Background: Overnight sleep disruptions in hospitalized patients is a well-known contributor to increased delirium risk and decreased patient satisfaction. Despite decades of work to reduce such disruptions, they persist. Medication dosing is a common type of disruption. Behavioral economics suggest defaults offer a powerful nudge to change outcomes. There is no evidence that default dosing times in EHRs were designed to minimize overnight disruptions. As part of a larger effort to reduce medication dosing burden on nurses, we sought to explore the feasibility of adjusting default times to not only reduce nursing burden but also reduce overnight disruptions to patients.
Purpose: We sought to explore default dosing time change on common medication frequencies to both reduce nursing burden and overnight dosing that would interrupt patient sleep.
Description: We exported all medication doses across 11 hospitals in June 2025. We then filtered to the most frequent standing doses, excluding PRNs and “once” dosings as sleep-friendly defaults are not applicable. This left 253,597 medication doses. We then created a heat map to visually analyze the data along with objective measures to assess efficacy of interventions. In current state, medications were administered during 11 of the 24 potential med passes throughout the day with a burden of 7 med passes on the day nurse and 4 on the night nurse with 74% of all administrations falling to the day nurse. We identified “key times” as breakfast, lunch, dinner, evening, and before bed – in current state 75% of medication administration was during key times. Finally, 4.9% of all doses occurred during sleep time (define as 2200-0600).We analyzed several different potential future states across the same metrics to find a better solution. We trialed two AI-generated solutions (OpenAI, GPT-4o) and a human-generated solution; the latter proved to be most effective across all measures.The future state included the following changes to key dosing times: daily 0900->0800; q12Hs 0900/2100->0800/2000; BID 0900/1700->0800/2000; q8Hs 0000/0800/1600->2200/0600/1400; TID 0900/1600/2100->0800/1400/2200; q6Hs 0000/0600/1200/1800->0200/0800/1400/2000; and QID 0900/1300/1700/2100->0800/1200/1700/2200.With these changes, total number of hours requiring administration drops from 11 to 8. Day nurse med pass drops from 7 to 4 with night nurse med pass maintaining at 4 sweeps. Total % of medications administered became more equitable with % day nurse dropping from 75% to 66%. There is increased consolidation during key times from 75% to 87%. Finally, percent of doses during patient sleep time drops from 4.9% to 2.2%. This correlates with 6,847 fewer overnight medication dosings per month — >82,000 per year across 11 hospitals.
Conclusions: By evaluating common medication default frequency times and optimizing to reduce nursing burden and patient overnight disruption, these new default times can serve a dual goal of offloading nurses toward other high-yield clinical activities while dramatically reducing unnecessary overnight disruptions to patients. We are working to make the necessary EHR change to implement these changes and will study important patient sensitive outcomes. We encourage other institutions to explore their own defaults as if they were adopted from foundation builds, they likely have defaults in need of adjustment to align with patient-centric dosing times.