Optimal surveillance regimes for individuals with ocular hypertension (OHT): modelling and economic evaluation

DESIGN: Several approaches will be used across three interlinked work packages to achieve our objectives: expert and consumer opinions; population preferences; systematic reviews; individual patient data analysis (and possible meta-analysis); and decision analytic modelling. SETTING: Primary care optometry and hospital eye clinics. HEALTH TECHNOLOGIES BEING ASSESSED: Tonometers (technologies used to measure intraocular pressure) including contact, non-contact tonometry and self-measurements tonometers; automated perimetry (technology to measure the visual field). TARGET POPULATION: Adults of 18 years and over with raised intraocular pressure. SEARCH STRATEGY: Sensitive electronic searches to identify reports of published and ongoing studies, abstracts from conference proceedings and other grey literature sources. Further focused searching will be developed for the economic model. REVIEW STRATEGY: 1) Assess identified prediction models according to the data and methods used in their development, and the quality of validation studies. 2) Update an existing systematic review of the performance of alternative tonometers to determine the optimal technology, in terms of measurement error. 3) Systematic identification of parameters required for the economic modelling. MODELLING EVALUATION OF ALTYERNATIVE SURVEILLANCE REGIMES: Four relevant individual patient datasets have been identified and access has already been agreed. Estimates of the within-person variability, between-person variability, and the natural trend of intraocular pressure within specified periods are needed to ascertain OHT. Furthermore, estimates of the variability of relevant clinical measures of glaucoma are required to determine the adequate frequency and number of repeated measurements necessary for monitoring. These estimates will be obtained from the identified databases. We will independently validate the identified prediction model in several databases. Models will be compared both in terms of their calibration (using calibration plots) and discriminating ability (using ROC analysis). SURVEY STRATEGY: We will survey a sample of the adult UK population and services users from the observational cohorts (Nottingham and Dunfermline) to determine preferences for the process and outcomes of testing, including consideration of the frequency of testing, using a Discrete Choice Experiment. MEASUREMENT OF COST AND OUTCOMES: An economic model will be developed based on discrete event simulation methods. The model structure and parameters will be based on the findings of evidence synthesis, prediction model, surveys and discussion with a convened advisory panel of experts and consumer representatives.