How we computerized our microbiology reporting; for a relatively small investment, this lab developed a computer system with broad reporting capabilities and automatic billing.
Our annual microbiology volume jumped from 14,000 specimens in 1983 to 20,000 in 1985. This growth, along with a hiring freeze at our 197-bed hospital, challenged the laboratory's ability to keep pace with paper-work and telephone inquiries.
Handwritten reports, billing sheets, and accession logs took hours to prepare each day. Everyone wanted information right away. Phone calls about the status of cultures pulled technologists from their testing duties, sending them to different bench areas or the files in search of a particular specimen's work card. The billing department needed codes for lab work the same day a specimen was received since some patients are discharged before completion of testing.
Microbiology as also being asked to provide more information based on results, such as the hospital's antibiotic susceptibility patterns. Too often, I as chief microbiologist had to do such tedious chores at home in the evening to meet the inspecting agencies' requirements for current data.
We decided to evaluate our information system and see whether we could save technologist time by computerizing. Whatever the change, we did not want it to compromise the clear, timely, accurate reports expected by our hospital staff and physicians' offices.
The first step: a review of work flow from specimen receipt until the final written laboratory report was filed. All specimens received in our laboratory are assigned an accession number. We were using two numerical accession logs, one for inpatient specimens and the other for everything else. In these logs, we recorded patient demographics, source of the culture, date of specimen receipt, ordering physician, and the billing codes.
The entries were more or less legible, depending on the penmanship of each technologist. Pages were prenumbered by hand or with a number-stamping device on slow afternoons. With slack periods becoming rarer, we often fell behind on this kind of housekeeping.
Separate logs helped us track the changing volume of specimens received from different sources and served as a way to check the individual billing codes needed for microbiology's three services--our own hospital, outpatient testing, and the tests we performed for another community hospital in exchange for lab work they did for us.
The test-ordering system at one time relied on nurses generating separate requisitions for each major laboratory department--chemistry, hematology, microbiology, and the blood bank. The microbiology form--a seven-part, odd-size combination--served not only to request tests but also as a work card and ultimately as a report form.
We produced billing data on single sheets of paper, one for each specimen received. Encoding was simple. Technologists merely had to verify that the proper box had been checked on the form. The billing clerks, however, had to handle all these sheets of paper containing sometimes illegible demographics. This made their computer data entry inefficient.
To assist the hospital and outpatient staffs, the laboratory finally adopted a single menu for test ordering. While this was convenient for the physicians and nurses doing the ordering, the microbiology staff had to spend more time handwriting patient demographics on the work cards and report forms, which nursing previously prestamped for us. And we were still stuck with the cumbersome single-sheet method for billing.
We kept the microbiology work cards for each specimen in numerical order while testing was in progress. This enabled us to match culture plates with the records. Once tests were completed, the work cards were filed alphabetically for quick retrieval of all results on a particular patient.
Easily reviewed files helped us avoid duplication of biochemical tests and susceptibility procedures. Unfortunately, the files were sometimes out of order, and technologists seldom had time to hunt for misplaced cards. So some tests were repeated unnecessarily --an unwanted expense.
Reporting results also took considerable time. Each report was removed from the backing card, sorted according to hospital ward or physician in private practice, and then sent on for charting or placed in an envelope for courier distribution outside the hospital. When workloads were particularly heavy, we barely met the distribution deadlines. Couriers were delayed, which meant reports would not reach physicians promptly.
Then there was the time spent at the copying machine. We had to duplicate some reports for consulting physicians and nursing homes, as well as our infection control officer and pharmacist. Pathologists wanted copies of all significant inpatient results.
In compliance with regulatory requirements, we frequently revised written procedures to reflect changes in methods and reagents. Since the micro lab could not hire a secretary, the chief microbiologist found herself spending more time at the typewriter and less time running the department.
We concluded that we needed: 1) a word processing program, which would let us update procedures without retyping the entire text, and 2) a computer program for microbiology data management and reporting that would reduce manual sorting and distribution time.
The system had to be easy to operate, and it had to provide clear, concise reports of test results. We wanted to bypass unnecessary reporting steps for simple tests if a yes or no would suffice to convey results. We also sought a simple reporting method on special tests for viruses, such as herpes simplex, and on frequently performed tests for bacterial toxin.
Although there were many word processing programs for all types of computer systems, the microbiology software we looked at was disappointing. No commercial program met all of our needs.
Most "inexpensive' hardware/ software packages retailed for $30,000 or more, and many of the programs were primarily designed to interface with a susceptibility system. The programs were not flexible enough to handle data entry for complex cultures or tests and could not recall patient information by name, a must if we were to answer phone inquiries quickly. Instead, the programs required an accession number to call up patient data, which meant going back to the log book.
Larger, more versatile systems retailed for $50,000 and up, far beyond the reach of our budget. Even some of these sophisticated systems did not provide all of the features we needed. For example, most were not able to generate separate billing codes for the hospital's inpatients, for outpatient work performed for the doctors' private patients, and for testing done for inpatients at another community hospital.
Since none of the commercial programs met our requirements, we decided to do the job ourselves. A computer programmer with a microbiology background --he was married to a member of our staff--teamed up with out technologists to develop the system.
I met three times with the programmer over a six-month period to outline the basics. He had helped us on other projects and had a good idea of what we needed. So did I in terms of what data the physicians required and what type of accessioning method would be easiest for the technologists.
We then spent 20 weekend and evening sessions over the next two years, analyzing the program to see what worked and what needed fixing. All of this could have been accomplished in six months by a full-time hospital computer programmer.
The laboratory purchased an IBM PC XT with 512K memory and a 10-megabyte hard disk for large amounts of data storage and retrieval. The program is written in BASIC and functions with either CP/M or an MS-DOS operating system. Our microbiology system is flexible enough to interface with a larger IBM-compatible laboratory system, should one be installed. We also purchased an on-site service contract for hardware maintenance.
Software was delivered in modules, and we went on line with a skeleton reporting program for eight months. During this time, all of the staff members who used the computer--from clerks on up to the chief microbiologist-- helped work out the bugs and fine-tune the program. Then we added the software for work lists, workload recording, and billing.
Our sole microcomputer gets maximum use throughout the day and well into the evening. Staff members take turns entering the various types of data, and it requires some fancy choreography to prevent a bottleneck. For example, as soon as the cultures are read each morning, a technologist enters all of the findings. Results are added or updated periodically, as needed. Billing data goes into the system automatically when a specimen is first accessioned. Each morning, we pull a copy of all work performed the previous day and send it along to the billing department. Routine bacteriology reports are also entered each morning.
In the afternoon, the technologists input the reports for herpes, Chlamydia, and Clostridium difficile toxin tests. Reports for acid-fast cultures, fungus cultures, and parasitology are entered on the evening shift, and new cultures are accessioned. The evening techs also tie up any loose ends, leaving the system free to begin the entire process over again the next morning.
We anticipated the needs of the laboratory staff, the physicians, and others who would actually use the laboratory reports. Since we wanted a system that was easy to learn and operate, the program was designed to be menu-driven. Even new staff members are able to enter demographics and test data after only an hour or two of instruction.
Technologists wanted minimal keyboard entry so that they could spend most of their valuable time identifying microorganisms. We met this reasonable request by designing special formats for each major type of test. Programming took a little longer, but it was worth the investment.
For example, our blood culture routine bypasses the initial Gram stain done on all other specimens. A single keystroke yields positive or negative test information for simple yes/no tests, including screens for group A beta hemolytic streptococci.
More complex tests, such as the herpes virus culture, require up to three updated reports if specimens show evidence of cytopathic changes in tissue culture. Our program follows distinct paths to confirmatory test results for various procedures.
We simplify entry for routine bacterial cultures by encoding organism names or test results in one to three keystrokes. For example, we type E for Escherichia coli and KOX for Klebsiella oxytoca. The computer translates the codes for the report and, of course, gives the correct spelling every time. For "normal flora' or "no growth' results, we need only type NF or NG, respectively. User-defined abbreviations can be changed or updated at will.
An abbreviated reporting format for susceptibility tests saves space on our half-page reports (Figure I). Although we often test bacteria against many drugs, the report shows only the most effective and least costly antibiotics. This helps keep patient costs down. Periodic antibiograms are generated from the stored results for our inpatient and outpatient isolates.
Entry of common results is made even easier by use of the carriage return: For example, our mycobacterial and mycology and blood cultures are usually negative for the suspected organisms. A carriage return on a blood culture yields a report of "no growth.'
For the initial smear of acid-fast cultures, a carriage return provides "no acid fast bacilli seen.' A touch of the Y key then reports that culture results will follow. Once the culture is in progress, a simple carriage return yields an interim result of "negative for acid-fast bacilli.' A carriage return in the mycology routine provides the result "no fungal elements seen' for the smear, and a subsequent carriage return as the culture progresses notes "no fungus isolated' on the report.
The microbiology laboratory produces a preliminary report within a few hours on most tests. When bacteria are examined, a final report usually follows within a day or so. When we know the identity of the microorganisms and are awaiting susceptibility results, we may generate an interm report. It's an easy matter to change the status of the report and add updates without reentering all of the preliminary information.
To upgrade a report from preliminary to interim, we type "I' at the status prompt and enter any new information. When the results haven't changed from one day to the next but we want to indicate that the report is now complete, we press carriage return past the results and add an F at the status prompt.
The system not only provides concise, readable hard-copy results but also eliminates our mad scrambles--prompted by phone inquiries--to locate work cards in progress or filed for storage. Any patient results stored in the computer's memory can be recalled in seconds. We merely enter all or part of the patient name, the hospital number, or other identifying information, such as the laboratory accession number.
Up to 6,000 patient results can be stored in the active system for ready access, but we find it more practical to keep 1,200 to 1,600. Older results are archived to floppy diskettes every two weeks for permanent storage.
The computer eases distribution of reports by organizing groups of printouts according to ward or physician. (In the future, we could link the system to offsite printers in physicians' offices for even faster service.) We can also print by range of dates if we want to pull all reports issued during a particular time period. With triplicate tractor paper for reports, we can send copies to the infection control officer, consulting physicians, nursing home stations, the pharmacy, or pathologists, if necessary.
We no longer send the billing office single sheets for each test. Billing lists are automatically printed with up to 16 patient names and codes on a single page. We can organize these lists by ward or service for more efficient encoding in billing. The lists can also be printed alphabetically, by patient name.
As Figure II shows, the new system saves the laboratory $1.43 per microbiology report, mainly in labor time. That works out to nearly $29,000 per year.
Before we computerized, it took five to seven minutes to write out (and sometimes write out again) the demographics for each specimen and record the test results. We also spent countless minutes locating results to answer phone queries. We can now do the same amount of work in less than two minutes per specimen and respond to inquiries in seconds. In addition, we once spent an extra minute or so filing work cards for bench use and then refiling them alphabetically for storage. We now file the cards just once, numerically, with much better accuracy.
The microbiology system also has a word processing program. This enables us to modify and write procedures much more easily and efficiently.
Since the system features off-the-shelf hardware, it can be used with such software as spread-sheets and database management programs. This will further facilitate workload recording and preparation of key hospital reports, such as summaries of significant bacterial isolates.
How much did all our efforts cost? For less than $13,000, we purchased the hardware and developed the software and taught technologists how to operate the computer. The system paid for itself within the first year, and we expect the savings to continue even as our volume increases.
Table: Figure I Computerized report on routine culture, susceptibilities
This report uses such abbreviations as GPC for gram-positive cocci, SEC, for squamous epithelial cells, and CEF for cefalosporin.
Table: Figure II Cost per lab report: Manual vs. computer system