ADS Capstone Chronicles Revised

‭3‬

‭(FDA,‬‭2018b).‬‭Healthcare‬‭professionals‬‭and‬ ‭consumers‬ ‭can‬ ‭voluntarily‬ ‭report‬ ‭adverse‬ ‭events‬ ‭related‬ ‭to‬ ‭suspected‬ ‭drug‬ ‭ingestion‬ ‭and‬ ‭these‬ ‭are‬ ‭called‬ ‭direct‬ ‭reports.‬ ‭In‬ ‭contrast,‬‭drug‬‭manufacturers‬‭and‬‭distributors‬ ‭are‬ ‭mandated‬ ‭reporters‬ ‭and‬ ‭must‬ ‭submit‬ ‭reports‬ ‭under‬ ‭certain‬ ‭circumstances.‬ ‭If‬ ‭a‬ ‭consumer‬‭or‬‭healthcare‬‭professional‬‭sends‬‭a‬ ‭report‬ ‭to‬ ‭a‬‭pharmaceutical‬‭entity,‬‭that‬‭entity‬ ‭is‬ ‭required‬ ‭to‬ ‭submit‬ ‭an‬ ‭exact‬ ‭copy‬ ‭of‬ ‭the‬ ‭report‬ ‭to‬ ‭FDA.‬ ‭In‬ ‭some‬ ‭cases,‬ ‭this‬ ‭creates‬ ‭duplicate‬ ‭reports‬ ‭in‬ ‭FDA‬ ‭records‬ ‭(FDA,‬ ‭2018b).‬‭The‬‭mandated‬‭reports‬‭fall‬‭under‬‭two‬ ‭categories:‬ ‭expedited‬ ‭reports‬ ‭are‬ ‭those‬ ‭that‬ ‭“contain‬ ‭at‬ ‭least‬ ‭one‬ ‭adverse‬ ‭event‬ ‭that‬ ‭is‬ ‭not‬ ‭currently‬ ‭described‬ ‭in‬ ‭the‬ ‭product‬ ‭labeling‬ ‭and‬ ‭for‬ ‭which‬ ‭the‬ ‭patient‬ ‭outcome‬ ‭is‬‭serious”‬‭(FDA,‬‭2023b),‬‭and‬ ‭nonexpedited‬ ‭reports‬ ‭which‬ ‭include‬ ‭serious/expected‬ ‭events,‬ ‭nonserious/unexpected‬ ‭events,‬ ‭and‬ ‭nonserious/expected‬‭events.‬‭The‬‭fourth‬‭type‬ ‭of‬‭report‬‭is‬‭a‬ ‭BSR‬ ‭which‬‭stands‬‭for‬‭biologic‬ ‭safety‬ ‭report.‬ ‭Figure‬ ‭1‬‭displays‬‭the‬‭increase‬ ‭in‬ ‭adverse‬ ‭event‬ ‭reporting‬ ‭since‬ ‭the‬ ‭early‬ ‭2000s,‬ ‭broken‬ ‭down‬ ‭by‬ ‭report‬ ‭type‬ ‭(figure‬ ‭from FAERS dashboard; FDA, 2023b).‬ ‭Figure 1‬ ‭Report Type Frequency‬

‭downloads.‬ ‭For‬ ‭example,‬ ‭in‬ ‭2022,‬ ‭there‬ ‭were‬‭2.3‬‭million‬‭adverse‬‭events‬‭reported,‬‭of‬ ‭which‬‭175,000‬‭resulted‬‭in‬‭death‬‭(figure‬‭from‬ ‭FAERS‬ ‭dashboard;‬ ‭FDA‬ ‭2023b;‬ ‭Figure‬ ‭2).‬ ‭At‬ ‭the‬ ‭time‬ ‭of‬ ‭this‬ ‭project,‬ ‭the‬ ‭database‬ ‭contains‬ ‭28.7‬ ‭million‬ ‭reports‬ ‭of‬ ‭which‬‭15.9‬ ‭million‬ ‭are‬ ‭serious‬ ‭and‬ ‭2.6‬ ‭million‬ ‭are‬ ‭related to death (Figure 2; FDA, 2023b).‬ ‭Figure 2‬ ‭ADR Outcomes by Seriousness‬

‭The‬ ‭FAERS‬ ‭reporting‬ ‭has‬‭some‬‭limitations.‬ ‭First,‬ ‭the‬ ‭data‬ ‭is‬ ‭correlational‬ ‭so‬ ‭causation‬ ‭cannot‬ ‭be‬ ‭inferred.‬ ‭The‬ ‭reports‬ ‭are‬ ‭not‬ ‭verified‬ ‭in‬ ‭any‬ ‭way‬ ‭and‬ ‭underlying‬ ‭health‬ ‭conditions‬ ‭are‬ ‭not‬ ‭assessed.‬ ‭Second,‬ ‭true‬ ‭rates‬ ‭of‬ ‭occurrence‬ ‭cannot‬ ‭be‬‭inferred‬‭from‬ ‭this‬ ‭data‬ ‭due‬‭to‬‭low‬‭reporting‬‭rates.‬‭Serious‬ ‭or‬ ‭death-related‬ ‭incidents‬‭are‬‭reported‬‭more‬ ‭often‬ ‭than‬ ‭nonserious‬ ‭events,‬ ‭but‬ ‭overall,‬ ‭under-reporting‬ ‭rates‬ ‭are‬ ‭around‬ ‭94%‬ ‭(Hazell‬ ‭&‬ ‭Shakir,‬ ‭2006).‬ ‭Some‬ ‭reporting‬ ‭information‬ ‭is‬ ‭ambiguous‬ ‭or‬ ‭missing‬‭which‬ ‭detracts‬‭from‬‭the‬‭usefulness‬‭of‬‭the‬‭database.‬ ‭Additionally,‬ ‭consumers‬ ‭may‬ ‭not‬‭know‬‭that‬ ‭this‬ ‭reporting‬ ‭system‬ ‭exists,‬ ‭and‬ ‭may‬ ‭not‬ ‭realize‬‭that‬‭an‬‭adverse‬‭event‬‭is‬‭drug‬‭related,‬ ‭so‬ ‭true‬ ‭population‬ ‭rates‬ ‭go‬ ‭undetected.‬ ‭To‬ ‭overcome‬ ‭these‬ ‭limitations,‬ ‭additional‬ ‭sources‬ ‭of‬ ‭data‬ ‭can‬ ‭be‬ ‭combined‬ ‭with‬ ‭FAERS‬ ‭data‬ ‭to‬ ‭create‬ ‭a‬ ‭complete,‬ ‭robust‬ ‭dataset‬‭for‬‭machine‬‭learning‬‭and‬‭subsequent‬ ‭insights.‬ ‭2.1 Problem‬ ‭With‬ ‭the‬ ‭rise‬ ‭in‬ ‭chronic‬ ‭disease‬ ‭(CDC,‬ ‭2024)‬ ‭and‬ ‭new‬ ‭drug‬ ‭developments,‬ ‭the‬ ‭impact‬ ‭of‬ ‭drug‬ ‭adverse‬ ‭events‬ ‭represents‬ ‭a‬ ‭large‬‭public‬‭health‬‭burden‬‭that‬‭will‬‭continue‬

‭The‬ ‭FDA‬ ‭FAERS‬ ‭was‬ ‭modernized‬ ‭and‬ ‭developed‬ ‭in‬ ‭2012‬ ‭to‬ ‭track‬ ‭post-marketing‬ ‭adverse‬‭reactions‬‭to‬‭FDA-approved‬‭products‬ ‭(e.g.,‬ ‭drugs,‬ ‭devices,‬ ‭cosmetics)‬ ‭in‬ ‭the‬ ‭community‬ ‭(FDA,‬ ‭2018b).‬ ‭Reports‬ ‭are‬ ‭submitted‬ ‭electronically‬ ‭and‬ ‭the‬ ‭data‬ ‭of‬ ‭adverse‬ ‭events‬ ‭related‬ ‭to‬ ‭pharmaceutical‬ ‭drugs‬‭(i.e.,‬‭prescribed,‬‭over-the-counter,‬‭and‬ ‭therapeutic‬ ‭biologics)‬ ‭is‬ ‭subsequently‬ ‭made‬ ‭publicly‬ ‭accessible‬ ‭via‬ ‭application‬ ‭programming‬ ‭interfaces‬ ‭(APIs)‬ ‭or‬ ‭file‬

153

Made with FlippingBook - Online Brochure Maker