Improving the Management of Patients’ Own Medications in Hospitalized Older Adults: A Multidisciplinary Before-And-After Intervention Study Authors: Xiaoyan Chen, Lin Liu, Jie Zhang, Linghua Zheng, Xiaona Chen, Mingxia Zhang Published: 2026 ABSTRACT AbstractPurposePatients’ own medications (POMs) may pose safety risks when their use is not integrated into hospital medication-management processes. This study evaluated a multidisciplinary intervention for POM management in hospitalized older adults.Patients and MethodsThis single-center before-and-after intervention study was conducted from March 2024 to February 2025. The intervention comprised five coordinated components: physician-led assessment, authorization, order entry, guidance, and monitoring; pharmacist-led verification, order review, and record maintenance; nurse-led verification, supervision, and education using color-coded labels (green, approved for use; red, not to be used; yellow, awaiting confirmation); IT-led integration of POM information into the electronic medication administration record; and procurement-led upgrades to storage equipment. The primary outcome was the incidence rate of POM-related medication errors per 1000 patient-days. Secondary outcomes were correct POM documentation within 24 hours of admission and correct POM storage at the pre-discharge assessment.ResultsThe study included 162 patients, with 81 in each period. Before the intervention, 174 POM-related medication errors occurred over 1628 patient-days (106.9 per 1000 patient-days), compared with 29 errors over 1581 patient-days after the intervention (18.3 per 1000 patient-days). The incidence rate ratio was 0.172 (95% CI, 0.116–0.254; P<0.001), representing an 82.8% reduction. Correct POM documentation within 24 hours increased from 53.1% (43/81) to 92.6% (75/81; χ2=31.951, P<0.001; OR=11.047, 95% CI, 4.319–28.253; RR=1.744, 95% CI, 1.408–2.160). Correct POM storage at the pre-discharge assessment increased from 32.1% (26/81) to 84.0% (68/81; χ2=44.707, P<0.001).ConclusionThe intervention was associated with improvements in all three measured POM outcomes at this center. Multicenter studies with longer follow-up are needed to assess generalizability and sustainability.Keywords: older adults, self-administered medication, medication management, patient safety, multidisciplinary collaboration Introduction Patients’ own medications (POMs) are drugs that patients bring into the hospital and use during their stay rather than medications dispensed by the hospital pharmacy.1 POMs are common among hospitalized older adults, with reported prevalence ranging from 27% to 45%.2 Although POM use may reduce public pharmaceutical costs,3 inadequate integration into hospital medication systems creates safety risks. Reported hazards include administration by family members, use of unlabeled pillboxes, and bypassing pharmaceutical oversight, which can contribute to missed doses, incorrect doses, duplicate therapy, inappropriate storage, and use of expired medications.4,5 Older adults frequently have multimorbidity and polypharmacy and may also have age-related cognitive or functional limitations. Despite relevant guidance, implementation remains inconsistent, partly because the responsibilities of physicians, nurses, and pharmacists are not always clearly delineated. Our institution previously had no standardized POM management process. Accordingly, this study aimed to evaluate whether a multidisciplinary POM management intervention was associated with changes in the POM-related medication error rate, the accuracy of POM documentation within 24 hours of admission, and correct storage at the pre-discharge assessment. Materials and Methods Study Design and Setting This single-center, before-and-after intervention study was conducted in the Department of Rehabilitation Medicine at Hangzhou Wuyunshan Hospital (Hangzhou Institute of Health Promotion) from March 1, 2024, to February 28, 2025. Patients admitted from March through August 2024 comprised the pre-intervention group, and patients admitted from September 2024 through February 2025 comprised the post-intervention group; thus, the groups consisted of different patients observed in two consecutive 6-month periods. Each participant was observed throughout hospitalization until discharge, and no post-discharge follow-up was conducted. The study was reported in accordance with the SQUIRE 2.0 (Standards for Quality Improvement Reporting Excellence) guidelines. Participants Inclusion criteria were: (1) age ≥65 years; (2) diagnosis of a chronic condition, including coronary heart disease, hypertension, diabetes, chronic heart failure, emphysema, chronic cholecystitis, lung cancer, or osteoporosis; and (3) use or carriage of POMs during hospitalization. Exclusion criteria were: (1) a diagnosis of Alzheimer’s disease without dedicated caregiver support; and (2) inability to manage POMs safely, as determined by a physician based on the patient’s ability to accurately state the medication name, dose, and method of administration, correctly select and handle the medication, and report being able to take the medication as prescribed. Participants were discontinued from the study and excluded from the final analysis if their hospital stay was fewer than 5 days or if they withdrew consent during the study. No a priori sample-size calculation or power analysis was performed. The sample size was determined by the predefined 6-month study periods; all consecutive eligible patients admitted during each period were included, yielding 81 patients in each group. Pre‑Intervention Group (March–August 2024) During the pre‑intervention period, routine care was provided as follows: physicians inquired about patients’ medication histories and prescribed POMs when deemed necessary. No pharmacist was involved in the process. POMs were stored in patients’ bedside cabinets without quantity or dosage labeling, and no electronic tracking records were maintained. Post-Intervention Group (September 2024–February 2025) A multidisciplinary team was established, comprising physicians, nurses, pharmacists, information technology (IT) engineers, and equipment procurement officers. Intervention 1: Physician-Led Prescribing and Authorization Physicians assessed patients’ medication regimens after admission. When continued use of a POM was considered necessary, the physician assessed the patient’s capacity for self-administration, provided information about POM use and obtained the patient’s agreement, entered the corresponding medication order, documented the decision in the medical record, and provided medication-use guidance and monitoring. Capacity for self-administration was assessed by asking the patient to state the medication name, dose, and method of administration and by observing the patient selecting and handling the medication. The patient was considered capable of self-administration when these descriptions and operations were accurate and the patient reported being able to take the medication as prescribed. No formal scale, checklist, or separate capacity-assessment record was used. Cognitive status, visual function, and manual dexterity were not assessed using separate standardized instruments; these factors were considered only insofar as they affected the patient’s responses or observed medication handling. Intervention 2: Pharmacist‑Led Admission Medication Reconciliation Pharmacists checked and registered all POMs within 24 hours of admission, updated and maintained medication information, reviewed POM orders issued by physicians, and provided necessary pharmaceutical counseling to patients. Intervention 3: Nurse‑Led Inspection, Supervision, and Education A three‑color coded labeling system (red, yellow, and green) was implemented for POMs. Red labels were affixed to expired, damaged, or unsafe medications, with instructions for patients to take them home or discard them. Yellow labels were applied to medications without an active order or pending confirmation; these were temporarily placed in a dedicated storage cabinet and designated for withheld use. Green labels were attached to medications that met prescription requirements and quality standards; these were placed in designated POM boxes for routine administration per orders. Nurses supervised medication adherence based on daily medication lists, delivered educational sessions, and observed for adverse reactions. Intervention 4: IT Department-Led Electronic Integration A “Patients’ Own Medications” module was incorporated into the electronic medication administration record (eMAR). Under the hospital rule used during the post-intervention period, the module was required whenever a patient used POMs and was activated for all 81 post-intervention patients. It automatically generated medication barcodes and administration lists after the physician issued the POM order. The activation criterion and use of the module did not vary by physician or department. The IT department conducted regular system inspections and maintenance. Intervention 5: Equipment Procurement Department‑Led Storage Upgrades Dedicated POM storage cabinets and medication boxes were installed in the wards, and the procurement department periodically evaluated their utilization and functionality. Outcome Measures Incidence of POM-Related Medication Errors During Hospitalization Medication errors were identified through direct observation by the responsible nurse, who classified each observed event. Duplicate or excessive administration was defined as an actual administration frequency above that prescribed (eg, twice daily when once daily was ordered) or an actual single dose/quantity above that prescribed (eg, two tablets when one tablet was ordered). Underdosing was an actual single dose/quantity below that prescribed, and a missed dose was an actual administration frequency below that prescribed. Incorrect timing was administration more than 1 hour later than the prescribed clock time. An improper administration method was any method inconsistent with the medical order or medication instructions, including incorrect dosage-form handling (eg, swallowing a medication intended to be chewed or dissolved in the mouth, or splitting or crushing an extended-release tablet), taking a medication after food when fasting administration was required, taking together or in the wrong sequence medications that were required to be separated or ordered sequentially, or using an inappropriate liquid vehicle. No independent second review was performed. Identified errors were reported to the supervising physician and head nurse. Incidence was calculated as (total number of errors / total patient-days) × 1000 and expressed as events per 1000 patient-days. Accuracy of POM Records within 24 hours of Admission A record was classified as correct only when all specified information—medication name, strength, dose, method/route, and frequency—was recorded within 24 hours and agreed with both the pharmacist’s POM verification/registration and the medications observed at the bedside. The pharmacist first checked and registered the POMs within 24 hours; the physician then issued the POM order, the pharmacist reviewed it, and the nurse cross-checked the order against the pharmacist’s registration, printed the medication list, compared it with the actual bedside medications, and provided medication education. The rate was calculated as (number of patients with all POM information correctly documented within 24 hours / total number of enrolled patients) × 100%. Proportion of Patients with Correct POM Storage at the Pre-Discharge Assessment The outcome was based on one inspection by the responsible nurse on the day before discharge. Correct storage required every POM to comply with the hospital’s pharmaceutical storage standards, including designated placement, intact packaging, clear labeling, valid expiration dates, and correct application of the three-color coding system. The assessment was recorded in the medication-quality management file. In the post-intervention group, pharmacists also inspected POMs at admission, and a nurse serving as the medication quality-control team leader conducted monthly spot checks; however, these checks were management and quality-control activities and were not counted as outcome assessments. The rate was calculated as (number of patients whose POMs all met the storage standards at the pre-discharge assessment / total number of enrolled patients) × 100%. Statistical Analysis Statistical analyses were performed using IBM SPSS Statistics version 27.0. Continuous variables following a normal distribution are presented as mean ± standard deviation (SD), and categorical data as counts (percentages). Between‑group comparisons for continuous variables were conducted using independent‑samples t‑tests, and for categorical variables using chi‑square (χ2) tests. The incidence of POM‑related medication errors (events per 1000 patient‑days) was compared using Poisson regression. In the model, the number of errors in each group served as the dependent variable, group assignment as the fixed factor, and the natural logarithm of total patient‑days in each group was included as an offset to adjust for differences in exposure time between groups. Effect sizes are reported as incidence rate ratios (IRRs) with 95% confidence intervals (CIs). The significance level was set at α = 0.05. Between-group comparisons of the 24-hour POM record accuracy rate were performed using the chi-square test. Effect sizes are reported as odds ratios (ORs) with 95% CIs. A supplementary risk ratio (RR) was also calculated with the pre-intervention group as the reference; therefore, an RR > 1 indicates a higher probability of correct documentation in the post-intervention group. The significance level was set at α = 0.05. Results Baseline Characteristics of Participants Before and After the Intervention A total of 162 geriatric patients were enrolled between March 2024 and February 2025, with 81 patients in the pre‑intervention group and 81 in the post‑intervention group. No statistically significant differences were observed between the two groups regarding age, sex, educational level, total number of chronic diseases, total number of patients’ own medications (POMs), or primary caregiver status (all P > 0.05). Detailed baseline characteristics are presented in Table 1. Table 1Baseline Demographic and Clinical Characteristics of the Pre-Intervention and Post-Intervention GroupsCharacteristicPre-Intervention (n = 81)Post-Intervention (n = 81)P valueAge, years, mean ± SD81.75 ± 9.8983.84 ± 7.560.136Sex, male/female, n46/3551/300.423Education, high school or above/junior high school or below, n62/1957/240.374Length of hospital stay, days, mean ± SD20.1 ± 7.719.52 ± 6.690.353Number of chronic conditions, mean ± SD3.14 ± 1.373.01 ± 1.160.539Primary caregiver, yes/no, n42/3944/370.753Number of POM types, mean ± SD2.52 ± 0.982.42 ± 0.930.838Note: Data are presented as mean ± SD for continuous variables and n for categorical variables. Between-group comparisons were performed using independent-samples t-tests for continuous variables and chi-square tests for categorical variables.Abbreviations: POM, patients’ own medications; SD, standard deviation. Baseline Demographic and Clinical Characteristics of the Pre-Intervention and Post-Intervention Groups Note: Data are presented as mean ± SD for continuous variables and n for categorical variables. Between-group comparisons were performed using independent-samples t-tests for continuous variables and chi-square tests for categorical variables. Abbreviations: POM, patients’ own medications; SD, standard deviation. Comparison of POM‑Related Medication Error Rates During Hospitalization In the pre‑intervention group, the total hospital stay amounted to 1628 patient‑days, during which 174 medication errors were recorded, yielding an incidence rate of 106.9 errors per 1000 patient‑days. In the post‑intervention group, the total hospital stay was 1581 patient‑days, with 29 errors documented, corresponding to an incidence rate of 18.3 errors per 1000 patient‑days. The intervention was associated with a significantly lower overall POM-related medication error rate (IRR = 0.172, 95% CI, 0.116–0.254; P < 0.001), corresponding to an 82.8% reduction in the incidence rate. By error type, duplicate or excessive administration decreased from 24 events (14.7 per 1000 patient-days) to 2 events (1.3 per 1000 patient-days), missed-dose or underdosing errors from 130 events (79.9 per 1000 patient-days) to 25 events (15.8 per 1000 patient-days), incorrect-timing errors from 14 events (8.6 per 1000 patient-days) to 1 event (0.6 per 1000 patient-days), and improper-administration-method errors from 6 events (3.7 per 1000 patient-days) to 1 event (0.6 per 1000 patient-days). The rate differences were statistically significant for the first three error types (P < 0.001, P < 0.001, and P = 0.012, respectively), but not for improper-administration-method errors (P = 0.103). Detailed results are summarized in Table 2. Table 2POM-Related Medication Error Incidence Rates Before and After the InterventionError TypePre-Intervention, n (rate)Post-Intervention, n (rate)Rate Reduction (%)IRR95% CIP valueDuplicate or excessive administration24 (14.7)2 (1.3)91.40.0860.020–0.363<0.001Missed dose or underdosing130 (79.9)25 (15.8)80.20.1980.129–0.304<0.001Incorrect timing14 (8.6)1 (0.6)92.60.0740.010–0.5590.012Improper administration method6 (3.7)1 (0.6)82.80.1720.021–1.4260.103Total174 (106.9)29 (18.3)82.80.1720.116-0.254<0.001Note: Rates are events per 1000 patient-days (1,628 patient-days before and 1581 patient-days after the intervention). Rate reduction was calculated as (1 - IRR) x 100. IRRs, 95% CIs, and P values were estimated using Poisson regression with the log of patient-days as an offset.Abbreviations: CI, confidence interval; IRR, incidence rate ratio; POM, patients’ own medication. POM-Related Medication Error Incidence Rates Before and After the Intervention Note: Rates are events per 1000 patient-days (1,628 patient-days before and 1581 patient-days after the intervention). Rate reduction was calculated as (1 - IRR) x 100. IRRs, 95% CIs, and P values were estimated using Poisson regression with the log of patient-days as an offset. Abbreviations: CI, confidence interval; IRR, incidence rate ratio; POM, patients’ own medication. Comparison of 24‑Hour POM Record Accuracy Between the Two Groups The accuracy rate of POM records within 24 hours of admission was 92.6% (75/81) in the post-intervention group, compared with 53.1% (43/81) in the pre-intervention group (χ2 = 31.951, P < 0.001). The unadjusted OR for correct documentation in the post-intervention group vs the pre-intervention group was 11.047 (95% CI, 4.319–28.253). Using the pre-intervention group as the reference, the supplementary RR was 1.744 (95% CI: 1.408–2.160), indicating that the probability of correct documentation was 1.744 times as high in the post-intervention group. The results are shown in Table 3. Table 3Correct POM Documentation within 24 Hours of Admission Before and After the Intervention, n (%)GroupTotal, nCorrect, n (%)Incorrect, n (%)Unadjusted OR (95% CI)RR (95% CI)Chi-squareP valuePre-intervention8143 (53.1)38 (46.9)Ref.Ref.31.951<0.001Post-intervention8175 (92.6)6 (7.4)11.047 (4.319–28.253)1.744 (1.408–2.160)Note: The unadjusted OR and RR compare the post-intervention group with the pre-intervention group for correct POM documentation.Abbreviations: CI, confidence interval; OR, odds ratio; POM, patients’ own medication; RR, risk ratio. Correct POM Documentation within 24 Hours of Admission Before and After the Intervention, n (%) Note: The unadjusted OR and RR compare the post-intervention group with the pre-intervention group for correct POM documentation. Abbreviations: CI, confidence interval; OR, odds ratio; POM, patients’ own medication; RR, risk ratio. Comparison of Correct POM Storage at the Pre-Discharge Assessment and Reasons for Noncompliance At the pre-discharge assessment, the proportion of patients with correct POM storage was 32.1% in the pre-intervention group and 84.0% in the post-intervention group; the corresponding noncompliance rates were 67.9% and 16.0%, respectively. Significant between-period differences were observed in the absence of a designated storage location, damaged packaging or loose tablets, and unclear or missing labels (all P<0.05). The difference in expired or deteriorated medications was not statistically significant (P=0.367). Detailed findings are presented in Table 4. Table 4POM Storage Management at Pre-Discharge Assessment and Reasons for Noncompliance, n (%)ItemPre-Intervention (n=81)Post-Intervention (n=81)Chi-squareP valueCorrect storage management26 (32.1)68 (84.0)44.707<0.001No designated storage location28 (34.6)9 (11.1)12.645<0.001Damaged packaging or loose tablets13 (16.0)2 (2.5)8.8900.003Expired or deteriorated medications4 (4.9)1 (1.2)-0.367aUnclear or missing labels10 (12.3)1 (1.2)7.9000.005Note: Storage status was assessed once, one day before discharge. Pearson’s chi-square test was used unless otherwise indicated. aFisher’s exact test was used because of sparse expected cell counts.Abbreviation: POM, patients’ own medication. POM Storage Management at Pre-Discharge Assessment and Reasons for Noncompliance, n (%) Note: Storage status was assessed once, one day before discharge. Pearson’s chi-square test was used unless otherwise indicated. aFisher’s exact test was used because of sparse expected cell counts. Abbreviation: POM, patients’ own medication. Discussion POM management in hospitalized older adults is a persistent patient-safety challenge. Multimorbidity, polypharmacy, and age-related cognitive and functional limitations can increase vulnerability to medication-related harm.6 In this single-center before-and-after study, a multidisciplinary intervention combining physician authorization, pharmacist verification, nurse supervision and education, eMAR integration, color-coded labels, and dedicated storage was associated with improvements in the POM-related medication error rate, 24-hour documentation accuracy, and correct storage at the pre-discharge assessment. The observed overall POM-related medication error rate decreased from 106.9 to 18.3 per 1000 patient-days (IRR = 0.172, 95% CI, 0.116–0.254; P < 0.001), corresponding to an 82.8% reduction in the incidence rate. The predominance of missed-dose or underdosing errors at baseline is consistent with literature describing adherence barriers, regimen complexity, forgetfulness, and functional limitations in medication self-management among older adults.7–9 Hammad et al reported that pharmacist-led medication reconciliation reduced medication discrepancies in older patients; however, their intervention and outcomes differed from ours, so the effect sizes are not directly comparable.10 Correct POM documentation within 24 hours increased from 53.1% to 92.6%, with an OR of 11.047 (95% CI: 4.319–28.253). With the pre-intervention group as the reference, the corresponding RR of 1.744 (95% CI: 1.408–2.160) indicates a higher probability of correct documentation after the intervention. Wang et al reported improved POM-order accuracy after a root-cause-analysis intervention,11 providing directionally consistent but setting-specific evidence. Medication reconciliation depends on a structured review of the patient’s complete medication profile,12,13 and pharmacist-led reconciliation can identify medication discrepancies at admission and discharge.10,14 Correct storage at the single pre-discharge assessment increased from 32.1% to 84.0%. Previous research has documented problems such as incomplete packaging, unclear labeling, and inappropriate household medication storage.15 Al-Shareef et al examined the disposal of expired and unused medicines in Riyadh.16 Marwa et al examined household disposal practices in Mwanza.17 Both studies were conducted in community settings rather than inpatient settings and therefore provide only indirect context for POM storage in the present study. In our study, dedicated cabinets and boxes, standardized inspections, and the three-color labeling process were implemented together; the before-and-after design cannot determine the effect of any single component. The practical contribution of the intervention is the integration of information collection, storage management, and medication administration across physicians, pharmacists, nurses, IT staff, and equipment procurement staff. This integrated workflow is compatible with the broader hospital medication-reconciliation literature.18 Because the intervention was implemented as a bundle, component-specific effects could not be determined. Future refinement may consider medication literacy and self-management capacity, which are relevant to hospitalized older adults.19,20 Study Limitations This study has several limitations. First, it was conducted in one department at a single center, limiting generalizability. Second, the pre-intervention and post-intervention groups comprised different patients enrolled in consecutive periods. The groups were compared only with respect to measured baseline characteristics; no matching, stratified analysis, or multivariable adjustment was performed. Temporal trends and unmeasured confounding therefore cannot be excluded. Third, medication errors were classified by the responsible nurse without an independent second review, and outcome-assessor blinding was not feasible. Fourth, the post-intervention period lasted 6 months, participants were followed only until discharge, and longer-term sustainability was not evaluated. Finally, no cost-effectiveness analysis was performed. Conclusions In this single-center before-and-after study, a multidisciplinary POM management intervention was associated with a lower POM-related medication error rate and higher rates of correct 24-hour documentation and correct storage at the pre-discharge assessment. Because the two groups comprised different patients and longer-term sustainability, costs, and implementation in other settings were not evaluated, these findings should be confirmed in multicenter studies with longer follow-up. Funding Statement Hangzhou Health Science and Technology Plan (ZD20230125). Clinical Trial Registration The National Medical Research Registry (ID: MR-33-23-035548). Data Sharing Statement Aggregated data supporting the reported results are included in this article and its tables. De-identified patient-level data and other unpublished materials are not publicly available. Requests for additional data may be submitted to the corresponding author for case-by-case consideration, subject to approval by the hospital and, where applicable, the ethics committee. Approval and release of additional data cannot be guaranteed. Patient Consent statement All participants were informed about the study protocol and provided written informed consent to participate in the study. All methods were performed in accordance with the relevant guidelines. All procedures were performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. Author Contributions Xiaoyan Chen and Lin Liu conducted the study, participated in data collection, and drafted the manuscript. Jie Zhang and Linghua Zheng contributed to the study design and performed the statistical analysis. Xiaona Chen and Mingxia Zhang participated in the acquisition, analysis, or interpretation of data and contributed to drafting the manuscript. Xiaoyan Chen,Lin Liu contributed equally to this work as the first author Jie Zhang, Linghua Zheng contributed equally to this work as the second author Xiaona Chen, Mingxia Zhang contributed equally to this work as the third authors. All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. Disclosure The authors declare that they have no competing interests. References Associated Data Data Availability Statement Aggregated data supporting the reported results are included in this article and its tables. De-identified patient-level data and other unpublished materials are not publicly available. Requests for additional data may be submitted to the corresponding author for case-by-case consideration, subject to approval by the hospital and, where applicable, the ethics committee. Approval and release of additional data cannot be guaranteed.