The fact that hospitals undergo physical change has become a subject of interest on an international level. These changes are of great concern to healthcare organizations—and to governments—because of the rising costs of construction and the need to control those costs as buildings adapt and expand over time. This interest is not new—understanding how facilities change through varying cycles came to our attention in the 1960s with renewed vigor when “systems” buildings became the subject of research. The U.S. Veterans Administration building system is a good case in point, with its innovative interstitial floor concept developed in the 1970s and used in a number of projects since.

That was also when a “life cycle” understanding of buildings emerged as a subject among building economists. Despite knowledge developed about the importance of the “long view” in life-cycle analysis, however, economic incentives still favor the short term. This is exacerbated by the fact that health care systems usually operate with one budget for new construction and another for maintenance and operations. Two “turfs” competing for resources do not necessarily support the kind of investment decisions needed to prepare buildings for the long haul.

While hospitals “churn” incessantly, little systematic research is being done on how hospitals change. This doesn’t mean that architects, contractors, product manufacturers, and clients have been able to avoid this challenge—many have worked hard to solve these problems ad hoc. Architects, consultants, and their clients couldn’t survive otherwise, and no doubt much has been learned in the way of principles and methods. However, few in practice have time to reflect upon and capture lessons acquired from daily work in any organized way. As it turns out, the more we focus on it, the more we find that our theory on the subject is weak.

In short, we don’t have a good way of explaining how hospitals are designed and how they change over time. What is generally heard is we should “design column-free spaces,” “make floor-to-floor height greater,” or, “don’t bury pipes in concrete.” These guidelines are important, but they don’t constitute theory.

This “theory deficit” is important and serious. When I say “theory,” I don’t mean what we want to happen (as in a manifesto), but what we understand about how things work. In that sense, good theory is the most practical thing we can have. Without it, our work is less well grounded than it should be.

Parallels to Other Building Types

There is evidence that hospital system clients, in the knowledge that their buildings are never truly finished, are asking architects and engineers new questions, with varying results. Some are building “shell” space because they have no way to predict what will later be filled in. Perhaps more importantly, smart clients are asking for demonstrations, beyond rhetoric and pat answers, of how their buildings will accommodate change when technical and organizational decisions are constantly in flux.

This is how office buildings, shopping centers, laboratories and, increasingly, housing developments around the world that follow “open building” principles address decision flexibility and long-term capacity for change in very matter-of-fact ways. Now, some healthcare projects are being designed similarly. The Banner Estrella Medical Center project in Phoenix designed by NBBJ and Orcutt/Winslow Partnership is one healthcare example pointing in the right direction, as is the Gonda Building at the Mayo Clinic in Rochester, Minnesota, designed by Ellerbe Becket. Both have established a loose-fit connection between the base building (defined in terms of accommodation capacity) and the structural, spatial, and equipment needs of the fit-out. There are a number of other exciting projects under way in Europe, including a new hospital in Belgium designed by the Austrian firm Baumschlager and Eberle, and the Erasmus MC University Hospital in Rotterdam designed by EGM Architects in the Netherlands. These projects also formally establish this loose-fit relationship between the base architecture and its changing fit-out.

The INO Hospital Project

One of the most clear-cut, methodologically rigorous approaches to designing change-ready hospitals is being used in a pioneering project in Bern, Switzerland: the INO project at the Insel University Hospital, managed by the Canton Bern Office of Properties and Buildings (OPB). An international peer review of this project took place in July 2006 in Bern, with sixteen architects, hospital facilities directors, and researchers from the United States, United Kingdom, and the Netherlands taking part with their Swiss counterparts. (Please contact the author to learn more about this.)

Although the problems faced by the OPB are highly technical, the first order of business at the INO was not to invent new technical systems (like the Veterans Administration Hospitals’ interstitial floors), but to rethink the paradigm of facility procurement and design management. This innovation is primarily organizational, but the consequences are architecturally and strategically compelling.

The INO project is a 50,000-square-meter addition to a large university hospital campus in Bern (figure 1). It’s one of the biggest hospital projects ever built in Switzerland. When the decision was made to go to the Canton Bern government for funding, the argument was that complex buildings such as this only become “whole” over time. The Canton Bern OPB had come to realize, after many conventionally procured buildings, that inevitably their buildings’ functions change to meet new medical procedures, new regulations, and new market and insurance conditions.

The Campus of the Insel University Hospital in Bern, Switzerland, with the new INO Phase I addition clearly visible in the foreground (the large building with many skylights and the curved facade). Phase II will almost double the footprint.

Recognizing these dynamics resulted in the management team, led by chief architect Giorgio Macchi, adopting an entirely new process for procuring the facility. A competition (a normal process of architect selection in many European countries) was held to select a design and construction firm for each of three distinct “system levels.” Ten firms were invited to submit proposals for the primary system.

The primary system’s base building is intended to last 100 years and is expected to accommodate changing departmental sizes and changing layouts of emergency, imaging, surgery, and pharmacy departments, zoned for specific floors (the building was not designed to have nursing units). The secondary system—which indeed has already accommodated changing sizes and layouts of departments, proving the capacity of the primary system—is intended to be useful for 20-plus years, and the tertiary system (equipment, finishes, and furnishings) is expected to have a useful life of 5 to 10 years. A full report on the procurement process is available at by clicking “INO.”

In the Canton Bern’s procurement process, a hierarchical structure of decision levels was developed as a management model. Higher-level decisions provide “capacity” for a range of lower-level functional scenarios (as in the “Scenario Buffered Buildings” approach outlined in Stewart Brand’s book How Buildings Learn). The principle is to partition the whole by what is called “systems separation.” One architect designs the base building, without detailed programmatic information about what will be filled in later. The primary system (base building) is not “neutral,” in that it constrains what can be installed inside by its plan form, column spacing, vertical mechanical shafts, main corridors, and so on (figure 2).

The primary system plan for the Phase I base building of the Insel University Hospital INO Project.

This is followed by a second competition for the selection of the secondary system architect, who engages in intense programming with the client’s hospital staff and, subsequently, proposes the secondary system to be installed in the base building. This strict separation is made to ensure that the base building is not designed for just one programmatic scenario. Later, another competition is held to select the firm to provide the tertiary system. In this process, the separate contracts were organized to reduce the typical dependencies among these three system levels.

figure 3shows the contents of each system level and distinguishes the spatial organization—the users’ main concern—from the technical support systems. Thus, in the INO model, the Systems Separation approach is both spatial and technical, with each level allowing some degree of adaptability or flexibility precisely because it isn’t integrated with the higher level. Both the spatial and technical systems taken together define a system level.

The basic organizational diagram showing the system separation on levels.

The approach yielded benefits even before the Phase I was completed. Functional layouts had already morphed because of changes in the organization’s priorities and client base, as well as medical procedures, hospital politics, and technology. The primary system has allowed that to happen with a minimum of fuss—figure 4 shows examples of how a given departmental layout can be accommodated initially or later changed in the same floor plate.

Two possible layouts (A, B) for the surgery floor.

True, there have been problems—particularly the lack of sufficient clarity of the contracts between the separate design firms. It became clear that defining the boundaries between system levels was more important and difficult to accomplish than was initially anticipated. In another project for a University building currently underway, the OPB has decided that it’s preferable to award one firm the responsibility for all systems levels, provided it offers clear evidence on how secondary system changes can be made without affecting the primary system.

Based on the successes achieved so far, though, future projects under the OPB’s jurisdiction will now be required to maintain a strict systems separation in terms of technical definition and contract scope—defining key interfaces, but leaving maximum decision flexibility to designers at the next level, even if one design firm is responsible for designing all levels.


This shift in approach toward “open building,” exemplified with methodological clarity by the INO project, is inevitable, not because of any theorist’s wishful thinking, but by brute necessity. No one can afford to invest in hospitals that cannot adjust. Few advocate the idea that we should simply demolish buildings after 30 years and build again to keep up with technological and organizational change. The implications of doing so are simply unacceptable if we want a stable urban form, a sustainable environment, or a sense of place.

New developments in building information modeling will provide much needed help in managing information flows among designers distributed among systems levels. For this to happen, however, the theory and methods (and habits) of open building need to be adopted by both clients and designers.

This will be no easy task, because there is ideological and practical resistance to doing so. The ideological reason to resist delivering hospitals via open building springs from habits of functionalist design and the idea of “whole building integration.” The practical reason to resist is more pragmatic: The delineation of levels of intervention and the natural distribution of design that follows requires new ways of working and coordination—it asks us to reconsider ingrained habits and conventions. Doing so is always painful and demands that extra effort be made before it pays off. Since we fear losing control, our instinctive reaction is to avoid such change as long as possible.

What does have to be avoided is the design of buildings that are dependant on only one scenario of functionality. The point is that we need to learn to make “agile” and “accommodating” designs at each level of intervention. This may not sound so difficult, but it is exactly opposed to the modernist/functionalist tradition. We have learned to define function (by detailed programming and human factors research now called “evidence-based design”) and then design the building to fit. We have also learned to model our behavior on the masters, claiming that good design is only possible if top-down control by one architect is maintained.

Unfortunately, this is still the way most architects are educated. In practice, of course, architects and their consultants work hard to design under these conditions of change; this is increasingly the norm. But they must apologize for what they do because there is little in the way of theory to explain these phenomena. This may explain why in print and at conferences, there is virtually no discussion of these issues. Until the profession—and the academy—embrace reality, we will be less able to provide the leadership necessary to design sustainable, humane, and “change-ready” healthcare facilities.HD

Stephen Kendall, PhD, RA, is Professor of Architecture and Director of the Building Futures Institute, Ball State University, Muncie, Indiana.


  1. Blyth A, Worthington J. Managing the Brief for Better Design. Spon, London, 2001.
  2. Boice JR. A History and Evaluation of the School Construction Systems Development Project 1961-1967. Building Systems Information Clearinghouse, Educational Facilities Laboratories, Inc., 1968.
  3. Brand S. How Buildings Learn: What Happens After They’re Built. Viking, New York, 1994.
  4. BuildingSystems Development and Stone, Marraccini&Patterson. Development Study—VA Hospital Building System. Washington D.C.:U.S. Government Printing Office, 1977.
  5. Hattis DB, Ware TE. The Peach Book: The PBS Performance Specification for Office Buildings, January 1971 (Revised September 1971); NBS Report 10 527.
  6. Hochbauamt Des Kantons Bern, (Ed.) Jurybericht: INO Inselspital Bern. Wettbewerb 1997, Primärsystem. Bern, 1997.
  7. Hochbauamt Des Kantons Bern (Ed.) Jurybericht: INO Inselspital Bern, Wettbewerb 1998, Sekundärsystem. Bern, 1998.
  8. Habraken NJ. The Structure of the Ordinary: Form and Control in the Built Environment. MIT Press, 1998.
  9. Kendall S. Open Building: A new paradigm in health care architecture. AIA Academy Journal 2004(October): available at 20041027.
  10. Macchi G. Prävention beim Bauen: Projektorganisatorische Vorkehrungen im Hinblick auf einen langfristig hohen Gebrauchswert. SIA, Schweizer Ingenieur- und Architekten-Verein (Ed.) Bauwerkserhaltung und Wirtschaftlichkeit – Perspektiven einer modernen Aufgabe, Referate der FEB-Studientagung vom Donnerstag, 23 Januar 1997, Dokumentation SIA D 0141. Zürich, 1997.
  11. Macchi G. Eine neue Werterhaltungspolitik für Bauverwaltungen. Hassler Uta (Ed.) Umbau: Über die Zukunft des Baubestandes. Tübingen; Berlin:Wasmuth, 1999.