The influence of the learning effect on the outcome of automated perimetry was studied as a function of eccentricity. The same comprised 20 patients with suspected glaucoma who were all naive to automated perimetry. Visual field examination of the right eye followed by that of the left eye was undertaken on each of three successive days and after a further interval of 12 days using a customized full-field program of the Humphrey Field Analyser 630 (stimulus size III). The program comprised 60 points out to an eccentricity of 60 degrees with an interstimulus grid of 12 degrees. Global, central, peripheral, superior and inferior mean sensitivity each significantly increased (P less than or equal to 0.01) from the first to the second right eye examinations and from the third to the fourth left eye examinations (P less than or equal to 0.01). The global short-term fluctuation, central mean defect and number of stimulus presentations decreased from the first to the second right eye examination (P 0.01). The order of examination between eyes and the interval between examination sessions influences the response recorded by automated perimetry.

The detection of change in the visual field is confounded by factors associated with the patient response. Sixteen patients who had previously undergone a training regime in automated perimetry over a short time period were followed up to evaluate the longer term learning and fatigue effects. The patients, all attending a glaucoma clinic, were originally naive to automated perimetry. The training period had comprised examination of the right followed by that of the left eye with the Humphrey Field Analyser 630 on each of 3 successive days and again after an interval of 12 days. The follow-up study comprised two examinations on 2 successive days after a follow-up period of between 5 and 15 months (mean 8.7 months). The learning effects present over the initial training period were not observed at the follow-up whilst the fatigue effects in the fellow eye were still apparent.

Light scatter was induced in 15 glaucoma patients, exhibiting clear media and moderate field loss, using cells containing varying concentrations of 0.5 microns diameter latex beads in distilled water. The right eye was examined with program G1 on the Octopus 201, and with a 45 degrees threshold profile on the Dicon AP3000, with and without a given cell. Forward light scatter due to the cell was assessed by measuring the depression of contrast sensitivity, with the Nicolet CS2000, under glare conditions. Perimetric mean sensitivity decreased linearly and loss variance decreased curvilinearly with increase in forward scatter. Threshold values for 26 glaucoma patients, determined in the absence of induced scatter, were then corrected for the effects of naturally occurring intraocular light scatter. The recalculated mean defect decreased linearly while loss variance remained essentially unchanged. Forward light scatter therefore predominantly exaggerates diffuse loss; the apparent underestimation of focal defects caused by induced scatter is partly a computational artefact resulting from inapplicable age-matched normal reference data.

The various stimulus parameters offered by two standard automated projection perimeters [Humphrey Field Analyser 630 (HFA) and Octopus 201], namely, stimulus size and location and the interaction of adaptation level and stimulus duration, were compared in a sample of 20 patients attending a glaucoma clinic using the visual field indices mean defect (MD), loss variance (LV), short-term fluctuation (SF) and corrected loss variance (CLV). LV and SF were greater with Octopus program 32 compared with Octopus program G1 (P < 0.02). No difference in the indices was found between stimulus sizes I and III for HFA program 30-2. MD was greater for program 30-2 compared with program 32 (P < 0.002) when expressed in terms of log (L/delta L) whereas LV (P < 0.02) and SF (P < 0.02) were greater for program 32. All differences were considered to be negligible in the clinical sense.

Hospital optometrists are involved in the care of diabetic patients in a variety of ways, e.g. performing screening as well as training, coordinating and acting in an advisory capacity in local schemes involving GOS optometrists. In response to local need and purchaser interest, an optometrically led diabetic eye screening service has been set up at Walton Hospital. All patients are referred from the Walton Hospital Diabetic Centre, screened by the hospital optometrists and either reviewed by the optometrists or referred to an ophthalmologist routinely for review or referred immediately if it is considered that laser treatment may be required. This paper reports on the protocols, implementation and first year of operation of the service. Patients referred to the clinic were those who had not had an eye examination within the last 12 months (excluding newly diagnosed insulin‐dependent patients) and all newly diagnosed non‐insulin dependent diabetics. An audit was performed over the first 3 months (n = 85). At the same visit, each patient was examined by an HES optometrist using a Volk lens with dilation, the patient was assessed by a consultant ophthalmologist with a specialty interest in diabetics and retinal photography was performed. The same consultant also assessed the slides. The mean age of the patients was 64 years (SD 12 years) and only 6.0% were frank type 1 insulin‐dependent patients. Laser treatment for retinopathy was required in 7.1% of patients and other treatable pathology was found in 9.4% of patients. When compared with the consultant's assessment, the optometrists' examination showed a sensitivity of 87.5% and specificity of 93.5% in the detection and appropriate referral of diabetic eye disease. Peripheral new vessels were missed in one patient due to the presence of lens opacities, and thus the patient was referred routinely instead of immediately. Using the photographic technique, a significant number of images were ungradeable due to poor quality, and the sensitivity was only 66.7%. The lower sensitivity resulted from the inability to detect macula oedema in the absence of hard exudates and peripheral proliferative retinopathy outside the field photographed. In addition, inadequate resolution confounded the detection of microaneurysms in non‐referrable cases. Over the following 9 months a further 182 patients were examined. Laser treatment for retinopathy was required in 6.6% of patients and other treatable pathology was found in 8.2% of patients. Additional funding to expand the service in 1995 has been successfully sought from the purchasers. The results and experiences from this study will be considered in relation to the HES optometrist working within a trust, the St Vincent's declaration and some economic aspects of diabetic healthcare.