Send message to Author

Short abstract:

This study considers how to balance potentially competing objectives of rapidly processing new information while challenging underlying negative assumptions about communities at risk that contributed to the opioid epidemic.

Long abstract:

The UCSF Opioids Industry Documents Archive (OIDA) contains internal corporate documents released in litigation, and was intended to generate new knowledge regarding the opioid epidemic that could identify structural factors underlying it and motivate changes in clinical practice and policy to prevent future public health harms. As of June 2023, OIDA had expanded to over 3 million documents, in the process creating an archive potentially “too numerous to be processed by human minds.” In past industry documents research in smaller archives, identifying relevant documents began with searches on key terms and proceeded with snowball searches, in which an initial keyword search is used to identify new relevant keywords. AI guided search methods have been proposed as an alternative strategy to review the large volume of documents contained in OIDA. However archivists have cautioned that AI guided search methods are not sufficiently sophisticated to completely replace human review, in part because they may reify underlying human biases. For example, authors of documents have referred to historically excluded groups in ways that are dismissive and harmful; including identifying sexual and gender minorities as “scum” (tobacco) and suggesting that women enjoy “Tupperware parties” (opioids). This study considers how to balance potentially competing objectives of rapidly processing new information while challenging underlying negative assumptions about communities at risk that contributed to the opioid epidemic. It calls for a reflective role for researchers and engagement with impacted communities in conducting research using these archival materials.