TARA for Medical Students, a Single-arm Mixed Methods Pilot Study

The primary aim of this study was to investigate the feasibility and acceptability of TARA as indicated prevention of symptoms of depression and anxiety and stress-related symptoms in self-selected medical students. The investigators also aimed to qualitatively explore the students' experiences of participating in TARA both during and after the program as well as at a two-year follow-up. The study was designed as a single arm pre-post intervention study.

Participants and procedures Participants were recruited from Umeå University School of Medicine. The University is located in the north of Sweden and enrolls 130 medical students per semester. Participants were recruited from the second and fourth semester through oral and written information at a mandatory class on professional development, including in total approximately 225 students. In addition to being a medical student there were no specific inclusion or exclusion criteria. Participants were allocated to one of two TARA-groups based on their personal weekday preference, since the groups are held on different days of the week.

Before inclusion, participants provided written informed consent. Study participation was voluntary and could be cancelled by the participant at any time.

Groups were held in premises provided by the university hospital. Each session for a given group was led by the same two facilitators, with expertise in contemplative practices and clinical psychology or psychiatry. The facilitators' role during sessions was to teach specific content and to model a collaborative, inclusive, non-judgmental, and supportive attitude. Continuous supervision and training of facilitators was provided by the principal investigator who also monitored fidelity to the manual, both in terms of content adherence and the process of delivery. The facilitators were videotaped to facilitate supervision and implementation of the protocol.

Assessment procedures At baseline, before the intervention, self-report of sociodemographic background and outcome measures were collected online, for details see outcome measures. After the intervention, and at a two-year follow-up outcome measures were repeated. Participants who did not provide data at the time of the data collection were reminded by the research team through email and/or telephone. The TARA-participants provided brief self-assessment and session-evaluation on paper at the beginning and end of each session.

A semi-structured group-interview about TARA was conducted halfway through the program. Participants who were willing to stay after the session were interviewed and the same participants were interviewed again upon completion of the program. The interviews were conducted by an independent researcher without any background or involvement in the intervention. A semi-structured interview guide created for this purpose was used. The questions were mainly about the participants' experiences of TARA. At two years follow-up another independent researcher conducted individual interviews with participants from both groups. These interviews were held online or at the interviewees' choice of location. All participants were invited by email. The semi-structured interview guide for the individual interviews were adapted for the current timeframe. Example questions asked in all interviews were: "What are your experiences of participating in TARA?" and "Can you mention something that was particularly hard or challenging?". The questions were broad or open. Experiences before, during and after TARA were covered. When necessary and/or appropriate questions were followed up using cued prompts "You said X, please tell me more about that", facilitating utterances, nodding, "Umm," etc., clarifying questions "Do you mean Y?", and relevant follow-up questions.

Analysis The dataset was checked for illogical values. For both categorical and continuous variables, responses outside the possible response categories were coded as missing. Descriptive statistics were calculated using standard measures. Sum scores, as well as means and standard deviations were calculated for each self-report measure. Missing data on the item level was imputed using Multiple Imputations with 5 iterations, to enable calculation of sum score despite missing items. The post intervention scores on each item on the secondary outcome measures - Hospital anxiety and depression rating scale- anxiety and depression sub scales, Perceived stress scale and Avoidance and Fusion Questionnaire for youth - were used as predictors in the imputation model and a random seed was applied. The data was checked for normality using histograms and Q-Q plots, as well as skewness and kurtosis. Boxplots were used to check for outliers. No interim analyses were conducted. Descriptive statistics was reported using standard measures.

To evaluate potential differences between the test-retest sample and the baseline sample the investigators used Fisher's test for nominal data, Mann-Whitney U test for ordinal data and non-normally distributed continuous data, and unpaired T-test for normally distributed continuous data.

Paired samples T-tests were conducted to compare self-rated symptom severity scores at baseline and post intervention as well as at baseline and the two-year follow up. All analyses were performed using SPSS statistics, version 26. All significance testing was two-tailed, with a significance level of 0.05. P-values were not corrected for multiple comparisons.

The interviews were audio-recorded, transcribed verbatim by the respective interviewer who had a high degree of familiarity with the data. The transcribed data was subjected to qualitative content analysis, which involved the systematic interpretation of the overt and underlying content and could be used to analyze participants' reflections, experiences, and attitudes. The text was first read several times separately by the investigators, then discussed to get sense of the text as a whole, and then divided into meaning-units relevant to the aim of the study. The meaning units were then coded and sorted into groups according to their variations, similarities, and differences. The investigators met regularly to discuss their interpretations and finally agreed on the structure of the data.

Time plan The investigators delivered TARA in the fall 2018. The final qualitative data collection was completed in fall 2020.