Low self-efficacy is a transdiagnostic risk factor for several mental disorders. Self-efficacy refers to one's belief that one is capable of performing a behavior necessary to successfully complete a task or achieve a goal. Consistent with theoretical models and empirical findings, individuals with low self-efficacy are more likely to perceive challenges as uncontrollable and threatening and thus are at increased risk for developing mental disorders during sensitive periods such as young adulthood. Self-efficacy interventions have been shown to be effective in promoting health behavior change, quality of life, and treatment adherence in patients with serious illnesses, as well as motivation and performance in students and employees. However, whether targeted self-efficacy training can prospectively prevent the onset of full-threshold anxiety, affective, and substance use disorders in young adults at increased risk for psychopathology remains an open question.
The aim of this randomized controlled trial is to test whether a brief cognitive-behavioral intervention in young adults with low self-efficacy can increase general self-efficacy (primary outcome of intervention effectiveness) and thus prevent the onset of DSM-5 mental disorders in the subsequent year (primary outcome of prevention effectiveness). In addition, the investigators examine whether improvements in domain-specific self-efficacy lead to subsequent improvements in general self-efficacy and thus to lower psychopathological symptoms (spillover effects).
Young adults (18-30 years) with low self-efficacy but no mental disorder will be included (N=378). The study will include screening, entry, baseline, post, and 12-month follow-up assessments plus additional course assessments in both groups. After the baseline assessment, participants will be randomized to an intervention or control group. The intervention group will receive group-based self-efficacy training (6 sessions of 75-90 minutes each). The control group will also meet in groups (6 sessions) but will only talk about psychological research findings unrelated to self-efficacy or cognitive-behavioral interventions without receiving any training. DSM-5 mental disorders will be assessed at study entry and follow-up with a structured diagnostic interview. Other outcomes will be assessed with established scales and ecological momentary assessments (EMA) at baseline, post and follow-up. Clinical outcomes include psychopathological symptoms (dimensional scores for anxiety, depression, anger, and somatic symptoms, as well as sleep disturbance) and mental disorders (DSM-5 categorical diagnoses of anxiety, affective, and substance use disorders). Intervention effectiveness will be tested using logistic/linear regression and multilevel analyses. Spillover effects between improvements in domain-specific/general self-efficacy and psychopathological symptoms over the course of the study will be examined using cross-lagged panel models.
Detailed Description Based on previous research, it is plausible that self-efficacy training may have positive mental health effects and thus successfully prevent the development of mental disorders over time: Consistent with social cognitive theory, self-efficacy training may reduce feelings of anxiety and thus reduce the risk of anxiety disorders. According to the theory of learned helplessness, self-efficacy training may reduce feelings of helplessness and depression and thus reduce the risk of affective disorders. In addition, and consistent with the self-medication hypothesis, individuals who receive self-efficacy training may be less likely to use tobacco, alcohol, or medication/illicit drugs to avoid or down-regulate unpleasant feelings, thereby reducing the risk of substance use disorders.
However, although self-efficacy interventions have been shown to be beneficial in other domains, it remains unclear whether self-efficacy training can effectively prevent the onset of full-threshold anxiety, affective, and substance use disorders in young adults. Given that selective and indicated prevention has been shown to be particularly effective in high-risk samples young adults with low self-efficacy and thus increased risk for mental disorders may benefit from targeted self-efficacy training.
Study aims
This randomized controlled trial in young adults (aged 18-30 years) with low self-efficacy but no 12-month mental disorder at study entry aims to test whether a short cognitive-behavioral intervention in young adults with low levels of self-efficacy (but no 12-month anxiety, affective, or substance use disorder, current intervention (seeking) or suicidality) can effectively increase self-efficacy and thus prospectively prevent the onset of DSM-5 anxiety, affective, and substance use disorders in the subsequent year. The following hypotheses will be tested (based on screening, entry, baseline, post, and 12-month follow-up assessments plus additional course assessments in both groups:
Hypotheses
1. Participants in the intervention vs. control group experience a greater increase in general self-efficacy from baseline to post (primary outcome of intervention effectiveness) and greater improvements in dimensional psychopathological symptoms and other dimensional outcome measures from baseline to post (secondary outcome of intervention effectiveness). Note: Other dimensional outcome measures include self-esteem, perceived control, social support, coping strategies, interpersonal competence, work engagement, and individual work performance.
2. Participants in the intervention vs. control group experience a lower risk of incident DSM-5 mental disorders from entry to follow-up (primary outcome of prevention effectiveness) and greater improvements in dimensional psychopathological symptoms and other dimensional outcome measures from baseline to follow-up (secondary outcomes of prevention effectiveness). Note: Not only binary diagnoses but also dimensional symptoms will be assessed to be able to examine more nuanced symptom changes as well as lagged effects between changes in self-efficacy and changes in mental health over time.
3. In the intervention group, improvements in domain-specific self-efficacy will lead to subsequent improvements in general self-efficacy over the course of the intervention (i.e., week 1-6 of the training; spillover hypothesis 1). Note: In the intervention group, participants choose to work on their self-efficacy in the domains of health, social relationships, or education/work. In these analyses, domain-specific self-efficacy refers to the domain that the particular participant works on during the intervention. The selected domains are recorded systemically.
4. In the intervention group, improvements in general self-efficacy will lead to subsequent improvements in dimensional symptom outcomes over the course of the study (i.e., from baseline to follow-up; spillover hypothesis 2).
Design The investigators will conduct a randomized controlled trial (in young adults aged 18-30 years with low levels of general self-efficacy but no 12-month anxiety, affective, or substance use disorder, current intervention (seeking), or suicidality with an intervention group and a control group. Participants in the intervention group will receive group-based self-efficacy training (8-12 participants per group; 6 sessions of 75-90 minutes each). Participants in the control group will also meet in groups (8-12 participants per group; 6 sessions) but will only talk about psychological research findings unrelated to self-efficacy or cognitive-behavioral interventions without receiving any training. The design contains three measurement points at which the main outcome variables self-efficacy and psychopathology are assessed: 1) a baseline assessment - immediately before the courses; 2) a post-assessment - immediately after the courses; 3) a 12-month follow-up assessment - 12 months after the post-assessment.
Procedure Individuals who meet the inclusion criteria will participate in the baseline assessment. After the baseline assessment, participants will be randomized to either the intervention or control group. Before each course session (6 times), both general and domain-specific self-efficacy will be assessed. Dimensional clinical outcomes will be measured at baseline, post-intervention, and at a 12-month follow-up. Additionally, a one-week EMA will be administered at these three time points.
The intervention will be delivered in an online group format (subgroups of 8-12 participants; 6 sessions of 75-90 minutes each) and led by a psychologist. The courses will be structured according to well-established self-efficacy interventions with proven effectiveness, targeting Bandura's 4 key sources of self-efficacy (i.e., mastery experience, vicarious experience, verbal persuasion, and physiological/emotional arousal).
Course sessions will be structured as follows: Opening, brief reflection on current progress toward the goal, discussion of homework (with special emphasis on participants' progress and sharing of experiences), introduction to the topic, supervised practice, answering open questions, and closing. Course sessions will be accompanied by weekly homework assignments for practice at home. Assignments will be prepared and discussed during each class session. Participants will also be asked to keep a homework and success diary (workbook) to document changes in thoughts, feelings and behaviors, questions/difficulties (to be discussed in the next session), and accumulating successes in daily life over time (to enhance vicarious experiences).
To increase adherence to the intervention, participants will receive a weekly text message with homework reminders and motivational support. Participants in the intervention group will be able to contact a psychologist between sessions to ask questions and receive additional support regarding the intervention and homework. The intervention will be tested in a pilot phase in which self-efficacy will be measured before, during, and after the intervention. In addition, feedback interviews on the intervention will be conducted during the pilot phase to assess its acceptability, and the intervention will be modified accordingly if needed.
Recruitment Participants will be recruited through personal contact and print/online advertisements in universities and vocational schools, businesses, coworking spaces, clubs, bars, cafes, restaurants, cinemas, theaters, fitness centers and sports clubs, pharmacies, general practitioners, health insurance companies, and public and social media. Recruitment sites will also include social and welfare institutions that cater to individuals from diverse backgrounds (e.g., low education or migration) to ensure an inclusive sample and to counteract common biases in intervention research (e.g., overrepresentation of highly educated groups). Participants will be recruited not only in Berlin and Potsdam, but also in other (metropolitan and rural) regions of Germany and via online media.
Inclusion and exclusion criteria are listed below. After the baseline examination, individuals who meet the inclusion criteria will participate in the baseline examination. After the baseline assessment, they will be randomized to the intervention or control group (balanced randomization \[1:1 ratio\] based on computer-generated permutated blocks). Participants in the intervention group are required not to receive any other psychological or psychopharmacological intervention at study entry and during the training. Participants in the control group are required not to receive any psychological or psychopharmacological intervention at study entry. However, they may or may not engage in any intervention over the study course (usual care). After completion of the study, they will have the opportunity to receive the same self-efficacy training as the intervention group.
Sample size calculations are based on the "weakest line" in the analyses: Any incident or recurrent mental disorder from entry to follow-up assessment in the intervention vs. control group. Calculations are based on data from the baseline and first follow-up assessment of the Early Developmental Stages of Psychopathology Study, a population-based study of adolescents and young adults from Germany. Based on meta-analytic evidence on indicated mental health prevention, the investigators expect the intervention group to improve from standardized self-efficacy scores below -1 to scores between -1 and -0.5. In the EDSP, 21% of those with baseline scores between -1 and -0.5 (\~ intervention group) and 41% of those with baseline scores below -1 (\~ control group) developed any incident anxiety, affective, or substance use disorder until follow-up (considering only individuals without psychopathology at baseline). With a statistical power of 0.9 and a dropout rate of 20% (from baseline to post and from post to follow-up, respectively), an incidence rate of 21% vs. 41% in the intervention vs. control group yields 189 individuals required per group at baseline (total N=378).
Because a 12-month follow-up period is a relatively short time frame to evidence group differences in onset rates of full-threshold mental disorders, also incident/recurrent sub-threshold disorders will be considered. Sub-threshold disorders are defined as disorders falling short of one diagnostic criterion (e.g., the time criterion). In these analyses, incidences of sub-threshold disorders not being present at the entry assessment will be additionally counted.
Data Exclusion/Missing Data To ensure response accuracy and validity, several control items will be embedded in the assessments. Observations will be excluded if accuracy scores are too low.
Full information maximum likelihood estimation and, if necessary, other imputation methods (e.g., multiple imputation) will be used to deal with missing data.
Analysis Data analyses will be performed using RStudio, Python, Stata, and Mplus. Data from the main assessment will be analyzed using linear (dimensional outcomes) and logistic (binary outcomes) regressions. To test whether outcome changes from baseline to post/follow-up vary by group, the difference in the respective outcome score (post/follow-up minus baseline) will be regressed on a group dummy (0=control, 1=intervention). To test whether the rates of incident/recurrent mental disorders from study entry to follow-up vary by group, the diagnostic outcome will be regressed on the group dummy. To test for clinically significant effects, changes in clinical features (e.g., symptom burden and impairment) due to the intervention will be additionally assessed using linear/logistic regressions. EMA data will be analyzed using multilevel analyses with measurement occasions (level 1) nested within persons (level 2). To test whether outcome changes from baseline to post/follow-up vary by group, the respective outcome will be simultaneously regressed on a time dummy (0=baseline, 1=post/follow-up), a group dummy (0=control, 1=intervention), and an interaction term (time\*group). Furthermore, multilevel models will be used to capture time-lagged associations between contextual factors and outcome changes in daily life. For example, the effect of daily hassles on momentary fluctuations in state self-efficacy and psychopathological symptoms in the intervention vs. control group will be examined. Dimensional outcomes with non-normally distributed residuals will be log-transformed (log(x+1)). To allow for comparisons across different measures and groups, all dimensional outcomes will be standardized (M=0, SD=1) based on the pooled standard deviation in the intervention and control groups at baseline (to account for potential baseline group differences). Analyses will be adjusted for gender and age. The alpha level will be set at .05. The "BY" method will be used to correct for multiple testing of dependent hypotheses. Spillover effects will be tested using multilevel models with time-lagged effects. If the power requirements are met random intercept cross-lagged panel models will be used. As a manipulation check, the investigators will test whether participants of the intervention group first increase in the domain of self-efficacy they work on (health, social relationships, or education/work) but not in the other domains. For example, individuals working on their health-related self-efficacy are expected to first increase in their health-related self-efficacy but not in the other 2 domains.
