5Ms Summary: A Single‐Center Retrospective Study in an Acute Geriatric Ward Authors: Yosuke Matsui, Tomomichi Sakai, Yuichiro Kusuhara, Wang Shujie, Tomihiko Tajima, Yosuke Yamada, Kazuhisa Watanabe, Chisato Fujisawa, Hirotaka Nakashima, Hitoshi Komiya, Hiroyuki Umegaki Published: 2026 The Geriatric 5Ms framework structures patient‐centered geriatric care [1], but routine implementation in Japan is rarely described. We embedded a 5Ms Summary in the electronic health record (EHR) of an acute geriatric ward. To our knowledge, this is the first quantitative report of routine inpatient 5Ms documentation in Japan. We examined domain completion and fulfillment to characterize incomplete documentation. At Nagoya University Hospital, the Summary was intended to be updated weekly, but editing access was limited to physicians under our institutional EHR configuration. We retrospectively evaluated the last Summary documented on or before discharge for each admission episode; patient outcomes were not examined. Of 139 admission episodes between September 1, 2024, and June 30, 2025, 135 (97.1%) had a 5Ms Summary and were included; four had none. Median patient age was 85 years (interquartile range [IQR], 81–89; range, 63–101), median length of stay was 18 days (IQR, 11–33.5), and 63 patients (46.7%) were women. Admission categories were emergency admission (105 [77.8%]), elective admission (16 [11.9%]), intrahospital transfer (8 [5.9%]), and interhospital transfer (6 [4.4%]). The most common primary diagnoses were pneumonia (35 [25.9%]), cerebral infarction (12 [8.9%]), and urinary tract infection (7 [5.2%]). For each patient, completion indicated whether a domain was documented at least once; fulfillment was the proportion of template‐defined sub‐items documented. Both measure documentation, not care quality or depth. The template contained 5 Multicomplexity, 3 Mobility, 5 Mind, 7 Medications, and 5 Matters Most items (25 total; Table S1). Pairwise completion differences were assessed using McNemar tests with Holm correction; Wilson 95% confidence intervals are shown in Figure 1A. Figure 1A shows completion, and Figure 1B shows per‐Summary fulfillment distributions. Completion ranged from 69.6% for Matters Most to 99.3% for Multicomplexity. Median per‐Summary fulfillment was 52.0% (IQR, 36–72); median domain fulfillment was 80.0% (IQR, 60–100) for Multicomplexity and 20.0% (IQR, 0–60) for Matters Most. Medications completion was 88.1%, whereas median fulfillment was 42.9%; prescription cascade and medications without a clear indication were rarely documented. Although ADLs and IADLs were documented at least once in 83.7% and 76.3% of Summaries, explicitly current values at discharge were recorded in only 10.4% and 3.7%, respectively. (A) Completion rate of each 5Ms domain. This bar graph shows the completion rate for each of the five domains in the final 5Ms Summary recorded before discharge for each patient (n = 135). Error bars indicate 95% Wilson confidence intervals. Completion was defined as 1 if one or more sub‐items were documented for a domain, and 0 if none were documented. Asterisks indicate statistically significant pairwise differences between domains (McNemar test with Holm correction, p < 0.05). This figure (A) presents the aggregate, domain‐level frequency with which each domain was documented at least once. (B) Distribution of fulfillment rates for each 5Ms domain. This figure displays the distribution of fulfillment rates for all final 5Ms Summaries recorded before discharge and for each domain, on a per‐record basis, using violin and box plots. The violin plot illustrates the distribution of fulfillment rates, while the box plot shows the interquartile range and median. “Total” was calculated using all 25 sub‐items as the denominator, and each domain was calculated using the prespecified sub‐items of that domain as the denominator. The fulfillment rate was defined as the number of documented sub‐items divided by the total number of prespecified sub‐items, multiplied by 100. In contrast to (A), (B) shows the distribution of fulfillment across individual records (n = 135); violin plots are used because fulfillment rates were non‐normally distributed and heterogeneous across records, which a single summary statistic would obscure. The Summary was designed to organize information from comprehensive geriatric assessment, not to serve as a separate assessment instrument. It could support multidisciplinary communication, but we did not evaluate its use by nonphysician clinicians or its effect on shared understanding. Matters Most had the lowest completion rate; its median fulfillment was also low. Relevant information may emerge gradually through repeated conversations and across professions. Our data cannot determine whether information was never documented or recorded elsewhere without being transferred to the Summary. Care aligned with patient priorities is associated with lower treatment burden and less unwanted health care [2]. For the advance care planning and goals‐of‐care components, EHR documentation can be difficult to record, locate, and carry forward [3]. Japanese qualitative work in multidisciplinary, multifacility palliative care similarly identified difficulty understanding patients' wishes and sharing information [4]. The US Age‐Friendly 4Ms have been translated into structured EHR documentation measures [5] and replicated with local adaptation [6]. Reported barriers include aggregating assessments and communicating across encounters and professions; proposed solutions include a single EHR location, improved cross‐encounter communication, and age‐specific prompts [7]. Limitations include the single‐center design and local workflow, staffing, and physician‐only editing access. Analyses were conditional on Summary creation and excluded four admission episodes without one. Only the last pre‐discharge Summary was evaluated, not weekly updating; retrospective classification may have misclassified ambiguous entries. Different numbers of items per domain (3–7) limit direct fulfillment comparisons. Documentation cannot establish care quality, prognosis, clinical outcomes, or multidisciplinary use. Analysis of the EHR‐embedded Summary revealed domain‐level differences; Matters Most had the lowest completion and low median fulfillment. These gaps do not establish that corresponding care was absent. Together with prior implementation studies, our findings support testing whether clearer responsibility for eliciting, summarizing, and updating information improves documentation of patients' priorities. Author Contributions Y.M., T.S., and H.U. conceived and designed the study. All authors contributed to data collection and analysis. Y.M., T.S., and H.U. drafted the manuscript, and all authors critically revised the manuscript and approved the final version. Ethics Statement The study protocol was reviewed and approved by the Institutional Review Board of Nagoya University Graduate School of Medicine (approval no. 2025‐0243). Conflicts of Interest The authors declare no conflicts of interest. Supporting information Table S1: Documentation rates of all 25 sub‐items of the Geriatric 5Ms Summary (n = 135).