Policy & Practice August 2018

OPIOIDS continued from page 22

One example of how Allegheny County uses integrated data is to help child welfare workers make informed maltreatment screening decisions on child abuse and neglect cases. The department uses an algorithm to produce a score predicting the likeli- hood a child will be removed from their home because of abuse or neglect. So far, the algorithm has been fairly accurate, and some argue it is saving lives. “Given the early results from Pittsburgh, predictive analytics looks like one of the most exciting innova- tions in child protection in the last 20 years,” Brett Drake, a professor at the Brown School of Social Work at Washington University in St. Louis, recently told New York Times Magazine. This is the kind of revolutionary thinking that is needed to confront the devastating opioid epidemic obliter- ating our country. Governments must act outside of their silos and share information, not only with other gov- ernments, but with the private sector as well. The bad news—challenges still exist and the next overdose death is just around the corner. The good news—we have more potential now than ever to share data and potentially prevent that next death.

there are examples of overcoming cultural challenges. In Allegheny County, PA, the Department of Human Services has brought together more than 20 sources of public and private data to serve more than 1.2 million residents of Greater Pittsburgh. The agency integrated data systems in two phases: (1) several different depart- ments merged under the umbrella of the Department of Human Services, allowing one director to make a decision on integrating many dif- ferent data sources, and (2) the agency aggressively sought to sign MOUs with other public entities to merge data and more effectively serve shared consumer populations. Erin Dalton, a deputy director of the Allegheny County Department of Human Services who oversees ana- lytics and technology, conveyed that the biggest barrier to data sharing is the culture of the social safety net eco- system. Leaders assume it isn’t possible or the right thing to do. In Allegheny County some interagency agreements took up to six years to complete, and laws were not changed nor were new laws written—people simply began interpreting existing laws differently. “Where there is a will, there is a way.”

way for data sharing. Essentially the way systems share informa-

tion is through Application Program Interfaces (APIs). It is not unusual for an addict to be involved with more than a dozen different programs or systems. Historically, data sharing was executed from agency to agency. So if there were 12 agencies, each agency would develop a separate policy and a separate set of APIs to communicate with the others (132 in total) and the majority of those would be duplicative. Today’s technology can replace that with a single core set of services that can be reused by all 12 agencies. But for this to work, one organization has to take a leadership role to identify and manage the services. This manage- ment must be rigorous—it identifies and tracks specific data; access, dates, security, testing, and so on. The ultimate outcome is effective gover- nance and achievement of data sharing that not only adheres to data privacy standards but gives greater confidence to the public and private agencies being asked to share data. Culture: Overcoming a Persistent Barrier Organizational culture continues to be a big barrier to data sharing. But

Policy&Practice   August 2018 38