From Evidence to Impact: Bridging the Implementation Gap in Geriatric Deprescribing

Orenstein et al. [1] present a prospective cohort study of considerable significance, leveraging 23 years of follow-up to demonstrate the high prevalence of potentially inappropriate prescribing (PIP) in community-dwelling older adults and its robust association with all-cause mortality, encompassing both potentially inappropriate medications (PIM) and prescribing omissions (PPO). This work provides some of the strongest evidence to date positioning PIP as a critical risk factor for adverse outcomes in aging populations. However, the focus must now shift from “identifying the problem” to effectively translating these findings into systematic deprescribing actions within clinical practice. Polypharmacy has become the norm in older adults, and recent studies indicate that PIP is strongly associated with adverse drug events, functional decline, hospitalization, and mortality [2]. The work of Orenstein et al. further demonstrates that PIMs exert greater harm among older adults with good baseline health, challenging the conventional focus on frail individuals and highlighting PIP as a pervasive silent threat. Although the updated 2023 Beers criteria [3] and the STOPP/START criteria [4] provide robust tools for identification, the implementation of systematic deprescribing remains inadequate. This gap between knowledge and practice has become a critical barrier to medication safety in older adults, underscoring the urgent need to translate recognition into effective action. Systematic reviews and trials show that interventions often yield inconsistent improvements in medication counts and PIMs, reflecting a persistent gap between recognition and implementation [5]. Barriers exist at multiple levels. At the physician level, clinical inertia, concern about adverse outcomes after discontinuation, and limited time are major obstacles [6]. At the patient level, medication dependence and fear of altering established regimens are common [7]. At the system level, fragmented care, lack of incentives, and absence of standardized pathways hinder progress [8]. Thus, listing problems alone is insufficient; implementation science is required to redesign processes and environments to achieve routine deprescribing. Transforming recognition into action requires parallel efforts in technology, teamwork, and education. First, technological support should move beyond simple alerts to advanced clinical decision support systems (CDSS) fully integrated with electronic health records, capable of generating individualized deprescribing and substitution recommendations [9]. Second, teamwork is essential; multidisciplinary models, particularly pharmacist-led medication reviews and medication therapy management (MTM), have proven effective in reducing PIP [10]. Third, education and culture must be strengthened by embedding deprescribing into medical curricula and promoting shared decision-making to enhance patient and family engagement. Together, these strategies establish a closed loop from identification to implementation, fostering the routine and precise integration of deprescribing into clinical care. Although the study by Orenstein et al. provides a strong signal, several methodological limitations merit attention for future research. First, the identification of PPO is prone to misclassification if clinical indications, contraindications, intolerance, patient preferences, and temporal context are not adequately characterized. This may blur the distinction between justified non-prescription and true omission, leading to biased effect estimates. Integrating outpatient records, hospital admissions, pharmacy dispensing, laboratory and imaging data, and nursing assessments, together with localized rules and validation, can substantially improve the accuracy and clinical relevance of research findings. Second, a single baseline measurement of renal function cannot capture the long-term dynamics of older adults and may lead to misclassification of PIM exposure, thereby compromising the accuracy of mortality risk estimates. Repeated follow-up assessments, time-dependent modeling, and sensitivity analyses can substantially reduce error and enhance both the robustness and clinical interpretability of research findings. Third, despite adjustment for common covariates, residual confounding factors such as frailty, cognitive decline, fall risk, prescribing habits, and socioeconomic status, as well as confounding by indication including disease severity and comorbidity complexity, may still bias estimates of the association between PIM or PPO and mortality. Incorporating detailed measures of clinical severity, applying high-dimensional propensity scores, and adopting target trial emulation could substantially reduce bias and enhance validity. Finally, the exclusive reporting of all-cause and non-cancer mortality limits mechanistic insight and clinical applicability. Outcome stratification, competing risk analysis, and mediation approaches may clarify drug-related causal pathways and provide stronger evidence for precise interventions. In summary, we commend the study by Orenstein et al. for reaffirming the serious risks of PIP. Future research should focus on implementation science to identify and validate effective deprescribing strategies. Clinicians must approach deprescribing with caution but firm commitment, while healthcare administrators should ensure sufficient resources and procedural support. Only through coordinated efforts can current evidence be effectively translated into tangible improvements in survival and quality of life for older adults. All authors contributed to the study conception and design. The first draft of the manuscript was written by Songhe Chen. Kai Chen conducted the literature search and revision. Ye Chen provided writing guidance, further critically revised the manuscript. All authors read and approved the final manuscript. The authors declare no conflicts of interest. The authors have nothing to report.