Validity And Reliability Of The Br(?)set Violence Checklist-Extended

Objective: To extend the Br?set Violence Checklist(BVC) and increase the sensitivity, make it more conform to the national conditions and social-cultural background in China, make it more practical in the clinical practice, and provide an effective violence risk assessment instrument for the psychiatric clinical staff in our country.Methods: There are two stages in this study.Stage?: An item pool was established by literature retrieval and group discussion. The Br?set Violence Checklist-Extended(BVC-E) was determined with Delphi method. 36 experts specializing in the fields of clinical psychiatry, psychiatric clinical nursing, psychiatric nursing management were selected. The expert questionnaires were delivered by express mail service(EMS) and E-mail in two-round expert consultation. The extended items of the scale were determined by results of the expert consultations.Stage?: The validity and reliability of the BVC-E were assessed. 6 experts with more than 30-year clinical working experience were invited to evaluate the content validity of the scale. 24 nurses who had been trained strictly to evaluate 10 cases independently, and the intraclass correlation coefficient(ICC) was used to evaluate the inter-rater reliability of the scale. 679 inpatients from the Sixth People's Hospital of Hebei Province were evaluated every 8 hours by BVC-E. The results of the reality of violent behaviors were used to test the predictive validity. The internal consistency reliability of the scale was evaluated by Cronbach'? coefficient.Results:Part?: 2 items were identified through two-round expert consultation. The BVC-E included 8 items which were confusion, irritability, boisterousness, verbal threats, physical threats, attacks on objects, history of violent attacks, involuntary hospitalization. The response rates of the two rounds were both 100%. The effective rates were separately 100% and 97.22%. The authority coefficients were 0.876 and 0.883. The coordination coefficients were 0.472 and 0.523.Part ?: 691 patients were evaluated. The basic information of 3 cases was incomplete and 9 cases were shedding, so the effective cases were 679. 6111 effective scales were collected in nine shifts. The overall evaluation time was 1-6(2.27±1.05) minutes.1. Validity:(1) Content validity: the item-level CVI(I-CVI) of BVC-E were confusion(0.83), irritability(1.00), boisterousness(0.83), verbal threats(1.00), physical threats(1.00), attacks on objects(1.00), history of violent attacks(1.00), involuntary hospitalization(0.83). The scale-level CVI(S-CVI) was 0.94.(2) Predictive validity: When the outcome variable was the occurrence of any verbal or nonverbal behavior that was threatening to self or others, the best predictive validity was eight hours, and the area under the ROC curve was 0.941(95%CI:0.929-0.953). When the cut-off was 3, sensitivity and specificity were 89.4% and 91.2% respectively, the sensitivity of BVC-E was higher than that of the original scale BVC(89.4% vs. 78.9%, P<0.001). When the outcome variable was the occurrence of any physical behavior that was threatening to the others, the best predictive validity was eight hours, and the area under the ROC curve was 0.974(95%CI:0.970-0.980). When the cut-off was 4, sensitivity and specificity were 92.7% and 93.5% respectively, the sensitivity of BVC-E was higher than that of the original scale BVC(92.7% vs. 74.5%,P<0.001).2. Reliability:(1) Inter-rater reliability: The interclass correlation coefficient(ICC) for the every item were: confusion(0.930), irritability(0.960), boisterousness(0.967), verbal threats(0.956), physical threats(0.951), attacks on objects(0.979), history of violent attacks(0.984), involuntary hospitalization(0.963). The overall ICC was 0.985.(2) Internal consistency reliability: The Cronbach'? coefficient was 0.732.Conclusion: The implementation process of Delphi method is scientific and reasonable. The validity and reliability of BVC-E are good, and the sensitivity of scale is improved after expanded. The scale is easily accepted by nursing staff, and it can be promoted as an effective psychiatric violence risk assessment tool in China.