Background: The accurate identification of polypharmacy (≥ 5 drugs) and potentially inappropriate medication (PIM) use is essential for optimizing geriatric care. Conventional approaches—including self-reported questionnaires, institutional prescr...
Background: The accurate identification of polypharmacy (≥ 5 drugs) and potentially inappropriate medication (PIM) use is essential for optimizing geriatric care. Conventional approaches—including self-reported questionnaires, institutional prescription records, and national claims data—are widely used but may not fully capture the actual medication exposure. Specimen-based drug detection using liquid chromatography–tandem mass spectrometry (LC–MS/MS) can provide additional, objective evidence of medication use.
Methods: We developed a laboratory-developed, specimen-based screening assay using LC–MS/MS for the qualitative detection of 100 target drugs. The panel was initially designed to support the evaluation of psychoactive substances and drugs of abuse in emergency settings and was subsequently expanded to include the therapeutic drugs relevant to adverse drug events in older adults, informed by the Beers Criteria and the list of“Medications with clear evidence of caution in the elderly” published by the Health Insurance Review & Assessment Service (HIRA). A total of 213 elderly outpatients were enrolled. Polypharmacy and PIM use were assessed using four tools, self-report, institutional EMR prescription records, HIRA claims data, and LC–MS/MS tests, and the influencing factors were examined. Using LC–MS/MS results as a reference for MS-target drug exposure, the concordance between tools, associated clinical factors, and the validity of the existing tools were evaluated, and the clinical utility of LC–MS/MS was further assessed by calculating the net reclassification improvement (NRI).
Results: For overall medication counts not restricted to the MS-target, the EMR records identified the highest number of medications, followed by HIRA data and self-reports. Agreement across tools was statistically significant (p < 0.001); Cohen’s kappa indicated moderate agreement between institutional EMR and HIRA data but poor agreement for self-reports. For MS-target drugs, polypharmacy and PIM exposure confirmed by LC–MS/MS were observed in 3.3% (7/213) and 12.2% (26/213) of participants, respectively. When MS-target drug exposure was evaluated per drug–specimen pair, the existing tools showed very low sensitivity (1.21–21.37%) with variable specificity (14.29–95.24%) against the LC–MS/MS reference results, and the MS-based approach demonstrated improvement compared with conventional tools as assessed by NRI. Discordant classifications suggested potential contributions from recall bias, over-the-counter (OTC) use or medication sharing, non-reimbursed prescriptions, medical-aid coverage not reflected in claims data, and pharmacokinetic constraints affecting detectability.
Conclusions: LC–MS/MS provided specimen-based evidence of MS-target drug exposure and revealed clinically meaningful discordance with conventional assessment tools. While limited panel coverage and diverse pharmacokinetics in outpatient constrain interpretation and preclude viewing MS as a standalone replacement, MS testing may serve as a valuable complementary approach to strengthen the medication-use assessment and support safer identification of polypharmacy and PIM exposure in older adults.