Named Drug Burden in FAERS Reports: Seriousness, Reporter Mix, and Geography
A review of 11.18 million reports tests the serious-report gradient across named-drug count, reporter qualification, geography, and serious sub-flags.
FDA adverse event reports name a median of two drugs and a mean of 4.56. Reports naming one drug carry a 46.4 percent serious share. The share rises to 61.9 percent for two to four drugs, 74.5 percent for five to nine, and 82.0 percent for ten or more. The gradient remains within United States reports and within the three largest reporter groups. Geography and reporter qualification change the level of every comparison, while named-drug count preserves the ordering. The association therefore survives two reporting-mix checks but does not establish that drug combinations caused serious outcomes.
1. Cohort and counting rules
The cohort contains 11,175,581 report rows received from 24 December 1986 through 30 June 2026. Drug count is the number of drug entries named on a submitted report. It does not establish prescriptions, concurrent exposure, adherence, or patient-level polypharmacy. The serious flag is set in the submitted record and is not an independent clinical adjudication. One report can name several drugs, reactions, and serious sub-flags.
2. Most reports name four or fewer drugs
Single-drug reports account for 4,939,274 records, or 44.2 percent of the cohort. Another 1,889,190 name two drugs, while 1,665,218 name three or four. The upper tail remains material: 859,659 reports name ten to nineteen drugs, and 364,241 name twenty or more. The maximum of 4,113 is retained as a recorded value without assigning a cause to it.
3. The crude serious share rises across every band
The serious flag appears on 2,292,220 single-drug reports and 1,003,625 reports naming ten or more drugs. Percentage comparison controls for the different band sizes. The serious share rises from 46.4 to 61.9, 74.5, and 82.0 percent without a reversal across the four prespecified bands. This is an association between report composition and a reporter-set flag.
4. Reporter qualification changes the level, not the ordering
Consumers file the largest volume in every band. Their serious share rises from 37.8 percent for one named drug to 70.3 percent for ten or more. Physician reports rise from 59.7 to 88.8 percent. Other health-professional reports rise from 48.1 to 89.6 percent. Reporter qualification therefore explains part of the overall level, but the named-drug gradient remains within each major reporter group.
5. Reporting geography changes every denominator
United States reports carry lower serious shares than other or missing-country reports in every drug-count band. Within United States reports, the share rises from 34.1 percent for one drug to 64.2 percent for ten or more. Outside that group, the share rises from 90.9 to 98.4 percent. The non-US level reflects different reporting regimes and case selection. Geography materially shifts the percentages but does not remove the within-group gradient.
6. Hospitalization increases with named-drug burden
Hospitalization appears on 12.8 percent of single-drug reports and 38.3 percent of reports naming ten or more drugs. Death flags rise from 7.8 to 9.8 percent. Life-threatening flags rise from 1.5 to 6.1 percent, while disabling flags rise from 1.0 to 3.6 percent. Other-serious flags rise from 32.6 to 62.8 percent. The components overlap because one report can carry several flags.
7. Supported interpretation
The record supports three conclusions within its reporting limits. Reports naming more drugs carry higher serious shares across the full cohort. The ordering remains within major reporter groups and within United States reports, so those two measured mix effects do not explain the entire association. The archive cannot separate clinical complexity, reporting completeness, selection, and causal drug interactions. Comparative drug analysis should therefore stratify or adjust for named-drug burden rather than treat every report as equivalent.
8. Limitations
Reports can be incomplete, duplicative, revised, and influenced by litigation or publicity. Drug entries can include suspect, interacting, and concomitant roles. The count does not distinguish those roles here. Reporter qualifications and countries contain missing values, and the two stratifications do not control age, disease severity, indication, exposure duration, or reporting organization. Serious sub-flags are reporter supplied and overlap. Later quarterly extracts can revise counts and report contents.