December 2008
Volume 49, Issue 12
Free
Glaucoma  |   December 2008
Screening Tests for Detecting Open-Angle Glaucoma: Systematic Review and Meta-analysis
Author Affiliations
  • Graham Mowatt
    From the Health Services Research Unit, Institute of Applied Health Sciences, College of Life Sciences and Medicine, University of Aberdeen, Aberdeen, Scotland, United Kingdom; the
  • Jennifer M. Burr
    From the Health Services Research Unit, Institute of Applied Health Sciences, College of Life Sciences and Medicine, University of Aberdeen, Aberdeen, Scotland, United Kingdom; the
    Eye Clinic, Aberdeen Royal Infirmary, Aberdeen, Scotland, United Kingdom; and the
  • Jonathan A. Cook
    From the Health Services Research Unit, Institute of Applied Health Sciences, College of Life Sciences and Medicine, University of Aberdeen, Aberdeen, Scotland, United Kingdom; the
  • M. A. Rehman Siddiqui
    Eye Clinic, Aberdeen Royal Infirmary, Aberdeen, Scotland, United Kingdom; and the
  • Craig Ramsay
    From the Health Services Research Unit, Institute of Applied Health Sciences, College of Life Sciences and Medicine, University of Aberdeen, Aberdeen, Scotland, United Kingdom; the
  • Cynthia Fraser
    From the Health Services Research Unit, Institute of Applied Health Sciences, College of Life Sciences and Medicine, University of Aberdeen, Aberdeen, Scotland, United Kingdom; the
  • Augusto Azuara-Blanco
    Eye Clinic, Aberdeen Royal Infirmary, Aberdeen, Scotland, United Kingdom; and the
  • Jonathan J. Deeks
    Department of Public Health and Epidemiology, University of Birmingham, Birmingham, United Kingdom.
Investigative Ophthalmology & Visual Science December 2008, Vol.49, 5373-5385. doi:10.1167/iovs.07-1501
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Graham Mowatt, Jennifer M. Burr, Jonathan A. Cook, M. A. Rehman Siddiqui, Craig Ramsay, Cynthia Fraser, Augusto Azuara-Blanco, Jonathan J. Deeks; Screening Tests for Detecting Open-Angle Glaucoma: Systematic Review and Meta-analysis. Invest. Ophthalmol. Vis. Sci. 2008;49(12):5373-5385. doi: 10.1167/iovs.07-1501.

      Download citation file:


      © 2016 Association for Research in Vision and Ophthalmology.

      ×
  • Supplements
Abstract

purpose. To assess the comparative accuracy of potential screening tests for open angle glaucoma (OAG).

methods. Medline, Embase, Biosis (to November 2005), Science Citation Index (to December 2005), and The Cochrane Library (Issue 4, 2005) were searched. Studies assessing candidate screening tests for detecting OAG in persons older than 40 years that reported true and false positives and negatives were included. Meta-analysis was undertaken using the hierarchical summary receiver operating characteristic model.

results. Forty studies enrolling over 48,000 people reported nine tests. Most tests were reported by only a few studies. Frequency-doubling technology (FDT; C-20-1) was significantly more sensitive than ophthalmoscopy (30, 95% credible interval [CrI] 0–62) and Goldmann applanation tonometry (GAT; 45, 95% CrI 17–68), whereas threshold standard automated perimetry (SAP) and Heidelberg Retinal Tomograph (HRT II) were both more sensitive than GAT (41, 95% CrI 14–64 and 39, 95% CrI 3–64, respectively). GAT was more specific than both FDT C-20-5 (19, 95% CrI 0-53) and threshold SAP (14, 95% CrI 1-37). Judging performance by diagnostic odds ratio, FDT, oculokinetic perimetry, and HRT II are promising tests. Ophthalmoscopy, SAP, retinal photography, and GAT had relatively poor performance as single tests. These findings are based on heterogeneous data of limited quality and as such are associated with considerable uncertainty.

conclusions. No test or group of tests was clearly superior for glaucoma screening. Further research is needed to evaluate the comparative accuracy of the most promising tests.

Glaucoma describes a group of eye diseases in which there is progressive damage to the optic nerve, leading to impaired vision and, in some cases, blindness if untreated. Glaucoma is the leading cause of irreversible blindness worldwide, 1 2 with open-angle glaucoma (OAG) the most common form. 1 Late detection is a major risk factor for blindness. 1 3 4 5 It is estimated from population surveys that in developed countries, more than 50% of prevalent OAG is undetected, 6 and this estimate is likely to be higher in developing countries. Recent evidence suggests that treatment is effective at delaying progression 7 8 ; thus, population-based screening of OAG is under consideration. 6 9 10 11 For screening to be considered, several criteria have to be met regarding the condition, the test, and the screening program. 9  
Tests for glaucoma involve an assessment of structural changes at the optic nerve head, functional visual loss by visual field testing, and the level of the intraocular pressure (IOP). There are many potential tests or combinations of tests for detecting glaucoma; however, to date, no single test or combination of tests has been identified as optimal in screening for glaucoma. 
The purpose of this study was to assess the comparative accuracy of candidate screening tests. 
Methods
Search Strategy
Highly sensitive electronic searches, using both controlled vocabulary and free text terms, were undertaken. We searched the following electronic databases: Medline (1966 to week 3, November 2005), Medline In Process (February 23 and December 6, 2005), Embase (1980 to 2005 week 49), Science Citation Index (1981 to December 3, 2005), Biosis (1985 to November 30, 2005), and Cochrane Central Register of Controlled Trials (CENTRAL; The Cochrane Library, Issue 4, 2005). In addition, full-text electronic searches of the American Journal of Ophthalmology (1998 to November 2005), Ophthalmology (1998 to November 2005), British Journal of Ophthalmology (1998 to November 2005), Investigative Ophthalmology and Visual Science (1998 to November 2005), and the Journal of Glaucoma (2001 to November 2005) were undertaken. Searches were restricted to English language publications. The reference lists of included studies were scanned to identify additional potentially relevant reports. Full details of the sources searched and search strategies used are available elsewhere 6 or can be obtained by contacting the authors. 
Inclusion and Exclusion Criteria
We included studies that assessed the accuracy of tests for detecting OAG in people older than 40 years who were likely to be representative of a screening situation (i.e., no selection and no previous tests had been performed) or of a group of patients with suspected glaucoma (i.e., patients identified from prior testing as possibly having glaucoma or as having, e.g., high IOP, or another risk factor for glaucoma but with an unconfirmed diagnosis). Both randomized (where participants were randomized to one or more tests) and observational (both cohort and case–control) studies were included. The reference standard was either confirmed OAG on follow-up or ophthalmologist-diagnosed OAG, as reported by the study. This latter reference standard required a clinical judgment by an ophthalmologist, including an evaluation of the optic nerve and a measure of visual function. In addition, the study had either to report or to allow the calculation of true and false positives and negatives. 
Non-English-language reports were excluded, as were conference abstracts. Case reports and studies investigating technical aspects of a test were excluded. Case–control studies in which the control group consisted of people with no ocular disease or specifically excluded people with other ocular disease, so that the spectrum of disease and nondisease was unlike that to be encountered in a screening situation, were also excluded. The spectrum of disease expected would be similar to the spectrum of the disease of the general population (e.g., more patients with mild glaucoma, fewer patients with severe glaucoma). 
The candidate tests fell within three broad categories: (1) structure (ophthalmoscopy, optic disc photography, retinal nerve fiber layer [RNFL] photography, Heidelberg retinal tomography [HRT] version II [Heidelberg Engineering, Heidelberg, Germany], GDx VCC retinal nerve fiber layer [RNFL] analyzer [Carl Zeiss Meditec, Oberkochen, Germany], optical coherence tomography [OCT], and retinal thickness analyzer [RTA]); (2) function (oculokinetic perimetry [OKP], white-on-white standard automated perimetry [SAP] including suprathreshold and threshold, short wave-length automated perimetry [SWAP], frequency-doubling technology [FDT], and motion-detection perimetry [MDP]); and (3) IOP (Goldmann applanation tonometry [GAT]; noncontact tonometry [NCT]; TonoPen; Reichert, Vienna, Austria). 
Data Abstraction and Quality Assessment
Two reviewers undertook single-data extraction of the included studies. In the event of uncertainty, the other reviewer provided advice and validated the data extraction. 
Two reviewers independently assessed the quality of the included studies using a version of QUADAS adapted for assessing reports of the accuracy of screening tests for OAG. QUADAS is a quality-assessment tool for use in systematic reviews of diagnostic studies. 12 Disagreements were resolved by consensus or arbitration by a third reviewer. A higher quality study was considered one for which “yes” was checked in response to questions 1 (patient spectrum representative), 3 and 4 (partial and differential verification bias avoided), and 6 and 7 (test review bias and diagnostic review bias avoided) on the adapted QUADAS checklist. 
Statistical Methods
After data extraction a common (most frequently reported) cutoff for each test was selected after discussion by two ophthalmologists (JMB, MARS). Summary receiver operating characteristic (SROC) curves were produced for each test where two or more studies reported estimates of sensitivity and specificity at the common cutoff. Meta-analysis models were fitted using the hierarchical summary receiver operating characteristic (HSROC) model 13 in WinBUGS 1.4. 14 Normally distributed random effects were assumed with noninformative uniform priors. No adjustment was made for the correlation between results from paired studies, as the level of information required is rarely reported. Summary sensitivity, specificity, and diagnostic odds ratios (DORs) at the operating point were reported for each model as the median and 95% credible interval (CrI). A DOR is a single indicator of test performance and is the ratio of the odds of testing positive in those with the disease relative to the odds of testing positive in those without the disease. 15 It can be calculated from the sensitivity and specificity: DOR = [sensitivity/(1 − sensitivity)]/[(1 − specificity)/specificity]. 
Credible intervals are the Bayesian equivalent of confidence intervals. A simplified model, which assumed a symmetrical ROC shape, was used where limited data caused convergence problems under the full model. Sensitivity analysis was undertaken by examining separately the results of the higher quality studies, using HSROC analysis where more than one higher quality study reported the same test. 
Comparisons between tests were made in two ways: First, studies in which participants were directly compared who either received all tests or were randomized to different tests were identified, and the direct comparisons inspected. Second, an indirect comparison between tests was made, for all tests reported by two or more studies were modeled together in a single HSROC model to formally compare test performance. Pair-wise differences in sensitivity and specificity between tests were assessed from the median difference and corresponding 95% CrI. 
Results
Trial Flow
Figure 1shows the flow of studies through the review. Out of a total of 5918 titles/abstracts screened, 877 potentially relevant full text articles were obtained, with 40 studies, published in 46 reports, meeting the inclusion criteria. 
Study Characteristics and Methodological Quality
Twenty studies were population-based and representative of a screening setting 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 whereas 20 studies were considered representative of patients with suspected glaucoma referred from primary care, of which 8 were cohort studies 40 41 42 43 44 45 46 47 and 12 were case–control studies. 48 49 50 51 52 53 54 55 56 57 58 59 60 61 Seven studies 18 34 40 43 44 48 58 used the first and best reference standard of OAG confirmed on longitudinal follow-up, whereas the remainder used ophthalmologist-diagnosed OAG. The characteristics of the included studies are shown in Table 1
The 40 studies enrolled more than 48,000 people, with more than 39,000 included in the analysis. The studies took place from 1963 to 2004. In 26 studies reporting the sex of the participants, 51% were women. The median (range) age of participants across studies was 60.5 years (13–97 years). The reports included several major population-based prevalence surveys, such as the Baltimore Eye Survey, 25 31 the Blue Mountains Eye Study, 23 the Crete, Greece, Glaucoma Study, 27 the Dalby Population Survey, 17 the Egna-Neumarkt Study, 18 the Framingham Eye Study, 43 the Glaucoma Screening Study (GLASS), 24 26 the Groningen Longitudinal Glaucoma Study, 53 54 59 the Rhondda Valley Study, 22 the Rotterdam Study, 38 the Segovia Study, 16 and the Visual Impairment Project. 37  
The included studies reported the following tests: ophthalmoscopy (seven studies); optic disc photography (six studies); RNFL photography (four studies); HRT II (three studies); OKP (four studies); SAP (14 studies); FDT (eight studies); GAT (nine studies); and NCT (one study). No reports of GDx VCC, OCT, RTA, SWAP, MDP, or TonoPen were identified that met our inclusion criteria. 
Figure 2summarizes the results of the quality assessment for the 40 included studies. Study quality was variable; only eight studies 20 21 30 34 38 39 45 46 met the specified criteria for higher quality studies. 
Quantitative Data Synthesis
Individual Tests.
The sensitivity and specificity of the individual tests included in the HSROC meta-analysis models are shown in Figure 3and Appendix 1, which also includes DORs. 
DORs ranged from 10 for FDT C-20-5 to 181 for FDT C-20-1, with higher DORs indicating a better ability to differentiate between diseased and nondiseased. There was statistical heterogeneity (variability in outcome beyond what would be expected by chance) across studies for most tests. Ophthalmoscopy, retinal photography (optic disc photography and RNFL photography), GAT, SAP (threshold and suprathreshold), and FDT C-20-5 were all relatively poorly performing tests based on lower DORs (range, 10–30). 
Eight studies met the criteria for higher quality studies, including six population-based studies and two cohort studies, and test accuracy data are detailed in Table 2 . For both SAP threshold and FDT C-20-5, higher quality studies reported lower sensitivity and specificity when compared with all studies, whereas two FDT C-20-5 studies not meeting the criteria for higher quality reported very high sensitivity (98% and 100%, respectively). For optic disc photography, compared with all studies, the higher quality studies reported similar sensitivity (74% versus 73%) but lower specificity (82% versus 89%). For HRT II, compared with all studies, the higher quality studies reported higher sensitivity (93% versus 86%), but slightly lower specificity (85% versus 89%). 
Seven studies reported test accuracy in different stages of glaucoma. 24 51 52 54 55 60 61 Of those reporting the same tests for different stages of glaucoma, Ieong et al. 55 reported a sensitivity of 72% for SAP (suprathreshold) for early stage glaucoma, whereas Enger and Sommer 51 and Katz et al. 24 both reported a sensitivity of 97% for SAP (threshold) for early/moderate stage glaucoma. 
Studies Directly Comparing Tests.
Six studies directly compared two or more of the following tests for detection of OAG: optic disc photography, HRT II, SAP, FDT, and GAT. 23 30 34 36 46 55 Table 3shows the common cutoff selected, sensitivity, specificity, DORs, and relative DORs for these studies. In each study SAP (either suprathreshold or threshold) was included as a comparator. DORs for the tests ranged from 4 for SAP threshold 46 to 75 for HRT II 30 (Table 3) . In terms of relative DORs, compared with SAP, GAT performed better in one study 36 but worse in another 23 (statistically significant), HRT II performed better than SAP in one study 30 (statistically significant) but worse in another, 55 FDT C-20-5 30 and FDT C-20 matrix 46 performed better than SAP, whereas optic disc photography 34 showed a broadly similar performance. 
Indirect Comparisons in a Single HSROC Model.
The results of the indirect comparisons in a single HSROC model are shown in Table 4 . From the large number of comparisons undertaken, six showed a statistically significant difference between tests (four in terms of sensitivity and two in terms of specificity). There was evidence that, at the common cutoff, FDT C-20-1 was significantly more sensitive than both ophthalmoscopy (30, 95% CrI 0–62) and GAT (45, 95% CrI 17–68), and that both SAP threshold (41, 95% CrI 14–64) and HRT II (39, 95% CrI 3–64) were significantly more sensitive than GAT. There was also evidence that GAT was significantly more specific than both FDT C-20-5 (19, 95% CrI 0–53) and SAP threshold (14, 95% CrI 1–37). Other differences in accuracy between tests may well exist that could not be detected due to the high level of uncertainty. The wide credible intervals reflected the small number of studies reporting each test and the generally high level of heterogeneity. Because of the imprecision in the estimates, no test (or group of tests) was clearly more accurate, based on a 5% significance level. Further analysis, at 10% and 20% levels of significance, identified additional statistically significant comparisons (Table 4) . For example, in terms of sensitivity, at a 10% significance level, FDT C-20-1 was better than SAP suprathreshold and at a 20% level better than optic disc photography, RNFL photography, and FDT C-20-5. OKP was better than GAT at a 10% level and HRT II better than ophthalmoscopy at a 20% level. In terms of specificity, at a 20% level FDT C-20-1 was better than SAP threshold and FDT C-20-5. 
Discussion
To our knowledge, this is the first systematic review of screening and diagnostic tests in glaucoma and includes 40 studies enrolling over 48,000 people and reporting nine tests. Most tests were reported by only a few, mostly heterogeneous, studies. The included studies reported tests of structure (ophthalmoscopy; optic disc photography, RNFL photography, and HRT II), visual function (FDT, OKP, and SAP) and IOP (GAT and NCT). Other tests were considered, including those of structure (GDx VCC, OCT, and RTA), visual function (SWAP and MDP), or using TonoPen to measure IOP. However, no studies using these tests met our inclusion in reporting of test accuracy outcomes. 
A systematic review of test accuracy is unlikely to identify the best test but can identify more promising ones. It is difficult to rank tests on paired values of sensitivity and specificity, as a highly specific test may be associated with a low sensitivity and vice versa. The choice of test depends on the importance of the tradeoff between missed cases, and false positives. OAG affects an estimated 2% of the adult population. A test of low specificity would be likely to overburden a health service with people who do not have glaucoma and cause unnecessary anxiety for a many individuals, and equally a test of low sensitivity would miss treatable disease which might be unacceptable to society. The DOR, a single measure of test accuracy, is a useful measure for comparing accuracy of several tests in a meta-analysis. 15 Based on a DOR ≥50, FDT C-20-1, OKP (both tests of visual function), and HRT II (a test of glaucomatous optic neuropathy) merit further evaluation of their performance as screening tests for glaucoma. It should be noted that these findings are relevant to the common cutoff point selected for each test; selection was based on the most frequently reported cutoff and when several cut-offs were reported, the cutoff most likely to represent early glaucoma. Furthermore, these findings are based on heterogeneous data of limited quality and as such are associated with considerable uncertainty. 
Methods of meta-analysis of diagnostic accuracy which combine studies in which both sensitivity and specificity vary have been available since 1990 and are continuing to evolve. 13 62 63 64 65 66 These methods are based on the idea of a tradeoff relationship between sensitivity and specificity, as occurs when studies vary in threshold, and seek to estimate the shape and position of the underlying receiver operating curve. From the estimate of this curve, it is possible to identify operating points. The approach adopted in our review identifies the average operating point for each test and makes comparisons between them, based on those studies reporting each test that share a common cutoff point. The Cochrane Collaboration are commencing publication of systematic reviews of diagnostic test accuracy, and the analytical approach we have followed is the one that they are recommending. 67 Estimation of a summary point specific to a test being used at a common threshold obtains the best estimate of test accuracy in parameters that are clinically meaningful. The tradeoff between sensitivity and specificity is important in judging the performance of a test and is best depicted by an ROC curve across different cutoff points. However, the included studies did not usually provide information across the whole range of cutoff points to allow such analyses to be undertaken. 
We used a Bayesian Hierarchical SROC model, as standard methods for meta-analysis do not address the threshold effect and are therefore not appropriate. 68 Several different levels of analyses were undertaken, including an analysis in which all tests were modeled simultaneously by using this Bayesian approach. This method allowed indirect comparison of sensitivities and specificities to be made, in addition to allowing DORs to be calculated, which is one of the advantages of the Bayesian method adopted. To produce results that are comparable to those from standard methods of meta-analysis we did not use informative priors. 
In addition to providing sensitivity and specificity estimates we also reported the DOR results. Some meta-analysis models can only provide the DOR estimate, and therefore we included this measure for comparability. One strength of the DOR is that it is a mathematically robust measure (like the standard odds ratio) and represents diagnostic accuracy as a single value. However, a disadvantage is that different combinations of sensitivity and specificity can lead to the same DOR. 
To be included, studies had to meet specific inclusion criteria. The validity of indirect comparisons does depend on assumptions regarding the characteristics of the included studies; however, the indirect method is formally performing the comparison that users of the report are likely to make when assessing the pooled results for the individual tests. As such, this method of indirect comparisons serves an important purpose and reaffirms the lack of certainty about which test is indeed the best. 
There are many potential sources of bias in primary diagnostic accuracy studies. Despite the huge volume of literature, no good-quality studies were found that provided a positive response to all questions on the modified QUADAS checklist. Based on limited evidence, of the tests reported by higher quality studies, including the three tests that were considered to merit further evaluation, estimates of sensitivity and specificity varied according to study quality. 
There is no universally agreed on optimal reference standard for the diagnosis of OAG, although progressive structural optic neuropathy has been proposed as the best possible reference standard. 69 70 In this review either of two reference standards was considered. There was no obvious pattern in the sensitivity and specificity of the tests in the seven studies 18 34 40 43 44 48 58 in which the first and best reference standard of OAG was confirmed on longitudinal follow-up compared with the remainder in which ophthalmologist-diagnosed OAG was used. Although the latter is suboptimal compared with the former, it is the accepted reference standard in clinical practice. However, establishing a reference standard in glaucoma is problematic, as in some people, optic disc damage precedes visual field loss, whereas in others the reverse is the case. 
The accuracy of a test may vary according to the population in which it is performed. Samples with higher prevalence often arise through preferential inclusion of suspect cases, which shifts the disease severity to include more moderate and severe disease, and since it is easier to differentiate between severely diseased and nondiseased persons, a test would be expected to report improved (apparent) sensitivity and specificity. Therefore, studies with a significantly higher prevalence than expected in a screening population should be interpreted with this limitation in mind. 21 28 29 30 36 39 These studies, including two that met the criteria for higher quality studies, 30 39 tended to recruit their participants through media advertising rather than contacting individuals in a predefined population and can be considered to be more representative of screening in higher risk populations. 
Twenty of the 40 studies included were hospital based, which by nature, is an enriched population likely to include a disproportionate number of participants with high IOP and with previous experience taking tests, potentially leading to overly optimistic performance estimates. 71 72 73 74 Most of the case–control studies identified applied stringent criteria for inclusion, such as visual acuity of 6/9 or no other ocular disease and as such were highly prone to bias. 75 To minimize this spectrum bias, case–control studies (n = 57) in which the participants were considered unrepresentative of a case mix found in a general population where OAG screening would be performed were excluded from the review. 
In the meta-analysis models for the individual tests, statistical heterogeneity was evident across most studies. Empirically, there was no obvious single cause of the heterogeneity, but potential contributory factors include differences in populations, study design, setting, prevalence, and severity of glaucoma within studies. Other factors include differences in reference standard, and in tests included within the same category (e.g., different types of perimetry and ophthalmoscopy have a large number of variants, potentially leading to heterogeneity in discriminatory power across studies reporting those tests), and the extent to which studies were affected by other potential biases (e.g., partial and differential verification bias, incorporation bias, test, and diagnostic review bias). 
Limitations
Relatively few studies were identified for each test, and it was not possible to perform sensitivity analysis based on study design. The common cutoff chosen for each test was the one most frequently reported across the included studies for that test, although this may not be the most appropriate. Most of the studies were poorly reported, an issue that has been highlighted in recent literature. 76 77 78 79 Only 6 of the 40 studies directly compared two or more tests. It was not possible to provide summary results of studies that directly compared tests because of small numbers. Studies not providing sufficient information to allow the calculation of 2 × 2 tables were excluded, although they may have contributed information on sensitivity and specificity. 
Systematic reviews provide a robust and rigorous evaluation of the available evidence, but according to their nature, as new studies are published, the review requires updating. Since the completion of our meta-analysis, further studies have been published on the performance of the tests included in this review. These include population-based studies in the United States, United Kingdom, Hungary, Japan, and China. These studies provide additional information on the performance of FDT perimetry alone, 80 81 82 in combination with GDx VCC, 83 and combined with an IOP measurement 84 and data on the performance of HRT II in an elderly population in the United Kingdom 85 and in a community screening program in Japan, 86 comparing HRT II with nonmydriatic fundus photography. Although systematic reviews rapidly become out of date, which is a limitation, one strength of a systematic review is that the methods are transparent and reproducible such that the review can be updated as further data become available in the future. Priorities for future research and optimal study designs can also be identified. 
Implications for Practice and Recommendations for Research
Ideally, a screening test for OAG should be safe, easy to administer and interpret, portable, quick, and acceptable to the people who are to be tested and should be sufficiently valid to distinguish between those who do and do not have OAG. Many potential screening tests for glaucoma are available. Of the many candidate tests, no one test or group of tests was clearly more accurate. Based on limited data, relatively poorly performing tests, ophthalmoscopy, standard automated perimetry, retinal photography, and Goldmann applanation tonometry were identified. 
Frequency doubling technology, (C-20-1), HRT II, and OKP were identified as having better diagnostic performance than other candidate tests, although these findings were based on poor-quality evidence. Further investigations should evaluate the most promising tests in directly comparative studies in a relevant population. 
Appendix 1
 
Figure 1.
 
Flow of studies through the review process.
Figure 1.
 
Flow of studies through the review process.
Table 1.
 
Characteristics of the Included Studies
Table 1.
 
Characteristics of the Included Studies
Study Index Test(s) Test(s) Performers and Interpreters Reference Standard Enrolled (n) Analyzed (n) Mean Age y (range) Sex Country Period
Population-Based Studies (Cross-sectional)
Anton 2004 16 GAT Ophthalmologists Ophthalmic examination 569 510 (40–79) M: 232; F: 278 Spain (Segovia Study) N/S
Bengtsson 1980 17 GAT Ophthalmologists Ophthalmic examination 1938 1511 (55–69) N/S Sweden (Dalby Population Survey) 1977–1978
Bonomi 2001 18 GAT Ophthalmologists Follow-up confirmation 5816 4297 eyes of 4297 people (40–80+) M: 1882; F: 2415 Italy (Egna-Neumarkt Study) N/S
Detry-Morel 2004 20 FDT C-20-5 Residents in training, paramedical staff Ophthalmic examination 1802 3211 eyes of 1620 people 63 (22–97) M: 680; F: 940 Belgium Oct 1999
Harasymowycz 2005 21 HRT II Ophthalmic photographer Ophthalmic examination 303 264 right eyes, 265 left eyes of 271 people 62.2 (SD 11.6) M: 90; F: 179 Canada Aug 2003–Feb 2004
Hollows 1966 22 GAT Ophthalmologists Ophthalmic examination 4608 4231 55 (40–74) Approx: M: 3639; F: 592 UK (Rhondda Valley Study) Summer 1963
Ivers 2001 23 SAP suprathreshold; GAT N/S Ophthalmic examination 4433 3654 (both tests) (49–97) M: 1582; F: 2072 Australia (Blue Mountains Eye Study) 1992–1994
Katz 1991 24 SAP threshold N/S Ophthalmic examination 355 355 eyes of 355 people Cases: 61; controls: 53 N/S USA (Glaucoma Screening Study) 1981–1992
Katz 1993 25 SAP suprathreshold N/S Ophthalmic examination 5308 4733 (40–80+) M: 2109; F: 3199 USA (Baltimore Eye Survey) Jan 1985–Nov 1988
Kozobolis 2000 27 GAT Uncertain Ophthalmic examination 1300 1107 (40–80+) M: 463; F: 644 Greece (Crete, Greece Glaucoma Study) Feb 1993–June 1998
Mansberger 2005 28 FDT C-20-5 N/S Ophthalmic examination 296 251 eyes of 251 people 45 (30–65) M: 117; F: 174 India N/S
Mundorf 1989 29 SAP suprathreshold N/S Ophthalmic examination 145 145 71 M: 40; F: 105 USA N/S
Robin 2005 30 Ophthalmoscopy; HRT II; SAP threshold; FDT C-20-5 Appropriately trained staff Ophthalmic examination 704 261 eyes of 261 people (all tests) 65 M: 281; F: 378 Australia Nov 2001
Weih 2001 37 Ophthalmoscopy N/S Consensus by panel of ophthalmologists, based on results of ophthalmic examination 4744 4636 59 (SD 12) M: 2230; F: 2514 Australia (Visual Impairment Project) 1992–1996
Wolfs 1999 38 Optic disc photography Technicians Ophthalmic examination 6777 5143 eyes of 5143 people (55 and over) N/S Netherlands (Rotterdam Eye Study) N/S
Yamada 1999 39 OKP; FDT C-20-1 Technicians Decision of glaucoma specialists, based on ophthalmic records 259 175 eyes of 175 people (OKP); 240 eyes of 240 people (FDT) FDT: 59.6 (SD 14.7); OKP: 58.8 (SD 15.6) M: 108; F: 135 USA N/S
Population-Based Studies (Cohort)
Christoffersen 1995 19 (Patient source; general practice) OKP GPs, medical secretaries Ophthalmic examination 195 187 57 (40–84) M: 51; F: 136 Norway N/S
Vernon 1990 32 (Patient source: general practice) Ophthalmoscopy; SAP suprathreshold; NCT Ophthalmoscopy: experienced ophthalmol-ogists; NCT/SAP: non-ophthalmically trained staff Ophthalmic examination 988 854 (ophth); 855 (SAP); 874 (NCT) 65 M: 374; F: 500 UK N/S
Wang 1998 36 (Patient source: general practice) Ophthalmoscopy; SAP suprathreshold; GAT (RNFL photography) N/S Ophthalmic examination 530 from primary care clinic) 400 (ophth); 214 (SAP); 357 (GAT) [136 (RNFL photo)] (40–65+) M: 111; F: 294 USA Jul 1991–Feb 1992
Population-Based Studies (Case–Control)
Vitale 2000 34 (Patient source: Cases and controls: sample of patients with and without glaucoma from the Baltimore Eye Study Follow-up Study) Optic disc photography; SAP suprathreshold Experienced technicians Follow-up confirmation 249 182 (disc photo); 228 (SAP); 68 M: 100; F: 149 USA (Baltimore Eye Study Follow-up) 1994
Already-Suspect Population (Cohort Studies)
Ekstrom 1993 40 (Patient source: people previously examined in a population-based glaucoma survey) GAT N/S Follow-up confirmation 760 413 (65–74) M: 364; F: 396 Sweden (Tierp Glaucoma Survey) Mar 1984–Mar 1986
Hammond 1979 41 (Patient source: eye clinic) Ophthalmoscopy Nurses skilled in use of the ophthalmoscope Ophthalmic examination 219 188 (21 and over) N/S USA N/S
Khong 2001 42 (Patient source: eye clinics) FDT C-20-5 N/S Ophthalmic examination 228 113 68.5 (23–91) M: 104; F: 119 Australia Dec 1999–Jan 2000
Leibowitz 1980 43 (Patient source: Framingham Eye Study) GAT Generally performed by 2nd or 3rd year residents in ophthalmology Follow-up confirmation 2631 574 (<65–75+) M: 272; F: 302 USA (Framingham Eye Study) Feb 1973–Feb 1975
Marraffa 1989 44 (Patient source: eye clinic) SAP suprathreshold Ophthalmologists Follow-up confirmation 104 182 eyes of 104 people 54.3 (18–76) M: 45; F: 59 Italy N/S
Schultz 1995 45 (Patient source: clinical practices of glaucoma specialist, cataract surgeon, and general ophthalmol-ogist) Optic disc photography Carried out: N/S Interpreted: 3rd-year ophthalmology residents Ophthalmic examination 258 365 eyes of? people (<40–>70) M: 112; F: 144; Unknown: 2 USA N/S
Spry 2005 46 (Patient source: hospital eye service) SAP threshold; FDT C-20 matrix SAP: clinic staff trained in visual field testing; FDT: N/S Ophthalmic examination 48 48 (both tests) 67.3 (SD 13.5) M: 24; F: 24 UK Oct 2003–Jan 2004
Theodossiades 2001 47 (Patient source: glaucoma clinics) Ophthalmoscopy Optometrists Ophthalmic examination 50 50 eyes of 50 people N/S N/S UK N/S
Already-Suspect Population (Case–Control Studies)
Airaksinen 1984 48 (Patient source: not stated) RNFL photography N/S Follow-up confirmation 142 132 eyes of 132 people Glaucoma: 62 (SD 20.5) Normal: 54 (SD 16.9); OHT: 57 (SD 12.7) N/S Canada + Finland N/S
Anton 1997 49 (Patient source: cases and controls: glaucoma unit) SAP threshold Uncertain Ophthalmic examination 180 180 eyes of 180 people Glaucoma: 61 (SD 8); Normal: 59 (SD 9) N/S Spain N/S
Damato 1989 50 (Patient source: Cases: not stated Controls: dermatology ward, hospital staff, relatives/friends of patients, patients with unilateral non-glaucomatous disease affecting the fellow eye) OKP Staff experienced in perimetry Ophthalmic examination 102 102 eyes of 102 people Glaucoma: 57.3; Normal: 54.4 N/S UK N/S
Enger 1987 51 (Patient source: Cases and controls: nerve fiber layer study) SAP threshold N/S Ophthalmic examination 112 170 eyes of 112 people Glaucoma: 61 (28–80); Normal: 51 (26–75) N/S USA N/S
Harper 1994 52 (Patient source: not stated) OKP; SAP suprathreshold Uncertain Ophthalmic examination 212 193 (OKP); 212 (SAP) Glaucoma: 67.8 (43–85); Normal: 61.5 (41–85) N/S UK N/S
Heeg 2005 53 (Patient source: Cases: glaucoma outpatient department Controls: old people’s homes, blood bank, other public places) FDT C-20-1; FDT C-20 full threshold N/S Ophthalmic examination 1112 208 (FDT C-20-1); 1112 (FDT C-20 full threshold) Glaucoma: 65 (13–91); Normal: 63 (33–94) Eligible: Glaucoma: M: 509; F: 542 Normal: M: 118; F: 119 Netherlands (Groningen Longitudinal Glaucoma Study) Jul 2000–Jun 2001
Ieong 2003 55 (Patient source: Cases: glaucoma subjects Controls: partners of cases, optometrist practice) HRT II; SAP suprathreshold Optometrists Ophthalmic examination 66 66 eyes of 66 people (both tests) Glaucoma: 69; Normal: 60 Glaucoma: M: 16; F: 13 Normal: M: 16; F: 21 UK N/S
Johnson 1999 56 (Patient source: not stated) FDT C-20-1 N/S Ophthalmic examination 108 160 eyes of 108 people Glaucoma: 64 (35–85); Normal: 46 (18–81) USA N/S
Quigley 1980 57 (Patient source: Cases and controls: Ophthalmol-ogical institute) Optic disc photography; RNFL photography Ophthalmologists Ophthalmic examination 175 294 eyes of? people (both tests) Readable photos: Glaucoma: 52.7 (SD 2.78); Glaucoma suspect: 45.2 (SD 1.56); Normal: 37.9 (SD 2.8) Unreadable photos: Glaucoma: 62.5 (SD 4.0); Glaucoma suspect: 59.6 (SD 6.3); Normal: 50 (SD 12.1) M: 86; F: 89 USA Jan 1978–Apr 1979
Sommer 1979 58 (Patient source: Cases and controls: glaucoma clinic) Optic disc photography; RNFL photography N/S Follow-up confirmation Unclear 223 eyes of ? people (both tests) N/S N/S USA N/S
Wollstein 2000 60 (Patient source: Cases: glaucoma clinic and ocular hypertension clinic Controls: spouses or friends of patients, responders-an advertisement) Optic disc photography Photos taken by trained technicians; assessed by glaucoma consultants, glaucoma fellow, clinical glaucoma technician Ophthalmic examination 123 123 eyes of 123 people Glaucoma: 65.1 (SD 10.06); Normal: 57.1 (SD 12.52) N/S UK N/S
Wood 1987 61 (Patient source: not stated) Ophthalmoscopy Ophthalmologists; junior doctors Ophthalmic examination 22 43 eyes of 22 people (32–75) N/S UK N/S
Figure 2.
 
Results of the quality assessment of the 40 included studies.
Figure 2.
 
Results of the quality assessment of the 40 included studies.
Figure 3.
 
Summary of sensitivity and specificity of tests included in the HSROC meta-analysis models.
Figure 3.
 
Summary of sensitivity and specificity of tests included in the HSROC meta-analysis models.
Table 2.
 
HSROC Analysis: All Studies Compared with Higher Quality Studies
Table 2.
 
HSROC Analysis: All Studies Compared with Higher Quality Studies
Optic Disc Photography HRT II FDT C-20-5 SAP Threshold
Sensitivity % (95% CrI) Specificity % (95% CrI) Sensitivity % (95% CrI) Specificity % (95% CrI) Sensitivity % (95% CrI) Specificity % (95% CrI) Sensitivity % (95% CrI) Specificity % (95% CrI)
All studies 73 (61–83) 89 (50–99) 86 (55–97) 89 (66–98) 78 (19–99) 75 (57–87) 88 (65–97) 80 (55–93)
Higher quality 74 (30–95) 82 (45–97) 93 (58–99) 85 (47–97) 72 (26–96) 60 (17–92) 73 (28–95) 64 (22–92)
Table 3.
 
Sensitivity, Specificity, DOR, and Relative DOR at the Common Cutoff for Studies Directly Comparing Tests
Table 3.
 
Sensitivity, Specificity, DOR, and Relative DOR at the Common Cutoff for Studies Directly Comparing Tests
Study Test Common Cutoff Sensitivity % (95% CrI) Specificity % (95% CrI) DOR (95% CrI) RDOR (95% CrI)
Vitale 2000 34 SAP supra Three adjacent points missed 50 (37–63) 83 (76–88) 5 (3–9) 1
Optic disc photo VCDR >0.6 77 (62–89) 59 (50–67) 5 (2–11) 0.99 (0.36–2.75)
Ieong 2003 55 SAP supra Optometrist’s judgment 72 (53–87) 95 (82–99) 46 (9–237) 1
HRT II Global/one of six segments abnormal 69 (49–85) 95 (82–99) 39 (8–198) 0.85 (0.08–8.54)
Robin 2005 30 SAP threshold AGIS score ≥3 (common cutoff) 63 (38–84) 74 (68–80) 5 (2–13) 1
HRT II ≥1 Borderline or 1 severe abnormality 95 (74–100) 81 (75–85) 75 (10–574) 15.01 (1.57–143.82)
FDT C-20-5 One abnormal point 84 (60–97) 55 (49–61) 7 (2–23) 1.31 (0.27–6.43)
Spry 2005 46 SAP threshold GHT outside normal limit and/or P < 0.05 with the PSD global index in one/both eyes 80 (52–96) 52 (34–69) 4 (1–18) 1
FDT C-20 matrix 100 (78–100) 27 (13–46) 12 (1–222) 2.83 (0.11–72.91)
Ivers 2001 23 SAP supra Three or more points missing 89 (80–94) 73 (71–74) 20 (10–39) 1
GAT IOP >22 mm Hg 14 (7–23) 98 (97–98) 6 (3–12) 0.31 (0.12–0.78)
Wang 1998 36 SAP supra Absolute or relative defects ≥ 17 70 (57–80) 67 (59–74) 5 (2–9) 1
GAT IOP > 21 mm Hg 28 (17–40) 96 (93–98) 9 (4–19) 1.89 (0.70–5.13)
Table 4.
 
Summary of Sensitivity, Specificity and DOR for Tests Included in the HSROC Meta-analysis Models
Table 4.
 
Summary of Sensitivity, Specificity and DOR for Tests Included in the HSROC Meta-analysis Models
Test Number of Studies Common Cutoff Sensitivity % (95% CrI) Specificity % (95% CrI) DOR (95% CrI)
Ophthalmoscopy 5 VCDR ≥0.7 60 (34–82) 94 (76–99) 26 (6–110)
Optic disc photography 6 VCDR ≥0.6 73 (61–83) 89 (50–99) 22 (3–148)
RNFL photography 4 Diffuse and/or localized defect 75 (46–92) 88 (53–98) 23 (4–124)
HRT II 3 ≥1 Borderline or outside normal limits 86 (55–97) 89 (66–98) 51 (11–246)
FDT
 C-20-1 3 1 Abnormal point 92 (65–99) 94 (73–99) 181 (25–2139)
 C-20-5 5 1 Abnormal point 78 (19–99) 75 (57–87) 10 (0.7–249)
OKP 4 1 Abnormal point 86 (29–100) 90 (79–96) 58 (4–1585)
SAP suprathreshold 9 ≥3 Points missing 71 (51–86) 85 (73–93) 14 (6–34)
SAP threshold 5 AGIS score ≥3 88 (65–97) 80 (55–93) 30 (6–159)
GAT 9 IOP >21 mm Hg 46 (22–71) 95 (89–97) 15 (4–49)
The authors thank the members of the OAG project group for assistance and comments at all stages of the review. Other members of the OAG Screening Project Group were: Rodolfo Hernandez, Luke Vale, Tania Lourenco, Cynthia Fraser, John Cairns, Richard Wormald, Stephen McPherson, David Wright, Kannaiyan Rabindranath, Rod Taylor, Norman Waugh, and Adrian Grant. 
QuigleyHA, BromanAT. The number of people with glaucoma worldwide in 2010 and 2020. Br J Ophthalmol. 2006;90:262–267. [CrossRef] [PubMed]
ResnikoffS, PascoliniD, Etya'aleD, et al. Global data on visual impairment in the year 2002. Bull World Health Organ. 2004;82:844–851. [PubMed]
GrantWM, BurkeJF, Jr. Why do some people go blind from glaucoma?. Ophthalmology. 1982;89:991–998. [CrossRef] [PubMed]
LeskeMC, HeijlA, HusseinM, BengtssonB, HymanL, KomaroffE. Factors for glaucoma progression and the effect of treatment: The Early Manifest Glaucoma Trial. Arch Ophthalmol. 2003;121:48–56. [CrossRef] [PubMed]
WilsonR, WalkerAM, DuekerDK, CrickRP. Risk factors for rate of progression of glaucomatous visual field loss: a computer-based analysis. Arch Ophthalmol. 1982;100:737–741. [CrossRef] [PubMed]
BurrJM, MowattG, HernandezR, et al. The clinical effectiveness and cost-effectiveness of screening for open angle glaucoma: a systematic review and economic evaluation. Health Technol Assess. 2007;11:1–190.
FlemingC, WhitlockE, BielT, SmitB. Primary care screening for ocular hypertension and primary open-angle glaucoma: Evidence Synthesis No 34 [document on the Internet]. 2005;Agency for Healthcare Research & Quality, US Preventive Services Task Force Available at: http://www.ahrq.gov/clinic/uspstf05/glaucoma/glaucsyn.pdf. Accessed April 2008.
MaierPC, FunkJ, SchwarzerG, AntesG, Falck-YtterYT. Treatment of ocular hypertension and open angle glaucoma: meta-analysis of randomised controlled trials. BMJ. 2005;331:134–136. [CrossRef] [PubMed]
SpryPG, SparrowJM. An evaluation of open-angle glaucoma against the NSC criteria for screening viability, effectiveness and appropriateness [document on the Internet]. 2005;UK National Screening Committee Available at: http://www.library.nhs.uk/screening/ViewResource.aspx?resID=61002 & tabID=288. Accessed October 2008.
FinOHTA. Cost-effectiveness of glaucoma screening (project). 2005;Finish Office for Health Technology Assessment Available at: http://finohta.stakes.fi/EN/research/index.htm. Accessed April 2008
Screening for glaucoma: an update-non-systematic review (project). 2007;Agence d’Evaluation des Technologies et des Modes d’Intervention en Sante (AETMIS) Avalable at: http://www.aetmis.gouv.qc.ca. Accessed April 2008.
WhitingP, RutjesAW, ReitsmaJB, BossuytPM, KleijnenJ. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol. 2003;3:25. [CrossRef] [PubMed]
RutterCM, GatsonisCA. A hierarchical regression approach to meta-analysis of diagnostic test accuracy evaluations. Stat Med. 2001;20:2865–2884. [CrossRef] [PubMed]
SpiegelhalterD, ThomasA, BestN. WinBUGS: Bayesian inference using Gibbs sampling. User manual, version 1.4. 2003;MRC Biostatistics Unit Cambridge, UK.
GlasAS, LijmerJG, PrinsMH, BonselGJ, BossuytPM. The diagnostic odds ratio: a single indicator of test performance. J Clin Epidemiol. 2003;56:1129–1135. [CrossRef] [PubMed]
AntonA, AndradaMT, MujicaV, CalleMA, PortelaJ, MayoA. Prevalence of primary open-angle glaucoma in a Spanish population: the Segovia study. J Glaucoma. 2004;13:371–376. [CrossRef] [PubMed]
BengtssonB. Findings associated with glaucomatous visual field defects. Acta Ophthalmol (Copenh). 1980;58:20–32. [PubMed]
BonomiL, MarchiniG, MarraffaM, MorbioR. The relationship between intraocular pressure and glaucoma in a defined population: data from the Egna-Neumarkt Glaucoma Study. Ophthalmologica. 2001;215:34–38. [CrossRef] [PubMed]
ChristoffersenT, ForsT, WaageS, HoltedahlK. Glaucoma screening with oculokinetic perimetry in general practice: is its specificity acceptable?. Eye. 1995;9:36–39. [PubMed]
Detry-MorelM, ZeyenT, KestelynP, CollignonJ, GoethalsM, Belgian Glaucoma Society. Screening for glaucoma in a general population with the non-mydriatic fundus camera and the frequency doubling perimeter. Eur J Ophthalmol. 2004;14:387–393. [PubMed]
HarasymowyczP, PapamatheakisD, FansiAK, GressetJ, LeskMR. Validity of screening for glaucomatous optic nerve damage using confocal scanning laser ophthalmoscopy (Heidelberg Retina Tomograph II) in high-risk populations: a pilot study. Ophthalmology. 2005;112:2164–2171. [CrossRef] [PubMed]
HollowsFC, GrahamPA. Intra-ocular pressure glaucoma and glaucoma suspects in a defined population. Br J Ophthalmol. 1966;50:570–586. [CrossRef] [PubMed]
IversRQ, OptomB, MacaskillP, CummingRG, MitchellP. Sensitivity and specificity of tests to detect eye disease in an older population. Ophthalmology. 2001;108:968–975. [CrossRef] [PubMed]
KatzJ, SommerA, GaasterlandDE, AndersonDR. Comparison of analytic algorithms for detecting glaucomatous visual field loss. Arch Ophthalmol. 1991;109:1684–1689. [CrossRef] [PubMed]
KatzJ, TielschJM, QuigleyHA, JavittJ, WittK, SommerA. Automated suprathreshold screening for glaucoma: the Baltimore Eye Survey. Invest Ophthalmol Vis Sci. 1993;34:3271–3277. [PubMed]
KatzJ, QuigleyHA, SommerA. Repeatability of the Glaucoma Hemifield Test in automated perimetry. Invest Ophthalmol Vis Sci. 1995;36:1658–1664. [PubMed]
KozobolisVP, DetorakisET, TsilimbarisM, et al. Crete, Greece glaucoma study. J Glaucoma. 2000;9:143–149. [CrossRef] [PubMed]
MansbergerSL, JohnsonCA, CioffiGA, et al. Predictive value of frequency doubling technology perimetry for detecting glaucoma in a developing country. J Glaucoma. 2005;14:128–134. [CrossRef] [PubMed]
MundorfTK, ZimmermanTJ, NardinGF, KendallKS. Automated perimetry, tonometry, and questionnaire in glaucoma screening. Am J Ophthalmol. 1989;108:505–508. [CrossRef] [PubMed]
RobinTA, MullerA, RaitJ, KeeffeJE, TaylorHR. Performance of community-based glaucoma screening using frequency doubling technology and Heidelberg retinal tomography. Ophthalmic Epidemiol. 2005;12:167–178. [CrossRef] [PubMed]
TielschJM, KatzJ, SinghK, et al. A population-based evaluation of glaucoma screening: the Baltimore Eye Survey. Am J Epidemiol. 1991;134:1102–1110. [PubMed]
VernonSA, HenryDJ, CaterL, JonesSJ. Screening for glaucoma in the community by non-ophthalmologically trained staff using semi automated equipment. Eye. 1990;4:89–97. [CrossRef] [PubMed]
VernonSA, JonesSJ, HenryDJ. Maximising the sensitivity and specificity of non-contact tonometry in glaucoma screening. Eye. 1991;5:491–493. [CrossRef] [PubMed]
VitaleS, SmithTD, QuigleyT, et al. Screening performance of functional and structural measurements of neural damage in open-angle glaucoma: a case-control study from the Baltimore Eye Survey. J Glaucoma. 2000;9:346–356. [CrossRef] [PubMed]
WangF, QuigleyHA, TielschJM. Screening for glaucoma in a medical clinic with photographs of the nerve fiber layer. Arch Ophthalmol. 1994;112:796–800. [CrossRef] [PubMed]
WangF, TielschJM, FordDE, QuigleyHA, WheltonPK. Evaluation of screening schemes for eye disease in a primary care setting. Ophthalmic Epidemiol. 1998;5:69–82. [CrossRef] [PubMed]
WeihLM, NanjanM, McCartyCA, TaylorHR. Prevalence and predictors of open-angle glaucoma: results from the visual impairment project. Ophthalmology. 2001;108:1966–1972. [CrossRef] [PubMed]
WolfsRC, RamrattanRS, HofmanA, de JongPT. Cup-to-disc ratio: ophthalmoscopy versus automated measurement in a general population: The Rotterdam Study. Ophthalmology. 1999;106:1597–1601. [CrossRef] [PubMed]
YamadaN, ChenPP, MillsRP, et al. Screening for glaucoma with frequency-doubling technology and Damato campimetry. Arch Ophthalmol. 1999;117:1479–1484. [CrossRef] [PubMed]
EkstromC. Elevated intraocular pressure and pseudoexfoliation of the lens capsule as risk factors for chronic open angle glaucoma-a population based 5 year follow up study. Acta Ophthalmol (Copenh). 1993;71:189–195. [PubMed]
HammondEA, BegleyPK. Screening for glaucoma: a comparison of ophthalmoscopy and tonometry. Nurs Res. 1979;28:371–372. [PubMed]
KhongJJ, DimitrovPN, RaitJ, McCartyCA. Can the specificity of the FDT for glaucoma be improved by confirming abnormal results?. J Glaucoma. 2001;10:199–202. [CrossRef] [PubMed]
LeibowitzHM, KruegerDE, MaunderLR, et al. The Framingham Eye Study monograph: an ophthalmological and epidemiological study of cataract, glaucoma, diabetic retinopathy, macular degeneration, and visual acuity in a general population of 2631 adults 1973–1975. Surv Ophthalmol. 1980;24:335–610. [CrossRef] [PubMed]
MarraffaM, MarchiniG, AlbertiniR, BonomiL. Comparison of different screening methods for the detection of visual field defects in early glaucoma. Int Ophthalmol. 1989;13:43–45. [CrossRef] [PubMed]
SchultzRO, RadiusRL, HartzAJ, et al. Screening for glaucoma with stereo disc photography. J Glaucoma. 1995;4:177–182. [PubMed]
SpryPGD, HussinHM, SparrowJM. Clinical evaluation of frequency doubling technology perimetry using the Humphrey Matrix 24–2 threshold strategy. Br J Ophthalmol. 2005;89:1031–1935. [CrossRef] [PubMed]
TheodossiadesJ, MurdochI. What optic disc parameters are most accurately assessed using the direct ophthalmoscope?. Eye. 2001;15:283–287. [CrossRef] [PubMed]
AiraksinenPJ, DranceSM, DouglasGR, MawsonDK, NieminenH. Diffuse and localized nerve fiber loss in glaucoma. Am J Ophthalmol. 1984;98:566–571. [CrossRef] [PubMed]
AntonA, MaquetJA, MayoA, TapiaJ, PastorJC. Value of logistic discriminant analysis for interpreting initial visual field defects. Ophthalmology. 1997;104:525–531. [CrossRef] [PubMed]
DamatoBE, AhmedJ, AllanD, McClureE, JayJL. The detection of glaucomatous visual field defects by oculo-kinetic perimetry: which points are best for screening?. Eye. 1989;3:727–731. [CrossRef] [PubMed]
EngerC, SommerA. Recognizing glaucomatous field loss with the Humphrey STATPAC. Arch Ophthalmol. 1987;105:1355–1357. [CrossRef] [PubMed]
HarperRA, HillAR, ReevesBC. Effectiveness of unsupervised oculokinetic perimetry for detecting glaucomatous visual field defects. Ophthalmic Physiol Opt. 1994;14:199–202. [CrossRef] [PubMed]
HeegGP, BlanksmaLJ, HardusPL, JansoniusNM. The Groningen Longitudinal Glaucoma Study. I. Baseline sensitivity and specificity of the frequency doubling perimeter and the GDx nerve fibre analyser. Acta Ophthalmol Scand. 2005;83:46–52. [CrossRef] [PubMed]
HeegGP, StoutenbeekR, JansoniusNM. Strategies for improving the diagnostic specificity of the frequency doubling perimeter. Acta Ophthalmol Scand. 2005;83:53–56. [CrossRef] [PubMed]
IeongA, MurdochI, CousensS, HealeyP, TheodossiadesJ. Sensitivity and specificity of two glaucoma case-finding strategies for optometrists. Ophthalmic Physiol Opt. 2003;23:341–346. [CrossRef] [PubMed]
JohnsonCA, CioffiGA, Van BuskirkEM. Evaluation of Two Screening Tests for Frequency Doubling Technology Perimetry. September 1999;103.13th International Perimetric Society Meeting Garda, Italy.
QuigleyHA, MillerNR, GeorgeT. Clinical evaluation of nerve fiber layer atrophy as an indicator of glaucomatous optic nerve damage. Arch Ophthalmol. 1980;98:1564–1571. [CrossRef] [PubMed]
SommerA, PollackI, MaumeneeAE. Optic disc parameters and onset of glaucomatous field loss. II. Static screening criteria. Arch Ophthalmol. 1979;97:1449–1454. [CrossRef] [PubMed]
StoutenbeekR, HeegGP, JansoniusNM. Frequency doubling perimetry screening mode compared to the full-threshold mode. Ophthalmic Physiol Opt. 2004;24:493–497. [CrossRef] [PubMed]
WollsteinG, Garway-HeathDF, FontanaL, HitchingsRA. Identifying early glaucomatous changes: comparison between expert clinical assessment of optic disc photographs and confocal scanning ophthalmoscopy. Ophthalmology. 2000;107:2272–2277. [CrossRef] [PubMed]
WoodCM. Limitations of direct ophthalmoscopy in screening for glaucoma. BMJ. 1987;294:1587–1588.
GatsonisC, PaliwalP. Meta-analysis of diagnostic and screening test accuracy evaluations: methodologic primer. Am J Roentgenol. 2006;187:271–281. [CrossRef]
KardaunJW, KardaunOJ. Comparative diagnostic performance of three radiological procedures for the detection of lumbar disk herniation. Methods Inf Med. 1990;29:12–22. [PubMed]
MacaskillP. Empirical Bayes estimates generated in a hierarchical summary ROC analysis agreed closely with those of a full Bayesian analysis. J Clin Epidemiol. 2004;57:925–932. [CrossRef] [PubMed]
MosesLE, ShapiroD, LittenbergB. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat Med. 1993;12:1293–1316. [CrossRef] [PubMed]
ReitsmaJB, GlasAS, RutjesAW, ScholtenRJ, BossuytPM, ZwindermanAH. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol. 2005;58:982–990. [CrossRef] [PubMed]
Cochrane Diagnostic Test Accuracy Working Group [website on the Internet].The Cochrane Collaboration Available at: http://srdta.cochrane.org/en/index.html. Accessed April 2008.
Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0.0 (updated February 2008) [document on the Internet]. The Cochrane Library. Issue 2, 2008. ;John Wiley & Sons Ltd Chichester, UK.Available at: http://www.cochrane-handbook.org/. Accessed April 2008.
MedeirosFA, ZangwillLM, BowdC, SamplePA, WeinrebRN. Use of progressive glaucomatous optic disk change as the reference standard for evaluation of diagnostic tests in glaucoma. Am J Ophthalmol. 2005;139:1010–1018. [CrossRef] [PubMed]
WeinrebRN, KhawKT. Primary open-angle glaucoma. Lancet. 2004;363:1711–1720. [CrossRef] [PubMed]
WildJM, KimLS, PaceyIE, CunliffeIA. Evidence for a learning effect in short-wavelength automated perimetry. Ophthalmology. 2006;113:206–215. [CrossRef] [PubMed]
YeniceO, TemelA. Evaluation of two Humphrey perimetry programs: full threshold and SITA standard testing strategy for learning effect. Eur J Ophthalmol. 2005;15:209–212. [PubMed]
MatsuoH, TomitaG, SuzukiY, AraieM. Learning effect and measurement variability in frequency-doubling technology perimetry in chronic open-angle glaucoma. J Glaucoma. 2002;11:467–473. [CrossRef] [PubMed]
JosonPJ, KamantigueME, ChenPP. Learning effects among perimetric novices in frequency doubling technology perimetry. Ophthalmology. 2002;109:757–760. [CrossRef] [PubMed]
RutjesAW, ReitsmaJB, Di NisioM, SmidtN, van RijnJC, BossuytPM. Evidence of bias and variation in diagnostic accuracy studies. Can Med Assoc J. 2006;174:469–476. [CrossRef]
BochmannF, JohnsonZ, Azuara-BlancoA. Sample size of studies of diagnostic accuracy in ophthalmology: literature survey. Br J Ophthalmol. 2007;91:898–900. [CrossRef] [PubMed]
JohnsonZ, SiddiquiMAR, Azuara-BlancoA. The quality of reporting of diagnostic accuracy studies of optical coherence tomography in glaucoma. Ophthalmology. 2007;114:1607–1612. [CrossRef] [PubMed]
ParanjothyB, ShunmugamM, Azuara-BlancoA. The quality of reporting of diagnostic accuracy studies in glaucoma using scanning laser polarimetry. J Glaucoma. 2007;16:670–675. [CrossRef] [PubMed]
ShunmugamM, Azuara-BlancoA. The quality of reporting of diagnostic accuracy studies in glaucoma using the Heidelberg retina tomograph. Invest Ophthalmol Vis Sci. 2006;47:2317–2323. [CrossRef] [PubMed]
WangYX, XuL, ZhangRX, JonasJB. Frequency-doubling threshold perimetry in predicting glaucoma in a population-based study: The Beijing Eye Study. Arch Ophthalmol. 2007;125:1402–1406. [CrossRef] [PubMed]
MansbergerSL, JohnsonCA, CioffiGA. The results of screening frequency doubling technology perimetry in different locations of the community. J Glaucoma. 2007;16:73–80. [CrossRef] [PubMed]
IwaseA, TomidokoroA, AraieM, et al. Performance of frequency-doubling technology perimetry in a population-based prevalence survey of glaucoma: the Tajimi study. Ophthalmology. 2007;114:27–32. [CrossRef] [PubMed]
TothM, KothyP, VarghaP, HolloG. Accuracy of combined GDx-VCC and matrix FDT in a glaucoma screening trial. J Glaucoma. 2007;16:462–470. [CrossRef] [PubMed]
VistamehrS, ShelstaHN, PalmisanoPC, et al. Glaucoma screening in a high-risk population. J Glaucoma. 2006;15:534–540. [CrossRef] [PubMed]
HawkerMJ, VernonSA, TattersallCL, DuaHS. Linear regression modeling of rim area to discriminate between normal and glaucomatous optic nerve heads: the Bridlington Eye Assessment Project. J Glaucoma. 2007;16:345–351. [CrossRef] [PubMed]
OhkuboS, TakedaH, HigashideT, SasakiT, SugiyamaK. A pilot study to detect glaucoma with confocal scanning laser ophthalmoscopy compared with nonmydriatic stereoscopic photography in a community health screening. J Glaucoma. 2007;16:531–538. [CrossRef] [PubMed]
Figure 1.
 
Flow of studies through the review process.
Figure 1.
 
Flow of studies through the review process.
Figure 2.
 
Results of the quality assessment of the 40 included studies.
Figure 2.
 
Results of the quality assessment of the 40 included studies.
Figure 3.
 
Summary of sensitivity and specificity of tests included in the HSROC meta-analysis models.
Figure 3.
 
Summary of sensitivity and specificity of tests included in the HSROC meta-analysis models.
Table 1.
 
Characteristics of the Included Studies
Table 1.
 
Characteristics of the Included Studies
Study Index Test(s) Test(s) Performers and Interpreters Reference Standard Enrolled (n) Analyzed (n) Mean Age y (range) Sex Country Period
Population-Based Studies (Cross-sectional)
Anton 2004 16 GAT Ophthalmologists Ophthalmic examination 569 510 (40–79) M: 232; F: 278 Spain (Segovia Study) N/S
Bengtsson 1980 17 GAT Ophthalmologists Ophthalmic examination 1938 1511 (55–69) N/S Sweden (Dalby Population Survey) 1977–1978
Bonomi 2001 18 GAT Ophthalmologists Follow-up confirmation 5816 4297 eyes of 4297 people (40–80+) M: 1882; F: 2415 Italy (Egna-Neumarkt Study) N/S
Detry-Morel 2004 20 FDT C-20-5 Residents in training, paramedical staff Ophthalmic examination 1802 3211 eyes of 1620 people 63 (22–97) M: 680; F: 940 Belgium Oct 1999
Harasymowycz 2005 21 HRT II Ophthalmic photographer Ophthalmic examination 303 264 right eyes, 265 left eyes of 271 people 62.2 (SD 11.6) M: 90; F: 179 Canada Aug 2003–Feb 2004
Hollows 1966 22 GAT Ophthalmologists Ophthalmic examination 4608 4231 55 (40–74) Approx: M: 3639; F: 592 UK (Rhondda Valley Study) Summer 1963
Ivers 2001 23 SAP suprathreshold; GAT N/S Ophthalmic examination 4433 3654 (both tests) (49–97) M: 1582; F: 2072 Australia (Blue Mountains Eye Study) 1992–1994
Katz 1991 24 SAP threshold N/S Ophthalmic examination 355 355 eyes of 355 people Cases: 61; controls: 53 N/S USA (Glaucoma Screening Study) 1981–1992
Katz 1993 25 SAP suprathreshold N/S Ophthalmic examination 5308 4733 (40–80+) M: 2109; F: 3199 USA (Baltimore Eye Survey) Jan 1985–Nov 1988
Kozobolis 2000 27 GAT Uncertain Ophthalmic examination 1300 1107 (40–80+) M: 463; F: 644 Greece (Crete, Greece Glaucoma Study) Feb 1993–June 1998
Mansberger 2005 28 FDT C-20-5 N/S Ophthalmic examination 296 251 eyes of 251 people 45 (30–65) M: 117; F: 174 India N/S
Mundorf 1989 29 SAP suprathreshold N/S Ophthalmic examination 145 145 71 M: 40; F: 105 USA N/S
Robin 2005 30 Ophthalmoscopy; HRT II; SAP threshold; FDT C-20-5 Appropriately trained staff Ophthalmic examination 704 261 eyes of 261 people (all tests) 65 M: 281; F: 378 Australia Nov 2001
Weih 2001 37 Ophthalmoscopy N/S Consensus by panel of ophthalmologists, based on results of ophthalmic examination 4744 4636 59 (SD 12) M: 2230; F: 2514 Australia (Visual Impairment Project) 1992–1996
Wolfs 1999 38 Optic disc photography Technicians Ophthalmic examination 6777 5143 eyes of 5143 people (55 and over) N/S Netherlands (Rotterdam Eye Study) N/S
Yamada 1999 39 OKP; FDT C-20-1 Technicians Decision of glaucoma specialists, based on ophthalmic records 259 175 eyes of 175 people (OKP); 240 eyes of 240 people (FDT) FDT: 59.6 (SD 14.7); OKP: 58.8 (SD 15.6) M: 108; F: 135 USA N/S
Population-Based Studies (Cohort)
Christoffersen 1995 19 (Patient source; general practice) OKP GPs, medical secretaries Ophthalmic examination 195 187 57 (40–84) M: 51; F: 136 Norway N/S
Vernon 1990 32 (Patient source: general practice) Ophthalmoscopy; SAP suprathreshold; NCT Ophthalmoscopy: experienced ophthalmol-ogists; NCT/SAP: non-ophthalmically trained staff Ophthalmic examination 988 854 (ophth); 855 (SAP); 874 (NCT) 65 M: 374; F: 500 UK N/S
Wang 1998 36 (Patient source: general practice) Ophthalmoscopy; SAP suprathreshold; GAT (RNFL photography) N/S Ophthalmic examination 530 from primary care clinic) 400 (ophth); 214 (SAP); 357 (GAT) [136 (RNFL photo)] (40–65+) M: 111; F: 294 USA Jul 1991–Feb 1992
Population-Based Studies (Case–Control)
Vitale 2000 34 (Patient source: Cases and controls: sample of patients with and without glaucoma from the Baltimore Eye Study Follow-up Study) Optic disc photography; SAP suprathreshold Experienced technicians Follow-up confirmation 249 182 (disc photo); 228 (SAP); 68 M: 100; F: 149 USA (Baltimore Eye Study Follow-up) 1994
Already-Suspect Population (Cohort Studies)
Ekstrom 1993 40 (Patient source: people previously examined in a population-based glaucoma survey) GAT N/S Follow-up confirmation 760 413 (65–74) M: 364; F: 396 Sweden (Tierp Glaucoma Survey) Mar 1984–Mar 1986
Hammond 1979 41 (Patient source: eye clinic) Ophthalmoscopy Nurses skilled in use of the ophthalmoscope Ophthalmic examination 219 188 (21 and over) N/S USA N/S
Khong 2001 42 (Patient source: eye clinics) FDT C-20-5 N/S Ophthalmic examination 228 113 68.5 (23–91) M: 104; F: 119 Australia Dec 1999–Jan 2000
Leibowitz 1980 43 (Patient source: Framingham Eye Study) GAT Generally performed by 2nd or 3rd year residents in ophthalmology Follow-up confirmation 2631 574 (<65–75+) M: 272; F: 302 USA (Framingham Eye Study) Feb 1973–Feb 1975
Marraffa 1989 44 (Patient source: eye clinic) SAP suprathreshold Ophthalmologists Follow-up confirmation 104 182 eyes of 104 people 54.3 (18–76) M: 45; F: 59 Italy N/S
Schultz 1995 45 (Patient source: clinical practices of glaucoma specialist, cataract surgeon, and general ophthalmol-ogist) Optic disc photography Carried out: N/S Interpreted: 3rd-year ophthalmology residents Ophthalmic examination 258 365 eyes of? people (<40–>70) M: 112; F: 144; Unknown: 2 USA N/S
Spry 2005 46 (Patient source: hospital eye service) SAP threshold; FDT C-20 matrix SAP: clinic staff trained in visual field testing; FDT: N/S Ophthalmic examination 48 48 (both tests) 67.3 (SD 13.5) M: 24; F: 24 UK Oct 2003–Jan 2004
Theodossiades 2001 47 (Patient source: glaucoma clinics) Ophthalmoscopy Optometrists Ophthalmic examination 50 50 eyes of 50 people N/S N/S UK N/S
Already-Suspect Population (Case–Control Studies)
Airaksinen 1984 48 (Patient source: not stated) RNFL photography N/S Follow-up confirmation 142 132 eyes of 132 people Glaucoma: 62 (SD 20.5) Normal: 54 (SD 16.9); OHT: 57 (SD 12.7) N/S Canada + Finland N/S
Anton 1997 49 (Patient source: cases and controls: glaucoma unit) SAP threshold Uncertain Ophthalmic examination 180 180 eyes of 180 people Glaucoma: 61 (SD 8); Normal: 59 (SD 9) N/S Spain N/S
Damato 1989 50 (Patient source: Cases: not stated Controls: dermatology ward, hospital staff, relatives/friends of patients, patients with unilateral non-glaucomatous disease affecting the fellow eye) OKP Staff experienced in perimetry Ophthalmic examination 102 102 eyes of 102 people Glaucoma: 57.3; Normal: 54.4 N/S UK N/S
Enger 1987 51 (Patient source: Cases and controls: nerve fiber layer study) SAP threshold N/S Ophthalmic examination 112 170 eyes of 112 people Glaucoma: 61 (28–80); Normal: 51 (26–75) N/S USA N/S
Harper 1994 52 (Patient source: not stated) OKP; SAP suprathreshold Uncertain Ophthalmic examination 212 193 (OKP); 212 (SAP) Glaucoma: 67.8 (43–85); Normal: 61.5 (41–85) N/S UK N/S
Heeg 2005 53 (Patient source: Cases: glaucoma outpatient department Controls: old people’s homes, blood bank, other public places) FDT C-20-1; FDT C-20 full threshold N/S Ophthalmic examination 1112 208 (FDT C-20-1); 1112 (FDT C-20 full threshold) Glaucoma: 65 (13–91); Normal: 63 (33–94) Eligible: Glaucoma: M: 509; F: 542 Normal: M: 118; F: 119 Netherlands (Groningen Longitudinal Glaucoma Study) Jul 2000–Jun 2001
Ieong 2003 55 (Patient source: Cases: glaucoma subjects Controls: partners of cases, optometrist practice) HRT II; SAP suprathreshold Optometrists Ophthalmic examination 66 66 eyes of 66 people (both tests) Glaucoma: 69; Normal: 60 Glaucoma: M: 16; F: 13 Normal: M: 16; F: 21 UK N/S
Johnson 1999 56 (Patient source: not stated) FDT C-20-1 N/S Ophthalmic examination 108 160 eyes of 108 people Glaucoma: 64 (35–85); Normal: 46 (18–81) USA N/S
Quigley 1980 57 (Patient source: Cases and controls: Ophthalmol-ogical institute) Optic disc photography; RNFL photography Ophthalmologists Ophthalmic examination 175 294 eyes of? people (both tests) Readable photos: Glaucoma: 52.7 (SD 2.78); Glaucoma suspect: 45.2 (SD 1.56); Normal: 37.9 (SD 2.8) Unreadable photos: Glaucoma: 62.5 (SD 4.0); Glaucoma suspect: 59.6 (SD 6.3); Normal: 50 (SD 12.1) M: 86; F: 89 USA Jan 1978–Apr 1979
Sommer 1979 58 (Patient source: Cases and controls: glaucoma clinic) Optic disc photography; RNFL photography N/S Follow-up confirmation Unclear 223 eyes of ? people (both tests) N/S N/S USA N/S
Wollstein 2000 60 (Patient source: Cases: glaucoma clinic and ocular hypertension clinic Controls: spouses or friends of patients, responders-an advertisement) Optic disc photography Photos taken by trained technicians; assessed by glaucoma consultants, glaucoma fellow, clinical glaucoma technician Ophthalmic examination 123 123 eyes of 123 people Glaucoma: 65.1 (SD 10.06); Normal: 57.1 (SD 12.52) N/S UK N/S
Wood 1987 61 (Patient source: not stated) Ophthalmoscopy Ophthalmologists; junior doctors Ophthalmic examination 22 43 eyes of 22 people (32–75) N/S UK N/S
Table 2.
 
HSROC Analysis: All Studies Compared with Higher Quality Studies
Table 2.
 
HSROC Analysis: All Studies Compared with Higher Quality Studies
Optic Disc Photography HRT II FDT C-20-5 SAP Threshold
Sensitivity % (95% CrI) Specificity % (95% CrI) Sensitivity % (95% CrI) Specificity % (95% CrI) Sensitivity % (95% CrI) Specificity % (95% CrI) Sensitivity % (95% CrI) Specificity % (95% CrI)
All studies 73 (61–83) 89 (50–99) 86 (55–97) 89 (66–98) 78 (19–99) 75 (57–87) 88 (65–97) 80 (55–93)
Higher quality 74 (30–95) 82 (45–97) 93 (58–99) 85 (47–97) 72 (26–96) 60 (17–92) 73 (28–95) 64 (22–92)
Table 3.
 
Sensitivity, Specificity, DOR, and Relative DOR at the Common Cutoff for Studies Directly Comparing Tests
Table 3.
 
Sensitivity, Specificity, DOR, and Relative DOR at the Common Cutoff for Studies Directly Comparing Tests
Study Test Common Cutoff Sensitivity % (95% CrI) Specificity % (95% CrI) DOR (95% CrI) RDOR (95% CrI)
Vitale 2000 34 SAP supra Three adjacent points missed 50 (37–63) 83 (76–88) 5 (3–9) 1
Optic disc photo VCDR >0.6 77 (62–89) 59 (50–67) 5 (2–11) 0.99 (0.36–2.75)
Ieong 2003 55 SAP supra Optometrist’s judgment 72 (53–87) 95 (82–99) 46 (9–237) 1
HRT II Global/one of six segments abnormal 69 (49–85) 95 (82–99) 39 (8–198) 0.85 (0.08–8.54)
Robin 2005 30 SAP threshold AGIS score ≥3 (common cutoff) 63 (38–84) 74 (68–80) 5 (2–13) 1
HRT II ≥1 Borderline or 1 severe abnormality 95 (74–100) 81 (75–85) 75 (10–574) 15.01 (1.57–143.82)
FDT C-20-5 One abnormal point 84 (60–97) 55 (49–61) 7 (2–23) 1.31 (0.27–6.43)
Spry 2005 46 SAP threshold GHT outside normal limit and/or P < 0.05 with the PSD global index in one/both eyes 80 (52–96) 52 (34–69) 4 (1–18) 1
FDT C-20 matrix 100 (78–100) 27 (13–46) 12 (1–222) 2.83 (0.11–72.91)
Ivers 2001 23 SAP supra Three or more points missing 89 (80–94) 73 (71–74) 20 (10–39) 1
GAT IOP >22 mm Hg 14 (7–23) 98 (97–98) 6 (3–12) 0.31 (0.12–0.78)
Wang 1998 36 SAP supra Absolute or relative defects ≥ 17 70 (57–80) 67 (59–74) 5 (2–9) 1
GAT IOP > 21 mm Hg 28 (17–40) 96 (93–98) 9 (4–19) 1.89 (0.70–5.13)
Table 4.
 
Summary of Sensitivity, Specificity and DOR for Tests Included in the HSROC Meta-analysis Models
Table 4.
 
Summary of Sensitivity, Specificity and DOR for Tests Included in the HSROC Meta-analysis Models
Test Number of Studies Common Cutoff Sensitivity % (95% CrI) Specificity % (95% CrI) DOR (95% CrI)
Ophthalmoscopy 5 VCDR ≥0.7 60 (34–82) 94 (76–99) 26 (6–110)
Optic disc photography 6 VCDR ≥0.6 73 (61–83) 89 (50–99) 22 (3–148)
RNFL photography 4 Diffuse and/or localized defect 75 (46–92) 88 (53–98) 23 (4–124)
HRT II 3 ≥1 Borderline or outside normal limits 86 (55–97) 89 (66–98) 51 (11–246)
FDT
 C-20-1 3 1 Abnormal point 92 (65–99) 94 (73–99) 181 (25–2139)
 C-20-5 5 1 Abnormal point 78 (19–99) 75 (57–87) 10 (0.7–249)
OKP 4 1 Abnormal point 86 (29–100) 90 (79–96) 58 (4–1585)
SAP suprathreshold 9 ≥3 Points missing 71 (51–86) 85 (73–93) 14 (6–34)
SAP threshold 5 AGIS score ≥3 88 (65–97) 80 (55–93) 30 (6–159)
GAT 9 IOP >21 mm Hg 46 (22–71) 95 (89–97) 15 (4–49)
×
×

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.

×