Table 1.

Observational studies of HSU (n=17)

Study	Design	Data source	Participants					HSU outcomes	Predictors of increased HSU	Not predictive of increased HSU
			Composition	N	Age	Per cent f	Q ST
Button et al26	Cohort	NHS eating disorders clinic	Patients with eating disorder	147	p	96	9	Total HSU	–	▸ Type of eating disorder diagnosis
Byford et al27	Cohort	NHS primary care database	Patients with depression	88 935	44.4 (SD=16.75)	68	18	▸ A&E attendances ▸ GP phone calls ▸ GP visits ▸ Inpatient days ▸ Medication usage ▸ Other specialist contacts ▸ Psychiatrist contacts ▸ Psychologist contacts	▸ Non-remission (after antidepressant treatment)	▸ Remission (after antidepressant treatment)
Chollet et al28	Cohort	NHS primary care database	Patients with GAD	29 131	48.5 (SD=17.5)	67	18	Total HSU	▸ Aged 31–49 ▸ Aged 50–64 ▸ High previous HSU ▸ High previous medication use ▸ Male ▸ Two comorbidities	▸ Aged 18–30 ▸ Aged >65 ▸ Lower previous HSU ▸ Lower previous medication use ▸ Female; No, one, or three comorbidities
Coid et al29	Cross-sectional	Adult Psychiatric Morbidity Survey	UK residents (some with BPD)	8397	16–74 (M not stated)	53	18	▸ Community psychiatric nurse contacts ▸ Counsellor contacts ▸ GP contacts for psychological problems ▸ Psychiatric inpatient admission ▸ Psychiatrist contacts ▸ Total HSU	▸ Diagnosis of BPD	▸ No diagnosis of BPD
Coid et al30	Cross-sectional	Adult Psychiatric Morbidity Survey	UK residents with a PD	626	16–74 (M not stated)	56	17	▸ Community psychiatric nurse contacts ▸ Counsellor contacts ▸ GP contacts for psychological problems ▸ Medication usage ▸ Psychiatric inpatient admission ▸ Psychiatrist contacts	▸ Cluster A, B and C: PD diagnoses ▸ Comorbid mental disorder and substance abuse	▸ No comorbidity
Cooper et al31	Cross-sectional	Adult Psychiatric Morbidity Survey	UK residents (some with CMPs)	7461	16+ (M not stated)	51	18	▸ PCT attendance ▸ GP contact for psychological problems ▸ Medication usage	▸ Aged >35 ▸ ADLs ▸ Widowed/divorced/separated ▸ Elevated neurotic symptoms ▸ Female ▸ Non-white ethnicity	▸ Aged <35 ▸ No ADLs ▸ Marital status other than widowed/divorced/separated ▸ Non-elevated neurotic symptoms ▸ Male ▸ White ethnicity ▸ Any home ownership status ▸ Number of qualifications
Cooper et al32	Cross-sectional	Adult Psychiatric Morbidity Survey	UK residents (some with CMPs)	22 196	16+ (M not stated)	52	19	▸ PCT attendance ▸ GP contact for psychological problems ▸ Medication usage	▸ Aged 35–54 ▸ Aged 75+ ▸ Divorced/separated/ widowed ▸ Elevated neurotic symptoms ▸ Female ▸ Non-home owner ▸ Non-white ethnicity	▸ Aged 16–34 ▸ Aged 55–74 ▸ Marital status other than widowed/divorced/separated ▸ Male ▸ Non-elevated neurotic symptoms ▸ Home owner ▸ White ethnicity
Foster et al33	Cross-sectional	Government surveys: adults in private households; adults with psychosis in households and adults in mental institutions	Adults with psychosis	470	16–64 (M not stated)	NS	10	▸ GP contact for psychological problems ▸ Psychiatric inpatient admission ▸ Any service use for a psychological problem	▸ Aged 16–34 ▸ Aged 45–54 ▸ Elevated neurotic symptoms ▸ White ethnicity	▸ Aged 35–44 ▸ Aged 55–64 ▸ Family circumstances ▸ Gender ▸ Household type ▸ Living arrangements ▸ Marital status ▸ Non-elevated neurotic mental health symptoms ▸ Non-white ethnicity ▸ Occupation ▸ Physical illness ▸ Qualifications
Hayward et al34	Cohort	Postal survey of a general practice population	GP attendees	2662	51.3 (SD=17.18)	55	16	▸ GP contacts ▸ Medication usage	▸ Insomnia symptoms ▸ Comorbid anxiety or depression	▸ No insomnia symptoms ▸ No comorbidity
Keene and Rodriguez35	Cross-sectional	Databases: health authority, mental health population, and A&E population	Health and mental health service users	625 964	16+ (M not stated)	52	16	A&E attendances	▸ Four typologies: (1) Young, male frequent attendees with self-harm and other injuries; (2) Young females with self-harm; (3) Older patients with multiple medical conditions; (4) Very old patients with cardiac conditions and fractures	–
Knapp et al36	Cross-sectional	Maudsley (NHS) psychiatric hospital data; patient interviews	Former Maudsley child and adolescent psychiatric patients	149	Not stated	61	11	▸ Client Services Receipt Inventory24	▸ Comorbidity of childhood depression and conduct disorder	▸ No childhood comorbidity
Mohan et al37	Cohort	PRISM psychosis study set in Maudsley & Bethlem NHS trust area	White (group 1) and African-Caribbean (group 2) patients with psychosis	140	40.55 (SD=14.9)	49	18	Client Services Receipt Inventory24	▸ Receiving intensive community treatment (for African-Caribbean patients only)	▸ Ethnicity ▸ Receiving intensive community treatment (for White patients only)
Patel et al38	Cross-sectional data from a RCT	RCT data set in South London/Maudsley NHS trust area	Patients with schizophrenia	85	26 (SD not stated)	26	13	▸ ‘Other’ ▸ A&E attendances ▸ CMHT contacts ▸ Community psychiatric nurse contacts ▸ Day care attendances ▸ General medical ward attendances ▸ GP contacts ▸ Group PCT attendances ▸ Home carer visits ▸ Inpatient admissions ▸ Inpatient days ▸ Non-psychiatric outpatient attendances ▸ Nurse contacts ▸ Occupational therapist contacts ▸ Psychiatric outpatient attendances ▸ Psychiatrist contacts ▸ Psychologist contacts ▸ Sheltered workshop attendances ▸ Specialist education attendances ▸ Total inpatient service use	▸ Cognitive deficits	▸ Antisocial behaviour ▸ Depression symptoms ▸ No cognitive deficits ▸ Positive symptoms ▸ Social withdrawal
Torres et al39	Cross-sectional	Adult sychiatric morbidity survey	UK residents with OCD	114	16–74 (M not stated)	65	13	▸ Any community service attendance ▸ Counselling attendance ▸ GP contact for psychological problems ▸ Home carer visits ▸ Medication usage ▸ PCT attendance ▸ Psychiatric inpatient admission ▸ Community psychiatric nurse contacts ▸ Psychiatric outpatient attendances ▸ Psychiatrist contact ▸ Psychologist contact ▸ Support group attendances ▸ Total HSU (‘any kind of treatment’)	▸ OCD diagnosis ▸ OCD with comorbid anxiety or depression	▸ No OCD diagnosis
Ullrich and Coid40	Cross-sectional	Adult Psychiatric Morbidity Survey	UK residents with ASPD	245	16–74 (M not stated)	22	16	▸ Community psychiatric nurse contacts ▸ GP contacts ▸ Other nursing service contacts ▸ Outreach worker contacts ▸ Psychiatric inpatient admission ▸ Psychiatrist contacts ▸ Psychologist contacts ▸ Support group attendances ▸ Total HSU	▸ Comorbid Axis 1 mental disorders	▸ Comorbid personality disorders
Walters et al41	Cohort	Seven NHS general practices	Primary care patients with mild-to-moderate distress	250	46 (SD not stated)	71	20	GP contacts	▸ ICD-10 disorders (apart from mixed anxiety and depression)	▸ Mixed anxiety and depression
Wright et al42	Cross-sectional	NHS mental health services	Patients with functional psychosis and comorbid substance abuse	61	43.1 (SD not stated)	56	15	▸ Inpatient admissions ▸ Inpatient days	–	▸ Dual diagnosis

A&E, accident and emergency; ADLs, activities of daily living restrictions; Age, mean age (if not stated, where possible, age range is stated); ASPD, antisocial personality disorder; BPD, borderline personality disorder; CMHT, community mental health team; CMP, common mental health problem; DBT, dialectical behaviour therapy; F, female; GAD, generalised anxiety disorder; GP, general practitioner; HSU, health service utilisation; M, mean; NHS, National Health Service; NS, not stated; OCD, obsessive compulsive disorder; PCT, psychotherapy; PD, personality disorder; Q, quality assessment; RCT, randomised controlled trial; ST, STROBE statement (score range 0–22; 0 represents lowest quality and 22 represents highest quality).¹⁴