*5.7. Data Interpretation: Importance of Training*

The accuracy of the data interpretation increased with the level of experience/training of the non-medical staff. As the optometrists from the SC-MEH [40,42] and SC-QMC [39] got extra training in these tasks, their interpretation of the fundus photographs and VF was more accurate than in the SC-MC. The lack of training could also explain why the agreement on evaluating OCT between optometrists from the SC-MC and GEs was less good, and worse than the agreement between GEs [33]. The optometrists from the SC-MEH and SC-QMC assessed the optic disc through slit-lamp examination and showed a high agreement with the GE because they were trained to use these devices [39,40,42]. None of the VC-articles investigated the accuracy of the non-medical staff.

#### *5.8. Quality of Management Decisions*

In their decisions on progression/referral, the non-medical staff of the SC-GFU followed the referral criteria strictly [36,37]. More importantly, adherence to these criteria increased when, besides referring patients to the GE directly [37], they could also ask for the GE's advice [36]. The level of agreement on progression between optometrists and the GEs from the SC-MC was only "fair", but similar to the level of agreement between two GEs. A point of concern was that almost 1/3 of the glaucoma cases being progressive would not be referred to the GE. Possible reasons were an incorrect interpretation of data (see above) and not using all data when making decisions. However, optometrists tended to be overcautious in general. In both the SC-MEH and GCC, the optometrists classified more patients as progressive or at higher risk than the GE [42,51]. Most likely, the reason was to make a safer decision. As a consequence, the optometrists tended to order more additional tests than the GEs [39,40,42].

Decisions on discharge/follow-up were also safeguarded. In the GCC, decisions of the non-medical staff in this respect were supervised virtually by the GE [51]. In the GFU, the non-medical staff could not discharge and could only decide to keep or shorten the interval as planned [36–38]. Similarly, in the GAC the non-medical staff could not discharge, and the GE would (virtually) assess the patient within a maximum time interval of three months [48]. The agreement with the face-to-face diagnosis was also high. In the GSC out of the 16 patients for whom the diagnosis differed between the face-to-face and virtual review, only three patients required medical intervention [49]. Two of these patients were diagnosed as having OHT, one of whom had an IOP at the face-to-face consultation which was twice as high as what had been found in the virtual review and in the referral letter. The third patient had narrow, occludable angles requiring prophylactic laser iridotomy [49]. In the SMS, the sensitivity of detecting unstable cases was dependent on the expertise of the GE; a higher sensitivity was noted for the consultant than for the fellow, both in the interand intra-observer agreement analyses [56]. The arbitrary stable/unstable classification, which was used in the SMS for deciding on the time to the next FU appointment, was explained as a possible reason for the low sensitivity [56]. However, since only stable and low-to-moderate risk patients could enter the clinic, the actual number of missed unstable patients was low, and even lower with advanced VF loss [56]. Due to the wide confidence intervals, it was suggested to perform more extensive studies to provide a

more accurate answer on the virtual clinic's sensitivity [56]. In the SC-MEH studies, the optometrist(s) and the GE independently decided on the follow-up interval and the further treatment [40,42]. The agreement was high and comparable to the agreement between two GEs. In SC-QMC study, where the GE had to state (dis)approval with the optometrists' decisions, the agreement was even higher [39].

#### *5.9. Acceptance*

All findings showed that the acceptance of care provided by a SC clinic or a VC was at least as good as in the StC clinic. In the SC-GFU, the GE scored higher on "knowledge" and "information received"; however, the difference was too small to be relevant [37]. In the study of Spackman et al., 98% of patients felt that the glaucoma VC was the same or better than the StC [55].

Acceptance of SC and VCs by the GEs was overall good; some medical staff however found their work in SC to be tedious [37].
