Table 2. Study characteristics.

Reference	Population	Method of record selection	Method of review	More than one adverse event per patient?	Time frame of included events	Kappa value for interobserver agreement	Endpoints
1. Brennan et al25,32	Acute-care hospital patients (no psychiatric patients)	Random sample of hospitalisations from 51 hospitals	Two-stage record review: first, screening for one of 18 criteria by trained nurses; second, review by two physicians	Not specified	Occurred before and during and detected during index admission	0.61	Incidence, negligence, outcome, location, type of event
2. O’Neil et al23,28	Hospital patients	All admissions to the medical service of one hospital over a 4-month period	1. Two-stage record review: first, screening for one of 15 criteria by medical-record analysts; second, review by physician	No	Not specified	0.57	Incidence, preventability
			2. Reviewing reported incidents
3. Wilson et al3	Acute-care hospital patients (no psychiatric or day-care patients)	Random sample of admissions from 28 hospitals	Two-stage record review: first, screening for one of 18 criteria by trained nurses; second, review by two medical officers	No	Occurred before* and during and detected during or after index admission	0.55	Incidence, preventability, outcome, provider of care, location, type of event
4. Thomas et al2	Hospital patients (no psychiatric, rehabilitation or drug/alcohol treatment patients)	Random sample of discharges from 28 hospitals	Two-stage record review: first, screening for one of 15 criteria by trained nurses; second, review by physician	No	Occurred before and during and detected during index admission	0.4	Incidence, negligence, outcome, provider of care, location, type of event
5. Vincent et al26,33	Acute hospital patients (general medicine, general surgery, orthopaedic surgery, obstetrics)	Records randomly drawn from two hospitals	Two-stage record review: first, screening for one of 18 criteria by trained nurses; second, review by clinician	Yes	Not specified	Not specified	Incidence, preventability, outcome, provider of care, type of event
6. Davis et al24,30	Hospital patients (no psychiatric, day-care or rehabilitation patients)	Random sample of admissions from 13 hospitals	Two-stage record review: first, screening for one of 18 criteria by trained nurses; second, review by medical officer	Not specified	Occurred before* and during and detected during index admission	0.47	Incidence, preventability, outcome, provider of care, location, type of event
7. Baker et al31	Hospital patients (no paediatric, psychiatric, obstetric or rehabilitative)	Random sample of admissions from 20 hospitals	Two-stage record review: first, screening for one of 18 criteria by trained nurses; second, review by physician	Yes	Occurred before and during and detected during or after index admission†	0.47/0.45/0.69‡	Incidence, preventability, outcome, provider of care, type of event
8. Sari et al29	Hospital patients (surgery, general medicine, elderly care, orthopaedics, urology, oncology, ENT, ophthalmology)	Random sample of admissions in one hospital	1. Two-stage record review: first, screening for one of 18 criteria by trained nurses; second, review by physician	Yes	Not specified	0.76	Incidence
			2. Reviewing reported incidents

*Only included if (partly) responsible for index admission.

†Only included if occurred/was detected during a hospital admission.

‡Kappa values for interobserver agreement of judgement on injury/causation/preventability.