011 min
Why hospitals needed evidence-based design
Picture a hospital design team planning a new surgical wing. They must choose between two-bed rooms and single rooms. They must decide how large the windows are, where the nurses sit, and which floor to lay. Each choice will shape the building for decades, and each one costs money.
For most of the twentieth century, choices like these came from three places. The first was precedent: the last hospital the firm built. The second was taste: what the lead architect, the chief executive or the donor liked. The third was the building code. A code sets the legal minimum for fire safety, structure and room size. It says nothing about whether patients recover faster.
After the Hill-Burton Act of 1946 paid for a wave of new hospitals in the United States, many were built around medical equipment and efficiency. They had deep floor plans, central air and artificial light. Few teams checked afterwards whether the building helped or hurt the people inside it.
None of these sources answers the question a hospital board should ask. Will this room help patients get better, and how will we know? Evidence-based design is the healthcare field's answer to that question. It treats each design choice as something that can be tested, the way medicine tests a treatment.
023 min
Where evidence-based design comes from
The idea that a building changes how sick people do is older than the term. During the Crimean War in the 1850s, Florence Nightingale nursed British soldiers in crowded wards at Scutari, in what is now Istanbul. She later wrote that fresh air, quiet, light, warmth and clean water were part of nursing itself. Her 1859 book Notes on Hospitals shaped the pavilion hospital: long, narrow wards with windows on both sides so air could move through.
The study most people cite as the start of the field came more than a century later. In 1984 the environmental psychologist Roger Ulrich published a short paper in Science. He went back through the records of a 200-bed hospital in suburban Pennsylvania from 1972 to 1981. On one side of a surgical wing, some windows looked onto a small group of trees and others onto a brown brick wall. The rooms were otherwise almost the same. Ulrich compared patients recovering from gall bladder surgery. Patients who looked at the trees left hospital after 7.96 days on average, against 8.70 days for those who faced the wall.
How Ulrich's study ruled out other causes
A difference in stay could come from many things besides the window, such as older patients or a different year. Ulrich could not assign rooms at random, because staff put each patient in whatever room was free. So he paired patients instead. Each pair was matched on sex, age within five years, being a smoker or not, being obese or of normal weight, the general nature of any earlier hospital stay, year of surgery within six years, and floor level. One person in each pair had the tree view and the other had the wall. The same nurses worked both kinds of room on each floor. This is weaker than a randomized controlled trial, but it removes the most obvious other explanations.
What Ulrich counted as recovery
Length of stay was only one measure. Ulrich also read the nurses' notes and counted negative remarks about each patient's mood and progress, and he sorted painkiller doses by strength. The tree-view group also drew fewer negative remarks in the nurses' notes and needed fewer strong painkiller doses. Counting notes and doses showed that a building can be judged with records a hospital already keeps.
From one study to a method
The method around the study took shape over the next twenty-five years. The Center for Health Design, a nonprofit, was founded in California in 1993. In 2000 it started the Pebble Project, in which hospitals agreed to build new features and then report what changed. The name of the method came from medicine. In 1996 David Sackett and colleagues described evidence-based medicine as the careful, explicit use of the best current evidence when deciding how to treat each patient. The architect D. Kirk Hamilton argued that design should work the same way. He wrote in 2003 that evidence-based designers make key decisions together with an informed client, based on credible research and on reviews of finished projects. In 2009 the Center launched EDAC, a credential for people who apply the method.
033 min
How evidence-based design works
Ulrich's study showed that a room could change recovery, but it did not tell a design team what to do next. Evidence-based design turns findings like his into a project routine. Back in the surgical wing, the team does not start by drawing. It starts by writing down what the building should change: fewer falls, fewer infections, less pain medication, less walking for nurses.
The Center for Health Design describes the routine in eight steps:
- Set goals for the building in terms that can be measured.
- Find research that bears on those goals.
- Judge that research: how strong it is, and whether its setting is like this hospital.
- Turn what the research supports into design ideas.
- Write a hypothesis for each idea: what it should change, and by how much.
- Measure the current building before anything changes. This is the baseline.
- Check during design and construction that the features survive budget cuts and changes.
- After people move in, measure again and compare with the baseline. This is a post-occupancy evaluation.
Other firms use shorter versions. The interior designer Rosalyn Cama uses four steps: gather information, map the goals, write hypotheses and design, then measure and share the results. The order is the same in each version: evidence, prediction, building, measurement.
The research methods are ordinary ones. Teams read the published studies, interview and survey staff and patients, watch how people move through the current building, and build full-size mock-ups of typical rooms to test before construction.
Hamilton's four levels of practice
Not every project does all of this. Hamilton described four levels of commitment:
| Level | What the team does |
|---|---|
| 1 | Reads current research and uses it to make design decisions |
| 2 | Predicts the outcome of each decision and measures it after opening |
| 3 | Reports the results publicly, beyond the project team |
| 4 | Publishes in peer-reviewed journals, often with academic researchers |
Most practices work at levels 1 and 2. Hamilton argued that a practitioner who has measured outcomes should share them with the wider field, which is what levels 3 and 4 ask.
Why the prediction comes before the building
Level 2 is where the method becomes more than reading. Hamilton's second level asks designers to predict outcomes in advance and then measure them, so a team cannot quietly play down a failure. Without a written prediction, a team that opens a new ward will find something that improved and call it the goal. That is confirmation bias: noticing the results that fit what you already believed. So the surgical-wing team writes before construction that single rooms will reduce infections caught on the ward. The post-occupancy evaluation now has a clear test, and a clear way to fail.
043 min
What evidence-based design research has found
A prediction is only as good as the research behind it, so the next question is what that research says. By the early 2000s there was a lot of it. A 2004 review led by Roger Ulrich and Craig Zimring found more than 650 rigorous studies linking hospital design to safety, stress and quality of care. The findings group around a few features:
- Daylight and views of nature. Linked to lower stress, less pain medication and better sleep, partly through the circadian rhythm, the 24-hour body clock that light sets.
- Single-patient rooms. Linked to fewer infections, more privacy and space for family.
- Noise. Quieter units are linked to better sleep for patients and less stress for staff.
- Staff safety and work. Ceiling lifts reduce back injuries from moving patients, and layouts that let nurses see more rooms cut walking and fatigue.
- Wayfinding and finishes. Clear routes, colour contrast and floors that are easy to clean reduce confusion and change how clean patients judge a hospital to be.
Two later studies show what a careful test of one feature looks like.
Sunlight and pain medication in Pittsburgh
Ulrich used old records. A team at the University of Pittsburgh, led by Jeffrey Walch, ran a study that followed patients forward in time, published in 2005. They followed 89 patients after spinal surgery on one hospital unit. Some rooms were on the bright side of the building and some on the dim side, and the bright rooms got 46% more natural light. Room colours and gowns were the same on both sides. Per hour, patients on the bright side took 22% less pain medication than those on the dim side. Their medication cost 21% less, and they reported less stress at discharge. They also reported slightly less pain, but that difference was not statistically significant.
Single rooms and infection in a Montreal ICU
The Pittsburgh study tested light. The next one tested walls. In 2011 Dana Teltsch and colleagues at the McGill University Health Centre in Montreal studied an intensive care unit that changed from multi-bed rooms to single rooms. They compared infection rates before and after the change, and used the ICU of a nearby hospital that did not change as a comparison. The bacteria they tracked, MRSA, C. difficile and VRE, spread easily between hospital patients and are hard to treat. After the move to single rooms, the rate at which patients picked up MRSA, C. difficile or VRE fell by about half. Patients also stayed in the ICU about 10 percent less time.
Why a window view could change recovery
Neither number explains why the room mattered. Ulrich gave a reason in his 1984 paper. Views of nature tend to calm fear, hold a person's attention and interrupt worrying thoughts, so they may help a stressed patient recover from anxiety. A patient who is less anxious may ask for less pain relief and go home sooner. Light may also work through the body clock. A reason like this matters for design, because it lets a team predict when a feature should work and when it should not.
052 min
How evidence-based design shows up in real buildings
Findings like these are now built into hospitals, and the clearest projects can say which study each feature came from. The Royal Jubilee Hospital Patient Care Centre in Victoria, Canada, is one of the most cited. Most of its bedrooms are single rooms. Ceiling-mounted lifts run from each bed to the washroom, so nurses do not lift patients by hand, a common cause of back injury. The team took colour contrast from dementia research, so older patients can see where walls, ceilings and handrails start. Ulrich has called it one of the world's leading evidence-based hospitals.
A smaller case shows the method at the start of a project. At Emory University in Atlanta, the director of the neurological ICU was not satisfied with the plans for a replacement unit. He asked Craig Zimring's group at nearby Georgia Tech what evidence linked ICU design to patient outcomes. Graduate students reviewed the research, the findings fed an intensive design session with the architects, and the unit was redesigned.
The method is also claimed more often than it is done. A firm that reads research but never measures its own buildings is working at level 1. It cannot tell whether its hospitals did what it predicted. In 2012 the Center for Health Design said that few care providers yet practised evidence-based design in full.
The routine has spread beyond hospitals. Office designers apply it to daylight, air quality and noise. School designers apply it to classroom layout and seating. Home designers apply it to thermal comfort and acoustics. The evidence in those settings is thinner than in hospitals. A designer should check whether a hospital finding fits a classroom before borrowing it.
Evidence-based design in interface work
Interface designers use the phrase more loosely, for decisions backed by research rather than opinion. The hospital routine still transfers. Say a product designer is asked to redesign a sign-up form. They write the outcome first: fewer people give up on the form. They record today's rate as the baseline. They read existing research on form design and run usability testing on the new version. They predict the change, then check it with an A/B test after launch.
The difference is speed. A form can be tested in a week, while a hospital wing takes years to build and cannot be built twice. That is why a hospital team depends on published studies, and a product team can depend more on its own tests.
062 min
Applying evidence-based design to a project
Back in the surgical wing, the team now has research, a method and real projects to learn from. These are the moves that turn that into practice:
- Write three to five outcomes the building should change, and say how each one will be measured.
- Collect baseline numbers from the current building now: falls, infections, staff injuries, patient survey scores. Without them, the evaluation after opening has nothing to compare with.
- For each major feature, write down the study it rests on and how close that study's setting is to yours. A study of gall bladder patients in 1970s Pennsylvania is not a study of a children's ward.
- Build a full-size mock-up of a typical room and let nurses and patients use it before construction.
- Agree who will run the post-occupancy evaluation and when, and put its cost in the budget.
- Share the results, including the features that did not work.
The method backfires in two ways. A team can follow published research so closely that it ignores what this hospital's staff and patients need. That is why Hamilton's definition puts an informed client in the decision. A team can also design only for outcomes that are easy to count, such as length of stay, and neglect outcomes that are hard to count, such as dignity or how a family feels in the room.
Some features cost more to build, such as single rooms and ceiling lifts. The case for them is lower running costs: fewer infections, fewer injuries and fewer patient transfers. State that business case to the client in numbers before construction, and check it after opening with the same measures.
072 min
Where evidence-based design is contested
The strongest objection to evidence-based design is about the evidence itself. In the 1960s and 1970s, the architectural psychology movement was criticised for architectural determinism. That is the belief that a building directly causes behaviour, when a study has only shown that the two go together. It is the old problem of correlation and causation. Reviews of the field still note that the evidence for many design features is weak, and sometimes points in opposite directions. Many studies are small, cover one hospital, and cannot assign patients to rooms at random. Most of the research also comes from North America and Europe. Designers elsewhere, for example in India, have asked for studies of their own hospitals and climates.
A plain brick wall is not every building
Ulrich himself limited his result. He noted that the wall in his study was a plain, almost featureless one. Ulrich wrote that his result does not cover every built view, or patients such as long-stay ones whose problem is boredom rather than anxiety. For a bored patient on a long stay, he suggested, a busy city street might help more than trees. His sample was also small: 46 patients, all after the same operation, in one hospital.
Single rooms are not the whole infection story
The Montreal result has its own limit. The Montreal researchers said hand hygiene, isolation precautions, careful antibiotic use and cleaning still matter alongside the rooms. A single room separates patients, but staff hands still move between rooms. A hospital that builds single rooms but washes hands less often may not see the same drop.
Evidence means more than journal papers
A common misreading is that evidence-based design means using only peer-reviewed studies. Hamilton's definition is wider. It counts reviews of finished buildings as evidence, and it puts the client in the decision. Many design questions have no published study, and every building is different. So the method asks for the best evidence available and an honest note of how strong it is. It does not ask for a trial behind every choice.
While this is unsettled, a practitioner can still act. Use the strongest studies for the largest decisions, label weaker evidence as weaker, and measure your own buildings so the next team has better evidence than you had.
081 min
Evidence-based design vs. nearby concepts
Evidence-based design is most often confused with research-informed design, and the difference between them is measurement.
| Concept | What it is | How it differs from evidence-based design |
|---|---|---|
| Research-informed design | Design that draws on research for ideas, a term that came from education | Uses research as input but makes no prediction and measures no outcome |
| Evidence-based medicine | Treatment decisions based on the best current evidence for each patient | The model evidence-based design copied. It decides treatment, not buildings |
| Post-occupancy evaluation | Measuring how a building performs after people move in | One step inside evidence-based design, not the whole method |
| Biophilic design | Bringing nature, daylight and natural materials into buildings | A set of features. Evidence-based design is a method that may choose them, or not |
| Building codes | Legal minimums for safety and size | Set the legal minimum, not an outcome to improve |
| Data-driven design in software | Product decisions based on usage data and experiments | Same logic, but faster tests and less reliance on published studies |
The table shows one pattern. Each neighbour has one part of the method. Evidence-based design is the full loop: research, a written prediction, the building, and a measurement that becomes evidence for the next project.
?5 questions
Questions people ask
Is evidence-based design only for hospitals?
Who can practise evidence-based design?
Can a small project use evidence-based design?
Is Ulrich's 1984 window study still valid?
Does evidence-based design mean every decision needs a study?
§12 sources
Sources for evidence-based design
Ulrich, R. S. (1984). View through a window may influence recovery from surgery. Science, 224(4647), 420–421. (full text: )
Sackett, D. L., Rosenberg, W. M., Gray, J. A., Haynes, R. B., & Richardson, W. S. (1996). Evidence based medicine: what it is and what it isn't. BMJ, 312(7023), 71–72.
Hamilton, D. K. (2006). Evidence-based design supports evidence-based medicine in the ICU. ICU Management & Practice, 6(3).
Walch, J. M., Rabin, B. S., Day, R., Williams, J. N., Choi, K., & Kang, J. D. (2005). The effect of sunlight on postoperative analgesic medication use. Psychosomatic Medicine, 67(1), 156–163.
Show all 12 sourcesShow fewer sources
Splete, H. For pain relief, look on bright side of hospital. A news report on the Walch study, The Hospitalist / MDedge.
Teltsch, D. Y., Hanley, J., Loo, V., Goldberg, P., Gursahaney, A., & Buckeridge, D. L. (2011). Infection acquisition following intensive care unit room privatization. Archives of Internal Medicine.
McGill University Health Centre (2011, January 11). Research shows single-patient rooms reduce hospital infections in ICU. News release.
Stall, N. (2012). Private rooms: evidence-based design in hospitals. CMAJ, 184(2), 162.
ArchPsych. Evidence-based practice: from research to design implementation. A summary of Hamilton's four levels of practice.
BHDP. Applying evidence-based design to drive continuous improvement in healthcare spaces. Describes the Center for Health Design's eight-step process and Rosalyn Cama's four steps.
OpenExamPrep. Definition, history and core principles of evidence-based design. An EDAC study guide with the field's timeline.
Wikipedia. Evidence-based design.
