Purchase this article with an account.
C. A. Castelberg, M. Zulauf, J. M. Wild; Residual Learning Effect Limits the Utility of Short-Wavelength Automated Perimetry (SWAP) Using the Swedish Interactive Threshold Algorithm (SITA). Invest. Ophthalmol. Vis. Sci. 2007;48(13):1626. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To determine, in normal individuals naive to perimetry, in individuals with ocular hypertension (OHT) who were experienced in standard automated perimetry (SAP), and in individuals with open angle glaucoma (OAG) also experienced in SAP, the magnitude of the learning effect for short-wavelength automated perimetry (SWAP) using the Swedish Interactive Threshold Algorithm (SITA) compared to that for SAP using SITA Standard and SITA Fast.
The cohort comprised 82 individuals (30 normal, 26 with OHT and 26 with OAG). Those with either OHT or OAG had previous experience, on average, of 3 examinations per eye with SAP. In the current study, each individual attended for perimetry with the Humphrey Field Analyzer on 5 occasions, each separated by one week. At each visit, both eyes were examined using Program 24-2 and each of the three algorithms; the right eye was always examined before the left eye. The order of algorithm was randomized within individuals between visits over each of the 5 visits. Separate analyses of covariance (ANCOVA) were undertaken for Mean Deviation (MD) and Pattern Standard Deviation (PSD) with age, diagnostic group and order of algorithm as between-subjects factors and eye and visit as within-subjects factors.
The MD improved over the 5 visits (p<0.001) regardless of eye (p=0.405); the improvement was greatest for SITA SWAP (p<0.001) particularly over the initial visits (p=0.011) and for those undertaking SWAP prior to SAP (p=0.025). On average, the MD for SWAP improved from -2.21dB to -0.96dB for the normal individuals, from -4.00dB to -1.89dB for those with OAG and from -1.69dB to -0.51dB for those with OHT. The corresponding improvement across the three groups for SITA Standard was 0.22dB, 0.57dB and 0.10dB, respectively, and for SITA Fast 0.24dB, 0.53dB and 0.41dB. Similarly, PSD improved over the five visits (p=0.007) and the improvement was greatest for SWAP (p<0.001) and for the second eye examined (p=0.013). The differences in the improvement between diagnostic groups failed to reach statistical significance (MD, p=0.074; PSD, p=0.063).
The residual learning effect for SITA SWAP is consistent with that reported elsewhere for the Full Threshold algorithm and further limits the clinical utility of SWAP compared to SAP.
This PDF is available to Subscribers Only