Psychodynamic Psychotherapy with Adolescents: Endings and their Association with Sociodemographic Characteristics and Symptoms

ABSTRACT

Child and adolescent psychopathology prevalence is a topic of discussion worldwide, affecting 14% of this population globally. Psychodynamic psychotherapy is one approach for the treatment of those conditions. However, besides qualified provision of psychodynamic psychotherapy being hard to reach, dropout is also a relevant issue, occurring in up to 72% of cases. The present study aimed to assess the predictive role of sociodemographic, family, and symptom variables in relation to treatment endings (i.e. non-adherence, dropout, completion) for 747 adolescents who received psychodynamic psychotherapy, through multinomial logistic regression and binary logistic regression. Family income, source of referral, motivation, and baseline clinical complaint were significantly different among ending types. Dropout cases presented higher rule-breaking behavior and externalizing problems scores when compared to treatment completers. Patients with higher family income, motivation for treatment, and presenting internalizing symptoms were more likely to be treatment completers.

Introduction

Adolescence is a developmental stage characterized by significant life changes and complex transformations. Concerning this population, current research has indicated that mental disorders affect a significant number of young people, with a global average prevalence rate of 15.8%. The prevalence rate tends to increase proportionally with age, with the average prevalence among preschoolers being 10.2% and among adolescents 16.5%. In Brazil, studies reported prevalence rates of 7 to 12.7% (Polanczyk et al., 2015; Thiengo et al., 2014). According to the “World Mental Health Report” from the World Health Organization, published in 2022, 14% of adolescents in the world experience mental health problems. Facing the high prevalence of psychopathologies during adolescence, it is crucial to offer treatments that cater to this population’s specific needs (World Health Organization, 2022).

Recent reviews have indicated that psychodynamic psychotherapy is effective in reducing internalizing and externalizing symptoms, both in its short- and long-term formats (Midgley et al., 2021). Psychodynamic or psychoanalytic psychotherapy (in the present study treated as synonyms, as in Gabbard, 2007) are treatments based on concepts and methods drawn from psychoanalytic theory. They encompass a range of treatments that focus on (a) the identification of relational patterns of an individual with oneself and others (including the therapist); (b) the identification of defensive patterns; (c) emotional expression; and (d) the discussion of past experiences that impact one’s current life (Leichsenring et al., 2023). The goals of psychodynamic psychotherapy include – but are not limited to – symptom remission.

Although different studies have provided important evidence that psychodynamic treatments are effective for young people, treatment dropout is still a common issue in this population. Treatment dropout is characterized by a client unilaterally ending their treatment prematurely, before fully addressing the issues that initially prompted them to seek therapy and before completing the prescribed treatment (Swift & Greenberg, 2015). It has been examined in different countries and contexts (Block & Greeno, 2011; de Haan et al., 2013; Gastaud & Nunes, 2009; O’Keeffe et al., 2018), and has been found to occur in 16 to 72% of the cases (de Haan et al., 2013; O’Keeffe et al., 2018).

Therapy ending can be understood as having three different types: non-adherence, dropout, or completion (Gastaud & Nunes, 2010). Non-adherence encompasses a psychotherapy process that ends before the dyad can establish clear treatment goals, during the initial sessions. Dropout refers to cases in which the young person unilaterally ends the treatment (Des Essarts et al., 2022; Lhullier et al., 2006). Finally, treatment completion signals instances when the treatment ends with a mutual agreement between the adolescent and the therapist, regardless of whether they have met the treatment goals or not (Des Essarts et al., 2022; Lhullier et al., 2006). In the psychodynamic psychotherapy with children and young people, this dynamic is even more complex, since parents are active actors in the treatment process. Parents and caregivers then play an essential role in determining the type of treatment ending, as their understanding of shared objectives is added to the understanding held by their child and the therapist (Block & Greeno, 2011; Des Essarts et al., 2022).

Despite being scarcer than the studies addressing treatments with adults, some authors have analyzed variables associated with psychodynamic psychotherapy dropout with adolescents (Baruch et al., 2009; Benetti et al., 2017; Block & Greeno, 2011; de Haan et al., 2013; Delgado & Strawn, 2012; O’Keeffe et al., 2018, 2019; Ollendick et al., 2008). Concerning demographic variables, there is no clear consensus in the literature. In some studies, gender and age have been found as predictors for dropout, with being older and male as risk factors for dropping out or not adhering to psychotherapy (de Haan et al., 2013; Midgley & Navridi, 2007; O’Keeffe et al., 2018; Pelkonen et al., 2000). However, Baruch et al. (2009) study points to the opposite scenario, with their findings indicating that older adolescents are more likely to continue attending therapy. By analyzing adolescents’ socioeconomic status, Baruch et al. (2009) also identified that adolescents experiencing homelessness were more likely to drop out of therapy.

Family variables have also been examined in terms of their association with treatment endings in children and adolescents. Among those variables, having a young mother and living in a single-parent household without the presence of a father significantly predicted dropout. Moreover, “poor parenting,” characterized by caregivers displaying critical comments, negative emotions, and hostility toward their child, was also associated with psychotherapy dropout in a meta-analysis performed by de Haan et al. (2013). Nevertheless, in a study carried out by Midgley and Navridi (2007), the family context did not seem to be always associated with psychotherapy ending with children and adolescents.

Finally, adolescents’ symptom severity and psychiatric comorbidities also seem to be associated with treatment dropout. Young people who present delinquent, antisocial, and overall externalizing behavior, as well as meet diagnostic criteria for attention-deficit hyperactivity disorder (ADHD) may be more at risk of dropping out (Baruch et al., 2009; de Haan et al., 2013; Pelkonen et al., 2000). Conversely, internalizing problems such as depression and anxiety seem to be protective factors against dropout (Baruch et al., 2009), or are unrelated to therapy ending (O’Keeffe et al., 2018). However, when these internalizing problems are seen in high severity (de Haan et al., 2013) or alongside other conditions such as ADHD (Benetti et al., 2017), they can be risk factors for dropout. The combination of emotional and behavioral problems, such as anxiety and aggressiveness, depression and conduct problems, general mental health problems, and school complaints can also lead to higher odds of dropout in adolescents (Gauy & Rocha, 2014).

It is observed that there is no clear consensus about individual factors in adolescents that are associated with better treatment outcomes or endings (Des Essarts, 2020; O’Keeffe et al., 2018). Few studies have addressed adolescent variables in clinical samples, with specific psychiatric conditions, and focused specifically on this age group (Block & Greeno, 2011; de Haan et al., 2013). Being a stage of development marked by the return of complex childhood experiences and the need for maturing one’s sexual identity and personality, adolescence delineates a challenging time for the individual. Understanding that adolescents experience frequent changes in their intentions, problems, and desires, can potentially affect their need and willingness to engage in psychotherapeutic treatment (Lewkowicz & Brodacz, 2015; Macedo, 2012).

Being able to identify risk factors for treatment dropout among adolescents can provide insights for the promotion of young people’s engagement in psychotherapy, as well as to inform practitioners on how to prevent it in this population (O’Keeffe et al., 2018). Therefore, the present study aimed to analyze the association between sociodemographic, family, and clinical variables and treatment dropout/ending with adolescents in the context of a community clinic that offers psychodynamic psychotherapy. Following Gastaud and Nunes’ typology (2009), treatment endings were considered as (1) completion; (2) dropout; and (3) non-adherence.

Method

The present study employed an observational, descriptive, correlational, and explanatory design (Sampieri et al., 2014). We adopted an indirect observational approach, based on clinical archives (Shaughnessy et al., 2014).

Participants

Participants were 747 adolescents from 12 to 17 years old who sought treatment in a community clinic. Out of all the patients who spontaneously sought or were referred to this institution, these young people encompassed the ones who were referred to psychodynamic psychotherapy after a routine screening process.

Instruments

Sociodemographic form

This form is routinely filled in the first interview with the adolescent’s parents or caregivers and covers questions concerning the young person and their parents’ educational level, occupational information, and family income.

Screening form

This form is filled by the psychologist carrying out the screening process during the first interview with the adolescent’s parents or caregivers. It includes items about the source of referral, the adolescent’s and their family’s initial complaint that led them to seek treatment, the cases’ overall clinical history, domestic violence history, previous and current treatments, diagnostic hypothesis, and therapeutic recommendation/referral. It also includes an assessment, made by the therapist, of the patient’s motivation for psychotherapy. This assessment is based on a five-point Likert scale, with 1 being “not at all motivated,” 2 “little motivated,” 3 “medium,” 4 “very motivated” and 5 “completely motivated.”

Child behavior checklist for ages 6–18

(CBCL/6–18; Achenbach, 1991; Achenbach & Rescorla, 2001) – An internationally validated instrument that is vastly used to assess clinical issues in childhood and adolescence, ranging from 6 to 18 years old. It includes the following domains: anxiety/depression, withdrawal, somatic complaints, social problems, thought problems, attention problems, rule-breaking behavior, and aggressive behavior. The two latter encompass externalizing problems and the three former internalizing problems. The instrument also includes items that examine the individual’s involvement in extracurricular activities, and their social and school competence, as well as providing a global score named “total problems.” The CBCL was filled by the young person’s mother, father, or other primary caregiver after the first screening interview, then in even 6-month intervals, and at the treatment end, as a routine measure at this institution.

Family adaptability and cohesion evaluation scales

(FACES III; Olson, 1986) – This scale assesses the risk for psychopathology considering two family functioning dimensions: cohesion (emotional closeness among family members) and adaptability (degree of flexibility the family manifests for changing rules and roles). The FACES III was filled by the young person’s caregiver after the first screening interview.

Ending form

This form is filled by the young person’s therapist at the end of the treatment, containing information about the treatment duration and the type of ending (i.e., non-adherence, dropout, or completion). This instrument also addresses the reasons attributed to the ending.

Data collection procedures

In this study, we analyzed the dataset of a training institution for psychoanalytic psychotherapists. The training lasts for three years and covers seminars on psychoanalytic theory and supervised therapy practice. The institution maintains a community service clinic. The patients referred to the institution are attended by trainees or members of the clinical staff. All users and their caregivers signed a consent form approved by a Research Ethics Committee, authorizing the use of their registers in research. The dataset was revised and analyzed, and we excluded the cases in which there was missing data regarding variables of interest for the present study.

Data analysis

We carried out descriptive analyses (means, standard deviations, percentages) concerning the adolescents’ variables. We assessed the variables’ normality through the Kolmogorov-Smirnov test. To investigate the association between therapies’ endings and sociodemographic, family, and symptom variables, we carried out the Kruskal Wallis test, and for categorical variables (sex, age group, household income, referral source, stated initial complaint, and motivation) we used the chi-square test. To assess the predictive power of the variables in relation to the therapies’ endings, we used multinomial logistic regression. Finally, we merged the “non-adherence” and “dropout” endings and carried out a binary logistic regression using the forward conditional method. All analyses were carried out through SPSS v. 22.

Ethical procedures

The present study is part of a wider project that was examined and approved by the Research Ethics Committee based at the authors’ university (protocol number 4.468.268, CAAE 40,230,920.00000.5344). There was no interference in the patients’ treatments and the institution’s routine assessment.

Results

The participants’ sex and age had a somewhat even distribution, as shown in Table 1. On the other hand, 96.4% of families had a household income of less than seven minimum wages.

Table 1. Sociodemographic characteristics, source of referral, initial complaint, and motivation for psychotherapy.

Variables	Number of patients (%)
Sex
Female	396 (52.9%)
Male	351 (47.1%)
Age group
12 to 14 years	375 (50.2%)
15 to 17 years	372 (49.8%)
Household income
1 a 3 minimum wages	574 (76.9%)
4 a 7 minimum wages	146 (19.5%)
More than 7 minimum wages	25 (3.6%)
Referral source
Health professional/school	618 (82.8%)
Family	52 (7%)
Adolescent	39 (5%)
Other	38 (5%)
Motivation
Low or moderate	530 (71%)
High	214 (29%)
Stated initial complaint
Externalizing complaint	504 (67.5%)
Internalizing complaint	243 (32.5%)

The most common source of referral was the patients’ school or health professionals, while only 12% of the cases families or adolescents sought treatment spontaneously. Most young people reported externalizing complaints and low or moderate motivation for treatment. Table 1 synthesizes sociodemographic characteristics, source of referral, initial complaint, and motivation for psychotherapy concerning the adolescents included in this study.

Regarding the CBCL-assessed symptoms, 53% presented with clinical scores for anxiety/depression, 39.3% for social problems, and 38% for thought problems. When examining the symptoms at a grouped level, 61% had externalizing symptoms, 67,9% internalizing symptoms, and 81% total problems. The complete data concerning the adolescents’ clinical symptoms are presented in Table 2.

Table 2. Clinical symptoms according to the parent-informed CBCL.

CBCL symptoms	Number of patients (%)
Social competence	293 (39.3%)
Anxiety/depression	395 (53%)
Thought problems	283 (38%)
Withdrawal	210 (28%)
Somatic complaints	60 (8%)
Attention problems	180 (24%)
Rule breaking	224 (30%)
Aggressive behavior	220 (29%)
Grouped symptoms	Number of patients (%)
Externalizing problems	455 (61%)
Internalizing problems	507 (67.9%)
Total problems	605 (81%)

Concerning the ending types and reasons attributed to those, 92.8% of adolescents had a unilateral psychotherapy ending (i.e., non-adherence or dropout). Regarding the reason for ending therapy, 28.8% reported financial problems and 24.2% stated a lack of motivation, considering the family or the young person. The data on treatments’ endings and reasons attributed to those are presented in Table 3.

Table 3. Type and reason for treatment endings.

Type of ending	Number of patients (%)
Non-adherence	327 (43.3%)
Dropout	367 (49.5%)
Completion	53 (7.2%)
Reason for ending	Number of patients (%)
Financial issues	215 (28.8%)
No-show	189 (25.4%)
Family/adolescent low motivation	180 (24.2%)
Treatment considered no longer required	98 (13.3%)
Patient believed they had achieved their goals	62 (8.3%)

The three ending categories – non-adherence, dropout, and completion – were compared with sociodemographic characteristics, referral source, initial stated complaint, and the adolescents’ motivation for treatment at baseline through the chi-square test. The chi-square test of independence evidenced that there was an association between household income [X² (4) = 292.74; p < .001], referral source [X² (4) = 291.88; p < .001], motivation [X² (2) = 82.23; p < .001] and initial stated complaint [X² (2) = 64.44; p < .001] and the different ending categories. The adjusted residuals indicate that there is a difference between the observed frequencies in this sample and the expected values. The data is displayed in Table 4.

Table 4. Comparison between sociodemographic variables, referral source, initial stated complaint, and patients’ motivation between different treatment endings.

	Frequency (expected values)
	Non-adherence	AR	Dropout	AR	Completion	AR	Cramer’s V (p)	p	X^2
Household income							0.447 (.00**)	.00**	292.74
1 to 3	255 (252)	4.0	310 (279)	5.5	9 (43)	−11.4
4 to 7	67 (58)	1.6	44 (65)	−4.0	22 (9)	4.5
7 or higher	9 (42)	−4.3	22 (10)	−4.0	23 (2)	15.8
Referral source							0.448 (.00**)	.00**	291.88
School/Health professional	278 (273)	1.1	32 (37)	7.7	9 (48)	−16.7
Family/adolescent	32 (37)	−1.2	10 (43)	−7.6	45 (6)	16.8
Motivation							0.249 (.00**)	.00**	82.23
Low to moderate	203 (201)	0.3	260 (234)	4.5	11 (39	−8.7
High	76 (78)	−0.4	66 (91)	−4.4	43 (15)	8.8
Initial stated complaint							0.297 (.00**)	.00**	64.44
Internalizing	84 (106)	−3.6	123 (126)	−0.6	45 (18)	7.8
Externalizing	225 (202)	3.6	244 (240)	0.6	9 (35)	−7.8

Note. AR = Adjusted residuals; X² = Chi-square; **: Significance p < .001.

We carried out the Kruskal Wallis test to examine if the family variables contributed to differentiating the three types of therapy endings. None of the three variables – family cohesion, adaptability, and risk – differentiated the three ending groups.

We used the same test to compare the three ending groups and symptom variables. According to this analysis, some symptom domains had a significant difference between ending groups, especially the non-adherence cases concerning completers, as shown in Table 5.

Table 5. Comparison between endings and baseline CBCL symptoms.

	Non-adherence (Mean Rank)	M(SD)	Dropout (Mean Rank)	M(SD)	Completion (Mean rank)	M(SD)	X^2	p
CBCL 6/18: “Activities”	254,55	37,8(8,93)	247,6	38,19(20,6)	292,76	40,55(9,88)	3,303	,192
CBCL 6/18: “Social competence”	255,55	39,4(10,06)	251,7	40,61(27,05)	298,81	51,97(64,59)	3,535	,171
CBCL 6/18: “School competence”	241,45	42,9(36,02)	239,3	40,70(21,23)	257,50	40,7(10,69)	,515	,773
CBCL 6/18: “Total social competence”	214,57	34,1(8,52)	204,9	33,34(9,51)	267,67	38,2(9,36)	7,308	,026*
CBCL 6/18: “Anxiety/depression”	284,89	65,8(10,66)	270,9	64,78(9,55)	266,25	64,59(10,47)	1,166	,558
CBCL 6/18: “withdrawal”	272,66	69,9(54,61)	282,8	73,50(63,61)	256,54	65,16(10,10)	1,234	,540
CBCL 6/18: “Somatic complaints”	280,92	65,1(33,54)	279,0	64,40(30,69)	234,92	60,45(8,80)	3,122	,210
CBCL 6/18: “Social problems”	283,68	63,1(9,21)	275,8	67,32(56,70)	240,52	60,85(8,49)	2,634	,268
CBCL 6/18: “Thought problems”	279,38	65,3(40,66)	276,9	71,11(76,07)	257,74	61,28(9,57)	,663	,718
CBCL 6/18: “Attention problems”	281,23	70,2(60,89)	280,4	66,15(47,55)	224,08	60,26(10,50)	4,925	,085
CBCL 6/18: “Rule-breaking”	281,81	65,5(57,02)	283,7	69,45(77,43)	199,54	56,23(6,25)	10,659	,005*
CBCL 6/18: “Aggressive behavior”	281,27	65,6(40,75)	283,1	62,94(10,43)	206,25	58,3(9,25)	8,866	,012*
CBCL 6/18: “Internalizing problems”	278,55	69,4(47,34)	276,2	68,81(43,79)	246,21	81,80(111,94)	1,488	,475
CBCL 6/18: “Externalizing problems”	279,67	64,4(54,97)	283,5	61,14(10,29)	191,46	55(10,07)	12,376	,002*
CBCL 6/18: “Total problems”	277,20	64,4(9,60)	283,3	69,08(52,54)	206,52	61(9,197)	8,465	,015*

Note. M = mean; SD = standard deviation; X² = Chi-square *: Significance < 0,05.

The mean rank was significantly higher in the non-adherence and dropout endings for the domains “rule breaking,” “aggressive behavior,” and “total problems” in relation to treatment completers. In those symptoms, the difference between the ones who unilaterally ended therapy and the ones who consensually ended it was significant, being the highest symptom scores seen in the dropout cases. The “social competence” domain scores were significantly higher among completers in comparison to non-adherence and dropout.

To examine if any variable would predict a specific ending, we carried out a multinomial logistic regression using the three ending categories used in this study – non-adherence, dropout, and completion. According to this analysis, no variables (including sociodemographic, family, and symptom characteristics) presented significant predictive value to explain any of the treatment endings.

Non-adherence and dropout encompass unilateral endings (i.e., decided upon by the young people and/or families without agreeing with the therapist) and, in our study, presented similar results about reasons for ending treatment and similar CBCL mean scores. With that in mind, we also dichotomized this variable in terms of unilateral endings (i.e., non-adherence and dropout) versus agreed/bilateral endings (i.e., treatment completion).

After dichotomizing the variable, we carried out a binary logistic regression (through the forward-conditional method) to investigate what sociodemographic and symptom characteristics could adequately predict treatment endings. Following this analysis, we found a statistically significant model [X² (8) = 176.934; p < .001, R² Negelkerke = 0.756) that explained 75.6% of the ending variable, being 96.8% of the cases correctly classified (r = .986).

With this model, some variables were included in the final regression equation, while others were excluded by the test. Table 6 displays the results on what variables predict treatment completion.

Table 6. Variables that significantly predict treatment completion.

			95% Confidence interval
Variable	Sig.	Exp(B)	Inferior	Superior
Household income
From 4 to 7 minimum wages	.003*	8.427	2.77	34.182
More than 7 minimum wages	.000**	118.803	13.807	1022.273
Referral source
Family or adolescent	.000**	52.240	13.314	204.918
Motivation
High motivation	.004*	8.760	1.975	38.856
Initial stated complaint
Internalizing problems	.003*	9.251	2.124	40.290
Constant	.000	.000	–	–

Note. *: Significance < 0,05; **: Significance < 0,001.

Those with a household income from four to seven minimum wages were 8.4 times more likely to complete their treatment compared to the ones with lower household incomes. The patients whose family income were larger than seven minimum wages were 118.8 times more likely to complete their treatments. Self-referred patients or patients who sought therapy following their family advice were 52.24 times more likely to complete their treatment compared to the ones referred by other sources. Being highly motivated for treatment, according to therapists’ ratings, made patients 8.76 times more likely to end their treatment in relation to the ones considered less motivated at baseline. Finally, young people who stated internalizing problems as their initial complaint were 9.25 more likely to complete their treatment compared to the ones who reported externalizing problems.

Discussion

The present study aimed to analyze different variables (i.e., sociodemographic, family, and symptom characteristics) in relation to psychodynamic psychotherapy endings with adolescents in a community-based clinic. Household income, referral source, motivation, and initial stated complaint were found to be associated with different ending categories. However, family variables such as cohesion, adaptability, and risk did not differentiate therapy endings.

Firstly, concerning the characteristics of our sample, we identified high percentages of dropouts and non-adherence, with 92.8% of cases not achieving an ending that was consensual between patient and therapist. These rates are consistent with other recent studies carried out in developed countries, such as the ones published by de Haan et al. (2013), and O’Keeffe et al. (2018). We also identified that our participants were predominantly referred by schools or health professionals. Only a small percentage of adolescents sought psychotherapy spontaneously, or through their caregivers’ initiative. Furthermore, most of these young people lived low-income households.

After observing the dropout rates in our sample, we then assessed what patients were more likely to drop-out from or complete their psychodynamic treatments, and what variables predicted the type of treatment ending. According to our analyses, all variable domains (i.e., household income, referral source, motivation, and initial stated complaint) were associated with and predicted treatment endings.

When specifically examining household income, we found that this variable was not only associated with different treatment endings, but also acted as a predictor, establishing a causal relationship. Adolescents living in higher income households were significantly more likely to complete their treatments in comparison to their less affluent counterparts. As observed in Brazilian studies with adult patients (e.g., Pessota et al., 2020), and studies addressing youth in different countries (Baruch et al., 2009 in England, and de Haan et al., 2013 in the Netherlands) our findings suggest that the association between household income and dropout can also be seen among adolescents in Brazil. Overall, these results indicate that current financial stress can be a risk factor for therapy dropout (Xiao et al., 2017).

Financial instability should be recognized as a major factor contributing to interruptions in psychotherapy, which may not be related to any internal resistance from the patient. These are external and factual (and even political) circumstances impacting the treatment that therapists need to be mindful of in their practice. By having this awareness, clinicians could then avoid misinterpreting interruptions as solely stemming from intra-psychic factors.

Therapists should take into account the patient’s surroundings and overall life circumstances can exacerbate symptoms and make it difficult for the adolescent to prioritize and be available for treatment. Especially in Low- and Middle- Income Countries (LMIC), adolescents might be more prone to early school dropout, face the need to work to help support the family, exposure to violent environments and/or communities, or factors such as teenage pregnancy, drug abuse, and conflicts with the law (de Ribera et al., 2019; Kieselbach et al., 2022; Mussida et al., 2019; Ozeylem et al., 2021; Pradhan et al., 2015).

Concerning source of referral, we identified that adolescents who were referred to psychotherapy by their school or health professionals were less likely to complete their treatments in comparison to the ones who were self-referred or referred by their family. This finding diverges from Gastaud and Nunes (2009), in which children referred by physicians and other psychologists were more likely to remain in treatment. However, it is worth noting the developmental differences between children and adolescents. While children are usually more dependent on adults, adolescents are developing their sense of agency and might prefer to take a more active stance in relation to their treatment (Block & Greeno, 2011).

Our findings concerning the source of referral are aligned with the analyses on patient motivation: young people who were considered “highly” motivated according to their therapists were more likely to complete their treatments. This reiterates that being able to identify one’s own distress and seek help can be a protective factor to psychotherapy dropout. This finding is consistent with a previous study (Des Essarts et al., 2022), but we should note the limitation of the present study, regarding how motivation was assessed (from the therapist’s subjective point of view).

Finally, we examined the association and causal relationship between initial complaints and treatment endings. In this analysis, we identified some significant differences between the CBCL symptoms’ mean scores and treatment endings. The difference is in accordance with previous publications (Baruch et al., 2009; de Haan et al., 2013; Pelkonen et al., 2000; Pessota et al., 2020) that found that externalizing problems are a significant predictor of therapy dropout. These results might also indicate what clinical conditions are prone to less favorable prognosis and higher dropout rates, therefore raising questions about possible adaptations for a technique that is better tailored for these patients.

Clinical scores in the CBCL “social competence” domain were the only significantly superior among adolescents who completed their treatments in relation to cases of non-adherence and dropout. This might indicate that having more flexible social and interpersonal skills can be linked to an enhanced capacity to deal with the challenges of a therapy process, especially the ones involving the therapy alliance and its ruptures (Cirasola et al., 2021; Des Essarts et al., 2022).

The symptom variables, despite presenting significant differences between treatment endings, did not present significant predictive power. However, when examining the patients’ stated initial complaint, we identified that the ones presenting with internalizing problems were 8 times more likely to complete their treatment in relation to the ones who reported externalizing problems, in consonance with previous studies (Baruch et al., 2009; Pelkonen et al., 2000). Even if our findings did not present significant results within the regression models, the descriptive and correlational findings converge with the literature, by indicating that externalizing problems in adolescents can be risk factors for dropout (Baruch et al., 2009; de Haan et al., 2013; Pelkonen et al., 2000; Pessota et al., 2020), especially in cases that include rule-breaking behavior.

Symptoms of externalization in adolescents can pose challenges in exploring the patient’s role in their distress, creating difficulty in establishing a stable therapeutic alliance that can withstand external stressors. It’s the therapist’s responsibility to address fluctuations in patient motivation, helping them recognize how their suffering interferes with their life and goals, thus facilitating their engagement in treatment (Rosa et al., 2018). That finding could lead psychoanalytic psychotherapists to adapt their clinical techniques, aiming to make young people with externalizing symptoms engaged in the treatments. This could lead to better outcomes and lower risks of dropout in these processes.

The study also demonstrates that psychotherapists must not observe isolated symptoms, but the patient’s symptoms, psychological structure, and overall context in a more holistic way. This awareness could make the therapist able to create an alliance with the patient’s protective characteristics in order to advance the treatment, as well as be aware of those that will put the psychotherapy process at risk.

The family variables did not present any predictive power in our study, as described in Midgley and Navridi (2007). Despite those family characteristics not being directly associated with treatment ending, as in de Haan et al. (2013) study, our findings indicate that families play an important role in the prognosis of adolescent psychotherapy (Block & Greeno, 2011). While adolescents’ higher motivation for treatment is associated with more favorable endings, it is possible that more active family involvement in the psychotherapy process might be associated with lower dropout rates.

Final considerations

The current study investigated the predictive role of sociodemographic, family, and symptom variables in the treatment endings of psychodynamic psychotherapy for adolescents. Household income, referral source, adolescents’ motivation for treatment, and their initial stated complaint were predictors of the type of ending. In that sense, these findings indicate what patients are more or less likely to engage in psychodynamic treatments or drop out.

When examining our findings alongside the broader literature on adolescent psychotherapy dropout, we can identify a clearer picture of protective and risk factors. Having a combination of internalizing problems, high household income, being motivated for treatment, and entering through one’s own or family’s desire may lead to a higher chance of having better outcomes from psychotherapy. However, it is worth noting that the current evidence base is still inconclusive regarding gender, age, isolated symptoms, and family characteristics making further investigations necessary.

It is also worth mentioning this study’s limitations. First, our dataset is drawn from one specific community clinic based in southern Brazil. Findings including other centers and other regions could provide a more thorough description of how those variables impact psychotherapy. Despite addressing both patient and family variables, the retrospective nature of this study meant that we could not examine therapist’s variables that could be key in understanding therapy endings. Likewise, we also could not analyze how the therapeutic alliance established between young people and therapists was associated with treatment endings. Furthermore, the assessment of the patient’s motivation was done through the therapists’ perspectives, which might not capture accurately the young people’s actual motivation. Further investigations including the adolescents’ perspective, alongside a more inclusive investigation on ethnicity, sexual orientation, and gender identity, can provide a more reliable depiction of the way adolescents feel within the psychotherapy process and how those variables affect their outcome.

Overall, identifying what patients are more prone to non-adherence or dropout of psychotherapy might inform clinicians about cases that require special attention. This identification could lead to a deeper understanding of those cases and might contribute to adaptations in the technique for these populations.

Patient anonymization statement

Potentially personally identifying information presented in this article that relates directly or indirectly to an individual, or individuals, has been changed to disguise and safeguard the confidentiality, privacy and data protection rights of those concerned, in accordance with the journal’s anonymization policy.

Informed consent

All participants in this study provided informed consent for their involvement.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This study was supported by the National Council for Scientific and Technological Development, CNPq, Brazil, Process [313011/2020-6].

Notes on contributors

Eduardo Brusius Brenner

Eduardo Brenner, MSc. Trainee in Freudian and Lacanian psychoanalysis at the Psychoanalytic Association of Porto Alegre (APPOA).

Guilherme Fiorini

Guilherme Fiorini, MSc, PhD. Research Officer at the Child Attachment and Psychological Therapies Research Unit (ChAPTRe) and Research Tutor at the MSc in Early Child Development and Clinical Applications, at Anna Freud and University College London (UCL).

Vera Regina Röhnelt Ramires

Vera Regina Röhnelt Ramires, MSc, PhD. Associate professor at the Graduate Program in Psychology, Atitus Educação. Psychotherapist of Children, Adolescents, and Adults