A comparison of two fast threshold strategies, SPARK and SITA in threshold perimetry

  • Say Kiang Foo

Student thesis: Doctoral ThesisOphthalmic Doctorate


This thesis investigated the extent of differences between a newer fast threshold strategy, SPARK and the “gold standard” strategy, Swedish Interactive Threshold Algorithm (SITA).
A between-visit comparative study in subjects with glaucoma and suspected glaucoma, subjects with cataract and healthy controls showed SPARK had a lower inter-test variability and was less influenced by the perimetric-experience of subjects compared to SITA. Pointwise analysis found reduction of superior peripheral sensitivity with SPARK in the second visit. Lesser unreliable visual field (VF) results were obtained with SPARK possibly due to its fixation loss monitoring methods and shorter test duration.
SPARK produced about 40% of time-saving and higher estimated sensitivity as compared to SITA in normal subjects. Higher estimated sensitivity may be attributed to reduced fatigue effect, instrumentation and applied algorithms. An acceptable agreement was achieved for mean sensitivity (MS) between the strategies and the bias decreased towards old age.
High bias and large limits of agreement of mean deviation (MD) were found between SPARK and SITA in glaucoma patients and the agreements of global indices between strategies were not achieved. Overestimation of MD and underestimation of pattern standard deviation (PSD) with SPARK led to underestimation of glaucoma severity level and poorer diagnostic sensitivity compared to SITA. The overestimation of MD was also found in cataract patients which led to the masking of diffuse loss. A depressed normative database was possibly used in SPARK algorithm. Between-strategy agreement was not found for the global indices in cataract patients.
Pointwise bias differences of threshold estimates between SPARK and SITA were found to have an almost similar progressive pattern with highest bias in nasal field decreasing towards the temporal field in either glaucoma, cataract or healthy subject which demonstrated a possible systemic difference between strategies. Hence, SPARK is yet to be an alternative strategy to SITA
Date of Award17 Jun 2019
Original languageEnglish
SupervisorRebekka Heitmar (Supervisor) & Robert Cubbidge (Supervisor)


  • fatigue effect,
  • inter-test variability
  • global indices
  • mean deviation,
  • diffuse loss

Cite this