Table 1.

Key definitions of electronic health records (EHRs) and family violence

EHRs Routinely collected healthcare data stored and accessible electronically as part of routine care or as part of administrative databases for wider patient, practice and billing management purposes.
Coded indicators of EHRs Single or combination of codes aimed at identifying different forms of family violence (table 2). Codes must be based on a recognised classification system (eg, International Classification of Disease system) assigned to patients’ EHRs by a clinician or by a professional coder based on medical records (eg, discharge documentation).
Family violence outcomes Any of the below outcomes obtained independently from the coded indicator (ie, not using other codes in the EHR to validate the indicator):
Child maltreatment (CM) Any act of commission or omission by a parent or caregiver resulting in harm, the potential for harm or threat of harm to a child (0–12 years) or an adolescent (13–19 years) including neglect, psychological, physical, sexual and emotional abuse. Harm does not need to be intended.6–8 6–8
Neonatal abstinence syndrome (NAS) or fetal alcohol syndrome (FAS) Statutory guidelines definition of CM in the UK include any form of maternal omission or commission during pregnancy with potential for significant biopsychosocial harm to the future child such as NAS and FAS.8 This may include compliance with prescribed medication (clinician approved) and/or maternal substance misuse during pregnancy, resulting in drug withdrawal symptoms in newborns and/or harm to the future child's biopsychosocial functioning/appearance.
Intimate partner violence (IPV) Any incident of threatening behaviour, violence or abuse (psychological, physical, sexual, financial or emotional) between adults who are, or have been, intimate partners or family members.7 In this review, studies are predominately restricted to estimates of women aged 15–50 years.
Eligible methods for obtaining independent reference standards
Independent and blinded/non-blinded manual chart review of full medical, social and/or criminal records (either as extracted or by recoding charts for direct comparison with originally assigned codes).
The individual meets criteria defined by a validated instrument (eg, self-report survey) completed within 1 month of the coded EHR.
Clinician confirms case via a self-report survey administered to the service.
Linkage of the same individual to another independent non-health-related database that provides concordance of the diagnosis by a qualified professional (eg, social worker and police).
Linkage of the same individual to another healthcare database that provides concordance of the diagnosis (eg, similar diagnosis recorded in both hospital discharge and general practitioner records).
Level of certainty for distinguishing cases of family violence from non-cases in the information obtained by the reference standard Quality ratings correspond to the level of certainty that the reference standard is a true measure of CM, NAS, FAS or IPV. Ratings 1–5 apply to CM, 1–4 apply to NAS and FAS and 1–3 apply to IPV.	Quality rating
CM confirmed at case conference or family, civil or criminal court proceedings; admitted by the perpetrator; or witnessed abuse AND non-CM actively excluded by stated criteria (eg, witnessed accidental cause, caused by metabolic bone diseases etc).	5 (high)
CM confirmed by stated criteria including multidisciplinary assessment AND non-CM actively excluded by stated criteria.	4
CM/IPV defined by stated criteria AND source verifying non-CM/IPV merely stated.	3
CM/IPV stated but no supporting detail is given AND source verifying non-CM/IPV merely stated.	2
Suspected CM/IPV AND no criteria stated for verifying non-CM/IPV.	1 (low)

The criteria for assessing studies of EHRs was adapted from McCormick et al,¹³⁴ and the quality ratings for the level of certainty of CM in the reference standard were adapted from Kemp et al.²⁹ The complete quality assessment tool is provided in online supplementary table S3.