Cohort Profile Update: The Heinz C. Prechter Longitudinal Study of Bipolar Disorder

Abstract

Since its inception in 2006, the Heinz C. Prechter Longitudinal Study of Bipolar Disorder has continued to actively enrol and follow a cohort of individuals with bipolar disorder and unaffected psychiatric controls with deep phenotyping and ongoing monitoring of symptom severity. There are 1393 individuals in the cohort and 899 who continue to actively provide ongoing bimonthly outcomes assessments. The average follow-up period is 9 years (range 0–7 years). The study includes data from seven ontological classes of phenotypes: disease, neuropsychology, personality, motivated behaviour, sleep and circadian rhythms, life story and treatment outcomes patterns. The database of measures has been standardized and harmonized along with collected biological samples that can be shared and further utilized to research bipolar disorder. In addition to further longitudinal outcomes and symptom severity measures, new measures include wearable device data, mobile technology assessments, voice-derived emotion annotation, induced pluripotent stem cells and diverse molecular omics data. All clinical, longitudinal, biological and deoxyribonucleic acid samples are available through the Heinz C. Prechter Genetic Repository, distributed by the University of Michigan Central Biorepository. Initial diagnostic evaluation, omics data and longitudinal measures and outcome data, including coded subsets of the PRIORI speech data set, are available via request at the following e-mail address: [email protected], [email protected] conditional on data-use agreement. The Heinz C. Prechter Longitudinal Study of Bipolar Disorder is an open cohort of bipolar disorder (BD) that began in 2006 and continues to enrol new participants. The study includes individuals with BD and unaffected controls, both with deep phenotyping using a dimensional and multidisciplinary approach. The rationale for the establishment of this cohort was the recognized need for deep clinical phenotypic data with longitudinal symptoms and outcomes data. The complexity of BD along with the dynamic and variable nature of the clinical phenotype, which includes variability in the frequency of changing clinical states over time, was a further impetus for the longitudinal approach. Participants are opportunistically recruited into the cohort through advertisements on the web and in newspapers, outpatient specialty psychiatric clinics, community mental health centres, community outreach events and in the inpatient psychiatric unit at the University of Michigan from 2006, continuing until the present. Inclusion criteria into the cohort include: (i) BD I diagnosis with history of mania or schizoaffective disorder, manic type; or (ii) BD II diagnosis with history of major depressive episode or hypomania; and (iii) ≥18 years of age and willingness to participate in a longitudinal study. Participants were excluded during screening for enrolment if they had schizophrenia or schizoaffective disorder, depressive type; active substance dependence (that would impair their ability to provide accurate information); medical illnesses associated with depression (e.g. Cushing’s disease, stroke, etc.); or substantial intellectual impairment (IQ < 70). However, individuals initially ascertained as having BD but over time, following detailed clinical assessment, were found to have other diagnoses, e.g. non-affective disorder, other affective disorder and major depression disorder, have been retained and continue to be followed. Healthy and unaffected controls were recruited through community and campus advertising and included if they had no history of Diagnostic and Statistical Manual of Mental Disorders version IV (DSM-IV) axis I psychiatric illness and no family history of psychiatric diagnosis; no attempt was made to match for sex or age. Recruitment of controls was discontinued at n = 288 but they remain in the follow-up cohort, roughly matched by age and sex. The control group is representative of the geographical unaffected population. Further, as the number of follow-up measures increases in the BD cohort, each individual participant effectively becomes their own control by comparing current measures with previous. Compared with other large BD cohorts1,2 this cohort covers a longer time of ≤17 years and collects more frequent measures (every 2 months). Phenotyping is organized into multiple phenotypic classes that include disease, neurocognitive, personality, motivated behaviours, sleep and circadian rhythms, life story and treatment outcome patterns. Since the initial publication of the descriptive baseline data3 from 1111 participants, we have focused on the curation and harmonization of the longitudinal data for efficient sharing and analyses related to longitudinal outcomes while continuing a rolling enrolment focusing on diversifying the race and ethnicity of the cohort. There have been 282 additional participants enrolled. Over the years, there have been 254 participants who have chosen not to continue, 176 who have not responded within 3 years of their last completed measure, 15 who have completely withdrawn all data from the study and an additional 64 deaths. Currently there are 899 active participants with an average of almost 10 years of participation. In aggregate, there are ∼182 000 individual bimonthly self-report clinical assessment measures, 30 000 individual biannual assessments, 84 000 individual annual self-report or clinician-assessed measures, ∼1400 individual baseline neuropsychology assessments with 428 and 120 individual 5-year assessments and 10-year neuropsychology assessments, respectively. New individuals are added with the goal of maintaining an ‘actively’ followed cohort of between 900 and 1000 participants. Emphasis is currently placed on those with recent onset and minority individuals. Herein we (i) update the demographic description of the cohort and the current and new data collected that are available for sharing under institutional data-use agreements and (ii) review recent findings from the cohort, including a summary of participant satisfaction surveys, and provide links to the programme website for further details and bibliography. With this cohort update, we emphasize the ongoing and evolving longitudinal data collection that is accessible for sharing. The data set includes the initial baseline assessments and adds the longitudinal monitoring data from baseline enrolment, continuing from January 2006 through to 31 December 2022. The updated demographics for the Heinz C. Prechter Bipolar Longitudinal Study of Bipolar Disorder are listed in Table 1. The starting year for a given instrument (some assessment instruments were added after the study began), the frequency of collection, the number of participants, the number of collected measures and the average percent completion rate are shown in Figure 1 and Tables 2 and 3. Survey completion. Number of measurements, number of participants and average percent completion of each measure stratified by the frequency of measure collection (a, b, c, d, e). Brief Social Phobia Scale (BSPS), Columbia Suicide Severity Rating Scale (CSSRS) and Longitudinal Interval Follow-up Evaluation (LIFE) are captured every 2 years and the neuropsychology battery consisting of 16 tests and 8 cognition domains is captured at baseline, Year 1, Year 5 and Year 10. Altman Self-Rating Mania Scale (ASRM), Alcohol Use Disorders Identification Test (AUDIT), Buss-Durkee Hostility Inventory (BDHI), Brown-Goodwin Lifetime History of Aggressive Behavior (BGAH), Behavior Inhibition System (BIS), BSPS, COVID-19 Impact Scale (CIS), CSSRS, Childhood Trauma Questionnaire (CTQ), Diagnostic Interview for Genetic Studies (DIGS), Experiences in Close Relationships (ECRQ), Epworth Sleepiness Scale (ESS), Family Adaptability and Cohesion Evaluation Scale (FACES II), Fagerstrom Test for Nicotine Dependence (FTND), Generalized Anxiety Disorder Scale (GAD-7), Hamilton Rating Scale for Depression (HDRS), Life Events Checklist (LEC), Life Events Occurrence Survey (LEOS), Life Functioning Questionnaire (LFQ), LIFE, Munich Chronotype Questionnaire (MCTQ), Measures Related to Close Interpersonal Relationships (MRCIR), Patient Health Questionnaire-9 (PHQ-9), Pittsburgh Sleep Quality Index (PSQI), Rand 36 Item SF Health Survey (RAND-36), Health Survey Short Form-12 (SF-12), Structured Interview Guide for Hamilton Depression Rating Scale (SIGH-D), Seasonal Pattern Assessment Questionnaire (SPAQ), Young Mania Rating Scale (YMRS) The Heinz C. Prechter Longitudinal Study of Bipolar Disorder cohort descriptives as of 31 December 2022 NOS, not otherwise specified. Listing of survey instruments with the number of participants, distinct measurements, average percent complete, commencement year and collection frequency In addition to the availability of the ongoing, growing longitudinal clinical assessment data collected in the cohort, concomitant efforts have emphasized standardizing and harmonizing all data elements. Further, efforts are ongoing to maintain the engagement of participants as well as to expand and diversify the ethnic and racial backgrounds of the cohort participants. Participants are routinely notified and invited to participate in auxiliary studies in BD and, as these studies progress, the data generated typically become part of the collective Prechter data set. The COVID Impact Scale was implemented in 2020 and continues to be collected bimonthly to study the effects of the Sars-CoV-2 pandemic. A participant feedback survey is now collected on a rolling annual anniversary to monitor participant engagement and satisfaction, often making emendations as needed. From this feedback, for example, the self-report assessment line-up was reordered to improve the completion rate, sexual orientation categories were clarified and dated insensitive questionnaires removed, including the Barratt Impulsiveness Scale, Brown-Goodwin Aggression history, Buss-Durkee Inventory and the Brief Social Phobia Scale. Cohort participants have embraced mobile technology to collect research data. Mobile monitoring that includes ecological momentary assessment (EMA), a digital self-report survey of mood for BD (DigiBP)4 and an integrated self-management application based on a well-established approach (LifeGoals)5 has been undertaken. Additionally, the PRIORI (Predicting Individual Outcomes for Rapid Intervention) project, focused on the analysis of speech and acoustic patterns,6 has expanded to include multiple human annotations of speech segments to estimate levels of activation and valence in speech for training machine-learning models in analyses of mood and emotion. The PRIORI emotion data set is available for sharing. Research in the Heinz C. Prechter Bipolar Longitudinal Study of Bipolar Disorder has evolved in two main complementary directions: (i) the long-term outcomes patterns, including lifetime effects of the temperament, behaviours, sleep, life story, outcomes, and disease determinants of BD; and (ii) a focus on the functional, physical, emotional, molecular and neurocognitive aspects of the disorder. New research areas focus on longitudinal time-series studies emphasizing the high variability of symptoms and outcomes within BD. Bimonthly self-report assessment of symptom severity offers data-driven methods to categorize and stratify individuals according to the nature and intensity of symptoms and outcome variability. At the beginning of the COVID-19 pandemic, the effects of the pandemic and related social, health and occupational changes on the cohort of participants were studied. The ongoing longitudinal design is well suited to studying the impact of the COVID-19 pandemic based on frequently collected measures before, during and after as the pandemic waxes and wanes. An initial investigation, specifically studying the effects of a mandatory state-wide stay-at-home order, showed that the effects of the mandatory isolation were notable on all individuals in the study. However, those with BD were affected more, experienced more significant difficulty and did not recover as quickly as the unaffected controls.7 The COVID-19 Impact Scale is now integrated into the study and completed bimonthly.8 The newly captured data collected through this instrument will be used to monitor the immediate short-term influence of the pandemic as well as the functional outcome effects of the pandemic longitudinally. Participating, capturing and analysing data through mobile technologies is a fresh direction of the cohort and study. This includes expansion of the LifeGoals5 app to engage and monitor participants beyond our geographical area. Recruitment in LifeGoals and PRIORI are by invitation within our existing longitudinal member participants. The PRIORI app6 is expanded to include monitoring of the ambient audio environment and calls of the individual in a secure manner designed to not compromise privacy. There is no attempt made nor is there opportunity to identify anyone outside of the individual participant, as audio data are asynchronously encrypted and immediately deleted from the smartphone. Data are decrypted and processed within our controlled server; all programming and data structures have been rigorously reviewed by our information assurance and cybersecurity teams, and the outcomes of processing are computational and descriptive in nature. Biological mechanisms underlying BD will continue to be studied utilizing the high-throughput omics data in combination with cell and developmental biology approaches. Several cohort participants have provided fibroblasts (n = 30) and other biological materials to our biobank repository, including blood (n = 770) and saliva (n = 60) for genomics. Using the provided fibroblasts, 23 induced pluripotent stem cell lines (iPSCs) have been developed and altered using CRISPR genomic editing technologies. These iPSC cell lines have been differentiated into different neuronal cell types and organoids. Immunohistochemical localization of lineage-restricted proteins was carried out in both cells and histological sections of organoids. At several time points during development, these cell lines and the organoids derived from them have been used to generate bulk single cell- and single nuclei-RNA sequencing to study the relational time-course developmental gene-expression differences between those with BD and controls. Proteomics and phospho-proteomics have also been completed in a developmental time-course manner to understand signalling differences during the development of different cell types and between those with BD and controls. The Heinz C. Prechter Longitudinal Study of Bipolar Disorder is an open cohort of individuals with BD and other psychiatric illnesses and controls who agree to participate. Follow-up and enrolment are continuous and the original inclusion and exclusion criteria are unchanged.3 Currently, there are 1393 participants age at enrolment is The cohort includes individuals with of BD diagnosis within the that includes manic 36 with major depressive disorder with other mood with psychiatric illness and 288 controls The time for which participants have been in the study is 8 years with a of years and there is an rate of of 31 January 64 participants had during the study and an additional 254 have to not continue, of using data. The of participants to be for additional research or clinical An annual participant satisfaction survey was implemented in and a rate each of the feedback is in a by the study and there are also example, in of the participants that their in the research study was and of the participants with BD that the study the of their they and their An update on the descriptive and all measures collected in this cohort is provided in Tables 2 and 3 and shown in Figure 1. A new of measure is the outcomes measure, which has been participants new of measure to outcomes and is based on the of the cohort that is now part of a that The PRIORI data set includes a of distinct calls from of which speech audio segments are for emotion with segments new biological is collected and including fibroblasts (n = 30 cell (n = 23 cell organoids (n = 36 (n = cell types and saliva samples (n = From these materials for omics data have been newly in addition to the and single data. for the iPSC samples there is now and, single and data. health and information with the number of participants and distinct measurements, average percent complete, commencement year and collection the number of participants and collection frequency of different types of biological and the audio types including the number of participants, year distinct number of distinct number of segments and segments PRIORI and PRIORI are the two of the PRIORI deoxyribonucleic The Heinz C. Prechter Bipolar Longitudinal Study of Bipolar Disorder continues to collect and expand a of data with the opportunity to and at the of multiple and dimensional to the disorder. the findings is the ongoing and engagement the participants with BD and the multidisciplinary engagement of the community in Sleep and circadian outcomes those with BD. those with more frequent depressive symptoms over sleep is a of outcomes including neurocognitive and to be over with BD on but the of are to those of the unaffected There were differences in those with BD have and and with controls. is BD as and depressive The outcome of the COVID-19 pandemic and associated occupational and is of major in the cohort. are ongoing and an initial that focused on a period in the of Michigan found those with BD experienced a impact from the in and Further, those with BD those BD and were more individuals were affected The study and of the biological mechanisms underlying BD remain a major of the Heinz C. Prechter Bipolar Longitudinal Study of Bipolar There is that between BD and control and this is by the of that and of from those with BD and controls has in the BD group that be to and to controls has been by in the of individuals with to the and has in measures of and models have the and that a of a of in with in neuronal to the addition of to neuronal analyses are with in the and mechanisms with of a between the environment of the and the and of the cohort have focused on the for processing and along with and to to and in BD that is in the manic These and emphasize cognition in BD and that these become A history of has a significant on the and outcome of outcome based on longitudinal an affective mood states not follow a BD in the of with a and to states Mobile monitoring of outcomes is currently a focus and the long-term of mobile monitoring is a of the Prechter The research programme and a website with the that are based on the cohort. The The major of the Heinz C. Prechter Longitudinal Study of Bipolar Disorder is the detailed and of the clinical and biological data including data on medical and mental health This is a of having multidisciplinary from and cell and developmental Further, a of participant continues with a shared for in research to and for the treatment and of BD. A of self-report data has been on the participants this is a from the of the participant the data and, as an the group was to continue data collection in the A of data is that there will be variability based on (e.g. but this is in questionnaires by descriptive associated with the include the geographical from a and community in in the demographics of the cohort is and This is an given the between and BD. A related includes its cohort for the and the of with BD and controls, to the nature of clinical research and the from participants for longitudinal This the a group of participants and with more who are to participate in long-term studies and not the with BD of the study are now we to the geographical of our cohort and into that will the of our improve the of this cohort, the research programme has a to engage with on and to this cohort and more and while actively and focus to from not well in the cohort or in research on BD in The impact of this is starting to of the newly recruited participants 2020 are racial is that the diagnostic categories remain in the and have to be updated to However, there are no changes for the lifetime diagnosis of BD between and as our diagnostic the episode of depression and mania to the initial study Data on and were collected with standardized assessment as the as and no were made to collect information based on the to other as and the Bipolar the average age of into the Heinz C. Prechter Longitudinal Study of Bipolar Disorder is a age at episode of years, individuals with BD to engage in the study at of their the of the illness be in our which to the of longitudinal of other is in BD and the age of enrolment will more who are The Heinz C. Prechter Longitudinal Study of Bipolar Disorder to maintain active of individuals for their lifetime and to our engagement with those with levels of with BD and those at for the illness via All clinical, longitudinal, biological and deoxyribonucleic acid samples are available through the Heinz C. Prechter Genetic Repository, distributed by the University of Michigan Central Biorepository. Initial diagnostic evaluation, omics data and longitudinal measures and outcome data, including coded subsets of the PRIORI speech data set, are available via request at [email protected] conditional on data-use agreement. Data can be shared in an or a data be available for who can provide as in other An ongoing updated publication is provided at and further at the following e-mail address: [email The Heinz C. Prechter Longitudinal Study of Bipolar Disorder is reviewed and by the University of Michigan All participants provided to enrolment in the study. is reviewed with the participant for in the study. I of the data and data and and This research is by the Heinz C. Prechter Bipolar Research at the University of Michigan Depression and the from the of Health several in the cohort. With we the University of Michigan Prechter Bipolar Longitudinal Research Participants and the and of the Prechter Bipolar Research including and for their efforts in the collection and of the data used in this has for and and research from and are as on 9 and both by the of the University of Michigan and related to the technology the PRIORI on the for and All other no related to this

Topics & Concepts