3.4.1. Quality of Care (QoC)

Distinction is made between the QoC provided by the non-medical staff and the QoC provided by the SC clinic in general (see Table 4). Quality is measured by evaluating completeness, accuracy and management decisions.

**Table 4.** The quality of care provided by the non-medical staff and the quality of care provided by the shared care clinic in general.



#### **Table 4.** *Cont.*

Abbreviations: SC = shared care clinic; StC = standard care clinic; \* = statistical significant difference (*p* ≤ 0.05); \*\* = no statistical significant difference (*p* > 0.05); κ = kappa; IQR = interquartile range; GE = glaucoma expert; NMS = non-medical staff; Opto = optometrist; HCP = health care providers; IOP = intra-ocular pressure; VA = visual acuity; VF = visual field; Gonio = gonioscopy; OCT = optical coherence tomography; HRT = Heidelberg retinal tomography; GDx = GDx ECC scanning laser polarimetry; CCT = central corneal thickness; Combined compliance\* = combined completion of visual field, gonioscopy, measurement of central corneal thickness, and imaging (OCT or fundus photographs); SOF = suspicion of progression; VF: 41.2% \*\*\*: Out of the 34 patients who required a visual field examination on a yearly basis, 20 patients did not receive it in the SC-GFU; mo = month(s); AAO PPPg = American Academy of Ophthalmology Preferred Practice Pattern guidelines; ANGIG&RANZCO = recommendations of the Australian and New Zealand Glaucoma Interest Group and the Royal Australian and New Zealand College of Ophthalmologists; SC-RVAC = shared care clinic, established between the Royal Victorian Eye & Ear Hospital and the Australian College of Optometry; GFU = glaucoma follow-up unit; MC = Mayo Clinic's campus in Rochester; MEH = Moorfield Eye Hospital; QMC = Queen's medical centre; GHGC = Greenwich hospital glaucoma clinic; SGC = Stable Glaucoma Clinic in New Zealand; GMC = Glaucoma Management Clinic in Australia.

Performance of the Non-Medical Staff

Performance of the non-medical staff was evaluated by comparison with the "gold standard", which was the performance of the GE or a working protocol of the corresponding SC clinic.

• Completeness of data collection:

The non-medical staff working at the SC-GFU performed the required tests as per protocol in almost all visits [36–38]. VF was the only test with a poor compliance, i.e., in only 25.4% of the visits that required VF according to the protocol, the test was actually performed. Of note, also in the StC-GFU run by the GE, VFs were only performed in 16.9% of the visits where a VF was required according to the protocol [36].

• Accuracy of data collection:

Agreement between GEs and optometrists on IOP was evaluated in two studies [33,40] and was found to be good. Banes et al. [40] noted that optometrists tended to record lower IOP, but differences were small.

Agreement on structural glaucomatous damage was evaluated in three studies [33,40,42]. When performing slit-lamp examination, the optometrist's cup/disc ratio was comparable to that of GEs [40] and the optometrist's ability to decide whether or not an optic disc was glaucomatous was also found to be good (sensitivity and specificity ~83.0%) [42]. When evaluating fundus photographs on stability, the agreement between all HCP (GEs and optometrists) of the SC-MC [33] was found to be good and comparable to the agreement between GEs alone. Banes et al. [40] demonstrated the (dis)agreement rate to be independent of the cup/disc ratio values. Only Shah et al. [33] examined the agreement on OCT interpretation between all HCPs, including GEs and optometrists, and found it to be "fair". The study of Phu et al. [44] evaluated the agreement between optometrists and ophthalmologists on gonioscopy. The agreement in the exact assessment of the angle was "fair to moderate". Consistency with a final diagnosis, whether the angle was open or closed, was 93.4% [44].

Agreement on functional glaucomatous damage was evaluated in four studies [33,39,40,42]. The agreement on VF-status was "fair" [42], "moderate" [33] and "almost perfect" [39,40]. Banes et al. [42] pointed out that optometrists were more cautious than GEs, by classifying more eyes as being "progressive".

• Management decisions:

Several studies examined the non-medical staff's ability to make management decisions based on their interpretation of tests and examinations [33,36,37,39,40,42].

As for glaucoma status, in the SC-GFU, the non-medical staff referred half of the cases which met one of the back-referral criteria in the protocol back to the GE. Out of the cases that met the GDx- or VF-criterion (indicating suspected progressive damage), 92.0% and 75.0% of the cases, respectively, were actually sent back. These values amounted to 66.7% for the IOP-criterium (IOP > tIOP) and 36.0% for the VA-criterium (declined ≥ 2 lines) [37]. In the SC-GFU, the non-medical staff could also opt to seek advice of the GE when one the above criteria was met. In 100% of the cases that met the GDx- or the VF-criterium, the non-medical staff asked for advice or referred back. This value amounted to 84.6% and 68.2% for the IOP- and VA-criterium, respectively [36].

In the SC-MC, disease progression was defined as IOP > tIOP, progression on optic nerve photographs, OCT or VF [33,34]. Shah et al. [33] showed a "fair" level of agreement on glaucoma progression diagnosis between all HCPs (optometrists and GEs) and between GEs alone. The level of agreement between all HCPs was higher when relying on IOP or disc hemorrhages compared with the agreement when relying on OCT or VF [33]. Of all the available test data, the OCT and VF data were considerably less used by the optometrists than by the GEs. This discrepancy in use was also reflected in the high discrepancy in interpreting OCT and VF between all HCPs (agreement of 36.0%, κ = 0.26, for OCT, and agreement of 53%, κ = 0.45, for VF) [33].

Two other articles evaluated the agreement between HCPs on whether a patient should be discussed with the GE [39,42]. Ho et al. [39] found this agreement to be "almost perfect" between the GE and the non-medical staff. Banes et al. [42] found this agreement to be slightly smaller (72.0%), but equal to the agreement between two GEs on whether a patient should be discussed with them. Three studies [39,40,42] evaluated the agreement on disease status by using the proposed follow-up interval as a measure. A shortened interval

indicated that the disease status was judged to be worsening. Overall, the agreement was "almost perfect" [39,40]. Only Banes et al. [42] showed a "fair" agreement.

As for ordering tests, the non-medical staff of the SC-MH [40,42] and the SC-QMC [39] was allowed to do so. Ho et al. [39] showed a high agreement on ordering a VF at the next appointment between the optometrists and GE. Although Banes et al. [42] assessed lower values, the agreement was still good and similar to the agreement between two GEs. In both clinics, the optometrists tended to order more additional tests than the GEs [39,40,42].

The non-medical staff of the SC-MH [40,42] and the SC-QMC [39] were also able to decide on further treatment. In both clinics, agreement was high for both the medical and surgical treatments.

Performance of the SC Clinic

In this case, the "gold standard" corresponds to the StC or the guidelines used (Table 4).

• Completeness of data collection:

The Mayo Clinic showed an increase in compliance on initial testing to the American Academy of Ophthalmology Preferred Practice Pattern (AAO PPP) guidelines [45] after implementation of the SC-MC [35]. Similarly, the SC-RVAC [41] showed a high compliance to both the AAO PPP [45] and the ANGIG&RANZCO recommendations [28]. The compliance on rate of testing was weak, but similar, for VF in both SC-GFU and StC-GFU [36,37].

• Accuracy of data collection:

No difference was found between the results obtained by the SC-GFU [36,37] and the StC-GFU. One exception was VA, which declined in more visits of the StC-GFU than in the SC-GFU [36]. Holtzer-Goor et al. attributed this difference to the different protocols used in both clinics; the SC-GFU had to perform VA at every visit while the StC-GFU had to perform VA only when judged to be necessary (but at least once a year) [36]. In other words, the StC-GFU would mainly perform VA for those patients who mentioned having difficulties with their sight [36]. The implementation of SC-RVAC resulted in a 14.0% increase in correct diagnosis when assessing the optic nerve compared to the StC clinic [41].

• Management decisions:

No difference was found between the StC-GFU and the SC-GFU in the decision on the number of patients judged to be stable or progressive [36,38]. Holtzer-Goor et al. concluded that a SC scheme did not miss a significant number of cases of suspected progression [36]. Damento et al. assessed the decision on "disease status" in the Mayo Clinic by using the "number of patient visits" as a measure [34]. The rationale was that, if an HCP judged the disease status to be worsening, that HCP decided to shorten the follow-up interval, which resulted in more visits taking place in a certain amount of time. No difference was found in the number of patient visits between the SC-MC and the StC-MC [34].

Furthermore, the number of treatment changes was similar between the SC-GFU and the StC-GFU [36–38]. Moreover, no difference was found concerning the reason for change, i.e., IOP exceeding the tIOP, intolerance to the medication, structural or functional progression [36,37]. Likewise, the number of procedures carried out in the SC-MC and the StC-MC did not differ [34]. However, the number of procedures performed by the GE tended to increase after implementation of the SC scheme [34].

#### 3.4.2. Acceptance

#### Patients

Patient satisfaction was about the same in the SC-GFU and the StC-GFU [36,37]. No difference was noted in the dimensions "overall mark", "knowledge", "waiting area", and "information received". Patients scored the SC-GFU higher on 'taking sufficient time" and "giving sufficient information".

When comparing HCPs, Holtzer-Goor et al. assessed a higher score on the "overall mark" for the non-medical staff [36]. The GE got a higher score on the dimension "information received". Patients gave both the GE and the non-medical staff similar scores on "knowledge" and "waiting area". Lemij et al. [37] found similar scores as Holtzer-Goor et al. [36], but assessed a higher score for the GE on "knowledge" and "information received". In the SC-RVAC, almost 95% of the responders opted to be treated in the SC-RVAC rather than remaining on the waiting list of the StC-RVAC [41].

### Staff

All clinicians of the SC-RVAC found the SC clinic an excellent opportunity to exchange knowledge, and 82.0% wanted to stay working in the clinic [41]. Similarly, the GE and the ophthalmic technicians were very pleased to work in the SC-GFU [37,38]. The ophthalmic technicians indicated the patient contact and the increased responsibility to be the main reasons. However, the optometrists working in the SC-GFU found their work tedious, and thought the shared clinic was not working satisfactorily [37].
