Pragmatical assessment of six evaluation scales for subjective components of functional impairment in assault survivors based on EHRs mining and the identification of five profiles of victims: a controlled observational study (Preprint)

BACKGROUND Most researches considering relationships between violence and health focus on associations between a type of violence and a specific medical diagnosis. Little is known about the functional consequences of violence, directly assessed as a primary outcome. Yet, WHO acknowledged the importance of considering a functional approach of health (2001) when releasing the International classification of functioning, disabilty and health (ICF). The evaluation of victims’ functional impairment relies both on objective, subjective and contextual elements, as when they report symptoms of fear, pain or the circumstances of the assault. Using scales to quantify these subjective elements should facilitate their interpretation in a judicial context. OBJECTIVE To 1) elaborate a typology of violence based on medical observations reported and available through electronical medical records (EHRs) and 2) based on this typology, to characterize and 3) assess 6 scales measuring subjective elements of functional impairment among victims of violence. METHODS We conducted a retrospective, monocentric study included assaults’ survivors examined in a French department of forensic medicine over 12 months. Typology was built based on the mining of EHRs and the use of pattern recognitions algorithms. The optimal number of types of violence was searched requiring a approach based on robustness and stability of clusters, using a sample/re-sample and multi-metric scheme. We then paired patients based on their belonging to the same homogeneous profile and evaluated the intra-rater and inter-rater reproducibility of the scales. RESULTS We found five remarkably stable profiles of situations of violence, consistent with clinical practice. All pain, fear and life-threat scales were significantly associated with higher functional impairment, suggesting that their measure contributes to the global assessment of functional impairment. Intra and inter reproducibility of scales among the same situations of violence profiles were measured from mild to good (coefficient of concordance 0.46 to 0.66 and 0.43 to 0.66, respectively). Victims reporting intimate partner violence showed higher scores in fear and perception of a life threat both during assault and medical interview while victims reporting battery by multiple, unknown assailants presented higher scores only in perception of a life threat during assault. CONCLUSIONS We could elaborate a well defined 5-profile typology of violence as encountered in medical practice, based on observations and robust EHRs mining techniques. These profiles provide a more comprehensive characterization of real-life situations of violence, which means they could be use as a typology for further research on violence rather than only focussing on a specific type of violence. Pain, fear and life-threat scales were related to functional impairment in accordance with expert knowledge, and showed in real life conditions a fair reproducibility for similar situations of violence. Subjective elements related to functional impairment in assault survivors can be quantified using Likert scales during medical interviews.