Manufacturer changes lead to clinically important differences between two editions of the TNO stereotest

Stereoacuity tests used in clinical practice should be repeatable and reproducible. However, it has been observed in a clinical setting that new editions of the TNO stereotest appear to give different values from those obtained using previous versions. The purpose of the present research was to investigate this observation.

One hundred and twenty-one Dutch subjects, 88 (73%) females and 33 (27%) males, with an average age of 34.0 years (range 18-55) had their stereoacuity measured using two different versions of the TNO stereoacuity test (TNO 13 and TNO 15). The TNO was tested in a counterbalanced order so that consecutive subjects started with alternate editions to avoid bias.

There was a significant difference (p < 0.001) between the median value for stereoacuity measured with TNO 13 (30 s of arc) and TNO 15 (60 s of arc). The bias between the two test versions was -0.23 Log arcseconds (95% limits of the differences: 0.15 to -0.60 Log arcseconds).

This study reveals that results obtained with two different editions of a commonplace stereoacuity test are not comparable. New versions come on the market at regular intervals and the assumption that they will give the same results as previous versions may not be valid. Besides the statistically significant difference between the TNO 13 and TNO 15, the Bland-Altman plot also showed a considerable bias and the 95% limits of the differences between the TNO 13 and TNO 15 are more than two steps on the Log arcsecond scale. This difference between two editions of the TNO stereotests is not clinically acceptable and therefore it is inappropriate to use the two versions of the test interchangeably. It is important in both research and clinical records to specify the edition of the TNO test used.