Origin and perspective | Definition/inclusion and method of review | Review stage 1 | Review stage 2 | Criterion/trigger | Sample size and time frame for inclusion |
---|---|---|---|---|---|
“Harvard Medical Practice Study” (HMPS)[1, 16, 17] with subsequent modifications [2–6, 18, 19] | An unintended injury or complication that results in disability at discharge, death or prolonged hospital stay and is caused by healthcare management rather than the patient’s underlying disease | Generally one reviewer per record | Mostly two physician reviewers per record | General for both methods: | Random, big samples to measure the incidence and to generalise the result |
Medicolegal and focus on negligence the first studies and thereafter quality improvement and preventability perspective | Includes both omission and commission | Screening for one of 18 criteria by trained nurses (can be other professionals) | Detailed independent review | An indication that patient harm may have occurred | An AE had to have occurred before and during and detected during and/or after index admission |
Adult, inpatients, often exclusion of e.g. psychiatric and rehabilitation patients | Comprehensive reading | Assess the AE by using different scales according to e.g. causation, severity, preventability, timing, causes, and types | Directs the medical reviewer to relevant parts of the records by the notes | ||
No assessment, generally only registration of found criteria, description of the potential AE, and a brief summary of the admission | Generally includes only one AE per patient i.e. the most severe | Some criteria/triggers are AEs by definition e.g. healthcare-associated infections | Different inclusion periods before and after index admission | ||
Research method | Two [1, 2] - three [6, 18] stage retrospective record review | No time limit | No time limit | Positive criteria/triggers may be without connection to patient harm i.e. false positive | |
HMPS method: | |||||
18 broad criteria | |||||
“Global Trigger Tool”[7] | Unintended injury resulting from or contributed to by medical care that requires additional monitoring, treatment or hospitalisation, or that results in death | Two reviewers per record | The team discuss the findings together | 54 triggers, mostly narrow | Random, small samples sufficient for the design of safety work over time |
Quality improvement tool for clinical practice | Includes commission, excludes omission | First screening independently for one of 54 triggers by trained nurses (can be other professionals), focus on triggers, no comprehensive reading, reads just relevant parts related to found triggers; second, consensus | One physician, who does not generally review the record but does authenticate the consensus findings of the AEs, the severity rating, and answer questions from reviewers in review stage 1 | The Swedish version [20] contains 53 triggers, the triggers “restraint use” in the care module and “other” in the medication module were excluded, and a trigger, “occurrence of any postoperative complication“, was added in the surgical module | 10 records every second week or 20 records every month per hospital |
Track AE rate over time in a hospital or a clinic | Adult, inpatients, exclusion of psychiatric and rehabilitation patients | Finds triggers, describes the potential AE, and categorise harm according to NCC MERP index [21] | The physician is the final arbitrator | Length of stay at least 24 hours | |
Two stage retrospective record review | No assessment of preventability | All identified AEs are included | An AE had to have occurred before and during and detected during and/or after index admission | ||
Maximum 20 minutes per record | No time limit | 30 days inclusion period before and after index admission | |||
The Swedish version includes the same preventability scale as used in HMPS methodology [2, 20] |