LAUSR

In the 19th century, Sir William Osler, one of the fathers of modern medicine, stated that ‘The good physician treats the disease; the great physician treats the patient who has the disease’, indicating that, as clinicians, we should pursue a patientcentred, rather than diseasecentred approach. In the attempt to provide more patientcentred holistic care, there has been increasing interest in patientreported outcomes (PROs) and patientreported outcomes measures (PROMs), which are any selfreported health status directly reported by patients and the tools (usually questionnaires) used to measure them, respectively. A variety of validated PROMs, either with generic or diseasespecific constructs, are available in ophthalmology to measure the effect of a given eye condition on the overall health function and diseasespecific features (eg, symptoms, treatment side effects, coping), respectively. The widespread use of PROMs across all fields of medicine has caused a paradigm shift in clinical research, with patients’ perception becoming a major determinant of therapeutic interventions’ outcomes. As a result, PROMs are increasingly used as endpoints in clinical trials across all medical fields, including, recently, glaucoma. The choice of primary outcome is critical in randomised controlled trials (RCTs), and determines the sample size. A poorly chosen primary outcome may invalidate the results of otherwise welldesigned and rigorously conducted trials, producing unreliable results and wasting resources. The primary outcome is the outcome measure that has the greatest importance according to the various stakeholders (investigators, patients, policymakers, funding bodies, pharmaceutical companies). Preventing visual disability from glaucoma and preserving the visualrelated quality of life (VR QoL) are arguably the ultimate therapeutic goals in glaucoma management. Glaucoma is an initially asymptomatic and usually slowly progressing disease, so attaining such a hard endpoint as symptomatic visual disability is (fortunately) infrequent, and besides, nearly always requires a long period of time, perhaps measured in decades. Waiting for patients to pass the visual disability threshold would be unethical, impractical and financially prohibitive for a randomised clinical trial. In the absence of appropriate hard endpoints, surrogate endpoints have been used in glaucoma research and clinical practice. According to the US Food and Drug Administration, surrogate endpoints may be used in medical research when the clinical outcome takes a very long time to occur, the clinical benefit of improving a surrogate endpoint is well understood, and conducing a clinical endpoint study would be unethical. In glaucoma research, all these criteria are true, and validated surrogate endpoints have many benefits in the setting of a clinical trial. 4 Compared with true clinical endpoints, less time is needed to appreciate the effect of intervention and a larger number of patients may reach the surrogate endpoints, reducing both RCT required duration and sample size. The benefits of lower IOP and slower functional and structural progression rates are wellestablished. 6 In the case of glaucoma, surrogate endpoints are also more ethical than true endpoint. The case for surrogate measures is not unique to glaucoma. Surrogate endpoints may be better than the true endpoint in many other chronic health conditions sharing similarities with glaucoma. A valid example is AIDS, a chronic condition caused by the retrovirus HIV. Like glaucoma, chronic HIV infection has a long asymptomatic course followed by an unfavourable outcome with severe acquired immunodeficiency and lifethreatening complications. CD4 count and HIV viral load are wellestablished surrogate biomarkers in the field and are widely used in HIV drug trials. Although they might not be the clinical outcomes that the patients perceive as the most important, such endpoints are chosen for practical purposes because the clinical endpoints of interest (AIDS and death) are now relatively rare and take many years to occur because of the introduction of highly active antiretroviral therapy (HAART), which has truly revolutionised the longterm prognosis of chronic HIV infection. A hypothetical randomised clinical trial HAART versus placebo having a PROM as a primary outcome might fail to show any benefit of HAART treatment in the trial time frame or, even worse, favour the placebo because of HAART’s considerable side effects, completely overlooking the established fact that HAART stopped HIV from being a death sentence. In conditions such as chronic HIV infection and glaucoma, the use of a validated surrogate is more effective, practical, ethical and logical than the hard clinical endpoint. Traditionally, visual field (VF) has been used as a primary study outcome, and many landmark RCTs have relied on glaucomatous VF progression as the main outcome measure. VF is not the perfect outcome measure for various reasons. There is no consensus on the best way to detect and quantify glaucomatous VF progression, and the various methods which have been described do not agree well with each other. Variability is an inherent property of VF examination, and the measurement of the true progression rates can be impaired by longterm fluctuations, especially in patients with very marked VF variability. As a result, RCTs based on a VF outcome must have a large sample size, frequent VF testing routine and reasonably long followup. Other proposed surrogate measures include structural measurements, such as optical coherence tomography (OCT)derived measurements. Structural measurements have low test–retest variability and higher signaltonoise ratio in early glaucomatous disease than functional tests but carry their own limitations. Because of the floor effect, they may not be optimal to track glaucomatous changes in advanced disease. Although structural measurements can predict VF outcomes, they are not a test of visual function, and the correlation between structure and function is moderate at best. 16 Both functional and structural measures status Moorfields Eye Hospital NHS Foundation Trust, London, UK Gloucestershire Hospitals NHS Foundation Trust, Cheltenham, UK