Identifying early signals prior to Bipolar Disorder diagnosis in a UK primary care patient cohort

Talk Code: 
Catharine Morgan
Darren M. Ashcroft, Roger T. Webb, Carolyn A. Chew-Graham, Anya Francis, Alison R. Yung
Author institutions: 
The University of Manchester, Keele University


Bipolar disorder (BD) is a serious mental illness characterised by mood instability. Delay in diagnosis is typically between 6 and 10 years with many affected individuals experiencing persistent untreated symptoms. The diagnostic delay is associated with poor outcomes including poor social adjustment, more hospital admissions and high prevalence of coexisting cardiovascular, endocrine/metabolic or neurological conditions. Unlike recognition and treatment of schizophrenia, there is no agreed strategy for improving early identification and treatment of BD. We aimed to investigate features of BD that precede its formal diagnosis to identify early and undetected signs of BD.


The Clinical Practice Research Datalink (CPRD) is an anonymised primary care electronic patient record database with linkage to secondary data, (Hospital Episode Statistics (HES) including A&E outpatient and hospital admissions). We identified all adult incident BD diagnoses made during years 2010-2017 inclusive by extracting Read codes from CPRD and ICD-10 codes from HES. We applied 1:20 (case: comparators without BD) matching by age, gender and GP practice. Health events of interest prior to the index date were extracted. These included other mental illness diagnoses, prescriptions (antidepressants, antipsychotics, benzodiazepines, Z-drugs, mood stabilisers, pregabalin, gabapentin, and strong opioids), other health events (such as substance abuse, self-harm/suicidal ideation, mood swings, and sleep disturbance) and service interactions (face-to-face consultations, missed appointments, A&E presentations, referral to mental health services). Annual episode incidence for cases and comparators and the odds ratios of cases presenting with each health event prior to the index date were reported.


There were 2,366 incident BD cases and 47,138 individuals in the comparator group (median age 40 years, 60.5% females). Cases had a higher incidence of diagnosis of depression, personality disorder, schizophrenia and anxiety disorder even 10 years before diagnosis of BD compared to comparators. Cases were 8 times more likely to have received 3 different categories of prescription as early as 6 years prior to the index date compared to comparators (OR: 8.4 [95% CI 6.8, 10.6]; p<0.001). The number of face-to-face consultations was higher in all years prior to the index date, with a median [IQR] of 8[15] for cases vs 4[9], p<0.0001 for comparators. At 5 years prior to BD diagnosis, cases were 5 times more likely to miss 6 scheduled appointments in a year compared to comparators.


Potentially useful signals to raise awareness in primary and specialist care are: more than three psychiatric drug prescriptions during the same year, multiple GP attendances in one year, increasing frequency of non-attendance of scheduled appointments. Our advisory board members, including individuals with lived experience, carers, clinicians and third sector services, were in agreement that these signs could aid earlier detection of BD, leading to more timely and appropriate care.

Submitted by: 
Catharine Morgan
Funding acknowledgement: 
This research is funded by the National Institute for Health Research (Research for Patient Benefit programme, PB-PG-0815-20025). The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health