Early-BipoLife A1 is part of the German Research Consortium BipoLife (Ritter et al. 2016) with a focus on bipolar disorders (Pfennig et al. 2020). It is a Germany-wide multicentre study in which 10 universities and teaching hospitals with screening and treatment programmes for BD participated. The BipoLife project was funded by the German Ministry of Education and Research (BMBF, Grant number: 01EE1404A). Extended information of the overall study design has been published within the study protocol (Pfennig et al. 2020). The overall aim of the A1 project is improving early recognition and intervention in people at‑risk of developing a BD (Pfennig et al. 2020).

Early-BipoLife is a naturalistic, prospective-longitudinal observational cohort study. Recruitment took place in early recognition centres and if not already implemented, individualized treatment as usual was provided. The participants were seen at baseline (BL), and 4 follow-up assessments (FU1-FU4) during a course of at least 2 years from July 2015 until September 2018. In order to map the course in QoL, data from the following assessments in which QoL was evaluated: baseline (BL), follow up 2 after 12 months (FU2) and follow up 4 after 24 months (FU4) was included. FU1 took place after 6 months and FU3 after 18 months and were both performed as telephone-interviews without assessing QoL data. The overall time of an examination was 3–7 h. The interviewers conducted a comprehensive standardizes face-to-face diagnostic interview and were trained 2 days centrally and later supervised and retrained by their principal investigator locally. The examinators were physicians and psychologists, not blinded for the bipolar risk status. The ethics committee of the Technical University Dresden and all local ethics committees approved the study. Only participants who gave their written informed consent were included. For minors <18 years it was a requirement to obtain their informed assent and have one of their caretakers confirm the consent (Pfennig et al. 2020).

All participants received state of the art counselling and treatment tailored to their individual needs, selected by the expertise and experience of their local study centre as requested in a naturalistic design (Pfennig et al. 2020).

Participants aged 15–35 years with at least one risk factor for the development of a BD were recruited at the German hospitals in Berlin, Bochum, Dresden, Frankfurt/Main, Hamburg, Marburg, Brandenburg/Neuruppin and Tübingen (Pfennig et al. 2020). 1229 participants from these sites across Germany were included in the longitudinal study. The following risk factors compose the inclusion criteria: family history of BD, (sub)threshold affective symptomatology, hypomanic/mood swings, disturbances of circadian rhythm/sleep, ADHD diagnosis and depressive/dysthymic disorders (Pfennig et al. 2020). Exclusion criteria were a diagnosis of a BD, schizoaffective disorder, schizophrenia or a diagnosis of anxiety, obsessive–compulsive or substance dependence disorder. A limited ability to comprehend the study or acute suicidality also led to exclusion. The first data on 2-year findings were published recently (Martini et al. 2023). From the overall 1229 participants, 1038 filled in the QoL self-report questionnaire at baseline (BL) and were included in the present analysis as the self-rating and participation in the study were voluntary.

For a comparison with an appropriate norm, the data from the Hawthorne et al. (2006) data set was used. The sample was age matched (20–29 years) and includes a broad range of health conditions from full health to terminal illness (Hawthorne et al. 2006).

Clinical and socio-demographic data throughout the study were assessed using the Structured Clinical Interview (Wittchen et al. 1997; Fydrich et al. 1997) together with instruments that address symptoms of psychotic prodrome (Ising et al. 2012; Miller et al. 2003; Schultze-Lutter et al. 2012) and psychiatric treatment, physical illness and substance use utilizing case report forms (CRF).

The Early Phase Inventory for Bipolar Disorders (EPIbipolar) is a semi-structured interview and was developed by Pfennig and Leopold et al. in 2012 to capture specifically the risk status regarding an onset of bipolar illness (Leopold et al. 2012). The EPIbipolar was used to address the course of disease in terms of change of severity of the risk status. Risk factor categories and information of the patient’s history were determined through a systematic review of the literature and clinical experience and form the basis for the classification into three risk categories (no risk, low risk, high risk). Included criteria: e.g. family history of BD, subthreshold depressive and (hypo-)manic symptoms, substance misuse, a diagnosis of ADHD or behavioural problems/conduct disorder, pronounced creativity, critical life events, changes in sleep/circadian rhythm, mood swings or increased affective lability, fearfulness/anxiety, dissociative symptoms, and impairment in psychosocial functioning (Pfennig et al. 2020). For the risk quantification these risk factors are listed, weighted and assigned to one of the three final groups depending on fulfilment of the aforementioned risk factors (Leopold et al. 2012).

The WHOQOL-BREF as a Patient Reported Outcome (PRO) instrument was used to assess the global health status of patients independently of a disease. This questionnaire consists of 26 items in total (Angermeyer et al. 2000), which are assigned to four domains: physical health (energy level and the ability to perform daily activities), psychological well-being (emotional and cognitive health), social relationships (personal connections such as family and friends) and environment (safety and access to resources). All items are answered on a five-point Likert scale. There are four types of a five-point Likert interval scale, which reflect intensity, capacity, frequency and evaluation, and one of these was attached to each item. Items inquire ‘how much’, ‘how completely’, ‘how often’, ‘how good’ or ‘how satisfied’ the respondent felt in the last 2 weeks. Domain scores were built as sum scores and transformed on a 0 to 100 scale to enable comparisons between domains composed of unequal numbers of items. The higher the score the better the self-perceived QoL (WHOQOL-Group 1998). Norm data have been published for all domains by Hawthorne et al. (Aigner et al. 2006) and are provided in Table 3.

The FERUS (Fragebogen zur Erfassung von Ressourcen und Selbstmanagementfähigkeiten, in English: Questionnaire for Assessing Resources and Self-Management Skills) measures health-related resources and self-management skills, comprising of 66 items reflecting 7 scales (motivation to change, self-observation, active and passive coping, self-efficacy, self-verbalisation, hope and social support) and a total score (Jack 2001). This total score represents the overall self-management ability and was included in the present analyses. The FERUS is answered on an agreement scale from 1 "not true" to 5 "very true"; scores were built as sum scores. According to the FERUS, the higher value, the more pronounced the self-management resource is.

In order to get better insight into QoL in the population of individuals at risk of bipolar disorder, the WHOQOL-BREF was first evaluated via descriptive statistics [mean (M), standard deviation (SD)] for the 4 subdomains of QoL (physical health, psychological well-being, social relations and environment). A comparison with a norm population (Hawthorne et al. 2006) was calculated with a T-test for one sample, reporting the test statistics T, df and p for significant results at baseline and after 2 years. The changes in QoL over the 2 year-period were evaluated in a mixed model for repeated measures (MMRM) and are described below.

Beside the changes in QoL, the predictors were addressed as well in the MMRM for each of the four domains separately covered by the WHOQOL-BREF. The QoL measures at follow-up times (12 months (FU2) and 24 months (FU4) after baseline) were considered as repeated measures, the patients as the random effect, time as fixed effects, and the baseline (BL) values of the dependent variables as well as age and sex as covariate. Outcomes were changes from baseline in QoL scores of the WHOQOL-BREF (dependent variable). As potential sociodemographic predictors of QoL a positive family history for mental illness (yes/no) and employment status (no employment, paid education, part time or less, full time) were included as independent variables, because of their relevance as risk factor and predictor for poor outcomes in mental disorders (Solmi et al. 2023). Regarding clinical characteristics, the following variables have been taken into account as independent variables additionally, derived from the literature described above: substance disorder (easy/medium/severe), global assessment of functioning score (GAF, continuous), current psychiatric treatment (yes—outpatient, yes—partly inpatient, yes—inpatient, no), ever received psychiatric treatment (no consultation, consultation or short treatment, continuous treatment for months, continuous treatment over years), current medication (yes/ no). The risk factor for developing a BD was imaged via the EPIbipolar (no risk, low risk, high risk). To reflect the change in risk score from baseline to FU4, a new variable was created taking the change in EPIbipolar over the 2-year period into account (increasing risk, decreasing risk, constant risk). Both variables regarding the risk factor were included as predictor variables (independent variable) as well. In addition, the coping total sum scale of the FERUS “self-management”, assessed at baseline, was entered in the MMRM model as potential predictor too. Baseline values were used as covariates to minimize the variance. The main effects (F), significance levels (p) and 95% confidence intervals (CI) are reported. In addition, correlations (Pearson product-moment correlation coefficient, r) were calculated between the self-management scale and all subdomains of the WHOQOL-BREF at baseline and FU4 to map the relationship between coping resources and QoL in Early-Bipolife. The level of significance was set at p < 0.05, two sided. The data were analysed using IBM SPSS Statistics (Version 27) (Version 2020).