Artikel
Development and validation of the unruptured intracranial aneurysm treatment score (UIATS) dervied from multidisciplinary consensus
Suche in Medline nach
Autoren
Veröffentlicht: | 2. Juni 2015 |
---|
Gliederung
Text
Objective: In view of the existing uncertainty regarding repair versus conservative management of unruptured intracranial aneurysms (UIAs), we endeavoured to develop a UIA treatment scoring (UIATS) model that includes and quantifies key factors involved in clinical decision-making in the management of UIAs and to assess agreement for this model among experts in UIA management and research.
Method: An international, multidisciplinary (neurosurgery, neuroradiology, neurology, clinical epidemiology) group of 69 experts was convened to develop and validate the UIATS model using a Delphi consensus. For internal (39 panel members involved in identification of relevant features) and external agreement (30 independent external reviewers), 30 selected UIA cases were used to test on a 5 point Likert Scale (5 indicating strong agreement) the level of agreement of reviewers with UIATS management recommendations. We determined inter-rater agreement (IRA) with standardized coefficients of dispersion (vr*) (vr*=0 indicating excellent and vr*=1 poor agreement).
Results: The UIATS comprises 13 different categories and 29 different features. Agreement with UIATS (mean Likert scores) was 4·2 (95%CI:4·1-4·3) per reviewer for both reviewer cohorts; agreement per case was 4·3 (95%CI:4·1-4·4) for panel members and 4·5 (95%CI:4·3-4·6) for external reviewers (p=0·017). Mean Likert scores were 4·2 (95%CI:4·1-4·3) for interventional (n=56) and 4·1 (95%CI:3·9-4·4) for non-interventional (n=12) reviewers (p=0·290). Overall IRA (vr*) for both cohorts was 0·026 (95%CI:0·019-0·033).
Conclusions: Experts in the field have a high level of agreement and IRA with UIATS management recommendations. Clinicians can use the UIATS model to appreciate what experts in the field would advise patients with UIAs. Since many factors in this consensus scoring model are based on case-control studies or assumptions, more empirical data for pivotal factors included in the scoring model are urgently needed.