Before trying to explain how the instrument works or how to interpret the findings you should briefly review logarithms. The statement 10² =100, can be expressed in another form by using Logarithms, abbreviated LOG. Instead, of saying that "10 exponent 2 equals 100 we can say " the logarithm of 100 to the base 10 equals 2," (LOG₁₀ 100 = 2). We can take the following values representing intensity in apostilbs and convert them to logarithms.

 LOG₁₀10,000 =4
LOG₁₀1000 = 3
LOG₁₀ 100 = 2
LOG₁₀10 = 1
LOG₁₀ 1 = 0
LOG₁₀ 0.1= -1
Each Step Representing One Log Unit Changes

The intensity range of the Humphrey Field Analyzer is from 10,000 to 0.1 apostilbs (asb). The term apostilbs relates to luminance of a given test target being projected on to the interior of the white bowl. The reason these values in apostilbs are converted to logarithms and then to decibels is because these numerical values are smaller, easier to work with and give the same results. They took 10,000 apostilbs and gave it a value of (0) zero meaning there is zero retinal sensitivity. This target has the highest intensity, therefore; if a patient does not respond the scotoma (area of reduced retinal sensitivity) is considered to be absolute or blind. Then one log unit below the 10,000 asb or 1000 asb becomes equal to 10 or 10 decibels (dB). The value (10) is used instead of (1) because it is easier to express changes in whole numbers from (0) through 10 instead of fractions. The term decibels is a relative value which expresses the attenuation from maximum intensity of 10,000 asb.

Using This Concept
10,000 ASB = 0 dB
1000 ASB = 10 dB
100 ASB = 20 dB
10 ASB = 30 dB
1 ASB = 40 dB
0.1 ASB = 50 dB

One can also say that 1dB = 0.1 LOG UNIT and 10dB = 1 LOG UNIT. The following graphic may further help explain the concept:

There is a direct relationship to Goldmann notations. The Goldmann filter steps (a) through (e) are each equal to one (1) dB . Filter steps (1) through (4) are equal to five (5) dB each. Examples: III2e to III3e is and increase of five dB in brightness. Going from an III2a to III2b is a one (1) dB increase in brightness.

The Humphrey Field Analyzer uses a constant target size equal to a Goldmann "III" and varies the target brightness only. The standard field analyzer target size GOLDMANN "III" (4 mm²) can be changed if a larger, say GOLDMANN "V" (64 mm²) or smaller GOLDMANN "I" (1/16 mm²) target size if needed for severely disturbed fields or in clinical research. However, the Field Analyzer can cover the whole range of stimulus values, to about the equivalent of the Goldmann V4e, without changing the standard target size III spot. With the typical kinetic "MANUAL PERIMETER" one must change both the target size and the brightness or filters.

The Humphery Field Analyzer's statistical package (STATPAC) uses a model based on test results of patients with normal fields, retinal sensitivity, and pupil size for each different age group. It compares the patient's test results against this model to determine how their threshold results, for each tested point, compares or falls outside the normal population model.

Every patient's test begins with an initialization procedure that determines the threshold level for each of the four quadrants of their central field. This then determines their reference hill of vision after correcting for age and general responsiveness of the patient.

The values in parentheses shows that point has been tested twice and the system uses the average of the two values to make other calculations. The second most sensitive value is used to calculate the expected height of the central hill of vision, known as the central reference level. It is indeed possible that actual values measure might be below the calculated value or expected value. If the value is lower than the expected value for a given patients age this will be expressed in the Total Deviation as a negative number say (-1) showing it is 1 decibel below the expected value. These negative values become diagnostic when they reach (-5) or greater and more so if there are several grouped together. The parameters for most automated perimeters are loosely based on a hill of vision where the central five (5) degree's sensitivity, in decibels, is in the low to mid 30 decibels, from five (5) to thirty (30) degrees sensitivity is in the mid to upper 20 decibels and beyond thirty (30) degrees the sensitivity is in the teens to low 20 dB.

Reliability Factors:

1.) FIXATION LOSSES:
The Humphrey field analyzer periodically checks the patient's fixation by presenting stimuli within their blind spot (Heijl-Krakau Technique). When the number of fixation losses is greater than 20%, a symbol (XX) will appear next to the fixation losses to alert the doctor there is reason for concern.

 

2.) FALSE NEGATIVE ERRORS:

A brighter stimulus is presented at a test point in the field that was earlier reported, as being seen, having "Normal Sensitivity" but now the patient does not respond to the bright stimulus. High, false negative scores might indicate fatigue or inattentive patient.

 

3.) FALSE POSITIVE ERRORS:

The projector makes a noise when it moves and the patient responds to the sound though no stimulus has been presented. The patient is responding to outside factors or trying to out guess when the stimulus will be presented. A high false positive score indicates that the patient is "Trigger Happy.

Interpretation

1.) FLUCTUATION FACTORS:

Short-Term: Relative normality is less than 3dB when the same point is tested twice and is usually between 1dB and 2dB during a given test period. There are two reasons for an abnormal Short-Term Fluctuation (SF), inattentive patient or a patient with a diseased visual system.

 

LOW FLUCTUATION: < 1.5 dB

NORMAL FLUCTUATION 1.5dB TO 2 dB

MEDIUM FLUCTUATION >2 dB BUT < 3 dB

HIGH FLUCTUATION >3 dB

 

2.) Grayscale - Is for the patients benefit; for their interpretation or understanding. Represents tested points and non-tested intermediate points, which have been assigned values, interpolated from surrounding points. It tells the doctor nothing about the depth of a scotoma.

 

3.) Total Deviation - These numeric values represents the difference in decibels between the patient's test results and the expected age-corrected normal values at each test point in the visual field. The plot just below this finding are graytone (symbols) which shows the statistical significance for a given test value. These are based on the deviation from expected normal patient threshold profiles. The darker the pattern (symbol) the more significant the deviation from the expected threshold

 

4.) Pattern Deviation - This plot is similar to the Total Deviation except the STATPAC attempts to adjust the analysis of the test results for any overall changes in the height of the measured hill of vision caused by say cloudy media, cataracts or small pupils. Hence, this numeric pattern deviation plot shows the deviation in decibels from the age-corrected normal values, adjusted for any shifts in overall sensitivity. The plot just below this finding are again graytone (symbols) which show the statistical significance of the results at each point. The darker the pattern (symbol) the more significant the deviation from the expected threshold.

 

5.) Probability of Abnormality - The P value represents the probability whereto a patients findings have deviated from the expected normal values. The probability statements is based on the Hill Of Vision distribution seen in the normal population. This P value is computed from the total deviation and the pattern deviation plots. P<1% means that this deviation happens in less than 1% of the normal population and must be consider highly suspicious.

GLOBAL INDICES

1.) Mean Deviation or Defect (MD) - The (MD) is the mean difference in decibels between the "normal" expected hill of vision and the patient's hill of vision. if the deviation is significantly outside the norms, a P value will be given. Example: P< 0.5% means that less than 0.5% of the normal population showed a (MD) larger than the value found for this test. This index is a measure of overall depression, elevation of the field or significantly deep losses in one part of the field and not in others.

 

2.) Pattern Standard Deviation (PSD) - This is a measurement of the degree which the shape of the patient's measured field or hill of vision departs from the "NORMAL" age-corrected reference field model. The value is expressed in decibels and any value of 2dB or greater will have a (P) value next to it indicating the significance of the deviation.

 

3.) Short-Term Fluctuation (SF) - This is what the Field Analyzer has been testing all along. It is simply an index of the consistency of the patients responses during the field testing. This value is obtained when ten (10) pre-selected points are tested twice and the difference, in decibels, of the patient's responses are compared.

 

4.) Corrected Pattern Standard Deviation (CPSD) This is a calculated measurement in decibels of how much the total shape of the patient's hill of vision deviates from the shape of the "NORMAL" hill of vision for the patient's age, after being corrected for intra-test variability. In calculating the (CPSD) the STATPAC attempts to determine if the irregularities in the hill of vision are real by removing the short-term fluctuation (SF), which may mask a relative scotoma.

SUMMARY:
There are many different test types that you should know how to run, i.e., the 30 - 2 (thirty degree field) threshold test has been used extensively. However, it has become common knowledge that the 24 - 2 (twenty-four degree field) threshold test is faster, easier on the elderly patient and hence, the findings have been found to be more sensitive. The 120 point "Three Zone" Screener test each point 6dB brighter than the expected threshold, and points missed are retested at 10,000 asb. A point seen on retesting is recorded as a relative defect; if it is not seen, it is recorded as an absolute defect. Therefore, it tells you if the findings are within 6dB of the expected but also shows if a scotoma is relative or absolute in nature. There are an extensive number of different test that can be run and each should be chosen based on the information you are trying to obtain. The perimeter allows in depth statistical analysis, however, the interpretation and diagnosis still remains the domain of the clinician.

There are similar field tests like the 30-1, 24-1, 30/60-1. What are they and how do you use them? The 30-1 and 30-2 central threshold test point patterns are interlocking, as are the 24-1 and 24-2 central test patterns. Separately each pattern tests points every 6 degrees, but when test results are merged, using the Field Analyzer's merge feature, the effective grid density is 4.2 degrees. In like manner, the 30/60-1 and 30/60-2 peripheral test patterns each have a grid resolution of 12 degrees, but when merged the effective grid density is 8.4 degree resolution. The 30-1, 24-1, and 30/60-1 threshold tests include points on the horizontal and vertical meridians while the 30-2, 24-2, and 30/60-2 do not.
So should we be running these tests in combination with their interlocking field test. The following is a conversation I had with one of our former faculty members:

To: John:
The Humphrey Vision Analyzer's 24-2 field test points located 6 degrees apart and a 24-1 test points that fall every 6 degrees but at different locations. When the two test are combined or merged (over lapped forming a new single field) you get an effective grid density of 4.2 degrees. Would not the merged results be more accurate than the 24-2 alone? The same thing for a 30-1 and a 30-2; I know it would take longer to run both and might degrade the results. Do you know of any studies which indicate the 24-2 or 30-2 alone give just as accurate results?--Riley

To: Riley:
Good question. It would truly be more information, yet, more isn't always better in terms of fatigue factor. I can't site any study but it is interesting to think of this in light of what we know about increasing the accuracy of the test which is done by bringing the patient back two or three times until accuracy is established by looking at the mean deviation (two readings within 2 dB considered fairly consistent) (although I prefer less than 1 dB between visits). This probably doesn't answer your question directly but its an esoteric stab at it anyway!--John

The following are two test results: a 30 - 2 threshold test on a fifty seven (57)-year-old with suspect of glaucoma and a 24 - 2 threshold test on a twenty four (24)-year-old uncorrected 6 diopter myope which illustrates the importance of the pattern deviation.

CENTRAL 30 - 2 THRESHOLD TEST
NAME:   DIGGER, O'DELL			BIRTHDATE:  03-17-33   DATE:   07-17-90
STIM. III WHITE			Rx +5.00-1.25 X105	PUPIL 6mm	VA 20/20
LEFT
AGE	57
QUESTIONS ASKED  544
FIXATION LOSSES  3/28
FALSE POS ERRORS  1/2
FALSE NEG ERRORS  2/12
TEST TIME  00:16:25

TOTAL								PATTERN 
DEVIATION							DEVIATION

The first thing you should look at and understand is how the total deviation was derived. The patient's responses are on the upper left hand corner and the expected values for this patient of 57 years are to the right of it.

TOTAL DEVIATION = (PATIENT'S RESPONSE) - (EXPECTED NORMAL)
These are both represented in decibels and the difference between the two result in either a positive or negative decibel value. Positive means the patient performed above the expected and a negative means they performed that number of decibels below the expected.

PATTERN DEVIATION = (TOTAL DEVIATION) + (OVERALL SENSITIVITY CHANGES)
In this case the overall sensitivity was found to have a value of (2 dB) above the expected. You will find in most cases each corresponding point of the pattern deviation has changed from the total deviation by 2 decibels.

The Graytone Symbols just below each of the above mentioned shows two areas of concern. The upper field related part of each can be explained by the patient's upper lid and lashes. The Graytone Symbols associated with the Pattern Deviation shows a somewhat better picture. However, the inferior nasal field has a most definite arcuate fiber field defect. This is where objective patient related data becomes very important. Dose the patient have a cataract that could produce the defect? Dose the patient have an enlarged or notched optic nerve head? Does the patient have asymmetrical nerve heads and/or are the IOP's high?

YOU MUST LOOK AT THE FOLLOWING TOO:

FIXATION LOSSES OK

FALSE POS. ERRORS OK

FALSE NEG. ERROR OK
GLOBAL INDICES

MD QUESTIONABLE

PSD QUESTIONABLE

SF QUESTIONABLE

CPSD QUESTIONABLE

The above mentioned objective findings are going to be very important in making a diagnosis and determining how to manage this patient.

In the next case (24 - 2 threshold test) go through the same procedure. You should notice that there are a large number of centrally located Total Deviation values of (-5) or greater and would normally be of concern. However, when you look at the Pattern Deviation you find the analysis has picked up enough overall change in sensitivity that only one point remains. This becomes very obvious when you look at the Graytone Symbol patterns, the lower right is totally clear. Remember this was an uncorrected six diopter myopic patient who took this test. You should note the difference in the time it took to run this central 24 degree full threshold field and that of the central 30 degree full threshold field.

                  CENTRAL  24  -  2  THRESHOLD TEST
NAME: DIGGER, O'DELL		BIRTHDATE:  10-27-64	DATE:  01-30-91
STIM.  III WHITE		FIXATION TARGET  CENTRAL
STRATEGY  FULL THRESHOLD		Rx USED 0.00 	PUPIL DIAMETER 4mm
						REAL Rx -6.00D
            
RIGHT
AGE  24
QUESTIONS ASKED 313
FIXATION LOSSES 1/18
FALSE POS ERRORS 0/3
FALSE NEG ERRORS 0/7
TEST TIME 00:08:17

	TOTAL						PATTERN 		GLOBAL INDICES
	DEVIATION					DEVIATION		MD    -4.21 DB P<2%
										                             PSD    1.67 DB
										                             SF    1.43 DB
										                             CPSD  0.67 DB
            

Centralcecal scotoma involving the macula and blind spot

The two previous graphics are representations of the hill of vision and how the visual fields are changed as the result of damage to the retina or hill of vision.

Additional Information Regarding the Module 750. The instrument is smaller and more compacted than the 630 and has a touch sensitive screen or key board for inputting patient information and navigating around the different fields and data base. Patient information like record number, date of birth, patient's distant acuities, distant prescription, pupil size, etc., then the instrument calculates the needed prescription to run the selected test. Pretty cool interface, but does not have a 24-1, 30-1, 30/60-1 or 30/60-2 test. There is a 60-4 test which appears to have taken the 30/60 series place and I have not received information back telling what the 60-4 or more correctly what the (-4) part of the test stands indicates. Looking at the points tested with the 60-4 it would appear it test out to 60 degrees but not the central 30 degrees which is the same testing for which the 30/60 was designed.

The school received the new Short Wavelength Automated Perimetry (SWAP) program update in January of 1997. This Statpack contains the following: background of the bowl will be yellow illuminated (500 nm) the target size is a Goldmann (V) which is 64mm2 or 1.8 degrees in size. The target color is blue/violet 440 nm. Research using these parameters indicates visual field defects, in ocular hypertensives, are detectable three to five years sooner than the traditional white on white perimetry presently being used. It is interesting, according to Dr. Jane Grogg, the testing takes about 6 to 7 minutes longer to run and in patients with known reduced retinal sensitivity the SWAP really does not seem to give any more diagnostic information. Dr. Grogg points out the increased testing time adds to the patients fatigue.

Dr. Grogg likes a new program which is on our Humphrey's Visual Field unit and being used in the Ocular Disease Clinic is the Swedish Interactive Threshold Algorithm or (SITA). It is much faster than either the normal 24-2 or 30-2 full threshold fields or the SWAP. The one drawback she finds is it presently does not have a statpack which gives the mean deviation (MD) etc. information. The SITA program is totally new to me but the time for a normal full threshold fields is commonly 14 to 16 plus minutes per eye while the SITA is more like 6 to 7 minutes per eye which cuts down on patient fatigue and improves the reliability.

Both the SWAP and SITA programs run the typical 24-2 and 30-2 fields and the results are in decibels. Therefore, interpreting the results is the same as the standard Humphrey's Visual Fields.