H567/03 Section B Option 3 - Ulrich (1984) and design for health

Why a window view might matter

Ulrich investigated whether a window view of a natural setting was associated with recovery after surgery. Hospital patients often have limited access to outdoor environments and spend long periods in the same room. A view may offer pleasant interest, reduce stressful thoughts and help relieve anxiety. This provides a possible psychological pathway from environmental design to recovery; the study did not directly measure every step in that pathway.

The research relates to the psychological effects of the built environment. Building position and window placement determine what a bed-bound patient can see. The social area considers how the setting supports or restricts experience, alongside individual preferences and the wider hospital context. Biological recovery and psychological stress can interact, so an environmental influence is not a claim that surgery outcomes are explained by architecture alone.

The comparison concerned trees versus a largely featureless brick wall. It was not a comparison between all natural and all urban views. An engaging urban scene or a disliked natural view might have different effects. This specificity is essential when transferring the finding to a design proposal.

A retrospective matched comparison

Ulrich examined records from a suburban Pennsylvania hospital for patients who underwent cholecystectomy, a gallbladder operation, during 1972–1981. Operations between 1 May and 20 October were selected so that the trees had foliage. Patients younger than 20 or older than 69, those with serious complications and those with a history of psychological disturbance were excluded.

There were 46 selected patients: 23 tree-view patients and 23 wall-view patients, arranged into 23 matched pairs. Each group included 15 women and eight men. Patients had been assigned to rooms when rooms became vacant; the researcher did not randomly allocate patients to views for this study. It is a retrospective comparison of naturally occurring conditions, not a randomised laboratory experiment.

Pairs were matched on sex, age within five years, smoking, obesity status, previous hospitalisation, surgery year within six years and floor level. Second-floor patients were also matched by room colour. Matching by physician was achieved for only seven pairs; the remaining physician distribution was similar across groups. Matching reduces particular alternative explanations, but it cannot guarantee equivalence on unrecorded factors or within broad matching ranges.

The rooms were double occupancy and similar in dimensions, window size, furniture and bed arrangement. The same nurses worked on a given floor. Windows allowed patients in either bed to see out while lying down. Similar room features make the view a more plausible explanation of a difference, although naturally assigned rooms do not establish complete experimental control.

Procedure and recovery measures

A nurse with extensive surgical-floor experience extracted recovery information from the records without knowing the patient's window view. This blinding reduced the opportunity for expectations about trees to influence extraction or classification. The original ward nurses who wrote the records were not described as blind to the view, so blinding the extractor does not remove every possible source of judgement bias.

Five kinds of information were collected: days from surgery to discharge; number and strength of analgesic doses each day; anti-anxiety medication doses; minor complications such as persistent headache or nausea; and nurses' notes about condition and recovery. Using several outcomes avoids reducing recovery to one impression. Medication and discharge still depend on clinical decisions as well as the patient's state.

Nurses' notes through the end of the seventh recovery day were classified as positive or negative. Positive comments concerned, for example, good mood or movement; negative comments concerned distress or needing encouragement. These qualitative records were turned into counts. Clear category rules and more than one independent coder would strengthen reliability; the summary does not provide an inter-rater coefficient to memorise.

The setting offers ecological validity because these were actual surgical recoveries under ordinary care. Retrospective records can be incomplete or inconsistently written, however, and important psychological variables may not have been measured. Any use of medical records requires protection of confidentiality; this lesson does not claim that the guide documents a particular consent procedure.

Results and cautious conclusions

Tree-view patients had shorter average postoperative hospital stays: 7.69 days compared with 8.70 for wall-view patients. The reported difference was statistically significant, with $p=0.025$. The observed mean difference was $8.70-7.69=1.01$ days. This is a group average, not a promise that each patient will leave 1.01 days earlier.

Negative notes averaged 1.13 per tree-view patient and 3.96 per wall-view patient, a difference of 2.83 notes. That difference was significant ($p<0.001$). More positive notes were recorded for tree-view patients, but the positive-note difference was not significant. Distinguishing positive and negative outcomes prevents the unsupported statement that every measure of behaviour improved significantly.