Developing and testing the predictive validity of household firearm storage measures: insights from rural Alaska.

Introduction: Measuring change in firearm storage is paramount to evaluating if interventions influence storage. Yet, there is little empirical basis for how to measure this change. This methodology study compared three different firearm storage measures using data from the Family Safety Net trial (n=46), a randomised controlled trial among firearm-owning adults in a rural Alaska Native community to encourage unloaded and locked firearm storage.

Methods: The team compared baseline and 1-month follow-up measures of household firearm storage: (1) cumulative firearm accessibility, (2) average firearm accessibility, and (3) number of unlocked firearms. All measures included information about locking status; the cumulative and average firearm accessibility measures incorporated additional information about ammunition. Unadjusted and adjusted generalised estimating equations evaluated the associations of the storage measures with adult perceptions of youth firearm access. Receiver operating characteristic curves from unadjusted models compared the measures' abilities to predict adult perceptions of youth firearm access.

Results: All household firearm storage measures were sensitive to change between baseline and 1 month follow-up (ps<0.05). These measures demonstrated a similar and moderate ability to predict adult perceptions of youth access (area under the curve range: 0.68-0.76).

Conclusion: The number of unlocked firearms is an easily interpretable outcome when evaluating the effect of firearm storage interventions on storage practices, and it performed similarly well in predicting adults' perceptions of youth firearm access as compared with more complex measures. Research that evaluates the ability of firearm storage measures to predict actual youth access will inform interventions to prevent youth firearm access.