Background and objective This project links population data to the Wisconsin Violent Death Reporting System (WVDRS) to determine the extent to which firearm possession criteria are being followed as well as the potential impact of the adoption of proposed possession criteria.

Design and study population Criminal justice data for WVDRS homicide suspects and victims and suicide decedents 2008–2011 and a sample of matched control group of driver’s license holders (to characterise the state population) will be abstracted.

Methods Individual legal possession statuses (prohibited/not prohibited) under each current and expanded criterion will be determined. Proportions of interest will be calculated from two-way contingency tables, and tests between groups with categorical variables (eg, criterion is met or not) will be performed with Fisher’s exact or binomial proportion tests. Tests between groups with continuous variables (eg, number of misdemeanours) will be performed by zero inflated negative binomial regression. Area under the receiver operating characteristic curve will be used to quantify the prediction accuracy of specific univariate or multivariate logistic model for prediction. Inverse probability weighting will be used for analyses that extend from matched controls to the general state population of license holders.

Discussion Linked data sets and partnerships are challenging, but necessary for comprehensive public health research. Results of this study will contribute knowledge on the proportion of prohibited suspects and suicide decedents that used firearms in violent deaths and, if applying expanded criteria would have increased prohibited persons. This study will also investigate risk and protective factors for being a victim of homicide.