BUSI 611 Week 8 Discussion Board Forum 8 Example

Discussion Board Forum 8: Technology, Analytics and the Future of Health Care Operations

BUSI 611 Operations Management for Health Organizations, Discussion Board Forum 8

Thread by [Student Name]

The health system, its hospitals and the classmates are composites; the textbook, Scripture and the peer-reviewed sources cited are real.

Our health system runs three hospitals within 25 miles of one another: a 420-bed regional medical center and two community hospitals of about 150 beds each. On most weekdays the regional center is above 95% occupancy and boards admitted patients in its emergency department for an average of six hours, while one of the community hospitals often has 15 to 20 open beds. Transfers between our own hospitals are arranged by phone between charge nurses who cannot see each other's census, and discharges at the regional center are tracked on a whiteboard. Our chief operating officer has proposed a capacity command center to manage beds across all three hospitals, and the board has asked whether it would be worth the cost. Vendors have quoted between $1.5 million and $4 million for software, screens and renovation, plus ongoing staffing, so the decision is large enough to deserve evidence rather than enthusiasm.

Langabeer and Helton (2016) describe information technology and analytics as tools that make operations visible and support faster, better decisions, but they caution that technology produces value only when it is matched with redesigned processes and clear accountability. A command center fits that description when it does more than display data. It combines real-time information on beds, admissions, discharges and transfers with the people who can act on it, bed managers, transfer nurses, transport and environmental services, so that a decision about where a patient should go is made once, with the whole system in view.

The model has spread quickly. Franklin et al. (2023) surveyed capacity leaders at health systems in a national consortium, and 31 of 38 responded. Twenty-five were operating capacity command centers, which ranged from single-hospital to system-wide in scope. The most common reason for building one was reducing emergency department boarding, cited by 24 respondents, and the most common functions were bed management and interhospital transfers, present in all 25 centers. Eighteen tracked financial return on investment, and all 18 reported a positive return. The survey relied on a small convenience sample, so the evidence describes current practice more than it proves effect. Specific applications show what can be achieved. Kittur et al. (2025) describe how Tampa General Hospital's command center built an artificial intelligence tool that flags patients in the neuroscience intensive care unit who are ready to move to a lower level of care, used during multidisciplinary rounds. Six months after implementation, intensive care length of stay fell by about 7%, roughly eight hours, from 4.15 to 3.88 days, transfer orders were placed more than three hours earlier and unplanned returns to the unit did not increase.

I recommended that the board approve a phased command center with three functions at launch: system-wide bed management, transfers between our hospitals and real-time tracking of expected discharges. Advanced predictive tools can follow once the basic workflows are stable. The center should also have a medical director, a hospitalist with authority to settle disputes about where a patient goes, since many transfer delays are disagreements rather than missing information. It should report four measures monthly: emergency department boarding hours at the regional center, the number of transfers to our community hospitals, the hours from discharge order to bed ready and occupancy at each hospital. Earlier forums in this course showed that discharge timing, bed turnover and scheduling are process problems, and the center should be judged on whether it improves those processes, not on how impressive its screens look.

When Nehemiah's builders were spread thinly along Jerusalem's wall, far apart from one another, he told them to gather to the place where they heard the trumpet sound, so that the scattered workers could act together (English Standard Version Bible, 2001/2016, Nehemiah 4:19-20). A command center is a trumpet for a health system: it gives scattered units one place to see the whole wall and rally where the need is greatest.

Reply to Alyssa

Alyssa, your thread argued that a command center becomes truly useful only when it predicts demand, and you proposed a model that forecasts which emergency patients will be admitted so bed managers can start finding beds before the physician decides. Langabeer and Helton (2016) describe predictive analytics as one of the most promising operations tools, since earlier information gives managers more options. Recent research shows such models can be accurate. Zeinali et al. (2026) analyzed 998,109 encounters at 11 emergency departments in a South Carolina health system and trained machine learning models to predict admission using only data available before physician evaluation, including features drawn from triage notes by large language models. The strongest models scored 0.89 on the area-under-the-curve measure and caught more than 83% of eventual admissions, and adding language model features improved recall further.

Accuracy is not the only test. Ghazalbash et al. (2026) built models to predict delayed hospital discharge among older adults in Ontario using two decades of data and reached an area under the curve of 0.82 with an extreme gradient boosting model. They also tested fairness across sex, rural or urban residence and housing instability, found that bias mitigation improved calibration across groups and noted trade-offs among accuracy, fairness and explainability. The main drivers of predicted delay were functional and cognitive decline and regional differences in services.

For your proposal, I would suggest two safeguards. First, use the admission prediction to prompt early bed planning, not to change clinical decisions, so a wrong prediction costs only some planning effort. Second, check the model's accuracy separately for patient groups, since a model that underpredicts admissions for one group could delay their beds systematically. Results should be reviewed quarterly by a team that includes clinicians.

Chronicles singles out the tribe of Issachar for its people who could read their era and judge the right course for the nation (English Standard Version Bible, 2001/2016, 1 Chronicles 12:32), and a good model can help leaders understand the hours ahead, as long as it serves judgment rather than replacing it. There is also a workflow question your thread could explore. A prediction is useful only if someone acts on it, so the command center would need a rule for what a bed manager does when the model flags a likely admission, such as reserving a bed tentatively or alerting the receiving unit's charge nurse. Without that rule, the prediction becomes one more number on a screen. Testing the rule for a month on one unit, and counting how often early planning actually shortened boarding, would show whether the model earns its cost. Has your organization decided who would be accountable for a prediction model's performance after it goes live?

Reply to Ruben

Ruben, your thread argued that command centers are expensive dashboards that cannot fix the real causes of boarding, which you described as late discharges and too few staffed beds. You have a fair point. Langabeer and Helton (2016) emphasize that technology cannot substitute for process redesign, and earlier in this course we saw how stubborn discharge timing is. Epstein et al. (2025) found that most weekday discharges at 197 Florida hospitals still occurred after 3 p.m. in 2022, with only about 13% by noon, unchanged for more than a decade despite heavy pressure to free beds. A screen that shows late discharges does not make them earlier. Neither does it create staffed beds; if the community hospitals' open beds lack nurses, a command center can only move the waiting from one hallway to another.

The evidence also suggests that command centers are valued for coordination as much as for data. In the survey by Franklin et al. (2023), every center performed bed management and interhospital transfers, functions that depend on people from different departments working from the same information, and every center that measured financial return found a positive one. For a system like the one in your thread, much of the benefit may come simply from placing the transfer nurse, the bed manager and the environmental services lead in one room, which shortens the phone chains that slow every placement and lets disagreements be settled face to face.

I would suggest a test that addresses your concern directly: before buying software, co-locate the existing staff for three months with a shared census spreadsheet and measure boarding hours and transfer times. If coordination alone improves flow, the technology investment can be targeted at the remaining gaps; if it does not, the problem is capacity, and the money should go to staffed beds instead.

Jesus saw the crowds as sheep without a shepherd and had compassion on them (English Standard Version Bible, 2001/2016, Matthew 9:36), and patients boarding in hallways deserve leaders who look past the tools to the people waiting. If the trial does show a benefit, the next question is which technology would add the most to it. The Tampa General tool that Kittur et al. (2025) describe was built for one specific decision, moving intensive care patients down a level, and it worked because the team first defined the barriers to that decision. A system that starts by naming its most common flow decisions, and only then asks software to support them, is less likely to buy an expensive dashboard that no one uses. Your skepticism is useful precisely because it forces that discipline on the project. Would your hospital be willing to run a low-cost co-location trial before committing to a vendor?

References

English Standard Version Bible. (2016). Crossway. (Original work published 2001)

Epstein, R. H., Dexter, F., & Fahy, B. G. (2025). Most weekday discharge times at acute care hospitals in the state of Florida occurred after 3 PM in 2022, unchanged from before the COVID-19 pandemic. Journal of Medical Systems, 49(1), 31. https://doi.org/10.1007/s10916-025-02164-5

Franklin, B. J., Yenduri, R., Parekh, V. I., Fogerty, R. L., Scheulen, J. J., High, H., Handley, K., Crow, L., & Goralnick, E. (2023). Hospital capacity command centers: A benchmarking survey on an emerging mechanism to manage patient flow. Joint Commission Journal on Quality and Patient Safety, 49(4), 189-198. https://doi.org/10.1016/j.jcjq.2023.01.007

Ghazalbash, S., Zargoush, M., Guilcher, S. J. T., & Kuluski, K. (2026). Responsible AI for predicting delayed hospital discharge among older adults: Development and evaluation study for balancing accuracy, equity, and explainability. JMIR Medical Informatics, 14, e83244. https://doi.org/10.2196/83244

Kittur, K., Dombrowski, K., Salomon, K., Glover, J., Roy, L., Lund, T., Chiodo, C., Fugate, K., & Patel, A. (2025). Leveraging artificial intelligence to reduce neuroscience ICU length of stay. Journal of Healthcare Management, 70(2), 126-136. https://doi.org/10.1097/JHM-D-23-00252

Langabeer, J. R., II, & Helton, J. (2016). Health care operations management: A systems perspective (2nd ed.). Jones & Bartlett Learning.

Zeinali, F., Taaffe, K., Gaafary, C., Jackson, W., Ramsay, M., Hobbs, J., & Pirrallo, R. (2026). Predicting emergency department disposition using machine learning and large language models to support proactive capacity management: A multicenter retrospective study. BMC Emergency Medicine, 26(1). https://doi.org/10.1186/s12873-026-01487-x