There is substantial regional variation in Medicare spending for advanced cancer, yet no consistent association between mean regional spending and survival.

In jurisdictions throughout the United States, thousands of sexual assault kits (SAKs; also known as a “rape kits”) have not been submitted by the police for forensic DNA testing. DNA evidence may be helpful to sexual assault investigations and prosecutions by identifying perpetrators, revealing serial offenders through DNA matches across cases, and exonerating those who have been wrongly accused. This paper describes a longitudinal action research project conducted in Detroit, Michigan after that city discovered approximately 11,000 untested sexual assault kits in a police department storage facility. We conducted a root cause analysis to examine individual, organizational, community, and societal factors that contributed to the development of the rape kit backlog in Detroit. Based on those findings, we implemented and evaluated structural changes to increase staffing, promote kit testing, and retrain police and prosecutors so that cases could be reopened for investigation and prosecution. As we conducted this work, we also studied how this action research project impacted the Detroit criminal justice system. Participating in this project changed stakeholders’ attitudes about the utility of research to address community problems, the usefulness of DNA evidence in sexual assault cases, and the impact of trauma on survivors. The results led to new protocols for SAK testing and police investigations, and new state legislation mandating SAK forensic DNA testing.

No abstract

Statistical procedures for variable selection have become integral elements in any analysis. Successful procedures are characterized by high predictive accuracy, yielding interpretable models while retaining computational efficiency. Penalized methods that perform coefficient shrinkage have been shown to be successful in many cases. Models with correlated predictors are particularly challenging to tackle. We propose a penalization procedure that performs variable selection while clustering groups of predictors automatically. The oracle properties of this procedure including consistency in group identification are also studied. The proposed method compares favorably with existing selection approaches in both prediction accuracy and model discovery, while retaining its computational efficiency. Supplemental material are available online.