|
Citation
|
Rothman EF, Campbell JK, Hoch AM, Bair-Merritt M, Cuevas CA, Taylor B, Mumford EA. BMC Pediatr. 2022; 22(1): e337. |
|
Abstract
|
BACKGROUND: Dating abuse (DA) is prevalent and consequential, but no brief DA screening tools are available for use in pediatric or other settings. This study was designed to determine the sensitivity, specificity, and predictive values of the MARSHA-C, which is a three-item DA victimization screening tool.
METHODS: The participants were 224 U.S. youth ages 11-21 years old (20% male, 77% female, 3% non-binary gender). Youth completed an online questionnaire about adolescent relationship abuse. The survey included the Measure of Adolescent Relationship Harassment and Abuse (MARSHA), which is a comprehensive DA measurement instrument normed on a nationally representative sample. Of 34 DA victimization items from the MARSHA, the three most prevalent items were hypothesized to have good predictive validity of the full scale score as a brief, screening version (MARSHA-C). The sensitivity, specificity, positive predictive value, and negative predictive value of the MARSHA-C to identify victims of DA was calculated.
RESULTS: Using the MARSHA as the reference standard, the cutpoint of 1 on the MARSHA-C screening tool was identified as optimal. The MARSHA-C had a sensitivity of 84%, a specificity of 91%, and positive predictive value of 91%. Thus, for youth who endorse ≥ 1 MARSHA-C items, there is a 91% probability that they have experienced DA in the past year. Exploratory analyses by demographic subgroups suggest that the predictive validity of the MARSHA-C is approximately equivalent for females and males, younger and older adolescents, Asian, Black, Latinx, Multiracial and White youth, and heterosexual and lesbian, gay, and bisexual youth.
CONCLUSIONS: The MARSHA-C can be used to detect DA among 11-21-year-old youth via online surveys for research purposes, or in clinical care settings to facilitate proactive patient counseling or parent-oriented anticipatory guidance.
Language: en |