Clinical studies were undertaken in hospitals to investigate the effects of various fluorescent lamps currently available on the colour appearance of skin by determining their influence on clinical diagnoses and judgements of perceived confidence in those decisions. The Department of Health were interested in the suitability of high-efficacy, narrow-band emitting lamps which offered savings in power and hence savings in revenue costs. A range of pathologies having different skin colour appearances were seen in the Dermatology Department, John Radcliffe Hospital, Oxford, and pathologies having systemic colour effects in the Special Care Baby Unit, Guys Hospital, London. A lighting appraisal was also carried out in a continuing-care geriatric ward at Ashford General Hospital, Kent. The Oxford experiment lasted for seventeen weeks and involved twelve clinicians and 248 Caucasian patients. Forty-nine different pathologies were seen. The Guys experiment lasted seven and a half days during which time one clinician made forty observations involving seventeen babies, most of whom were Caucasian, having one or more of three different pathologies. The Ashford experiment lasted three weeks when fourteen senior nursing staff made general assessments of the lighting in the ward. Physical measurements of the colour of diseased and healthy skin were made during the Oxford and Guys experiments. This paper, the first of three, gives statistical analysis of the results indicating that clinicians and nursing staff are satisfied with lamps having a correlated colour temperature of around 4000 K and that certain narrow-band emitting fluorescent lamps can be used as alternatives to the lamp already approved and in service.

This paper (the second in a series of three) describes an investigation into the possibility of using colorimetric measures to quantify the differences between skin lesions and adjacent clinically normal skin as viewed under a variety of fluorescent lamps. The results show that it is possible to differentiate anaemia and cyanosis from clinically normal skin using various colour parameters. However, none of the colour parameters investigated could differentiate between all the dermatological pathologies observed.

A large sample of spectral relectance curves of clinically normal and pathological human skin was analysed by multivariate statistical procedures in a quest to identify underlying fundamental characteristics by which skin types could be described. The sample of dermatological skin conditions could be classified into five groups. A sample of skins from premature ill-babies could be classified into four groups with similar general characteristics to the dermatological sample. Discriminant analyses showed a poor relationship between these groups and pathology or the clinically described skin colour appearance. There was, however, evidence to suggest that the groups were related to the degree of pigmentation present in the skin. Special colour rendering indices were calculated for the nine cluster-group mean spectral reflectance curves. Two of these provided a simple means of distinguishing between a number of lamps which had been identified in earlier studies to be acceptable or unacceptable for use during clinical examination of patients. These lamps included certain new narrow-band emitting fluorescent lamps having a correlated colour temperature around 4000 K. A new specification for lamps to be used for general purposes in clinical areas in hospitals was established. It incorporates the existing limits of lamp chromaticity and the CIE Special Colour Rendering Indices for the two mean skin reflectance curves which enabled lamps to be distinguished from each other.

No abstract