Tray Accuracy and the Room-Service Decision at a 320-Bed Hospital
A 320-bed community hospital serves about 2,900 patient meals per day through a traditional trayline with a 12-minute assembly target. Over eight months, food-service patient satisfaction fell from the 78th to the 41st percentile on the hospital survey, tray accuracy is 82%, and 14% of trays arrive more than 15 minutes outside the requested window. Nutrition-assistant call-outs run 18% and overtime costs $9,800 per month; two vacancies remain unfilled. The director proposes a $310,000 conversion to room service with a hostess model, spoken menus, and on-demand cooking. The CFO will approve only if projected payback is 12 months or less and patient satisfaction improves within two quarters. The chef worries that on-demand cooking will worsen ticket times, and nursing wants guaranteed delivery for insulin-dependent patients. The board asks for a recommendation at the next quality committee meeting.
1. What is the most defensible first step: approve the full room-service conversion, reject it, or authorize a pilot? Explain the operational and financial reasoning.
Read the worked answer
Neither full approval nor outright rejection is defensible because the baseline data are confounded. Tray accuracy of 82% and 14% late trays coexist with 18% call-outs and two vacancies, so staffing and scheduling may explain much of the failure. A $310,000 conversion would consume capital before the root causes are isolated, and the CFO's 12-month payback requires about $25,800 per month in savings or new revenue, which is not supported by the current evidence.
A bounded pilot is the best first step. Select two units with similar acuity and meal volume, one medical and one surgical, and run room service there for 8 to 12 weeks while maintaining the trayline elsewhere as a control. Limit the initial menu to items that can be produced reliably with existing labor. Protect insulin-dependent patients with a scheduled-delivery rule and a documented backup tray.
The pilot should test three assumptions: that on-demand service improves accuracy and satisfaction, that hostess duties can be absorbed without increasing overtime, and that food and labor cost per patient day remain neutral. If the pilot fails, the hospital still has diagnostic data and has avoided a premature capital commitment. If it succeeds, the CFO has a defensible basis for phased rollout rather than a single large bet.
2. Design the measurement and decision-rule set for the pilot, including at least four metrics, data sources, and thresholds.
Read the worked answer
The measurement plan should separate process, outcome, and cost metrics, and should use the same definitions before and during the pilot. Process metrics: tray accuracy, defined as trays containing all ordered items and correct textures, with a target of at least 95%; late trays, defined as delivery more than 15 minutes outside the requested window, with a target below 5%; and average assembly-to-delivery time. Source these from the tray ticket system and a daily 30-tray audit stratified by unit and shift.
Outcome metrics: patient satisfaction with food temperature, taste, and timeliness, using the same survey items and a target of at least the 70th percentile by the end of the pilot; and complaint calls per 1,000 meals, with a target at or below baseline. Cost and labor metrics: overtime hours and dollars per patient day, meals per labor hour, and food cost per patient day, all compared with a pre-pilot baseline of at least eight weeks.
Decision rules should be written in advance: scale if accuracy is at least 95%, late trays below 5%, satisfaction at the 70th percentile or better, and overtime does not increase by more than 5%; modify if outcomes improve but cost or overtime worsens; stop if safety events, insulin delivery failures, or accuracy below 88% occur. Use run charts rather than a single end-point comparison, and review weekly with nursing, the chef, and finance.