- Email: [email protected]
Reliability of Simultaneous Visual Field Testing
Reliability of Simultaneous Visual Field Testing Benjamin C. Kramer, MD,1 David C. Musch, PhD, MPH,1,2 Leslie M. Niziol, MS,1 Jennifer S. Weizer, MD1 Objective: To test for differences in reliability and performance between traditional, isolated perimetry and simultaneous testing with 2 patients in the same room. Design: Comparative case series. Participants: A total of 471 eyes of 261 subjects. Methods: Consecutive patients undergoing Humphrey visual field (VF) testing in the Kellogg Eye Center glaucoma clinic were screened. Patients who underwent VF testing with another patient in the same room (“double field”) during the screening interval were included as subjects if a comparison isolated VF from the same patient (“single field”) was obtainable from the clinic’s records. An individual subject’s performance and reliability on his/her double and single fields were compared using a paired t test. In addition, the double fields were stratified by technician-to-patient ratio, and their VF indices were compared using an independent 2-sample t test. Main Outcome Measures: False negatives, false positives, fixation losses, mean deviation, pattern standard deviation, VF index, VF duration, and technician-to-patient ratio. Results: No significant differences between single and double fields were found in the reliability or performance parameters. Test duration was longer in double fields than in single fields (6.1 vs. 5.9 minutes, P ⬍ 0.001). There were no significant differences found in reliability or performance indices when the double-field data were stratified by technician-to-patient ratio (1:2 vs. 2:2). Conclusions: There is no decrement in VF performance or reliability when patients undergo simultaneous testing with another patient in the same room. Busy clinical practices may be able to minimize costs and maximize efficiency by having 1 technician simultaneously supervise more than 1 test-taker in the same space. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Ophthalmology 2012;119:304 –307 © 2012 by the American Academy of Ophthalmology.
Visual field (VF) testing guides many clinical decisions when treating adult glaucoma. It is accepted that accurate perimetric assessment requires an alert, focused patient familiar with the testing protocol. Such diverse stimuli as consuming alcohol,1 listening to a Mozart sonata,2 and watching an instructional video3 have been demonstrated to affect patient performance. With the transition from manual to automated perimetry, it was expected that VF testing would become less dependent on an adept, focused technician. The manual of operations for the Collaborative Initial Glaucoma Treatment Study, however, demands intensive supervision during automated VF testing.4 Perimetrists are instructed to monitor continually and to offer verbal feedback to patients, such as words of encouragement and reminders to blink and maintain head position. Several authors have suggested that such intensive technician monitoring is unnecessary.5,6 In a recent glaucoma text, the role of technicians is deemphasized: “only [ensure] that the patient understands the testing procedure, is comfortably positioned at the perimeter, and adheres to the requirements of the test.”7 Visual field testing is traditionally conducted in a quiet, isolated room containing a single patient and 1 perimetrist. Some clinics perform simultaneous VF testing with more than 1 patient in a single room to reduce costs and maximize efficiency. Hypothetically, the presence of a perimetrist offering continuous verbal feedback to 1 patient while an-
304
© 2012 by the American Academy of Ophthalmology Published by Elsevier Inc.
other is being tested simultaneously in the same space might induce concentration lapses. Also, significant sounds are generated during automated testing, including a series of noises intended to assess for false-positive responses, and these may be distracting to other test-takers present. This study was designed to test for differences in reliability and performance between traditional, isolated perimetry and simultaneous testing with 2 patients in the same room. We also sought to examine the effects of the technician-topatient ratio under these testing conditions.
Materials and Methods University of Michigan Institutional Review Board/Ethics Committee approval was obtained. Records of all consecutive patients undergoing Humphrey automated VF (Carl Zeiss Meditec, Dublin, CA) testing in the Kellogg Eye Center’s glaucoma clinic during a 7-month interval (October 2008 to April 2009) were retrospectively reviewed. Patients were identified for potential study enrollment if they performed a “double field” during the 7-month study period. A “double field” was defined as a Humphrey automated VF test performed during a testing session that overlapped another patient’s testing session in the same room by at least 1 minute. A “testing session” was defined as the time from the beginning of a patient’s first eye’s VF to the conclusion of a patient’s second eye’s VF, as determined by timestamps on the report (starting time ISSN 0161-6420/12/$–see front matter doi:10.1016/j.ophtha.2011.08.021
Kramer et al 䡠 Reliability of Simultaneous Visual Field Testing Table 1. Reliability and Performance in Single and Double Visual Fields
and other clinical parameters. A P value less than 0.05 was considered to be statistically significant.
Mean (SD)
Reliability FP (%) FN (%) FL (%) Performance MD (dB) PSD (dB) VFI (%)
Single VF
Double VF
P*
Results
3.2 (4.4) 4.0 (5.8) 11 (14)
3.4 (4.7) 4.2 (6.8) 12 (20)
0.47 0.64 0.37
⫺4.9 (6.0) 4.7 (3.8) 87.7 (15.9)
⫺5.1 (6.3) 4.6 (3.9) 87.3 (17.3)
0.22 0.32 0.25
Visual field tests performed on 471 eyes from 261 subjects (140 women and 121 men; mean ⫾ standard deviation [SD] age, 65.8⫾13.5 years) were included in the study. Fifty-one subjects contributed just 1 eye. There were 234 right eyes and 237 left eyes. The mean interval between single and double fields was 353 days (SD 141, range 5–721). In 56.5% of subjects, the single field occurred before the double field. There were no significant differences in the reliability indices between the single and double fields. By comparing the single fields with their corresponding double fields, respectively, the means ⫾ SDs were FP 3.2⫾4.4% versus 3.4⫾4.7% (P⫽0.47), FN 4.0⫾5.8% versus 4.2⫾6.8% (P⫽0.64), and FL 11.1⫾14.1% versus 12.0⫾20.2% (P⫽0.37). Also, there were no significant differences in the performance indices between the single and double fields. By comparing the single fields with their corresponding double fields, respectively, the means ⫾ SDs were MD ⫺4.9⫾6.0 dB versus ⫺5.1 ⫾ 6.3 dB (P⫽0.22), PSD 4.7⫾3.8 dB versus 4.6⫾3.9 dB (P⫽0.32), and VFI 87.7⫾15.9% versus 87.3⫾17.3% (P⫽0.25) (Table 1). Mean test duration was significantly shorter among single fields than double fields, lasting 5.9⫾1.5 minutes versus 6.1⫾1.5 minutes (P⫽0.001). When the double VFs were stratified by technician-topatient ratio (1:2 vs. 2:2) and compared, there were no significant differences in the reliability indices, performance indices, or test duration (Table 2). When all single and double fields were pooled, correlations among the performance, reliability, and clinical parameters were identified. Increased FN rate was associated with increased age, worse VA, longer test duration, and worsening global performance (MD, PSD, and VFI). Increased FP rate was associated with increased test duration and better MD, and increased FLs were associated with worse VA and longer test duration (Table 3).
dB ⫽ decibels; FL ⫽ fixation loss; FN ⫽ false negative; FP ⫽ false positive; MD ⫽ mean deviation; PSD ⫽ pattern standard deviation; SD ⫽ standard deviation; VF ⫽ visual field; VFI ⫽ visual field index. *P values are derived from paired t tests.
and test duration). If only 1 eye was tested, the testing session concluded after that eye’s VF was completed. Patients were allowed to contribute a double field from each eye if both eyes underwent VF testing during a testing session. If a patient had double fields from 2 different testing sessions during the study period, only the VF from the first testing session was included. Supervising technicians were familiar with Collaborative Initial Glaucoma Treatment Study VF protocols, but no specific supervision protocols were enforced. If the same technician supervised both patients during overlapping testing sessions, the technician:patient ratio was recorded as 1:2. If different technicians supervised both patients during overlapping testing sessions, the ratio was 2:2. All VFs were performed in the same room containing 2 Humphrey VF analyzers separated by a thin curtain that was not soundproof. Patient instructions for performing the VF were given by the technician in the VF room, whether or not another patient was concurrently undergoing VF testing in the same room. After identifying each double field in the study period, a comparison single field from each patient was sought. A “single field” was defined as a VF from the patient’s same eye, using the same testing protocol (Swedish Interactive Testing Algorithm standard 10-2, 24-2, 30-2, or Short Wavelength Automated Perimetry) as the corresponding double field and performed within 2 years of the double field. Single fields were performed in isolation with continuous technician supervision, per the clinic’s usual protocol. All patients who performed double fields during the study period and had a comparison single field available were included as subjects in this study. If more than 1 single field within 2 years of the double field was available, the most chronologically proximal examination was chosen. Clinical parameters recorded per patient were sex, age at time of double field, and Snellen visual acuity (VA) per eye at the time of the double field. Data collected per VF were performance indices such as mean deviation (MD), pattern standard deviation (PSD), and VF index (VFI), and reliability indices recorded were percent false positive (FP), percent false negative (FN), and percent fixation losses (FLs). Test duration and whether the right or left eye was being tested were also recorded for each VF. An individual’s double and single VF performance and reliability indices were compared using a paired t test. The double fields were stratified by technician-to-patient ratio, and their VF parameters were compared using independent 2-sample t tests. Pearson’s correlation coefficients were calculated between reliability indices
Discussion We compared the VF testing of the same patient under 2 conditions: individual testing in isolation versus simultaneTable 2. Reliability and Performance in Double Visual Fields, Stratified by Technician:Patient Ratio Mean (SD) Tech:Patient Ratio Reliability FP (%) FN (%) FL (%) Performance MD (dB) PSD (dB) VFI (%)
1:2 (n ⫽ 213)
2:2 (n ⫽ 258)
P*
3.1 (4.1) 4.6 (6.7) 14 (26)
3.6 (5.0) 3.8 (6.9) 11 (14)
0.30 0.19 0.09
⫺5.3 (6.7) 4.7 (3.9) 87 (19)
⫺4.8 (6.0) 4.6 (3.8) 88 (16)
0.39 0.59 0.39
dB ⫽ decibels; FL ⫽ fixation loss; FN ⫽ false negative; FP ⫽ false positive; MD ⫽ mean deviation; PSD ⫽ pattern standard deviation; SD ⫽ standard deviation; VFI ⫽ visual field index. *P values are derived from independent t tests.
305
Ophthalmology Volume 119, Number 2, February 2012 Table 3. Proportion of Variability (r2) in Reliability Indices Explained by Age and Visual Function Test Results Percent False Negative
Age logMAR VA VF duration MD PSD % VFI
Percent False Positive
Percent Fixation Loss
r2
P
r2
P
r2
P
0.21 0.16 0.47 ⫺0.33 0.32 ⫺0.32
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001
0.00 ⫺0.03 0.17 0.09 0.04 0.01
0.94 0.32 ⬍0.001 0.009 0.27 0.70
0.06 0.08 0.11 0.02 0.00 0.00
0.08 0.009 ⬍0.001 0.62 0.94 0.98
logMAR ⫽ logarithm of the minimum angle of resolution; MD ⫽ mean deviation; PSD ⫽ pattern standard deviation; VA ⫽ visual acuity; VF ⫽ visual field; VFI ⫽ visual field index.
ous testing with another patient in the same room. We failed to detect any difference in reliability and performance between these 2 circumstances. When double fields were stratified by technician-to-patient ratio, there still appeared to be no benefit to isolated testing. The only significant difference detected was slightly longer mean test duration for double fields than single fields (6.1 vs. 5.9 minutes). Hypothetically, longer testing might induce a fatiguerelated decrement in performance, but the extra 12 seconds is not likely clinically significant because we did not detect any worsening. In our study population, increased FNs were strongly correlated with increased age, poorer VA, longer test duration, and worse VF performance indices. Previous authors have observed these correlations, concluding that FNs are more indicative of glaucomatous field loss than patient concentration.8,9 To our knowledge, the generally accepted benefits of isolated VF testing have not been studied, although several authors have suggested that continuous technician supervision is unnecessary. Johnson et al5 studied technician VF supervision of 169 consecutive subjects with suspected optic nerve disease. After a brief practice VF test, 1 eye was randomized to continuous or intermittent supervision. No significant differences were found in any reliability or performance measure, so the authors concluded that both strategies were equally effective. Van Coevorden et al6 also examined the utility of continuous versus initial technician supervision in VF testing. In that study, 200 patients with glaucoma or neuro-ophthalmic disease had 1 eye undergo VF testing both with and without supervision, with testing order randomized. As in our study, no differences were found in the individual reliability or performance indices. When the manufacturer’s specified flag of unreliability was used (ⱖ20% FL, ⱖ33% FN or FP), a small subgroup of patients was found to benefit from supervision. Predictors of benefit from supervision were low educational level and prior testing with high FPs. Of note, the effect of supervision disappeared when correction was made for multiple statistical tests. Our study’s strengths include a large number of VFs and comparisons performed for the same patient; however, there are some limitations. Although most patients had double fields performed simply through the happenstance of clinic scheduling, technicians were at liberty to defer simultaneous testing on the basis of perceived patient ability. The
306
patients most likely to benefit from isolated testing, such as those with previous poor performance or reliability, may have been screened out from study inclusion. Second, the learning effect, whereby VF performance and reliability in novice subjects undergoing serial testing tend to improve, has been well documented.10 –12 We sought to mitigate the learning effect by including single fields from before or after the double VF, but the results were still slightly skewed toward the single VF first pattern (56.5%). Finally, there was variability in the amount of time overlap of VFs included as double fields. For example, 1 or 2 minutes of overlap during a 20-minute testing session may have been insufficient to induce a significant decrement in performance. We would argue that our large study population helps to offset this limitation. In conclusion, our study suggests that there is no decrement in VF performance or reliability when selected patients undergo simultaneous testing with another patient in the same room. Prospective study is needed to determine if our results can be generalized to all patients. Busy clinical practices may be able to minimize costs and maximize clinical efficiency by having 1 technician simultaneously supervise more than 1 test-taker in the same space.
References 1. Zulauf M, Flammer J, Signer C. The influence of alcohol on the outcome of automated static perimetry. Graefes Arch Clin Exp Ophthalmol 1986;224:525– 8. 2. Fiorelli VM, Kasahara N, Cohen R, et al. Improved automated perimetry performance following exposure to Mozart. Br J Ophthalmol 2006;90:543–5. 3. Sherafat H, Spry PG, Waldock A, et al. Effect of a patient training video on visual field test reliability. Br J Ophthalmol 2003;87:153– 6. 4. Collaborative Initial Glaucoma Treatment Study. Manual of Operations. 1997:7–22. 5. Johnson LN, Aminlari A, Sassani JW. Effect of intermittent versus continuous patient monitoring on reliability indices during automated perimetry. Ophthalmology 1993;100:76 – 84. 6. Van Coevorden RE, Mills RP, Chen YY, Barnebey HS. Continuous visual field test supervision may not always be necessary. Ophthalmology 1999;106:178 – 81.
Kramer et al 䡠 Reliability of Simultaneous Visual Field Testing 7. Allingham RR, Damji KF, Freedman S, et al. Assessment of Visual Fields. In: Pine JW Jr, ed. Shields Textbook of Glaucoma. 6th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011:103. 8. Bengtsson B, Heijl A. False-negative responses in glaucoma perimetry: indicators of patient performance or test reliability? Invest Ophthalmol Vis Sci 2000;41:2201– 4. 9. Birt CM, Shin DH, Samudrala V, et al. Analysis of reliability indices from Humphrey visual field tests in an urban glaucoma population. Ophthalmology 1997;104:1126 –30.
10. Heijl A, Lindgren G, Olsson J. The effect of perimetric experience in normal subjects. Arch Ophthalmol 1989;107: 81– 6. 11. Werner EB, Adelson A, Krupin T. Effect of patient experience on the results of automated perimetry in clinically stable glaucoma patients. Ophthalmology 1988;95:764 –7. 12. Wild JM, Dengler-Harles M, Searle AE, et al. The influence of the learning effect on automated perimetry in patients with suspected glaucoma. Acta Ophthalmol (Copenh) 1989; 67:537– 45.
Footnotes and Financial Disclosures Originally received: April 21, 2011. Final revision: July 26, 2011. Accepted: August 9, 2011. Available online: November 23, 2011.
Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Manuscript no. 2011-623.
1
Department of Ophthalmology and Visual Sciences, University of Michigan Medical School, Ann Arbor, Michigan.
2
Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor, Michigan. Planned poster presentation at: Association for Research in Vision and Ophthalmology meeting, May 5, 2011, Fort Lauderdale, Florida.
Supported by an unrestricted grant to the Department of Ophthalmology and Visual Sciences, University of Michigan, from Research to Prevent Blindness (RPB) and the RPB Lew R. Wasserman Merit Award (DCM). These funding organizations had no role in the design or conduct of this research. Correspondence: Jennifer S. Weizer, MD, Kellogg Eye Center, 1000 Wall Street, Ann Arbor, MI 48105. E-mail: [email protected].
307
Visual field (VF) testing guides many clinical decisions when treating adult glaucoma. It is accepted that accurate perimetric assessment requires an alert, focused patient familiar with the testing protocol. Such diverse stimuli as consuming alcohol,1 listening to a Mozart sonata,2 and watching an instructional video3 have been demonstrated to affect patient performance. With the transition from manual to automated perimetry, it was expected that VF testing would become less dependent on an adept, focused technician. The manual of operations for the Collaborative Initial Glaucoma Treatment Study, however, demands intensive supervision during automated VF testing.4 Perimetrists are instructed to monitor continually and to offer verbal feedback to patients, such as words of encouragement and reminders to blink and maintain head position. Several authors have suggested that such intensive technician monitoring is unnecessary.5,6 In a recent glaucoma text, the role of technicians is deemphasized: “only [ensure] that the patient understands the testing procedure, is comfortably positioned at the perimeter, and adheres to the requirements of the test.”7 Visual field testing is traditionally conducted in a quiet, isolated room containing a single patient and 1 perimetrist. Some clinics perform simultaneous VF testing with more than 1 patient in a single room to reduce costs and maximize efficiency. Hypothetically, the presence of a perimetrist offering continuous verbal feedback to 1 patient while an-
304
© 2012 by the American Academy of Ophthalmology Published by Elsevier Inc.
other is being tested simultaneously in the same space might induce concentration lapses. Also, significant sounds are generated during automated testing, including a series of noises intended to assess for false-positive responses, and these may be distracting to other test-takers present. This study was designed to test for differences in reliability and performance between traditional, isolated perimetry and simultaneous testing with 2 patients in the same room. We also sought to examine the effects of the technician-topatient ratio under these testing conditions.
Materials and Methods University of Michigan Institutional Review Board/Ethics Committee approval was obtained. Records of all consecutive patients undergoing Humphrey automated VF (Carl Zeiss Meditec, Dublin, CA) testing in the Kellogg Eye Center’s glaucoma clinic during a 7-month interval (October 2008 to April 2009) were retrospectively reviewed. Patients were identified for potential study enrollment if they performed a “double field” during the 7-month study period. A “double field” was defined as a Humphrey automated VF test performed during a testing session that overlapped another patient’s testing session in the same room by at least 1 minute. A “testing session” was defined as the time from the beginning of a patient’s first eye’s VF to the conclusion of a patient’s second eye’s VF, as determined by timestamps on the report (starting time ISSN 0161-6420/12/$–see front matter doi:10.1016/j.ophtha.2011.08.021
Kramer et al 䡠 Reliability of Simultaneous Visual Field Testing Table 1. Reliability and Performance in Single and Double Visual Fields
and other clinical parameters. A P value less than 0.05 was considered to be statistically significant.
Mean (SD)
Reliability FP (%) FN (%) FL (%) Performance MD (dB) PSD (dB) VFI (%)
Single VF
Double VF
P*
Results
3.2 (4.4) 4.0 (5.8) 11 (14)
3.4 (4.7) 4.2 (6.8) 12 (20)
0.47 0.64 0.37
⫺4.9 (6.0) 4.7 (3.8) 87.7 (15.9)
⫺5.1 (6.3) 4.6 (3.9) 87.3 (17.3)
0.22 0.32 0.25
Visual field tests performed on 471 eyes from 261 subjects (140 women and 121 men; mean ⫾ standard deviation [SD] age, 65.8⫾13.5 years) were included in the study. Fifty-one subjects contributed just 1 eye. There were 234 right eyes and 237 left eyes. The mean interval between single and double fields was 353 days (SD 141, range 5–721). In 56.5% of subjects, the single field occurred before the double field. There were no significant differences in the reliability indices between the single and double fields. By comparing the single fields with their corresponding double fields, respectively, the means ⫾ SDs were FP 3.2⫾4.4% versus 3.4⫾4.7% (P⫽0.47), FN 4.0⫾5.8% versus 4.2⫾6.8% (P⫽0.64), and FL 11.1⫾14.1% versus 12.0⫾20.2% (P⫽0.37). Also, there were no significant differences in the performance indices between the single and double fields. By comparing the single fields with their corresponding double fields, respectively, the means ⫾ SDs were MD ⫺4.9⫾6.0 dB versus ⫺5.1 ⫾ 6.3 dB (P⫽0.22), PSD 4.7⫾3.8 dB versus 4.6⫾3.9 dB (P⫽0.32), and VFI 87.7⫾15.9% versus 87.3⫾17.3% (P⫽0.25) (Table 1). Mean test duration was significantly shorter among single fields than double fields, lasting 5.9⫾1.5 minutes versus 6.1⫾1.5 minutes (P⫽0.001). When the double VFs were stratified by technician-topatient ratio (1:2 vs. 2:2) and compared, there were no significant differences in the reliability indices, performance indices, or test duration (Table 2). When all single and double fields were pooled, correlations among the performance, reliability, and clinical parameters were identified. Increased FN rate was associated with increased age, worse VA, longer test duration, and worsening global performance (MD, PSD, and VFI). Increased FP rate was associated with increased test duration and better MD, and increased FLs were associated with worse VA and longer test duration (Table 3).
dB ⫽ decibels; FL ⫽ fixation loss; FN ⫽ false negative; FP ⫽ false positive; MD ⫽ mean deviation; PSD ⫽ pattern standard deviation; SD ⫽ standard deviation; VF ⫽ visual field; VFI ⫽ visual field index. *P values are derived from paired t tests.
and test duration). If only 1 eye was tested, the testing session concluded after that eye’s VF was completed. Patients were allowed to contribute a double field from each eye if both eyes underwent VF testing during a testing session. If a patient had double fields from 2 different testing sessions during the study period, only the VF from the first testing session was included. Supervising technicians were familiar with Collaborative Initial Glaucoma Treatment Study VF protocols, but no specific supervision protocols were enforced. If the same technician supervised both patients during overlapping testing sessions, the technician:patient ratio was recorded as 1:2. If different technicians supervised both patients during overlapping testing sessions, the ratio was 2:2. All VFs were performed in the same room containing 2 Humphrey VF analyzers separated by a thin curtain that was not soundproof. Patient instructions for performing the VF were given by the technician in the VF room, whether or not another patient was concurrently undergoing VF testing in the same room. After identifying each double field in the study period, a comparison single field from each patient was sought. A “single field” was defined as a VF from the patient’s same eye, using the same testing protocol (Swedish Interactive Testing Algorithm standard 10-2, 24-2, 30-2, or Short Wavelength Automated Perimetry) as the corresponding double field and performed within 2 years of the double field. Single fields were performed in isolation with continuous technician supervision, per the clinic’s usual protocol. All patients who performed double fields during the study period and had a comparison single field available were included as subjects in this study. If more than 1 single field within 2 years of the double field was available, the most chronologically proximal examination was chosen. Clinical parameters recorded per patient were sex, age at time of double field, and Snellen visual acuity (VA) per eye at the time of the double field. Data collected per VF were performance indices such as mean deviation (MD), pattern standard deviation (PSD), and VF index (VFI), and reliability indices recorded were percent false positive (FP), percent false negative (FN), and percent fixation losses (FLs). Test duration and whether the right or left eye was being tested were also recorded for each VF. An individual’s double and single VF performance and reliability indices were compared using a paired t test. The double fields were stratified by technician-to-patient ratio, and their VF parameters were compared using independent 2-sample t tests. Pearson’s correlation coefficients were calculated between reliability indices
Discussion We compared the VF testing of the same patient under 2 conditions: individual testing in isolation versus simultaneTable 2. Reliability and Performance in Double Visual Fields, Stratified by Technician:Patient Ratio Mean (SD) Tech:Patient Ratio Reliability FP (%) FN (%) FL (%) Performance MD (dB) PSD (dB) VFI (%)
1:2 (n ⫽ 213)
2:2 (n ⫽ 258)
P*
3.1 (4.1) 4.6 (6.7) 14 (26)
3.6 (5.0) 3.8 (6.9) 11 (14)
0.30 0.19 0.09
⫺5.3 (6.7) 4.7 (3.9) 87 (19)
⫺4.8 (6.0) 4.6 (3.8) 88 (16)
0.39 0.59 0.39
dB ⫽ decibels; FL ⫽ fixation loss; FN ⫽ false negative; FP ⫽ false positive; MD ⫽ mean deviation; PSD ⫽ pattern standard deviation; SD ⫽ standard deviation; VFI ⫽ visual field index. *P values are derived from independent t tests.
305
Ophthalmology Volume 119, Number 2, February 2012 Table 3. Proportion of Variability (r2) in Reliability Indices Explained by Age and Visual Function Test Results Percent False Negative
Age logMAR VA VF duration MD PSD % VFI
Percent False Positive
Percent Fixation Loss
r2
P
r2
P
r2
P
0.21 0.16 0.47 ⫺0.33 0.32 ⫺0.32
⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001
0.00 ⫺0.03 0.17 0.09 0.04 0.01
0.94 0.32 ⬍0.001 0.009 0.27 0.70
0.06 0.08 0.11 0.02 0.00 0.00
0.08 0.009 ⬍0.001 0.62 0.94 0.98
logMAR ⫽ logarithm of the minimum angle of resolution; MD ⫽ mean deviation; PSD ⫽ pattern standard deviation; VA ⫽ visual acuity; VF ⫽ visual field; VFI ⫽ visual field index.
ous testing with another patient in the same room. We failed to detect any difference in reliability and performance between these 2 circumstances. When double fields were stratified by technician-to-patient ratio, there still appeared to be no benefit to isolated testing. The only significant difference detected was slightly longer mean test duration for double fields than single fields (6.1 vs. 5.9 minutes). Hypothetically, longer testing might induce a fatiguerelated decrement in performance, but the extra 12 seconds is not likely clinically significant because we did not detect any worsening. In our study population, increased FNs were strongly correlated with increased age, poorer VA, longer test duration, and worse VF performance indices. Previous authors have observed these correlations, concluding that FNs are more indicative of glaucomatous field loss than patient concentration.8,9 To our knowledge, the generally accepted benefits of isolated VF testing have not been studied, although several authors have suggested that continuous technician supervision is unnecessary. Johnson et al5 studied technician VF supervision of 169 consecutive subjects with suspected optic nerve disease. After a brief practice VF test, 1 eye was randomized to continuous or intermittent supervision. No significant differences were found in any reliability or performance measure, so the authors concluded that both strategies were equally effective. Van Coevorden et al6 also examined the utility of continuous versus initial technician supervision in VF testing. In that study, 200 patients with glaucoma or neuro-ophthalmic disease had 1 eye undergo VF testing both with and without supervision, with testing order randomized. As in our study, no differences were found in the individual reliability or performance indices. When the manufacturer’s specified flag of unreliability was used (ⱖ20% FL, ⱖ33% FN or FP), a small subgroup of patients was found to benefit from supervision. Predictors of benefit from supervision were low educational level and prior testing with high FPs. Of note, the effect of supervision disappeared when correction was made for multiple statistical tests. Our study’s strengths include a large number of VFs and comparisons performed for the same patient; however, there are some limitations. Although most patients had double fields performed simply through the happenstance of clinic scheduling, technicians were at liberty to defer simultaneous testing on the basis of perceived patient ability. The
306
patients most likely to benefit from isolated testing, such as those with previous poor performance or reliability, may have been screened out from study inclusion. Second, the learning effect, whereby VF performance and reliability in novice subjects undergoing serial testing tend to improve, has been well documented.10 –12 We sought to mitigate the learning effect by including single fields from before or after the double VF, but the results were still slightly skewed toward the single VF first pattern (56.5%). Finally, there was variability in the amount of time overlap of VFs included as double fields. For example, 1 or 2 minutes of overlap during a 20-minute testing session may have been insufficient to induce a significant decrement in performance. We would argue that our large study population helps to offset this limitation. In conclusion, our study suggests that there is no decrement in VF performance or reliability when selected patients undergo simultaneous testing with another patient in the same room. Prospective study is needed to determine if our results can be generalized to all patients. Busy clinical practices may be able to minimize costs and maximize clinical efficiency by having 1 technician simultaneously supervise more than 1 test-taker in the same space.
References 1. Zulauf M, Flammer J, Signer C. The influence of alcohol on the outcome of automated static perimetry. Graefes Arch Clin Exp Ophthalmol 1986;224:525– 8. 2. Fiorelli VM, Kasahara N, Cohen R, et al. Improved automated perimetry performance following exposure to Mozart. Br J Ophthalmol 2006;90:543–5. 3. Sherafat H, Spry PG, Waldock A, et al. Effect of a patient training video on visual field test reliability. Br J Ophthalmol 2003;87:153– 6. 4. Collaborative Initial Glaucoma Treatment Study. Manual of Operations. 1997:7–22. 5. Johnson LN, Aminlari A, Sassani JW. Effect of intermittent versus continuous patient monitoring on reliability indices during automated perimetry. Ophthalmology 1993;100:76 – 84. 6. Van Coevorden RE, Mills RP, Chen YY, Barnebey HS. Continuous visual field test supervision may not always be necessary. Ophthalmology 1999;106:178 – 81.
Kramer et al 䡠 Reliability of Simultaneous Visual Field Testing 7. Allingham RR, Damji KF, Freedman S, et al. Assessment of Visual Fields. In: Pine JW Jr, ed. Shields Textbook of Glaucoma. 6th ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011:103. 8. Bengtsson B, Heijl A. False-negative responses in glaucoma perimetry: indicators of patient performance or test reliability? Invest Ophthalmol Vis Sci 2000;41:2201– 4. 9. Birt CM, Shin DH, Samudrala V, et al. Analysis of reliability indices from Humphrey visual field tests in an urban glaucoma population. Ophthalmology 1997;104:1126 –30.
10. Heijl A, Lindgren G, Olsson J. The effect of perimetric experience in normal subjects. Arch Ophthalmol 1989;107: 81– 6. 11. Werner EB, Adelson A, Krupin T. Effect of patient experience on the results of automated perimetry in clinically stable glaucoma patients. Ophthalmology 1988;95:764 –7. 12. Wild JM, Dengler-Harles M, Searle AE, et al. The influence of the learning effect on automated perimetry in patients with suspected glaucoma. Acta Ophthalmol (Copenh) 1989; 67:537– 45.
Footnotes and Financial Disclosures Originally received: April 21, 2011. Final revision: July 26, 2011. Accepted: August 9, 2011. Available online: November 23, 2011.
Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Manuscript no. 2011-623.
1
Department of Ophthalmology and Visual Sciences, University of Michigan Medical School, Ann Arbor, Michigan.
2
Department of Epidemiology, School of Public Health, University of Michigan, Ann Arbor, Michigan. Planned poster presentation at: Association for Research in Vision and Ophthalmology meeting, May 5, 2011, Fort Lauderdale, Florida.
Supported by an unrestricted grant to the Department of Ophthalmology and Visual Sciences, University of Michigan, from Research to Prevent Blindness (RPB) and the RPB Lew R. Wasserman Merit Award (DCM). These funding organizations had no role in the design or conduct of this research. Correspondence: Jennifer S. Weizer, MD, Kellogg Eye Center, 1000 Wall Street, Ann Arbor, MI 48105. E-mail: [email protected].
307