At the Gurobi Decision Intelligence Summit in Vienna, POLYPOINT shared how they create fair, transparent healthcare staffing schedules that meet both hospitals’ and employees’ needs.
Most employers who manage shift work struggle with scheduling, but healthcare staffing is an especially complicated challenge. While primary care and outpatient clinics may keep regular business hours, hospitals and inpatient centers run 24 hours a day, 7 days a week, all year long.
Every stakeholder has different shift preferences and needs. POLYPOINT is a Swiss software company that delivers workforce management and resource allocation technology to hospitals and healthcare centers in Switzerland, Germany, and neighboring countries. They aim to reduce customers’ administrative burden and optimize workforce allocation, which ultimately improves patient outcomes.
At the 2025 Gurobi Decision Intelligence Summit in Vienna, Reinhard Bürgy, a Solution Architect with POLYPOINT, led a session on how the company uses a hybrid method (i.e., partly optimized by Gurobi, and partly manual) to create healthcare staffing schedules that are fair, transparent, and effective.
The Unique Challenges of Healthcare Staff Schedules
Prior to POLYPOINT, healthcare administrators developed schedules manually. They had staff preferences on paper and agreements in Excel sheets. Effective scheduling relied on administrators’ deep knowledge of each staff member’s and clinic’s idiosyncrasies.
Scheduling planners must consider many variables:
Minimum, Maximum, & Optimum Coverage: Units have minimum staffing requirements for nights, weekends, and holidays to ensure they preserve quality care during irregular hours. Long term, however, minimum staffing practices could lead to burnout or declines in patient care. Thus, each unit also has an optimum staff pattern to preserve employee and patient wellbeing.
Seasonality & Surges: Clinics may have predictable ebbs and flows in demand that affect shift requirements.
Skill Availability: Certain specialties may require that each shift have a nurse, technician, or doctor with a higher level of certification or seniority that only some staff members have.
Regulatory & Legal Compliance: Policies may require individual employees to work certain shifts or take a minimum amount of time off under specific conditions.
Robustness: Staff may call out on short notice (due to illness, etc.), and the schedule must be resilient to these changes.
Staff Preference: Few people enjoy working on weekends, nights, and holidays, but these shifts must have coverage. Each staff member has their own preferences for how they contribute to these “undesirable” shifts.
Transparency & Fairness: Few things erode staff morale as quickly as a schedule that is random or unfair. Employees should understand how their preferences were balanced with the unit’s needs, and “good” or “bad” shifts should be distributed equitably.
A unit’s scheduling problem involves assigning shifts and rest days to all staff members across an entire month. The number of possible schedules is astronomically large — but sheer size alone is not what makes the problem hard. It is the complex web of constraints (coverage requirements, skill availability, legal rules, fairness targets) that makes it impractical to find a mathematically proven optimum in reasonable computation time.
Instead of chasing an optimal (and perhaps impossible) schedule, POLYPOINT developed a Gurobi-powered tool to help administrators create one that is practical and satisfies key constraints. The tool works in a two-phase system:
Phase One, Work Block Generation: A work block is a sequence of consecutive shifts without a rest day in between. In this phase, the model generates a very large set of potential work blocks per staff member, considering all aspects that apply to a single block in isolation — for example, Jane’s preference for no more than three consecutive shifts, John’s night shift preference, or Julia’s contractual agreement to have Wednesdays off. This large set is then filtered down to at most 50,000 good work blocks per employee, and filtered again so that the total across all staff does not exceed 800,000.
Phase Two, Unit Schedules: The filtered work blocks are then fed into an integer program that assembles a complete schedule for the unit. This phase handles all constraints that span multiple work blocks — such as coverage requirements, fairness across the month, and rest day distribution. The resulting model typically has around 800,000 variables but only 5,000 to 10,000 constraints. The only hard constraint is that no employee can have overlapping work blocks. All other constraints are soft, with varying penalties so that desirable and undesirable shifts are distributed equitably.
This two-phase approach lets POLYPOINT’s clients generate high-quality schedules within an hour of computation time — a major improvement over previous manual approaches, which were both more time-consuming and harder to assess for quality and fairness. It is common that planning effort for shift planners is reduced by roughly two-thirds.
Why Some Manual Planning Can Be Better than None
Staff schedules are a unique challenge in the optimization world. When an auto manufacturer optimizes their assembly line, the gaskets do not have preferences for whether they go in a sedan or a truck. Human resources are human—staff care about their schedules. Their opinions and perceptions matter.
Because of this, POLYPOINT adopted a blended model. Bürgy told audiences, “Remember that a 100% perfect solution is often unattainable. The best approach is often a hybrid one, combining automatic with manual planning.”
Staff enter their schedule preferences and availability via a mobile app, then the two-phase model designs the solution. Ultimately, however, an administrator refines the final schedule manually using an interactive planning board to address unique circumstances and specific cases.
This hybrid approach respects staff members’ autonomy and need for transparency while helping administrators navigate the immense complexity of scheduling.
Navigating Human Complexity Humanely
POLYPOINT transformed an opaque, complicated, and perennial business problem into a simple routine. By recognizing the unique challenges of schedule optimization, they developed a fair, transparent process and achieved high-quality schedules for the sector.
Want to learn more about how they did it? Download their presentation here.
