- Email: [email protected]
Computerized Diagnostic Instruments for Ophthalmic Practice
Computerized Diagnostic Instruments for Ophthalmic Practice DAVID L. GUYTON, MD
Abstract: The microprocessor, the so-called "computer-on-a-chip," is providing automation of many of our diagnostic instruments, including refractors, keratometers, lensmeters, and perimeters. Microprocessors can monitor switches, control lights, drive motors, and perform complex mathematical calculations in a fraction of a second. These capabilities promise a standardization of measurement never before possible. Clinical tests will soon not only use instruments controlled by computers, but the tests themselves will be administered by computers. Cost benefit ratios are decreasing as this new technology becomes an expected part of ophthalmic practice. [Key words: automation, computer, instrument, lensmeter, microprocessor, perimeter, programmed, refractor.]
Nowhere is it more evident than in the exhibit hall of the Annual Meeting of the American Academy of Ophthalmology that computers have entered ophthalmic practice. They are not only on our bookkeeper's desk, but they are also hidden away inside our diagnostic instruments. They operate automated refractors, automated keratometers, automated lensmeters, and automated perimeters. It is the tiny microprocessor, the so-called "computeron-a-chip," which makes automation of our instruments possible (see Fig 1). This marvelous device is only eleven years old. It typically contains tens of thousands oftransistors, and yet is so small that it has to be packaged in a protective case so that it can be handled and connected properly. Microprocessors can respond to button pushes, can tum lights on and off, can drive motors, and can perform complex mathematical calculations in a fraction of a second. It is no wonder that microprocessors now serve as fast, efficient, and accurate brains for many of our instruments. Concentrating on three automated instruments will illustrate the capabilities that microprocessors provide. These three instruments are the Perimetron, an autoFrom the Wilmer Ophthalmological Institute, Johns Hopkins University School of Medicine, Baltimore, Maryland. Presented at the Eighty-seventh Annual Meeting of the American Academy of Ophthalmology, San Francisco, California, October 3Q-November
5, 1982.
115
mated perimeter from Coherent Medical, the Lens Analyzer, an automated lensmeter from Humphrey Instruments, and the SR-IV Programmed Subjective Refractor from AO. First, the Perimetron.
PERIMETRON The Perimetron I (Fig 2) was first announced in 1975, currently costs about $29,000, and is designed to perform either kinetic or static visual field tests with the help of an operator. The operator seats the patient, aligns the patient's eye with the instrument, activates a chosen testing program, and prompts 'the patient during the test to ensure valid observations. The results are plotted out with an X-Y plotter. An important feature of the Perimetron, which is made possible by the microprocessor, is its self-calibration. The brightness of the bulb is read with a photocell, and the voltage is automatically adjusted to give the standard background illumination. The alignment of the projector is checked automatically by projection of the test spot toward a photodetector located at a known reference spot in the visual field. If minimal misalignment has occured by moving the instrument from one room to another, the microprocessor will re-zero the instrument electronically to compensate for this small misalignment. The brightness of the test spot itself is also calibrated automatically by the built-in program.
116
OPHTHALMOLOGY •
INSTRUMENT AND BOOK SUPPLEMENT 1983
Fig I. A typical microprocessor, the INTEL 8086. The actual chip, containing tens of thousands of transistors, is only one-eighth as long as the rectangular package. This particular microprocessor is the "brain" of the Humphrey Lens Analyzer, Model 322.
All of these calibration steps, which are performed each time before the patient is tested, requiring only seconds, are simplified tremendously by the microprocessor. The cycle time for the Perimetron is two milliseconds. That is, each of the control switches is scanned by the microprocessor 500 times each second, with the position of each switch being read with each scan. This provides essentially instantaneous response to the input switches. Also, the output to the plotter is revised every two milliseconds, increasing the accuracy of the plot. The test spot is moved about, and the X-Y plotter is operated, by stepping motors. With each pulse received from the microprocessor, the stepping motor advances one step, usually a few degrees of rotation of the shaft. By delivering single pulses or series of pulses, the microprocessor can thereby actuate mechanical movement with very precise digital control. Stepping motors are obviously ideal for obtaining accurate control of moving parts by microprocessors.
Alt~ough each switch is scanned 500 times a second, the mIcroprocessor may be programmed not to respond to certain switches during the running of given p~o gra~s. Fo~ example, during a static perimetry test, the vanous sWitches that set the Perimetron for the kinetic scans will simply not function. This is an example of logic being programmed into the instrument, logic which is implemented by the microprocessor. Various programs are supplied with thePerimetron, and are selected by switches on the control panel. For example, the number eight program consists of two kinetic scans at the I2e and I4e isopter levels, and includes 72 points tested by static perimetry within the 30° central field. A blind spot routine is also included in this program, with these various tests being performed automatically by the microprocessor. Up until now, the Perimetronhas been the only automated perimeter to attempt kinetic scanning. The scanning procedure has been designed to try to simulate exactly what the operator of the Goldmann perimeter does. After a first isopter point is determined, scanning for an adjacent isopter point is performed along the adjacent radius, starting only 6° further out, avoiding the inefficiency of going all the way to the peripheral field and coming back in. If unusual responses are obtained, as expected with a step in the visual field, for instance, the kinetic scanning automatically shifts by 90° to attempt to pick up the edge of the visual field step. Such flexibility would be impossible without the microprocessor control. ' The displayed isopter levels may be expressed either in decibels, Goldmann perimeter units, or in apostilbs. The conversion factors for these different units are stored within the Perimetron's memory, and the microprocessor can simply calculate one unit from the next, a capability which is totally impractical for hard-wired instruments. An interesting use of the microprocessor is in the output of the Perimetron. The plotting of specific isopters must be done with specifically colored pens. When the plot requires a certain color, the X-V pen carriage is driven automatically by the microprocessor to lie adjacent to the well containing that colored pen. The operator plac~s the pen in the X -Y carriage, and plotting of that partIcular color begins, with little room for error. The preceding are only some of the Perimetron's features which are made possible by its microprocessor control. Next, the Humphrey Lens Analyzer.
HUMPHREY LENS ANALYZER
Fig 2. The Perimetron, a microprocessor-based automated perimeter designed to perform either static or kinetic perimetry.
The Humphrey Lens Analyzer (Fig 3) is available in several models, with the newest model being #322. All of the Humphrey Lens Analyzers are designed to give a~tomatic reading of sphere, cylinder, axis, and prism, With the operator merely having to position the spectacle lens or contact lens properly.
GUYTON • COMPUTERIZED DIAGNOSTIC INSTRUMENTS
The Lens Analyzer uses a simple but accurate method of determining the refractive properties of a lens by aiming very fine bundles of rays through the lens and measuring how the bundles of rays are deviated by the lens. Tedious calculations would be necessary if done by hand, and these have prevented this ray-trace method from being used in the past. With previous lensmeters, an optical focus endpoint has been used. Now that the microprocessor can handle these tedious calculations easily, the high accuracy of the ray-trace method of measuring lenses can be used. The automated lens power measurement with the Lens Analyzer is the single most important feature, for this avoids the common human errors when reading scales and when transcribing refractions from minus cylinder form to plus cylinder form and vice versa. The mathematical errors of spectacle measurement are thereby eliminated with the Lens Analyzer's microprocessor. The cycle time of the Humphrey Lens Analyzer is 0.1 second. That is, the readings are made, calculations are performed, and the output is updated 10 times each second. This is clearly made possible only by the unique capabilities of microprocessor control. Initial centering of the spectacle lens in the Lens Analyzer is facilitated by calculations made automatically by the microprocessor to display an alignment pattern which always moves to the right when the lens moves to the right, moves up when the lens moves up, and so forth. In other words, plus and minus lenses no longer give opposite direction of movement of the alignment pattern, and lowered-powered lenses decenter the alignment pattern as much as high-powered lenses, achieving more uniform centering across all spectacle lens powers. Two special detection features are built into the Model 322 Lens Analyzer. The first of these is called "wave detection." As the spectacle lens is moved about over the measuring nosepiece, the change in power is continually read, and if the power changes abruptly, the display activates, indicating that significant distortion exists in the lens. The second detection feature is the "non-toric" feature. If significant irregularity exists at anyone spot through a measured lens, the non-toric indicator illuminates to warn the operator of this problem. These detection functions are made possible by the use of the microprocessor.
117
Fig 3. The Humphrey Lens Analyzer, Model 322, provides fully automatic reading of sphere, cylinder, axis, and prism.
For routine refractions and for beginning operators, practically the entire refraction can be performed by successive pushes of a single REFRACT button. The microprocessor, in response to this button, steps the in-
SR-IV PROGRAMMED SUBJECfIVE REFRACfOR Let us now consider the SR-IV Programmed Subjective Refractor from A0 3 (Fig 4), an instrument in which I have a proprietary interest. The SR-IV is an automated subjective refractor designed to be operated by a technician to obtain endpoint subjective refractions. Microprocessor control within the SR-IV allows an "auto-sequence" to be programmed into the instrument.
Fig 4. The SR-IV Programmed Subjective Refractor, a microprocessor-based instrument for automated subjective refraction.
118
OPHTHALMOLOGY •
INSTRUMENT AND BOOK SUPPLEMENT 1983
Fig 5. Marg's computer-controlled refractor, one of the first diagnostic instruments to actually be administered by a computer.
strument to the next refraction step, presenting the appropriate target and activating the patient's knob for the appropriate adjustment, either sphere, cylinder, cross cylinder, or axis. Notice that cross cylinder adjustment is one of the modes of operation of the SR-IV. Cross cylinder adjustment simply means that when the cylinder power changes by two steps, the sphere changes one step in the opposite direction. This maintains the spherical equivalent of the refraction constant, with the microprocessor calculating and updating the sphere and cylinder values as power is changed. The cross cylinder adjustment feature is particularly valuable during cross cylinder testing of cylinder power, when the spherical equivalent should remain constant for proper cross cylinder accuracy. There is a considerable amount of refraction logic built into the SR-IV's program. For instance, a red-green target is included as the final step in the refracting sequence. Neither cylinder power nor cylinder axis adjustment is allowed by the microprocessor when the redgreen target is in place, for the red-green target should never be used for refinement of cylinder power or axis. Another example of built-in logic is the fact that the SR-IV program will not allow the astigmatic dial test to be used if cylinder power previously exists in the optics. This caused some consternation to one owner of the instrument, who suggested that the instrument be "improved" to allow astigmatic dial testing beginning with the old refraction. The microprocessor was smarter than the ophthalmologist in this case, for the astigmatic dial determination of axis is lotally erroneous in the presence of pre-existing correcting cylinder. During cross cylinder refinement of cylinder axis, it is well known that large axis steps should be taken with small cylinders and small axis steps should be taken with large cylinders. This is programmed into the SR-IV's microprocessor with axis step sizes of 10 to 10 0 being taken depending on the amount of cylinder power present in the SR-IV at the time. Thus when the operator
signals a change of axis, the microprocessor looks at the current value of the cylinder in the instrument and makes a decision regarding how much axis change will result in a quarter diopter worth of blur change. This variable axis stepping feature of the SR-IV, controlled by the microprocessor, thus provides a uniform cross cylinder refinement of cylinder axis never before achieved. . Store and recall capability are built into the SR-IV, with the microprocessor providing storage of one refraction, to be recalled at a later time, such that visual acuity can be compared quickly with two entirely different refractions. When the RECALL button is pushed, the refraction currently in the optics and in the display is "flip-flopped" with the refraction in the stored memory. Only with microprocessor control is such manipulation easily achieved. An optional Data Printer is available for the SR-IV which does more than print the values of the refractions obtained. The Data Printer contains its own microprocessor. Visual acuities without correction, with the old glasses, and with the new refraction are all stored and printed out. Spherical equivalent is calculated by the microprocessor and printed out along with the refractive findings. The refraction can also be converted from the spectacle plane to the contact lens plane simply by flipping the Contact Lens switch. When this switch is in its flipped position, the refraction is actually still measured in the spectacle plane, but all values of sphere power and cylinder power are converted mathematically by the microprocessor to zero vertex distance. The most important feature of the Data Printer, though, is the overrefraction calculation capability. If the prescription of the old glasses is entered into the SRIV, and the patient is refracted over the old glasses, the trigonometric sum of the old glasses plus the overrefraction is automatically calculated, within half a second, by the microprocessor in the Data Printer, and printed out in an easy-to-read form.
CONCLUSION These are examples of capabilities of some of our new automated instruments which are made possible by the microprocessor, the single device which is most responsible for the proliferation of new instruments on our exhibit floors in recent years. One further use of computers in diagnostic instruments should be mentioned. In 1977, Elwin Marg devised a computer-controlled refractor4 (Fig 5). This instrument was remote controlled, operated by a large computer sitting against the wall. It was not miniaturized, and was not commercially successful because of the expense involved, but this instrument was different from all of the previous and subsequent automated refracting instruments. The computer actually administered the test. With a series of tape-recorded messages,
GUYTON •
COMPUTERIZED DIAGNOSTIC INSTRUMENTS
the computer would communicate with the patient, and the patient would communicate with the computer with an array of buttons which he held on his lap. The patient would be seated behind the refractor by an assistant, and then left alone in the room with the computer, to be refracted. The binocular subjective refraction for both distance and near required approximately 20 to 25 minutes, with very accurate results. This was one of the first applications of a computer in actually administering a diagnostic test. There are always problems in making the computer as flexible as a human operator. But then again the computer is not late to work, does not expect vacation time, and best of all, does not ask for a raise. As the psychology of machine-patient interaction is better understood, we will undoubtedly see more and more of these computer-ad-
119
ministered tests, representing the next generation of computer-controlled equipment. The microprocessor revolution in diagnostic instruments is truly underway.
REFERENCES 1. Perimetron Operator's Manual. Coherent Medical Division, Palo Alto, California, 1982. 2. The Model 322 Laboratory Lens Analyzer Manual. Humphrey Instruments Incorporated. San Leandro, California, 1982. 3. Guyton DL. The American Optical SR-IV Programmed Subjective Refractor: Principles of design and operation. Am J Optom Physiol Opt 1982; 59:800-814. 4. Marg E. Computer-assisted eye examination: Background and prospects. San Francisco: San Francisco Press, 1980.
Abstract: The microprocessor, the so-called "computer-on-a-chip," is providing automation of many of our diagnostic instruments, including refractors, keratometers, lensmeters, and perimeters. Microprocessors can monitor switches, control lights, drive motors, and perform complex mathematical calculations in a fraction of a second. These capabilities promise a standardization of measurement never before possible. Clinical tests will soon not only use instruments controlled by computers, but the tests themselves will be administered by computers. Cost benefit ratios are decreasing as this new technology becomes an expected part of ophthalmic practice. [Key words: automation, computer, instrument, lensmeter, microprocessor, perimeter, programmed, refractor.]
Nowhere is it more evident than in the exhibit hall of the Annual Meeting of the American Academy of Ophthalmology that computers have entered ophthalmic practice. They are not only on our bookkeeper's desk, but they are also hidden away inside our diagnostic instruments. They operate automated refractors, automated keratometers, automated lensmeters, and automated perimeters. It is the tiny microprocessor, the so-called "computeron-a-chip," which makes automation of our instruments possible (see Fig 1). This marvelous device is only eleven years old. It typically contains tens of thousands oftransistors, and yet is so small that it has to be packaged in a protective case so that it can be handled and connected properly. Microprocessors can respond to button pushes, can tum lights on and off, can drive motors, and can perform complex mathematical calculations in a fraction of a second. It is no wonder that microprocessors now serve as fast, efficient, and accurate brains for many of our instruments. Concentrating on three automated instruments will illustrate the capabilities that microprocessors provide. These three instruments are the Perimetron, an autoFrom the Wilmer Ophthalmological Institute, Johns Hopkins University School of Medicine, Baltimore, Maryland. Presented at the Eighty-seventh Annual Meeting of the American Academy of Ophthalmology, San Francisco, California, October 3Q-November
5, 1982.
115
mated perimeter from Coherent Medical, the Lens Analyzer, an automated lensmeter from Humphrey Instruments, and the SR-IV Programmed Subjective Refractor from AO. First, the Perimetron.
PERIMETRON The Perimetron I (Fig 2) was first announced in 1975, currently costs about $29,000, and is designed to perform either kinetic or static visual field tests with the help of an operator. The operator seats the patient, aligns the patient's eye with the instrument, activates a chosen testing program, and prompts 'the patient during the test to ensure valid observations. The results are plotted out with an X-Y plotter. An important feature of the Perimetron, which is made possible by the microprocessor, is its self-calibration. The brightness of the bulb is read with a photocell, and the voltage is automatically adjusted to give the standard background illumination. The alignment of the projector is checked automatically by projection of the test spot toward a photodetector located at a known reference spot in the visual field. If minimal misalignment has occured by moving the instrument from one room to another, the microprocessor will re-zero the instrument electronically to compensate for this small misalignment. The brightness of the test spot itself is also calibrated automatically by the built-in program.
116
OPHTHALMOLOGY •
INSTRUMENT AND BOOK SUPPLEMENT 1983
Fig I. A typical microprocessor, the INTEL 8086. The actual chip, containing tens of thousands of transistors, is only one-eighth as long as the rectangular package. This particular microprocessor is the "brain" of the Humphrey Lens Analyzer, Model 322.
All of these calibration steps, which are performed each time before the patient is tested, requiring only seconds, are simplified tremendously by the microprocessor. The cycle time for the Perimetron is two milliseconds. That is, each of the control switches is scanned by the microprocessor 500 times each second, with the position of each switch being read with each scan. This provides essentially instantaneous response to the input switches. Also, the output to the plotter is revised every two milliseconds, increasing the accuracy of the plot. The test spot is moved about, and the X-Y plotter is operated, by stepping motors. With each pulse received from the microprocessor, the stepping motor advances one step, usually a few degrees of rotation of the shaft. By delivering single pulses or series of pulses, the microprocessor can thereby actuate mechanical movement with very precise digital control. Stepping motors are obviously ideal for obtaining accurate control of moving parts by microprocessors.
Alt~ough each switch is scanned 500 times a second, the mIcroprocessor may be programmed not to respond to certain switches during the running of given p~o gra~s. Fo~ example, during a static perimetry test, the vanous sWitches that set the Perimetron for the kinetic scans will simply not function. This is an example of logic being programmed into the instrument, logic which is implemented by the microprocessor. Various programs are supplied with thePerimetron, and are selected by switches on the control panel. For example, the number eight program consists of two kinetic scans at the I2e and I4e isopter levels, and includes 72 points tested by static perimetry within the 30° central field. A blind spot routine is also included in this program, with these various tests being performed automatically by the microprocessor. Up until now, the Perimetronhas been the only automated perimeter to attempt kinetic scanning. The scanning procedure has been designed to try to simulate exactly what the operator of the Goldmann perimeter does. After a first isopter point is determined, scanning for an adjacent isopter point is performed along the adjacent radius, starting only 6° further out, avoiding the inefficiency of going all the way to the peripheral field and coming back in. If unusual responses are obtained, as expected with a step in the visual field, for instance, the kinetic scanning automatically shifts by 90° to attempt to pick up the edge of the visual field step. Such flexibility would be impossible without the microprocessor control. ' The displayed isopter levels may be expressed either in decibels, Goldmann perimeter units, or in apostilbs. The conversion factors for these different units are stored within the Perimetron's memory, and the microprocessor can simply calculate one unit from the next, a capability which is totally impractical for hard-wired instruments. An interesting use of the microprocessor is in the output of the Perimetron. The plotting of specific isopters must be done with specifically colored pens. When the plot requires a certain color, the X-V pen carriage is driven automatically by the microprocessor to lie adjacent to the well containing that colored pen. The operator plac~s the pen in the X -Y carriage, and plotting of that partIcular color begins, with little room for error. The preceding are only some of the Perimetron's features which are made possible by its microprocessor control. Next, the Humphrey Lens Analyzer.
HUMPHREY LENS ANALYZER
Fig 2. The Perimetron, a microprocessor-based automated perimeter designed to perform either static or kinetic perimetry.
The Humphrey Lens Analyzer (Fig 3) is available in several models, with the newest model being #322. All of the Humphrey Lens Analyzers are designed to give a~tomatic reading of sphere, cylinder, axis, and prism, With the operator merely having to position the spectacle lens or contact lens properly.
GUYTON • COMPUTERIZED DIAGNOSTIC INSTRUMENTS
The Lens Analyzer uses a simple but accurate method of determining the refractive properties of a lens by aiming very fine bundles of rays through the lens and measuring how the bundles of rays are deviated by the lens. Tedious calculations would be necessary if done by hand, and these have prevented this ray-trace method from being used in the past. With previous lensmeters, an optical focus endpoint has been used. Now that the microprocessor can handle these tedious calculations easily, the high accuracy of the ray-trace method of measuring lenses can be used. The automated lens power measurement with the Lens Analyzer is the single most important feature, for this avoids the common human errors when reading scales and when transcribing refractions from minus cylinder form to plus cylinder form and vice versa. The mathematical errors of spectacle measurement are thereby eliminated with the Lens Analyzer's microprocessor. The cycle time of the Humphrey Lens Analyzer is 0.1 second. That is, the readings are made, calculations are performed, and the output is updated 10 times each second. This is clearly made possible only by the unique capabilities of microprocessor control. Initial centering of the spectacle lens in the Lens Analyzer is facilitated by calculations made automatically by the microprocessor to display an alignment pattern which always moves to the right when the lens moves to the right, moves up when the lens moves up, and so forth. In other words, plus and minus lenses no longer give opposite direction of movement of the alignment pattern, and lowered-powered lenses decenter the alignment pattern as much as high-powered lenses, achieving more uniform centering across all spectacle lens powers. Two special detection features are built into the Model 322 Lens Analyzer. The first of these is called "wave detection." As the spectacle lens is moved about over the measuring nosepiece, the change in power is continually read, and if the power changes abruptly, the display activates, indicating that significant distortion exists in the lens. The second detection feature is the "non-toric" feature. If significant irregularity exists at anyone spot through a measured lens, the non-toric indicator illuminates to warn the operator of this problem. These detection functions are made possible by the use of the microprocessor.
117
Fig 3. The Humphrey Lens Analyzer, Model 322, provides fully automatic reading of sphere, cylinder, axis, and prism.
For routine refractions and for beginning operators, practically the entire refraction can be performed by successive pushes of a single REFRACT button. The microprocessor, in response to this button, steps the in-
SR-IV PROGRAMMED SUBJECfIVE REFRACfOR Let us now consider the SR-IV Programmed Subjective Refractor from A0 3 (Fig 4), an instrument in which I have a proprietary interest. The SR-IV is an automated subjective refractor designed to be operated by a technician to obtain endpoint subjective refractions. Microprocessor control within the SR-IV allows an "auto-sequence" to be programmed into the instrument.
Fig 4. The SR-IV Programmed Subjective Refractor, a microprocessor-based instrument for automated subjective refraction.
118
OPHTHALMOLOGY •
INSTRUMENT AND BOOK SUPPLEMENT 1983
Fig 5. Marg's computer-controlled refractor, one of the first diagnostic instruments to actually be administered by a computer.
strument to the next refraction step, presenting the appropriate target and activating the patient's knob for the appropriate adjustment, either sphere, cylinder, cross cylinder, or axis. Notice that cross cylinder adjustment is one of the modes of operation of the SR-IV. Cross cylinder adjustment simply means that when the cylinder power changes by two steps, the sphere changes one step in the opposite direction. This maintains the spherical equivalent of the refraction constant, with the microprocessor calculating and updating the sphere and cylinder values as power is changed. The cross cylinder adjustment feature is particularly valuable during cross cylinder testing of cylinder power, when the spherical equivalent should remain constant for proper cross cylinder accuracy. There is a considerable amount of refraction logic built into the SR-IV's program. For instance, a red-green target is included as the final step in the refracting sequence. Neither cylinder power nor cylinder axis adjustment is allowed by the microprocessor when the redgreen target is in place, for the red-green target should never be used for refinement of cylinder power or axis. Another example of built-in logic is the fact that the SR-IV program will not allow the astigmatic dial test to be used if cylinder power previously exists in the optics. This caused some consternation to one owner of the instrument, who suggested that the instrument be "improved" to allow astigmatic dial testing beginning with the old refraction. The microprocessor was smarter than the ophthalmologist in this case, for the astigmatic dial determination of axis is lotally erroneous in the presence of pre-existing correcting cylinder. During cross cylinder refinement of cylinder axis, it is well known that large axis steps should be taken with small cylinders and small axis steps should be taken with large cylinders. This is programmed into the SR-IV's microprocessor with axis step sizes of 10 to 10 0 being taken depending on the amount of cylinder power present in the SR-IV at the time. Thus when the operator
signals a change of axis, the microprocessor looks at the current value of the cylinder in the instrument and makes a decision regarding how much axis change will result in a quarter diopter worth of blur change. This variable axis stepping feature of the SR-IV, controlled by the microprocessor, thus provides a uniform cross cylinder refinement of cylinder axis never before achieved. . Store and recall capability are built into the SR-IV, with the microprocessor providing storage of one refraction, to be recalled at a later time, such that visual acuity can be compared quickly with two entirely different refractions. When the RECALL button is pushed, the refraction currently in the optics and in the display is "flip-flopped" with the refraction in the stored memory. Only with microprocessor control is such manipulation easily achieved. An optional Data Printer is available for the SR-IV which does more than print the values of the refractions obtained. The Data Printer contains its own microprocessor. Visual acuities without correction, with the old glasses, and with the new refraction are all stored and printed out. Spherical equivalent is calculated by the microprocessor and printed out along with the refractive findings. The refraction can also be converted from the spectacle plane to the contact lens plane simply by flipping the Contact Lens switch. When this switch is in its flipped position, the refraction is actually still measured in the spectacle plane, but all values of sphere power and cylinder power are converted mathematically by the microprocessor to zero vertex distance. The most important feature of the Data Printer, though, is the overrefraction calculation capability. If the prescription of the old glasses is entered into the SRIV, and the patient is refracted over the old glasses, the trigonometric sum of the old glasses plus the overrefraction is automatically calculated, within half a second, by the microprocessor in the Data Printer, and printed out in an easy-to-read form.
CONCLUSION These are examples of capabilities of some of our new automated instruments which are made possible by the microprocessor, the single device which is most responsible for the proliferation of new instruments on our exhibit floors in recent years. One further use of computers in diagnostic instruments should be mentioned. In 1977, Elwin Marg devised a computer-controlled refractor4 (Fig 5). This instrument was remote controlled, operated by a large computer sitting against the wall. It was not miniaturized, and was not commercially successful because of the expense involved, but this instrument was different from all of the previous and subsequent automated refracting instruments. The computer actually administered the test. With a series of tape-recorded messages,
GUYTON •
COMPUTERIZED DIAGNOSTIC INSTRUMENTS
the computer would communicate with the patient, and the patient would communicate with the computer with an array of buttons which he held on his lap. The patient would be seated behind the refractor by an assistant, and then left alone in the room with the computer, to be refracted. The binocular subjective refraction for both distance and near required approximately 20 to 25 minutes, with very accurate results. This was one of the first applications of a computer in actually administering a diagnostic test. There are always problems in making the computer as flexible as a human operator. But then again the computer is not late to work, does not expect vacation time, and best of all, does not ask for a raise. As the psychology of machine-patient interaction is better understood, we will undoubtedly see more and more of these computer-ad-
119
ministered tests, representing the next generation of computer-controlled equipment. The microprocessor revolution in diagnostic instruments is truly underway.
REFERENCES 1. Perimetron Operator's Manual. Coherent Medical Division, Palo Alto, California, 1982. 2. The Model 322 Laboratory Lens Analyzer Manual. Humphrey Instruments Incorporated. San Leandro, California, 1982. 3. Guyton DL. The American Optical SR-IV Programmed Subjective Refractor: Principles of design and operation. Am J Optom Physiol Opt 1982; 59:800-814. 4. Marg E. Computer-assisted eye examination: Background and prospects. San Francisco: San Francisco Press, 1980.