A new health technology: where is the consensus on a clinically worthwhile benefit?

AIM: New therapies are often introduced into the NHS prior to full evaluation, leading to inequities in provision. Uncertainty exists regarding the value of photodynamic therapy in the treatment of neovascular age-related macular degeneration. We ascertained the availability of this treatment and the information used to inform clinical policy. METHODS: A postal survey of all clinical directors/lead consultants in the UK sought data on which (if any) patients were referred or treated with PDT by their unit, the sources of evidence informing clinical policy and the threshold of clinical benefit at which respondents would support the use of PDT. RESULTS: 123/152 questionnaires were returned. 42% of units make some provision for PDT on the NHS, including routine provision by 9%. 14.5% of units offer the option of care in the private sector, whilst 26.5% treated or referred no patients. The threshold at which respondents considered introduction of PDT would be justifiable varied widely. Respondents cited local literature review, advice from clinicians, guidance from the Royal College and information from the pharmaceutical industry as most influential in determining current policy. However, the National Institute for Clinical Excellence (NICE) and the Cochrane Library were anticipated as playing a greater role in shaping future practice. CONCLUSIONS: Substantial variation exists in the availability of PDT. Advocates of PDT may interpret our data as an indication of the NHS failing to provide an effective therapy equitably, whilst others may deduce that patients are receiving an under-evaluated treatment in routine clinical practice. The differing thresholds at which clinicians believe treatment would be justified may further exacerbate variations and the priority given to PDT.     

Texture segmentation: Do the processing units on the saliency map increase with eccentricity?

The saliency map is a computational model and has been constructed for simulating human saliency processing, e.g. pop-out target detection (e.g. Itti & Koch, 2000). In this study the spatial structure on the saliency map was investigated. It is proposed that the saliency map is structured into processing units whose size is increasing with retinal eccentricity. In two experiments the distance between a target in the stimulus and an irrelevant structure in the mask was varied systematically. Our findings had two main points. Firstly, in texture segmentation tasks the saliency signals from two texture irregularities interfere, when these irregularities appear within a critical spatial distance. Second, the critical distances increase with target eccentricity. The eccentricity-dependent critical distances can be interpreted as crowding effects. It is assumed that additionally to the target eccentricity, also the strength of a saliency signal can determine the spatial area of its impairing influence.     

Sixty strabismus cases operated with the Computerized Strabismus Model 1.0: When does it benefit, when not ?

While, in routine strabismus surgery, empirical guidelines and experience are the best in judging which eye muscles to operate, a complex case may need a unique surgical approach, the consequences of which cannot always be envisioned in detail. We sought to improve the results of surgery in these cases by preoperative simulation of each case with the Computerized Strabismus Model 1.0 (CSM). The basis of this model was laid by David A. Robinson. It has been improved by us over the past years to the point that it can be used clinically. Improvements concerned, for example, the mechanics of the eye muscles and the anatomy of insertions and origins. The ease of operation has been improved and the algorithms have been made so much faster that a full calculation for 9 positions of gaze now takes 10 seconds on a hand-held Hewlett Packard 200LX Palmtop. From 1994 onwards, all cases to be operated in our department which were more complex than straightforward horizontal rectus muscle surgery were simulated in the model preoperatively. The predictions of the model compared well with the actual result of surgery in most cases. The model was particularly good in handling complex and unique disorders of motility. However, the model could not reliably predict the effect of strabismus surgery in cases with mechanical restrictions of motility.