weekly and biweekly therapy
Flexibility is needed despite the mandated session limit policies in result in three trajectories of change: early improvement, slow change,
many clinics. Finally, the multiple trajectories start to diverge in the and worse before better. Compared to biweekly therapy, weekly
first several sessions. This is consistent with previous findings that therapy generally increases the chances of early improvement and
clients show multiple trajectories of symptom change in early shortens the duration of suffering for clients who change slowly but
psychotherapy and these early trajectories can significantly predict results in greater deterioration for individuals in the worse before
final treatment outcome (Lin et al., 2023). Therefore, therapists may better trajectory. These findings suggest that the impacts of session
compare each client’s symptom change to these trajectories in early frequency on treatment outcome varied across clients and highlight
therapy and adjust treatment strategies if they fall into less favorable the importance of adjusting session frequency individually. We also
trajectories. found that clients’ CCAPS profile at intake predicts the trajectories of
change in weekly and biweekly therapy. Some clients, for example,
Limitations and Future Directions those with eating concerns, may need more frequent treatment to
recover, whereas other clients may benefit equally from less frequent
This study has several limitations that warrant caution. First, we treatment. These findings can guide practitioners to allocate resources
identified clients attending approximately weekly and biweekly more efficiently by prioritizing session frequency for clients who are
therapy using a naturalistic data set. Although this method enhanced in dire need.
the generalizability of the findings, the weekly sample and biweekly
sample may differ in terms of presenting concerns and treatment
motivation, thereby confounding the effects of session frequency on References
change trajectories. Indeed, the weekly sample showed a higher level Allan, N. P., Gros, D. F., Myers, U. S., Korte, K. J., & Acierno, R. (2017).
of OQ at baseline compared to the biweekly sample. A potential Predictors and outcomes of growth mixture modeled trajectories across an
explanation for this difference is that therapists might have already exposure-based PTSD intervention with veterans. Journal of Clinical
increased session frequency for clients with more severe symptoms Psychology, 73(9), 1048–1063. https://doi.org/10.1002/jclp.22408
to better serve their needs. It remains unclear how the session Bailey, R. J., Erekson, D. M., Cattani, K., Jensen, D., Simpson, D. M., Klundt,
frequency was determined by the therapists and clients. To address J., Vogeler, H. A., Schmuck, D., Worthen, V. E., Caldwell, Y., Beecher,
M. E., Griner, D., & Hobbs, K. (2022). Adapting stepped care: Changes to
this limitation, future researchers may match client characteristics or
service delivery format in the context of high demand. Psychological
even utilize an RCT design to achieve a direct comparison of change
Services, 19(3), 494–501. https://doi.org/10.1037/ser0000564
trajectories in biweekly and weekly therapy. Second, we only Bruijniks, S. J. E., Lemmens, L. H. J. M., Hollon, S. D., Peeters, F. P. M. L.,
examined the associations between the CCAPS profile and change Cuijpers, P., Arntz, A., Dingemanse, P., Willems, L., van Oppen, P., Twisk,
trajectories. Potential predictors such as working status, clinical J. W. R., van den Boogaard, M., Spijker, J., Bosmans, J., & Huibers,
history, suicide risk, and diagnoses should be included in future M. J. H. (2020). The effects of once- versus twice-weekly sessions on
studies to enhance the predictive capabilities. The effects of treatment psychotherapy outcomes in depressed patients. The British Journal of
approaches should also be considered. Psychiatry, 216(4), 222–230. https://doi.org/10.1192/bjp.2019.265
Third, the present study focused on clients receiving 3–10 therapy Cuijpers, P., Huibers, M., Ebert, D. D., Koole, S. L., & Andersson, G. (2013).
sessions. It is unclear about clients’ change trajectories in long-term How much psychotherapy is needed to treat depression? A metaregression
analysis. Journal of Affective Disorders, 149(1–3), 1–13. https://doi.org/
treatment and how session frequency may affect them. Also, the
10.1016/j.jad.2013.02.030
findings may not well represent those who attended only one or two
Cuijpers, P., Reijnders, M., Karyotaki, E., de Wit, L., & Ebert, D. D. (2018).
sessions, which were the majority of clients in university counseling Negative effects of psychotherapies for adult depression: A meta-analysis
centers. The mean number of sessions in most classes was around of deterioration rates. Journal of Affective Disorders, 239, 138–145.
five. The trajectories for the latter session may change if these clients https://doi.org/10.1016/j.jad.2018.05.050
completed all sessions and therefore should be interpreted with Cuijpers, P., van Lier, P. A. C., van Straten, A., & Donker, M. (2005).
caution. Another limitation is that the data set did not include the Examining differential effects of psychological treatment of depressive
 TRAJECTORIES IN WEEKLY AND BIWEEKLY THERPAY 11

disorder: An application of trajectory analyses. Journal of Affective counseling center assessment of psychological symptoms-62 (CCAPS-62).
Disorders, 89(1–3), 137–146. https://doi.org/10.1016/j.jad.2005.09.001 Journal of Counseling Psychology, 58(1), 97–109. https://doi.org/10.1037/
Drake, R. E., & Latimer, E. (2012). Lessons learned in developing a0021282
community mental health care in North America. World Psychiatry, 11(1), Lutz, W., Hofmann, S. G., Rubel, J., Boswell, J. F., Shear, M. K., Gorman,
47–51. https://doi.org/10.1016/j.wpsyc.2012.01.007 J. M., Woods, S. W., & Barlow, D. H. (2014). Patterns of early change and
Erekson, D. M., Bailey, R. J., Cattani, K., Klundt, J. S., Lynn, A. M., Jensen, their relationship to outcome and early treatment termination in patients
D., Merrill, B. M., Schmuck, D., & Worthen, V. (2022). Psychotherapy with panic disorder. Journal of Consulting and Clinical Psychology,
session frequency: A naturalistic examination in a university counseling 82(2), 287–297. https://doi.org/10.1037/a0035535
center. Journal of Counseling Psychology, 69(4), 531–540. https:// Muthén, L. K., & Muthén, B. (2020). Mplus user’s guide.
doi.org/10.1037/cou0000593 Nagin, D. S., & Odgers, C. L. (2010). Group-based trajectory modeling in
Erekson, D. M., Lambert, M. J., & Eggett, D. L. (2015). The relationship clinical research. Annual Review of Clinical Psychology, 6(1), 109–138.
between session frequency and psychotherapy outcome in a naturalistic https://doi.org/10.1146/annurev.clinpsy.121208.131413
setting. Journal of Consulting and Clinical Psychology, 83(6), 1097– Nylund, K. L., Asparouhov, T., & Muthén, B. O. (2007). Deciding on the
1107. https://doi.org/10.1037/a0039774 number of classes in latent class analysis and growth mixture modeling: A
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Gallagher, R. P. (2014). National survey of college counseling. http://d-scho Monte Carlo simulation study. Structural Equation Modeling, 14(4), 535–
This document is copyrighted by the American Psychological Association or one of its allied publishers.

larship.pitt.edu/28178/ 569. https://doi.org/10.1080/10705510701575396

Gorman, K. S., Bruns, C., Chin, C., Fitzpatrick, N., Koenig, L., LeViness, P., Owen, J., Adelson, J., Budge, S., Wampold, B., Kopta, M., Minami, T., &
& Sokolowski, K. (2021). The association for university and college Miller, S. (2015). Trajectories of change in psychotherapy. Journal of
counseling center directors annual survey. https://www.aucccd.org/asse Clinical Psychology, 71(9), 817–827. https://doi.org/10.1002/jclp.22191
ts/2020-21%20Annual%20Survey%20Report%20Public%20Survey.pdf Palmer, R., Owen, J., & Frazier, P. (2021). Trajectories of changes in distress
Heckman, T. G., Heckman, B. D., Anderson, T., Bianco, J. A., Sutton, M., & in counseling center clients: A replication study. Psychotherapy Research,
Lovejoy, T. I. (2017). Common factors and depressive symptom relief 31(3), 359–371. https://doi.org/10.1080/10503307.2020.1757175
trajectories in group teletherapy for persons ageing with HIV. Clinical R Core Team. (2022). R: A language and environment for statistical
Psychology & Psychotherapy, 24(1), 139–148. https://doi.org/10.1002/ computing. R Foundation for Statistical Computing. https://www.R-proje
cpp.1989 ct.org/
Jacques, J. G., & Abel, N. R. (2020). Using the stepped care model to Reese, R. J., Toland, M. D., & Hopkins, N. B. (2011). Replicating and
empower university students with learning disabilities. Journal of College extending the good-enough level model of change: Considering session
Counseling, 23(1), 85–96. https://doi.org/10.1002/jocc.12151 frequency. Psychotherapy Research, 21(5), 608–619. https://doi.org/10
Jankowski, P. J., Sandage, S. J., Bell, C. A., Rupert, D., Bronstein, M., & .1080/10503307.2011.598580
Stavros, G. S. (2019). Latent trajectories of change for clients at a Rozental, A., Castonguay, L., Dimidjian, S., Lambert, M., Shafran, R.,
psychodynamic training clinic. Journal of Clinical Psychology, 75(7), Andersson, G., & Carlbring, P. (2018). Negative effects in psychotherapy:
1147–1168. https://doi.org/10.1002/jclp.22769 Commentary and recommendations for future research and clinical practice.
Joesch, J. M., Golinelli, D., Sherbourne, C. D., Sullivan, G., Stein, M. B., BJPsych Open, 4(4), 307–312. https://doi.org/10.1192/bjo.2018.42
Craske, M. G., & Roy-Byrne, P. P. (2013). Trajectories of change in Tiemens, B., Kloos, M., Spijker, J., Ingenhoven, T., Kampman, M., &
anxiety severity and impairment during and after treatment with evidence- Hendriks, G. J. (2019). Lower versus higher frequency of sessions in
based treatment for multiple anxiety disorders in primary care. Depression starting outpatient mental health care and the risk of a chronic course; a
and Anxiety, 30(11), 1099–1106. https://doi.org/10.1002/da.22149 naturalistic cohort study. BMC Psychiatry, 19(1), Article 228. https://
Kazak, A. E. (2018). Editorial: Journal article reporting standards. American doi.org/10.1186/s12888-019-2214-4
Psychologist, 73(1), 1–2. https://doi.org/10.1037/amp0000263 van de Schoot, R., Sijbrandij, M., Winter, S. D., Depaoli, S., & Vermunt,
Lambert, M. J., Burlingame, G. M., Umphress, V., Hansen, N. B., J. K. (2017). The GRoLTS-checklist: Guidelines for reporting on latent
Vermeersch, D. A., Clouse, G. C., & Yanchar, S. C. (1996). The reliability trajectory studies. Structural Equation Modeling, 24(3), 451–467. https://
and validity of the Outcome Questionnaire. Clinical Psychology & doi.org/10.1080/10705511.2016.1247646
Psychotherapy, 3(4), 249–258. https://doi.org/10.1002/(SICI)1099- Xiao, H., Carney, D. M., Youn, S. J., Janis, R. A., Castonguay, L. G., Hayes,
0879(199612)3:4<249::AID-CPP106>3.0.CO;2-S J. A., & Locke, B. D. (2017). Are we in crisis? National mental health and
Lin, T., Anderson, T., Austin, M., & Mischkowski, D. (2023). Early treatment trends in college counseling centers. Psychological Services,
trajectories of symptom change and working alliance as predictors of 14(4), 407–415. https://doi.org/10.1037/ser0000130
treatment outcome. Psychotherapy Research, 33(2), 185–197. https:// Youn, S. J., Castonguay, L. G., Xiao, H., Janis, R., McAleavey, A. A.,
doi.org/10.1080/10503307.2022.2080028 Lockard, A. J., Locke, B. D., & Hayes, J. A. (2015). The counseling center
Lin, T., & Farber, B. A. (2021). Trajectories of depression in psychotherapy: assessment of psychological symptoms (CCAPS): Merging clinical
How client characteristics predict clinical improvement. Journal of Clinical practice, training, and research. Psychotherapy, 52(4), 432–441. https://
Psychology, 77(6), 1354–1370. https://doi.org/10.1002/jclp.23119 doi.org/10.1037/pst0000029
Lindebø Knutsen, M., Sachser, C., Holt, T., Goldbeck, L., & Jensen, T. K. Zilcha-Mano, S., & Errázuriz, P. (2017). Early development of mechanisms
(2020). Trajectories and possible predictors of treatment outcome for of change as a predictor of subsequent change and treatment outcome: The
youth receiving trauma-focused cognitive behavioral therapy. case of working alliance. Journal of Consulting and Clinical Psychology,
Psychological Trauma: Theory, Research, Practice, and Policy, 12(4), 85(5), 508–520. https://doi.org/10.1037/ccp0000192
336–346. https://doi.org/10.1037/tra0000482
Littleton, H., Horsley, S., John, S., & Nelson, D. V. (2007). Trauma coping
strategies and psychological distress: A meta-analysis. Journal of
Traumatic Stress, 20(6), 977–988. https://doi.org/10.1002/jts.20276 Received February 28, 2023
Locke, B. D., Buzolitz, J. S., Lei, P. W., Boswell, J. F., McAleavey, A. A., Revision received August 8, 2023
Sevig, T. D., Dowis, J. D., & Hayes, J. A. (2011). Development of the Accepted August 27, 2023 ▪

View publication stats