June 2021
Volume 62, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2021
The agreement of STATPAC and visualFields statistical packages in visual field analyses
Author Affiliations & Notes
  • Herman Stubeda
    Dalhousie University Faculty of Medicine, Halifax, Nova Scotia, Canada
  • Lesya Shuba
    Department of Ophthalmology and Visual Sciences, Dalhousie University Faculty of Medicine, Halifax, Nova Scotia, Canada
  • Marcelo Nicolela
    Department of Ophthalmology and Visual Sciences, Dalhousie University Faculty of Medicine, Halifax, Nova Scotia, Canada
  • Balwantray C Chauhan
    Department of Ophthalmology and Visual Sciences, Dalhousie University Faculty of Medicine, Halifax, Nova Scotia, Canada
  • Jayme R Vianna
    Department of Ophthalmology and Visual Sciences, Dalhousie University Faculty of Medicine, Halifax, Nova Scotia, Canada
  • Footnotes
    Commercial Relationships   Herman Stubeda, None; Lesya Shuba, None; Marcelo Nicolela, Alcon (C), Alcon (S), Allergan (S); Balwantray Chauhan, Centervue (S), Heidelberg Engineering (C), Heidelberg Engineering (S), Novartis (S), Santen (S), Topcon (S); Jayme Vianna, Eadietech (C)
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2021, Vol.62, 630. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Herman Stubeda, Lesya Shuba, Marcelo Nicolela, Balwantray C Chauhan, Jayme R Vianna; The agreement of STATPAC and visualFields statistical packages in visual field analyses. Invest. Ophthalmol. Vis. Sci. 2021;62(8):630.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : STATPAC is a widely used statistical package for visual field analysis, however, its normative dataset and methods are proprietary. The visualFields (vF) statistical package is a tool analogous to STATPAC but with opensource normative dataset and methods, which provides an advantage for collaboration. The purpose of this study was to determine whether vF analyses agree with STATPAC.

Methods : We defined 3 separate statistical environments containing unique normative datasets: STATPAC, vF with its native normative dataset (vF-SUNYIU), and vF with a new normative dataset of healthy fields from previous studies at our centre in Halifax (vF-Halifax). We created a glaucoma dataset using fields from patients in our glaucoma clinic. All data from Halifax were 24-2 SITA Standard, using only one field per subject. The glaucoma fields were analysed with the 3 environments generating 3 sets of total deviation (TD), pattern deviation (PD), global mean deviation (MD), and pattern standard deviation (PSD) values, which were compared with Bland-Altman plots. We also applied five criteria for glaucomatous fields, i.e., Glaucoma Hemifield Test, Hoddap-Anderson-Parrish 2, Foster, United Kingdom Glaucoma Treatment Study, and Low-pressure Glaucoma Treatment Study within each environment and determined agreement in the criteria results with Kappa statistics.

Results : The Halifax normative dataset contained 163 subjects. The glaucoma dataset contained 1848 subjects with mean (standard deviation) age of 67.1 (12.1) years, and mean STATPAC MD of -3.36 (5.25) dB. The agreement in TD and PD values between STATPAC and the two vF environments was high in the -26 to +5 dB range, but there was notable disagreement below -26 dB, in which both vF environments underestimated field loss by on average 1dB (Fig.1 A&B). The agreement in MD and PSD values between STATPAC and the two vF environments demonstrated a trend where vF underestimated field loss by up to 1dB as field loss increased (Fig.1 C&D). In contrast, agreement in values between vF-SUNYIU and vF-Halifax was high and uniform (Fig.1 A-D). The inter-environment agreement of glaucoma criteria exceeded kappa 0.725 in all cases (Fig.2).

Conclusions : Although STATPAC and vF may have differences in computation of TD values for severely damaged locations, our results highlight a robust agreement, which suggests vF is a viable alternative to proprietary statistical packages.

This is a 2021 ARVO Annual Meeting abstract.

 

 

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×