Benchmarking is a practice that compares a metric back to a standard in a particular industry. In healthcare design, this process allows designers and architects to compare existing or proposed building performance data against best practices. The purpose is to understand differences in multiple designs, between facilities or competitors. Ultimately, benchmarking provides a quantitative tool to align decisions and selections with established project objectives.

Common planning benchmarks used in healthcare design involve square feet per key planning unit (KPU), such as a patient bed, imaging modality, or operating room. These benchmarks are based on codes, research, and precedent, and factor in a certain amount of space necessary to support one KPU in a hospital. However, these metrics should also consider functionality and operations. For example, a hospital that elects to have a patient/family-centered care designation requires a larger patient room and more family sleeping space, which drives up the typical benchmark ratios for square feet per patient room.

Benchmarks can be an incredibly helpful design tool, whether sizing an entire campus or informing a departmental study. However, when benchmarking is performed in a vacuum without a standardized process, the results can be misleading. Both the American Institute of Architects and Building Owners and Managers Association offer guidance on standardizing processes for calculating areas. The simple mechanics of determining the amount of space per KPU is very important; in order to compare designs, the process of performing the area calculations should be handled the same. Are all comparisons using the inside face of demising walls or do some go to the mid-point of the walls? Are shafts, stairs, and elevators included or excluded?

For example, take the traditional rule of thumb of 2,500 square feet per bed for overall sizing of a hospital. In a nutshell, this benchmark implies that all the support spaces necessary to deliver inpatient care, including food, logistics, and diagnostic and treatment areas, have been tallied and divided by the total number of beds supported. For a 120-bed community hospital, 300,000 square feet (2,500 square feet X 120 beds) may be the sweet spot for enough programmed area to support each bed, without being oversized. A 25-bed critical access hospital, however, may need to be more efficient than a benchmark of 62,500 square feet to meet the cost-based reimbursement models for Medicaid and Medicare. Similarly, a 380-bed teaching facility using this benchmark may not have enough space per bed to support team rounding, consultations, and the provider work spaces necessary for both patient care and academics. So how can we make benchmarking a less blunt, more meaningful tool for any scale?

 

Benchmarking in practice

Using the 2,500-square-feet-per-bed benchmark as an example, here are eight things to consider when evaluating healthcare design benchmarks:

1. Space in the right place. If an area has “hit the target” in terms of square footage, the space must be well designed and fully functional, right? Actually, this is the number one misconception of benchmarking. At such a high level, this benchmark only indicates there might be enough space. It doesn’t tell us if the space is in the right places to be functional. For example, if sterile processing is a large area, but the operating rooms are small and antiquated, the functionality may be compromised—even if the surgery department benchmark is right on target.

2. Quantity versus quality. Facility benchmarks based on area per KPU can be used to identify the appropriate amount of space, but they rarely communicate the quality of space, a shortcoming that should be recognized up front. For example, if the patient rooms are still semiprivate, the overall appropriate amount of space may be accurate, but the quality of the patient experience might suffer. Other tools like parametric modeling, which is a computer simulation that illustrates how several variables are related and interact with each other, can be incorporated to better articulate quality. To best understand the experience of a patient in accessing natural light from the bed, benchmarking will simply provide a number, such as area of window glazing per room, while parametric modeling can compare designs based on multiple facets, including area of windows as well as sill heights, solar orientation, and distance of the beds from the exterior wall.

3. Context. Are you using this square-foot-per-bed benchmark for an academic medical center or a community hospital? Is the facility a for-profit model in a national network or an independent not-for-profit in a community health system? Benchmarks should always be tied to comparable data sources. It isn’t accurate to benchmark the support spaces needed for residents by comparing a teaching facility with the needs of a community hospital. Different models will benchmark differently, and the targets for the best practice benchmarks must adjust accordingly.

4. Apples to apples. Statistics can be manipulated (intentionally or inadvertently) and benchmarks are no different. If multiple facilities are being compared, then the data that make up the total square footage must be derived in the same way. If one facility’s area-per-KPU number includes a central utility plant and the others exclude it, then the benchmark is not comparing apples to apples. If one hospital runs semiprivate rooms and another facility is all private, the total bed complement may be the same, but the space needed per bed would vary significantly. The detailed composition of the benchmark must be clear to ensure you’re comparing like entities.

5. Building era. The age of the facility can also impact benchmarks. A modern facility is probably more efficient in utilizing space for current operations than one built long ago. Therefore, a renovation benchmark might be scaled back from a new construction benchmark when existing facility constraints, such as a small column grid, make it impossible to reach best-practice targets.

6. Operations. As noted previously, a benchmark must take into account functionality and an understanding of how a space is intended to be used from an operational perspective. For example, a portion of the 2,500-square-foot-per-bed benchmark includes materials management. But if supplies are stored in an off-site warehouse, this requires less space than if a hospital is planning to store all materials on-site (though the benchmark may need to factor in space for staging and distribution). Registration is another good example: If bedside registration is implemented, the square footage for a centralized registration department could be reduced in the overall hospital benchmark.

7. Cultural variations. Just as a benchmark should adjust for the type of facility it represents, it should also vary based on cultural differences. In the Middle East, gender separation and duplication of spaces may affect the typical benchmark but could be offset somewhat by a ward arrangement. In the United States, some areas of the country tend to experience larger familial support needs for patients in the hospital, impacting the size of lobbies and waiting rooms to accommodate larger groups of family members, which may increase the overall benchmark.

8. Cataloged data. To determine whether 2,500 square feet per bed is an appropriate best practice for a project, an organization needs to collect and record numerous data points of area calculations for similar facilities or departments. Being able to see trends in the data and identify outliers in square feet allocated for projects requires a database of multiple projects that’s accessible and consistently generated. Typical data collected and maintained for a full hospital may include overall area per bed, grossing ratios, percentage of space allocated to engineering infrastructure, and total KPUs by service line. Additional department-specific data, such as square feet per KPU as well as departmental grossing ratios, are also helpful in understanding the amount of circulation needed for various departments. For example, the ability to query multiple academic medical center metrics would illustrate the challenge in planning around 2,500 square feet per bed for this facility type and help build a case for additional space allocation in future planning. Then, department-level benchmarking would help identify the specific units that may be undersized.

 

Use it wisely

Based on these considerations, a benchmark like 2,500 square feet per bed can vary significantly from hospital to hospital. An organization’s facility should be carefully compared to hospitals of a similar size and context, and the details around the composition of the benchmark should be known and documented to make sure that an apples-to-apples comparison is realistic. What may first appear to be a single, static target is actually a fluid number that adjusts within a range of square feet to reflect operational decisions, culture, and building era. This refined benchmark can then provide a quantitative comparison that meaningfully conveys best practice for a specific application, crafting a much better tool for project decision-making.

Katie Fricke, AIA, NCARB, LEED AP, is a healthcare planning principal for HDR (Charlotte, N.C.). She can be reached at katie.fricke@hdrinc.com.