Evidence‐based optometry: Is the von Graefe method still the standard?

BLOODLETTING AND TRADITION The huge outpouring of clinical research in the 20th century showed that many of the treatments used previously by doctors were inert, and any benefits were likely due to the natural progression of the disease and the placebo effect.1,2 Some treatments did more harm than good but persisted. As early as 1855, Dr. John Hughes Bennett used mortality figures to show how dangerous bloodletting was: 33% of patients treated with bloodletting for pneumonia died compared with 3% of those who were left to regain their health.3 Excessive bloodletting was thought to have caused or at least hastened the deaths of Pope Innocent VIII, King Charles II, Queen Anne, Mozart, and George Washington.3 Despite this, bloodletting continued for many decades and was still used in the United Kingdom in the 1950s.3 This was at least partly due to confirmation bias, which means that, once a clinician has a positive opinion of a diagnostic test or management, they remember patient outcomes that support that opinion while ignoring all negative outcomes as “exceptions.”1 There may also be a feeling among more senior clinicians that tradition is important: in a London hospital in 1955, Dr. D. P. Thomas was ordered to use leeches for bloodletting “to maintain some of the old customs.”4 Why is this relevant to optometry? It is because confirmation bias (and other cognitive biases)5 and tradition remain strong drivers to continue to use tests and management plans that clinical evidence have shown to be limited. The March 2025 editorial discussed the research evidence that highlighted the limitations of binocular cross-cylinder and negative relative accommodation - positive relative accommodation as tentative reading add tests compared with simply using a tentative add based on your patient's age6,7 plus their working distance. This month's editorial looks at another test that seems to be used mainly due to confirmation bias and tradition, the von Graefe phoria test. THE GOLD STANDARD? The von Graefe method/technique or von Graefe (hetero)phoria test is used to measure distance and near heterophorias subjectively in a phoropter, with dissociation being achieved using a prism. It is widely used and has been held in very high regard. For example, an article published in this journal in 1996 proposed the von Graefe phoria measurement as the gold-standard test of phoria measurement.8 The words used are worth repeating: “…we propose that the von Graefe phoria test is the valid method of phoria measurement. This test has face validity, in as much as efforts are made to assure that all stimuli to fusion are absent and that the patient is accommodating for the target distance. In addition, it provides a method of actual measurement as compared with an estimate as in the cover test. The purpose of the study reported here is to determine the validity of the estimated cover test, that is, the ability of this test to predict the results of the von Graefe phoria test.” With results that showed “the estimated cover test can grossly underestimate or overestimate the von Graefe measurement,” the authors concluded that “Although the estimated phoria test is useful as a screening test, it should not be used as a substitute for the von Graefe phoria measurement. In all cases, it should be followed by prism neutralization or by von Graefe phoria measurement.”8 At that time, the von Graefe method was the principal measurement of subjective post-refraction heterophoria discussed in standard optometry textbooks.9,10 VALIDITY AND RELIABILITY/REPEATABILITY ASSESSMENTS The next article on this topic, published in this journal 2 years later, assessed the interexaminer repeatability of the various phoria measurements.11 The 95% confidence limits of agreement between two examiners are shown in Table 1. By far, the least repeatable were the two variations of the von Graefe test, which were up to twice as variable as the cover test and three times as variable as the Thorington tests.11 In a later study published in 2002, interexaminer repeatability was again poorer with the von Graefe test (±5.2Δ), this time in comparison with the Howell phoria card (±3.5Δ) and the Bernell Muscle Imbalance Measure (±2.8Δ).12 Two later studies found that the Maddox rod test showed larger limits of interexaminer repeatability variable than the von Graefe ([±5.7Δ vs. ±5.3Δ]13 and [±3.5Δ vs. ±2.5Δ]14), but the von Graefe was again shown to be more variable than the cover test (±2.5Δ vs. ±1.7Δ)14 and modified Thorington ([5.3Δ vs. ±3.3Δ]13 and [2.5Δ vs. ±1.4Δ]14). In addition, the Maddox rod test showed good repeatability when used with a trail frame.13 TABLE 1 - 95% Confidence limits of interexaminer agreement for seven heterophoria measurement techniques Interexaminer repeatability von Graefe — continuous presentation ± 8.2Δ von Graefe—flashed presentation ± 6.7Δ Prism-neutralized objective cover test ± 3.6Δ Estimated cover test ± 3.3Δ Prism-neutralized subjective cover test ± 2.5Δ Thorington ± 2.5Δ Modified Thorington ± 2.3Δ Modified from Rainey et al.11 The poor repeatability of the von Graefe method may result from variable amounts of prism adaptation, phoropter-induced proximal accommodation, a head tilt behind the phoropter leading to an induced vertical deviation, or a reduction in peripheral fusion.13 In addition, it is a relatively lengthy procedure and can be difficult for patients to understand. COVER TEST AS THE GOLD STANDARD? The von Graefe method cannot be the gold standard test of heterophoria measurement given its poor repeatability. Although it is not possible to use the cover test reliably in a phoropter, it seems a more appropriate gold standard of phoria measurement given that it is repeatable, can be used in objective or subjective form and estimation or measurement (by prism) form, is the only test that can discriminate between a heterophoria and a heterotropia, and provides useful additional information about the control of any phoria via the fluency of recovery. If the 1996 study had used the cover test as the gold standard and assessed the von Graefe method, it would have reported that the von Graefe method “can grossly underestimate or overestimate” the gold standard cover test results, and this seems a more appropriate summary of the two tests. HOW BEST TO MEASURE HETEROPHORIA SUBJECTIVELY IN A PHOROPTER The research literature is clear that, by far, the most repeatable of the phoria tests that can currently be used in a phoropter is the modified Thorington, which suggests that this is the best test to use in clinics.11–15 It is also a quicker and simpler test than the von Graefe method. In a recently published optometry textbook, the von Graefe method still retains its prime position: it is described in the first two sections of the Functional Tests chapter in Clinical Procedures for Ocular Examination, whereas the modified Thorington languishes in section 13.16 However, at least the Thorington test is included, and the von Graefe section acknowledges that “Literature suggests that the von Graefe technique is the least repeatable of the oculo-motor deviation methods available….” I assume that the use of the von Graefe will diminish over time, and a better alternative, currently the modified Thorington, will replace it. It just may take longer than we expect.