Find information about the General Election on Tuesday, November 5, 2024.
Allegheny County, like many communities across the country, does not have sufficient long-term supportive housing to serve every person or family meeting the criteria for homelessness. Although the County’s Continuum of Care (CoC) includes approximately 210 bridge housing beds, 940 rapid rehousing beds and 1810 supportive housing beds, the demand for housing is greater than the supply. To allocate available housing, the CoC’s coordinated entry system has depended upon a widely adopted but not locally validated actuarial tool that relies upon self-reported information that is highly personal and dependent upon the person's memory and trust. As part of an ongoing effort to improve decisions at key points in its systems, Allegheny County worked with local stakeholders, research partners (Auckland University of Technology) and data science ethicists (Eticas) to develop the Allegheny Housing Assessment (AHA).
The AHA is a decision support tool designed to help prioritize admissions to supportive housing services for individuals or families experiencing homelessness. The tool uses administrative data from Allegheny County’s data warehouse to predict the likelihood of three types of events occurring in a person‘s life if they remain un-housed over the next 12 months: a mental health inpatient stay, a jail booking and frequent use (4 or more visits) of hospital emergency rooms. These events serve as indicators of harm if a person remains un-housed. Like the previous assessment tool, the AHA assigns a risk score that is used as part of the housing prioritization process, but it is far more accurate and equitable and doesn’t require the time or trauma associated with asking sensitive questions at the time of housing crisis.
The AHA continues to evolve as we learn more about the impact of the assessment and the practicalities of the business processes. A solicitation for an independent evaluation of the AHA will be issued in early 2021 and we will report on the results, which may also result in further modifications to the model and/or the business processes. All of the data and feedback will be used not just to better prioritize housing resources but also to shed new light on the unmet need for housing for high-risk individuals.