Microcomputers in microbiology: a matter of special needs.
This set us off toward two goals. In the short term, our lab wanted to acquire a microcomputer system that would furnish some date processing capability. Over the long term, we would work with manufacturers on more thorough computer programs.
Since our mainframe hospital computer is basically an information system--transferring lab results to the physician or the ward--we recognized early on the need for a microbiology-based system to collect and collate data that would otherwise require hours of manual labor to amass.
But precisely what data? And in what form? Over the last three years we have given serious thought to these questions. Our list of what we want a computer to do is endless, but the major areas we've targeted are laboratory management, quality control, identification of unusual results, inventory control, education, and interface with existing instruments.
In the area of lab management, we'd like a computer system that can give us lists of significant findings, epidemiology reports (including flags on nosocomial infections), summaries of antibiotic susceptibility results, workload recording data, and correlations of personnel positions with workload and productivity.
Our objective for QC was a system that would monitor media input and output and store performance testing checks and equipment maintenance logs. As to flagging unusual results, we identified three areas where a computer could save us time identifying unusual microbes, spotting unusual antibiotic susceptibility patterns, and alerting us to the circumstances of several specimens from the same site on a given patient in a single day.
Inventory control is self-explanatory. In the next area, education, we hoped to use the computer to train students and new personnel, as well as for in-services. Finally, we needed a system that would interface with our two major instruments, a Micro-Coder Antibiotic Susceptibility system (Micro-Media) and Bactec (Johnston Laboratories) for detecting positive blood cultures.
The next step in our quest for a computer system was to talk with the experts. We found out that when microbiologists talk with computer specialists, the conversation quickly gets complicated.
Microbiology has major requirements that differ from those of other lab disciplines. Our test results can't be reported in just a few words or numbers: We have a specific vocabulary that is not used in everday conversation.
Moreover, it usually takes a number of days or weeks to complete reports, and additions to specimen data are common. We must also be able to add antimicrobial susceptibility data to various parts of our reports at varying intervals. Because of these complex requirements, computer programs designed for microbiology must have much more flexibility and ability to add comments (free text) than those designed for other lab departments (see Figure I).
It quickly became evident that no commercially available programs would meet all, or even most, of our needs. The solution as we've indicated, was to develop new programs in cooperation with manufacturers. But this would be a long, tedious process.
Thus, we decided to look at some current commercial software in an area that could furnish immediate assistance to us: interfacing with our susceptibility and blood culture instruments.
* Susceptibility testing. No matter what kind of automated susceptibility instrument a lab uses (Figure II), there is probably a computer hookup available. As we investigated these systems, we discovered that they all offered the same basic advantages--for example, automatic transfer of minimum inhibitory concentration data from reader to computer. Another feature common to most systems was microbe identification and probability data, in conjunction with the MIC; alternative choices of microbes; and, if necessary, additional tests for definitive identification. In addition, most systems were able to store and retrieve MIC data for epidemiological purposes in the form of both a listing of microbes and antibiotic resistance patterns by body site or hospital location and a year-to-date summary trend of microbes and antibiotics. They could also produce antibiotic resistance patterns (antibiograms) for blood and urine levels.
To get an idea of how these features work in the microbiology lab, let's look more closely at one of the available MIC microcomputer systems: the MBS-1 TeleVideo (MicroNet, Madison, Wis.). We were interested in this system because it could interface with our Micro-Coder.
The MBS-1 is actually a total microbiology microcomputer system, but we wouldn't have used it that way. It is program-driven, which means it automtically initiates the desired program function based on the information that is entered. This speeds things slightly, because the operator doesn't need to refer to a menu for every action. The system also offers help screens to aid the novice.
As a total microbiology system, the MBS-1 has been designed to store, edit, and report results, including Gram stain findings, MIC data, and identification results. It can print both interim and final reports through use of free text and mnemonic codes. It also assigns specimen and isolate numbers and is able to print incidence reports, summary reports, MIC data, and antibiogram reports.
Since we were interested in interfacing the MBS-1 with our Micro-Coder I, a company representative showed us how to store and retrieve susceptibility results. First we had to enter demographic data into the computer. Next, we put the computer-assigned specimen identification numbers and MIC values into the Micro-Coder I, Finally, a data card was inserted into the Micro-Coder I to facilitate transfer of the data to the computer for storage and printing of reports. (The new Micro-Coder II will make it possible to transfer this data by simply pressing a "send" button.)
We felt that the final susceptibility results for the MBS-1 were good. However, it offered much more capability than we needed. For example, we wouldn't be able to use the system-generated reports because our reports are issued by the hospital's mainframe computer. This drawback could be eliminated by linking the MBS-1 to the hospital computer.
The manufacturer is working on some improvements, such as speeding up data entry and expanding the data base. These changes may make the MBS-1 very appealing as a total microbiology computer system, but we have chosen to wait for one more suitable for our specific needs.
None of the commercial systems that we investigated was flexible enough to meet all the nees of our lab. Some customizing would have been essential.
For example, MIC systems should allow the user to decide on MIC breakpoints for susceptibility and resistance based on the needs of the particular hospital. There is currently no uniform MIC breakpoint for all antibiotics, although the National Committee for Clinical Laboratory Standards is trying to correct this. Until it does, labs face the nuisance of adjusting for breakpoints set by the manufacturer.
Commercial software would also be much more useful if it allowed the user to determine how results will be reported. Again, there is no uniform reporting system, and each laboratory employs the method that is most helpful to the clinicians it serves. The computer system should be able to adjust to the lab's reporting system, rather than the outer way around.
By the same token, it should be up to the lab to decide whether route reporting (interpretation of susceptibility or resistance based on dosage, route of administration, and location of infection) should be printed on the clinician's copy. For now, most computer systems only provide dosage data for adults. This information can cause confusion if it is automatically printed on all report forms, because pediatric dosage differs from dosage for adults.
Another boon would be a more timely way to add new antibiotics to the computer data base. Our laboratory is continually asked to test new antibiotics. Unfortunately, it takes six to 12 months with most commercial systems to put the new drugs on the MIC panel and program the computer to store the data. Thus, most laboratories using commercial MIC systems resort to a backup method to test the newer antimicrobials, and for as long as a year, the new data is neither stored in the computer nor included in antibiograms.
* Blood cultures. We perform more than 12,000 blood cultures each year on our two Bactec 460s. Because a number of our patients are on antibiotics or have subacute bacterial endocarditis, we need a blood culture system, such as the Bactec, that is sensitive, specific, and rapid. Only a few systems will interface with the Bactec. The one that offered the most advantages to our laboratory was the MicrA (Argus, Baltimore).
Among other things, this system stores patients' names, ID numbers, bottle numbers, and test results from both blood culture instruments, so that one technologist can monitor both machines. The computer provides a hard copy of growth index (GI) readings for each bottle during each run. These readings can be attached to requisitions, curbing transcription errors and saving time. The computer also stores and retrieves the identity of positive isolates.
Among the features that we particularly like is the ability to set positive vial criteria, using both threshold GI and change in GI for each test run on each type of culture medium. The sensitivity of the method can be increased by setting a lower threshold during the first few days of sampling, when the background GI is low. Later in the sampling process, when the background GI increases, a higher threshold enhances the specificity of the test.
The MicrA's printout of test results -- including patient name, ID number, bottle number, and all GI readings -- can replace the lab's log book. There are also options in recalling test results, on the basis of the patient's name, ID number, or bottle number.
In addition, the MicrA is a valuable management tool. It prints out monthly analyses on cultures read, positive cultures, microbes isolated, time to detection, and quality control data.
We saw the system demonstrated several times and visited a nearby microbiology lab to watch it in an actual work situation. We noticed very few problems--primarily because the system is customized for each lab. As a result, we decided to purchase the MicrA and are now awaiting installation.
In the meantime, the manufacturer will use detailed information on our blood culture procedure to customize the program. This will make the system conform to our needs. Eventually, we hope to interface the MicrA with our mainframe hospital computer and eliminate double entry of results.
Pleased as we are with the MicrA, we would like to see a few areas improved. For example, this system does not allow us to add comments to the reports. Customizing partially compensates for this disadvantage because we can include standard comments for the most common microbes in the program. There is an absolute cut-off of 200 items in the combined microbe, Gram stain, and comment categories. The manufacturer is aware of this limitation and is developing a program that will allow up to 40 characters of free text for each culture set.
The systems we have discussed here are by no means all the commercial microcomputers available for microbiology. As we mentioned earlier, a number of other computer systems are available for susceptibility testing, and at least one more interfaces with the Bactec blood culture system (Bactec Data Management System).
There are also a few total microbiology computer systems in operation. Many others are being developed with features specific to microbiology, such as user-defined work cards, labels, media codes, and special keyboards.
We are presently looking for a data base format to enter and store susceptibility data in a microcomputer, and we plan to interface our microcomputers with the hospital-based computer.
In our effort to tap the vast potential of computerization for microbiology, we are gathering a lot of basic information from the computer experts and a lot of practical information and suggestions from other labs. We realize progress will be slow, but it's not impossible. We have only to look at the problem and design ways of solving it, rather than settling for systems that were developed for other functions.
A few manufacturers are finally responding by tailoring systems specifically to microbiology. We all long for the day when most of our procedures and data are stored, analyzed, and distributed by computer, just as is done today in other sections of the clinical laboratory.
|Printer friendly Cite/link Email Feedback|
|Author:||Harrell, Lizzie J.|
|Publication:||Medical Laboratory Observer|
|Date:||May 1, 1984|
|Previous Article:||Putting the magic to work.|
|Next Article:||Word processing with a microcomputer.|