Search

The present Q&A paper aims at providing an overview of artificial intelligence with a special focus on machine learning as a currently predominant subfield thereof. Machine learning-based applications have been discussed intensely in legal scholarship, including in the field of intellectual property law, while many technical aspects remain ambiguous and often cause confusion. This text was drafted by the Research Group on the Regulation of the Digital Economy of the Max Planck Institute for Innovation and Competition in the pursuit of understanding the fundamental characteristics of artificial intelligence, and machine learning in particular, that could potentially have an impact on intellectual property law. As a background paper, it provides the technological basis for the Group’s ongoing research relating thereto. The current version summarises insights gained from background literature research, interviews with practitioners and a workshop conducted in June 2019 in which experts in the field of artificial intelligence participated.

Background The HIV treatment cascade is a tool for characterizing population-level gaps in HIV care, yet most adaptations of the cascade rely on surveillance data that are ill-suited to drive quality improvement (QI) activities at the facility level. We describe the adaptation of the cascade in health care organizations and report its use by HIV medical providers in New York State (NYS). Methods As part of data submissions to the NYS Department of Health, sites that provide HIV medical care in NYS developed cascades using facility-generated data. Required elements included data addressing identification of people living with HIV (PLWH) receiving any service at the facility, linkage to HIV medical care, prescription of antiretroviral therapy (ART), and viral suppression (VS). Sites also submitted a methodology report summarizing how cascade data were collected and an improvement plan identifying care gaps. Results Two hundred twenty-two sites submitted cascades documenting the quality of care delivered to HIV patients presenting for HIV- or non-HIV-related services during 2016. Of 101 341 PLWH presenting for any medical care, 75 106 were reported as active in HIV programs, whereas 21 509 had no known care status. Sites reported mean ART prescription and VS rates of 94% and 80%, respectively, and 60 distinct QI interventions. Conclusions Submission of facility-level cascades provides data on care utilization among PLWH that cannot be assessed through traditional HIV surveillance efforts. Moreover, the facility-level cascade represents an effective tool for identifying care gaps, focusing data-driven improvement efforts, and engaging frontline health care providers to achieve epidemic control.