Psychotherapists are becoming busier every day and are constantly trying to manage the many different responsibilities they have with the increase in demand for psychological services. Responsibilities can include assessment, treatment planning, clinical preparation, individual therapy, group therapy, case management, case consultation, documentation, coordinating care, supervision, training, and outreach. One setting that has been heavily impacted by this increase in demand is university counseling centers. In fact, during the 2016 school year the average level of counseling center utilization grew by 30% while the average institutional enrollment only grew by 5% (Center for Collegiate Mental Health, 2017). Further, they reported 150,483 individual college students sought mental health treatment during the 2016 school year. However, they reported only 3,419 clinicians from 139 colleges were available to meet this need. On average, that’s a caseload of over 44 students per clinician throughout the school year.
How are we currently meeting this need?
While students are seeking treatment more frequently, clinicians at counseling centers are struggling to find ways to manage this increase in demand. There are several approaches to this problem. For example, some clinicians increase the time clients wait between sessions, seeing clients every two weeks instead of weekly, which enables them to be able to see twice the number of students. However, often, this adjustment doesn’t adequately meet the demands and clinicians find themselves having to schedule clients three to four weeks out to meet the demands of their growing caseloads. This approach is problematic given that research has shown that time between sessions matters, and clients who participate in weekly therapy have a faster recovery, while those who participate in therapy every other week make slower progress (Erekson, Lambert, & Eggett, 2015).
Another approach is to create waitlists. This allows clinicians to avoid taking on more clients than they can reasonably care for while maintaining a shorter interval that clients wait between sessions. However, this means that many clients must wait for extended periods, sometimes months, before talking to a therapist at all. This of course can also be problematic, especially with clients who have high risk concerns like suicidality.
Another method is to decrease the number of intakes being offered. When intakes are decreased, the impact can be the same as a waitlist, as intakes are not available for clients for up to months at a time, meaning that they again end up waiting before being seen. Other counseling centers will eliminate intakes altogether, informing prospective clients that they will not be conducting additional intakes until the following semester. At that point, the general practice is to provide students with resources they can pursue for services elsewhere. Finally, some counseling centers create more therapy groups to address the increasing demand for services. In a group, a single clinician can meet the needs of several clients in 90 minutes as opposed to meeting the needs of a single client in 50 minutes.
Individual Versus Group Therapy
There is an ongoing debate about the effectiveness of individual versus group therapy. Unfortunately, findings in the literature thus far haven’t been entirely clear. Some individual studies have reported greater effectiveness is shown during individual therapy (Barber & Gilbertson, 1996; Brown & Lewinsohn, 1984; Craigie & Nathan, 2009; Dowrick et al., 2000; Juarrieta et al., 2008; Kurzweil, 2012; Mohr, Boudewyn, Goodkin, Bostrom, & Epstein, 2001; Sharp, Power, & Swanson, 2004; Shechtman & Ben-David, 1999; Trowell et al., 2002), while some studies suggest group is more effective (Belloch et al., 2011; Cabedo et al., 2010; Kalavainen, Korppi, & Nuutinen, 2007; Marchand, Roberge, Primiano, & Germain, 2009; Scott & Stradling, 1990). Other studies report both are equally effective (Baines, Joseph, & Jindal, 2004; Barron & Leary, 1955; Cole, Boyer, Spanbauer, Sprague, & Bingham, 2013; Frisch, Hofecker-Fallahpour, Stieglitz, & Riecher-Rossler, 2013; Liber et al., 2008; Marques & Formigoni, 2001; Muris, Mayer, Bartelds, Tierney, & Bogie, 2001; Renjilian et al., 2001; Rose, O’Brien, & Rose, 2009; Rossello, Bernal, & Rivera-Medina, 2008; Sobell, Sobell, & Agrawal, 2009).
Additionally, there have been several meta-analytic studies comparing individual psychotherapy with group psychotherapy and the findings are suspect at best (Aderka, 2009; Cuijpers, van Straten, & Warmerdam, 2008; Driessen et al., 2010; Eddy, Dutra, Bradley, & Westen, 2004; Engels & Vermey, 1997; Federoff & Taylor, 2001; Gould, Buckminster, Pollak, Otto, & Yap, 1997; Hans & Hiller, 2013a; Hans & Hiller, 2013b; McDermut, Miller, & Brown, 2001; Powers, Sigmarsson, & Emmelkamp, 2008; Reynolds, Wilson, Austin, & Hooper, 2012; Sockol, Epperson, & Barber, 2011; Weisz, McCarty, & Valeri, 2006).
Within these studies, there are several methodological flaws to take into consideration. One major flaw was noted when a meta-analytic study compared individual psychotherapy outcomes in one study with group psychotherapy outcomes in another. This is problematic for several reasons. For instance, when comparing outcomes across studies, therapists are often compared who work at fundamentally different settings, thus negating the potential impact of the setting. There are also different diagnoses and even different treatments being given within the individual and group conditions. Additionally, researchers may not be clear as to which outcomes are primary or secondary. For example, studies looking at obsessive compulsive disorder may not identify outcomes specifically related to their key hypotheses, such as labeling the Yale-Brown Obsessive Compulsive Scale (Y-BOCS; Goodman et al., 1989) as a primary outcome and other outcomes, such as depression, as secondary. As such, comparing outcomes across studies leads to non-equivalence within the conditions, making it nearly impossible to accurately report findings.
Even in meta-analyses where this flaw is not present, there are similar non-equivalent methodological issues. Specifically, individual psychotherapy and group psychotherapy comparisons within the same study may be nonequivalent in other ways, such as different total number of sessions, different time dosing, different therapists throughout the treatment protocol, and differing diagnoses. Thus, even studies that compare individual and group therapy directly can make it difficult for researchers to accurately compare the two conditions, given the irregularities between conditions.
Lastly, a limitation of meta-analyses is the interpretation of the average effect size in making conclusions without considering the amount of variability of studies used in creating this average. If the aggregated studies produce small effect size variability, or low heterogeneity, then we can have greater confidence that the average effect size is truly representative. However, if there is a wide discrepancy between studies, or high heterogeneity, then less confidence can be placed in the average effect size.
Up until now, both equivalent and nonequivalent studies have been aggregated together to highlight the current literature findings on individual versus group psychotherapy effectiveness. Last year, we conducted a meta-analytic review that addressed these crucial issues (Burlingame et al., 2016). We screened over 500 articles and found that the most frequent reason for exclusion was the absence of direct format comparison and the publication being a review, meta-analysis, or unpublished study. Of the 500 articles, 68 articles included both individual and group treatments in the same publication, and were included in the meta-analysis. First, we hypothesized that individual and group format for identical treatments offered in the same study would demonstrate treatment equivalence. Second, we hypothesized that individual and group format for nonidentical treatments offered in the same study would show treatment differences. In other words, we statistically tested outcome differences between individual and group formats when treatment conditions were both equivalent and nonequivalent.
We conducted three separate comparisons to analyze these outcome differences and test our hypotheses. First, we performed an omnibus test of all treatments (both identical and nonidentical) and all outcomes (primary and secondary). Thus, meta-analysis of 67 studies that directly compared individual and group formats found equivalence with moderate heterogeneity when primary and secondary outcomes were both combined and separately analyzed. Therefore, when we examined both identical and nonidentical studies on all outcome measures we found no differences between the outcomes of group and individual formats. Both formats showed equivalent pre-post outcomes with high heterogeneity, which was partially explained by diagnosis. Patients presenting with depression, substance use, anxiety, or eating disorders showed the highest level of improvement.
Second, to test our first hypothesis we conducted a meta-analysis of 46 studies that met criteria for format equivalence and identical treatment comparison. When only primary outcomes were analyzed, format equivalence was found with no significant heterogeneity. We view this test of primary outcomes as the best test of our first hypothesis since it compares identical formats on the symptoms targeted for treatment. This finding provides strong support for the notion that no differences in outcome exist when identical individual and group treatments are compared across homogenous and diverse patient populations.
Lastly, to test our second hypothesis we conducted a meta-analysis of 21 studies analyzing format differences for nonidentical studies. Results showed group and individual format equivalence with significant heterogeneity. In contrast to identical treatments, meta-analysis of primary outcomes did not reduce the heterogeneity. However, allegiance was one moderator that explained this variability in heterogeneity. In other words, studies in which researchers expressed allegiance to either group or individual therapy produced effect sizes that favored the preferred format. Researcher allegiance is a common moderator in psychotherapy meta-analyses (Lambert, 2013), therefore, we expected it would appear in studies testing format differences. As such, when interpreting nonidentical treatment format studies, it is important to be cautious, particularly when the authors reveal an allegiance to a format. Stated differently, since the effects of allegiance can only be revealed through meta-analysis of multiple studies, clinicians and researchers should exercise caution in applying the findings of any nonidentical treatment study, especially when format allegiance is present.
Clinician and Patient Implications
This is the largest format comparison meta-analysis that we know of, and the overlap between our findings and past meta-analyses increases our confidence in the conclusion that when identical treatments, patients, and doses are compared, individual and group formats produce statistically indistinguishable outcomes. However, groups come with a unique set of additional responsibilities when compared to individual therapy. For example, additional tasks include finding enough clients to begin the group, pre-group screening sessions, progress notes for each group member per session, progress notes for the group as a whole, and managing attrition. Although there are logistical challenges, encouraging clinicians to run groups may help fulfill the aforementioned issue of demand clinicians are facing, especially given our conclusion that groups and individual therapy produce statistically indistinguishable outcomes.