Identification of mental health and quality of life outcomes in primary care databases in the UK: a systematic review.

OBJECTIVES: To summarise the definitions and combinations of codes used to identify outcomes of anxiety, depression, fatigue, cognitive dysfunction (including mild cognitive dysfunction and dementia), sexual dysfunction, pain, sleep disorders, and fatal and non-fatal self-harm in studies using electronic health records from primary care databases in the UK. DESIGN: Systematic review. DATA SOURCES: Medline, Embase and lists of publications of the main primary care databases in the UK. ELIGIBILITY CRITERIA: Included data from a UK primary care database and studied outcome(s) of interest. DATA EXTRACTION AND SYNTHESIS: We abstracted information on the outcomes definition and codelists. When necessary, authors were contacted to request codelists. RESULTS: 120 studies were eligible. Codelists were available for 17/42 studies of depression; 21/41 studies of fatal and non-fatal self-harm; 17/27 studies of dementia/cognitive dysfunction; 5/12 studies of anxiety; 4/8 studies of pain; 3/6 studies of fatigue and sexual dysfunction; 1/2 studies of sleep disorders. Depression was most often defined using codes for diagnoses (37/42 studies) and/or antidepressants prescriptions (21/42 studies); six studies reported including symptoms in their definition. Anxiety was defined with codes for diagnoses (12/12 studies); four studies also reported including symptoms. Fatal self-harm was ascertained in primary care data linked to the Office for National Statistics mortality database in nine studies. Most studies of cognitive dysfunction included Alzheimer's disease, and vascular and frontotemporal dementia. Fatigue definitions varied little, including chronic fatigue syndrome, neurasthenia and postviral fatigue syndrome. All studies of sexual dysfunction focused on male conditions, principally erectile dysfunction. Sleep disorders included insomnia and hypersomnia. There was substantial variability in the codelists; validation was carried out i21/120 studies. CONCLUSIONS: There is a need for standardised definitions and validated list of codes to assess mental health and quality of life outcomes in primary care databases in the UK.