The electronic databases Ovid, Embase, PubMed, Psych INFO and Cochrane were used to identify all published studies of “psychodynamic” or “psychoanalytic” psychotherapies. These include a family of psychotherapeutic approaches focused on the understanding of personal and relational patterns, the potential link between the person’s past and present experiences, the expression of feelings, and the exploration of defense patterns (Gabbard, 2010; Leichsenring et al., 2023).

The electronic search initially includes papers indexed by the aforementioned web-based databases as of January 2020 and was updated in July 2023. In order to obtaine a comprehensive list, we also manually screened relevant textbooks, systematic reviews, meta-analyses and reference checklists of studies included and consulted experts in the field. Supplementary material contains the full search strings. The search strategy included terms referring to young adult population (e.g., ‘young’, ‘adult’ and ‘youth’), mental health problems (e.g., psychiat* or mental* or psychol* or diagnosis) and psychoanalytic treatments.

We adopted the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (PRISMA; Page et al., 2021). Because of the lack of research on outcomes of psychotherapy for young adults generally, we included RCTs, quasi-experimental studies, and naturalistic evaluations whose quantitative measurement of therapy process or outcomes were reported at ‘baseline’ (T1) and at a later time-point (T2).

We did not adopt a sample size cut-off in the selection of the studies in order to capture a broader range of quality and reflect the heterogeneity that occurs in clinical practice (Turner et al., 2013). Inclusion criteria were: studies of psychodynamic or psychoanalytic psychotherapies with published articles in English and in peer reviewed journals, subjects were between 18 and 27 years-old at the start of therapy and receiving treatment for mental health issues, to capture the extended psychological challenges of adolescence and young-adulthood (Giedd et al., 1999). We excluded studies of young populations suffering from organic mental disorders, and where no quantitative measurement of therapy outcomes were provided.

We defined intervention as therapy that was psychoanalytic or psychodynamic (Gabbard, 2010). We included short-term psychodynamic therapy, typically between 12 and 24 weekly sessions as well as longer term psychoanalytic treatments, up to 8 years duration. Comparators were any form of psychotherapy or pharmacotherapy that has proven its efficacy against the corresponding mental disorders on the basis of published standards and guidelines (Chambless and Hollon, 1998; Nathan and Gorman, 2015). If the same study had published more than one article we included the publication of the main outcome instead of the follow-up, or secondary analyses.

Given the heterogeneity of design, populations and measures of the studies included, we synthesized the findings in a narrative form, and grouped the outcomes measures into five main thematic categories across the studies included (clinical symptoms, recovery/relapse, psychosocial functioning, personality structure and interpersonal relationships).

We applied the Randomized Controlled Trial of Psychotherapy Quality Rating Scale (RCT-PQRS; Kocsis et al., 2010), to rate the methodological quality of RCTs of psychotherapy. The RCT-PQRS was developed by an expert committee with different allegiances (e.g., CBT, psychodynamic therapy, pharmacology) and based on pre-existing quality measures of RCTs (Thoma et al., 2012). The scale was designed to evaluate the quality of psychotherapy trials and has shown good psychometric properties in terms of internal consistency and external validity (Jadad et al., 1996; Moher et al., 2001; Moncrieff et al., 2001). It includes 24 items each assessing a specific element of the study design and methodology, with scores ranging between 0 and 2, yielding a maximum score of 48. For the purpose of the study, we used the total score of the 24 items as the primary quality measure. A quality score of 24 or above is considered to represent a cut-off for a “reasonably well-done study” (Gerber et al., 2011).

For naturalistic studies, we used a quality assessment tool designed by the National Institute for Health Research (National Institutes of Health, 2014), which allows a critical appraisal and assessment of the internal validity of each study, whether the study findings can be attributed to the intervention. The tool includes a total of 14 questions to assess selection, information, measurement, and confounding bias (for example differences patients’ characteristics at the start of the intervention). Each study was rated as 1 (Yes) or 0 (No) on each question, based on whether the item criteria was met or not. A total score was calculated summing the single questions and each study was classified as ‘poor’ (total score below 5), ‘fair’ (between 6 and 10) or ‘good’ quality (total score above 10). A low score of the study quality corresponds to greater risk of bias and, viceversa, higher quality rating indicates a lower risk of bias.

We used the following inclusion criteria: RCTs of psychoanalytic therapy for young adults, comparing psychoanalytic therapy to another treatment with established efficacy or control conditions, and using quantitative outcome measures. The primary outcomes were “target symptoms,” which included various measures for the mental disorders being studied, such as anxiety symptoms in anxiety disorders or obsessive-compulsive behavior measures in obsessive-compulsive disorder. The secondary outcomes included psychosocial functioning (such as personality, social, and occupational functions).

We conducted two separate meta-analyses: (i) first, we compared the effect of psychodynamic psychotherapy to other treatments on outcomes, (ii) second, we compared the effect of psychodynamic therapy to control conditions (e.g., waitlist, treatment as usual) for chosen primary and secondary outcomes.

Data extracted were: study authors and year of publication; sample characteristics (size, psychiatric diagnosis, diagnostic measures); details of interventions (including length of treatment) and control group condition; primary and secondary outcome measures; follow-up duration, and effect sizes. In case of missing data, we contacted the corresponding authors.

Statistical analyses were carried out using Stata (Version 15; StataCorp, 2017). Due to between-study heterogeneity, we adopted a random effects model, which is recommended for meta-analysis of psychological therapies (Higgins and Green, 2008).

For continuous outcomes, pooled standardized mean difference (SMD) was calculated with 95% confidence intervals. SMD was utilized in order to pool together means across studies utilizing different outcome measures and was calculated by dividing the mean difference in outcome between participants allocated to psychodynamic psychotherapy and those allocated to the comparison intervention or control group by the pooled within-groups standard deviation. If the outcomes were expressed as an event proportion, they were converted to odds ratios and then subsequently converted to Hedges’ g (Hedges, 1981).

Cohen’s criteria for the interpretation of effects was used: 0.2 suggests a small effect, 0.5 a medium effect, and 0.8 for a large effect (Higgins and Green, 2008). The results are displayed using a forest plot.

Heterogeneity between studies was assessed with the Cochran’s Q test (DerSimonian and Laird, 1986). A statistically significant Q value indicates true heterogeneity in effect sizes beyond random error; the I² statistic was calculated to express the proportion of variation among studies that was due to heterogeneity (Higgins et al., 2003). Heterogeneity was categorized as low (0–40%), moderate (30–60%), substantial (50–90%), or considerable (75–100%; Deeks et al., 2019).

To further explore heterogeneity, we carried out meta-regression analyses to test effects of potential moderator variables such as: year of publication, quality score (total score of the RCT-PQRS), age, sex, outcome measure (self-report vs. interview), recruitment methodology (community compared with clinical populations compared with mixed samples), intent-to-treat compared with completer analyses, diagnosis, patient-per therapist ratio (to investigate the presence of bias from therapist effects), and average sample size per group to assess the small study bias (Nüesch et al., 2010). All moderators were otherwise entered separately to the meta-regression analysis due to the small number of included studies (Borenstein et al., 2010).

We conducted sensitivity analyses to examine the effect of outliers, defined as those studies displaying a 95% CI that did not overlap with the 95% CI of the pooled effect size. Publication bias was assessed by visual inspections of funnel plots. In addition, Egger’s test of publication bias was used to investigate whether there was a tendency for selective publication of studies (Egger et al., 1997).

A significance level of p = 0.05 was used for the random-effects model, homogeneity, publication bias, and meta-regression analyses.

The PRISMA flowchart showing the inclusion process is illustrated in Figure 1. A total of 23,759 related documents were initially retrieved. After reading the titles and abstracts, 469 studies remained to be read in full. According to the inclusion and exclusion criteria, the results were filtered in order to select the eligible studies. A total of 22 published papers from the database searches fulfilled the selection criteria. Of those, 14 studies were RCTs, and eight studies were naturalistic studies.

Data on a total of 2,649 young adults were included in the studies (2,063 in the treatment and 586 in the control conditions). A summary of the main characteristics of the studies included can be found in Table 1. Ten studies compared psychodynamic therapy with another psychotherapy treatment, including CBT, Counseling, Cognitive-Analytic Therapy (CAT), Family Therapy (FT), Dialectical behavior therapy (DBT). One study compared psychodynamic therapy with pharmacological intervention.

The review identified 15 studies evaluating psychodynamic treatment for participants with an emotional disorder, including depression (n = 8) and anxiety (n = 7). A total of six studies focused on individuals with eating disorders and a further six on individuals with personality disorders. One study included participants with psychosis.

With the exception of one study on group psychoanalytic psychotherapy, all studies included individual psychodynamic therapy. The majority of studies (n = 14) were conducted in European countries, with the remainder being conducted in Asia (n = 6), Australia (n = 1) and Canada (n = 1).

There was some consistency in the therapy aims. About half of the studies (n = 13) focused on symptom reduction as the primary target. Six studies included assessment of psychosocial functioning. The interventions ranged from once or twice a week Psychoanalytic Psychotherapy to Intensive short-term dynamic psychotherapy, Self- psychological treatment, Supportive Psychoanalytic Psychotherapy, Focal psychoanalytic psychotherapy, manualized treatments like Transference-focused psychotherapy, and Dynamic Interpersonal Therapy, Supportive-Expressive Therapy, Conversational model (psychoanalytic treatment for BPD), Brief Empathic Psychotherapy, Integrative therapy. Only one study included five times a week Psychoanalysis (Gerber et al., 2004). The length of the treatments varied in terms of duration spanning from a minimum of four sessions up to 8 years. A total of 15 studies included a follow-up period, ranging between 2 months to 5 years after therapy termination.

We grouped the outcome measures into five categories across the included studies, including clinical symptoms, recovery/relapse, psychosocial functioning, personality structure and interpersonal relationships. A total of 9 studies (both RCT and naturalistic), reported the Global Severity Index scale of the Symptom Checklist 90 Revised (SCL-90-R; Derogatis and Unger, 2010) as the main clinical outcome measure for psychiatric symptoms, and showed a significant decrease in psychopathological symptoms from baseline to end of psychodynamic treatment (Nemirovski Edlund and Carlberg, 2016; Falkenström, 2010; Riva Crugnola et al., 2020; Stefini et al., 2017). A naturalistic study, the Young Adult Psychotherapy Project (Philips et al., 2006), showed that improvements on the overall psychological functioning measured using the SCL-90 and the Global Assessment of Functioning Scale (GAF; American Psychiatric Association, 2013) were maintained also at follow-up 1.5 and 3 years after termination (Philips et al., 2006; Lindgren et al., 2010; Werbart et al., 2012, 2017). Similarly, a significant decrease in psychotic symptoms was maintained for 5 years after psychodynamic treatment in young people experiencing a first psychotic episode (Harder et al., 2014).

Several studies focused on emotional disorders and affective symptoms, such as depression (n = 6), anxiety (n = 4) and internalizing/externalizing problems (n = 1). Both self-report (using the Beck Depression Inventory, BDI; Beck et al., 1996) and interviewer-rated symptoms of depression (i.e., using the Structured Clinical Interview for DSM; American Psychiatric Association, 2013) were significantly lower after psychodynamic psychotherapy and at follow-up compared with the wait-list control group (Ajilchi et al., 2016; Walton et al., 2020). However, a study on post-partum depression in young mothers showed that the benefit of treatment on depression was no longer significant by 9 months post-partum (Cooper et al., 2003).

One study that evaluated the outcomes of psychoanalysis for young adults showed that improvement in levels of depression and anxiety and number of Axis I diagnosis scales achieved by the end of treatment termination was maintained 18 months after termination (Gerber et al., 2004). Young people who received community-based psychodynamic psychotherapy reported a decrease in internalizing and externalizing problems after 1 year in treatment (Baruch and Fearon, 2002).

Studies of short-term psychodynamic treatment also showed significant differences in the generalized anxiety and depression symptom scores between pre-treatment, post-treatment, and follow-up stages in the experimental group compared to the control group (Mowlaie et al., 2018; Orvati Aziz et al., 2020), with a significant reduction of fear and avoidance symptom scores in the young adults with social anxiety and social phobia in psychodynamic treatment compared to those in the control group (Moghadam et al., 2015; Rahmani et al., 2020).

A few studies focused on secondary outcomes, including dropouts from treatment (N = 1), suicide attempts (N = 2), self-harming (N = 2), and eating disorder psychopathology, including recovery (N = 1), inpatient admission/relapse (N = 1), vomiting or binging (N = 3) and weight gain (N = 1). A significant decrease over time in suicidal and/or self-harming episodes and psychiatric inpatient admissions was observed in two RCTs focusing on psychodynamic psychotherapy for young people with a diagnosis of borderline personality disorder (Doering et al., 2010; Walton et al., 2020).

In terms of eating disorders, there was a significant weight gain in the Psychodynamic Psychotherapy group after 1 year of treatment compared to the routine treatment group (Dare et al., 2001) and higher remission rates in young adults in psychodynamic treatment group compared to CBT (Stefini et al., 2017).

When compared to CBT, psychoanalytic psychotherapy was shown to be effective in reducing purging and binge eating (Garner et al., 1993; Stefini et al., 2017) and remission from bulimia occurred for 15% of participants in the psychoanalytic psychotherapy group at 2 years follow-up (Katznelson et al., 2020; Poulsen et al., 2014). However, the improvements in eating disorder symptoms and general psychopathology occurred more rapidly in the patients receiving the CBT intervention (Garner et al., 1993; Katznelson et al., 2020; Poulsen et al., 2014).

A minority of studies assessed psychosocial functioning (n = 6), including a global measure of functioning and social behavior. When compared to other talking therapies, Transference Focused Psychotherapy showed efficacy in the domain of social functioning in people with borderline personality (Doering et al., 2010). The treatment effect of psychodynamic psychotherapy on social functioning increased over 2 years after treatment but became non-significant at 5-year follow-up (Harder et al., 2014; Rosenbaum et al., 2012).

One study also assessed executive functioning, with the psychodynamic group significantly improving on executive functioning at post-treatment compared to young adults who were in the wait-list group (Ajilchi et al., 2016).

Another category was represented by outcomes related to the assessment of personality structure (Personality Disorder diagnosis n = 4, dissociation n = 1, sense of self n = 1, mindfulness n = 1, emotion regulation n = 1, defense mechanisms n = 1, affective meaning states n = 1). BPD severity, personality organization, dissociation or sense of self showed significant improvement over time at the end of psychodynamic treatment (Doering et al., 2010; Walton et al., 2020). Short term psychodynamic therapy had an effect on overall defensive functioning, but no significant change for overall coping functioning (Kramer et al., 2010, 2014).

A handful of studies focused on interpersonal aspects, including constructs such as therapeutic alliance (n = 3), attachment style and reflective function (n = 2), interpersonal problems (n = 2), experience of the therapy (n = 1), self and object representations (n = 1). The Young Adult Psychotherapy Project found positive changes with regard to reflective function, interpersonal problems and self and object representations (Katznelson et al., 2020; Lindgren et al., 2010; Philips et al., 2006; Werbart et al., 2017). However, another study found no moderation effect of attachment styles on the effectiveness of the psychodynamic intervention (Riva Crugnola et al., 2020).

Overall, the majority of the reviewed studies showed that psychoanalytic psychotherapy had a significant positive effect on young adults involved in the treatment, including symptoms reduction, remission, psychosocial and interpersonal functioning, and improvement on personality outcomes.

Around 75% of mental health disorders have their onset in young adulthood (Auerbach et al., 2018; Kessler et al., 2005). Yet, only a limited amount of evidence as per effective treatment options for these individuals is available. Our meta-analytic study of 14 RCTs has shown that psychoanalytic psychotherapy might be an effective treatment option alongside other psychotherapeutic approaches. It certainly is more effective than not receiving treatment at all (Hedges’g − 1.24 [95% CI: −1.97;–0.51], p < 0.001).

Many young adults who need support from mental health services are ‘lost’ to the system when they transition into adulthood (Mclaren et al., 2013; McNicholas et al., 2015; Singh et al., 2010). Those affected are often the most vulnerable and disadvantaged; getting lost during this transition increases the likelihood to be unemployed and out of education or training; it can also impact negatively on their physical health (Trotta et al., 2019). By showing that psychoanalytic therapies are effective in the treatment of the young adults, the findings from our review highlight the need to offer a range of treatment options, tailored to the young person’s needs.

A distinctive feature of psychodynamic interventions compared to other talking therapies is the focus on interpersonal relationships, the identification of recurring patterns in people’s functioning and in their feelings (Blagys and Hilsenroth, 2000; Luborsky et al., 1990). In understanding attitudes and feelings in the present, psychoanalytic approaches emphasize the importance of a developmental perspective as well as the role of insight into the past. There is a growing consensus on the importance of focusing on a young person’s internal world, on the unconscious factors that orient the perception of external reality and how these, in turn, might lead to the development and maintenance of pathological defenses. As such, psychoanalytic therapy employs a different mechanism of change than other therapies and therefore might be an important treatment options for some young people as this approach addresses specific aspects that other therapeutic modalities might not focus on (Blagys and Hilsenroth, 2000).

Taken together, these findings offer some initial evidence for the effectiveness of psychoanalytic treatment for young adults and highlight the need to conduct research to understand the efficacy of different treatment options for young adults and whether the benefits persist also in later follow-up. Therefore, this further highlights the importance of more robust and high quality studies focusing on young adulthood that aim to prevent and treat the progression of mental health and interpersonal difficulties later in life.