A pilot program for POL proficiency testing.

A state medical technology society enrolled physicians' office labs in a program to evaluate testing skills-and initial results proved the program was needed.

The Nebraska Society for Medical Technology is committed to improving the quality of testing in physicians' office laboratories. In 1985, NSMT started sponsoring workshops for POL personnel. ' Two years later, the society extended its involvement by introducing a pilot proficiency testing program for these laboratories.

Many physicians with office labs are apprehensive about the Health Care Financing Administration's plans for nationwide proficiency testing under the Clinical Laboratory Improvement Amendments of 1988. However, PT programs conducted by private nonprofit organizations, states, and approved accrediting bodies will be acceptable alternatives if they meet Federal standards. So the Nebraska society got a head start on what may be a trend toward many new proficiency testing programs.

As chairwoman of the NSMT committee on physicians' office proficiency testing, I oversaw the pilot PT venture. (I am also director of the American Society for Medical Technology's five-state Region IV and a member of the American Society of Clinical Pathologists' ad hoc committee on physicians' office laboratories.)

We launched a recruitment drive in November 1986 with letters to 726 primary care physicians-family practitioners, obstetricians and gynecologists, internists, and pediatricians culled from Nebraska Medical Association lists.

The letter included a postcard that could be returned to request additional information about proficiency testing or our educational programs. Forty-eight (7 per cent) of the inquiry cards came back. Along with more detailed information, we sent these physicians' offices a proficiency testing enrollment card.

We outlined a quarterly testing program for limited-testing laboratories, at a cost of $50 per year to cover the cost of purchasing and shipping materials. There were no strings attached. Results from individual POLs would not be reported to any third parties, and physicians could act on the findings as they saw fit. As we will see, findings in the first year indicated a strong need for "improved testing in physicians' office labs.

Seventeen offices signed up. This represented 2 per cent of the offices in the original mailing and 35 per cent of those that returned inquiry cards. Dispersed throughout the state, 15 of the POLs were in family practices, and two served internists.

Six Omaha area hospital and reference labs agreed to serve as referees. All of these labs were either accredited by the College of American Pathologists or in institutions accredited by the Joint Commission on Accreditation of Healthcare Organizations.

Control materials were provided to evaluate the most common constituents: glucose, potassium, uric acid, cholesterol, and total bilirubin in chemistry; hemoglobin, hematocrit, and red and white cell counts in hematology; and pH, protein, glucose, ketone, occult blood, bilirubin, urobilinogen, specific gravity, nitrites, and leukocyte esterase in urinalysis.

Participants and referees received identical kits containing all the materials needed to test for these constituents, along with specimen handling instructions and a report form. (If a lab didn't offer a particular procedure, it checked off the box, "test not performed," on the report form.) I packaged the kits myself and shipped them by United Parcel Service.

Commercial vendors supplied the control materials, generally at a discount. We sent level I of Fisher SeraChem unassayed chemistry control in the first quarter and level II in the third quarter. Fisher also provided compiled data summary sheets for the target ranges. The Para 4 hematology material, with assigned assayed values that we used for target ranges, came locally from Streck Laboratories in Omaha. The second quarter specimen was in the normal range, and the fourth quarter specimen was in the high abnormal range. For urinalysis, we used EM Diagnostic Systems' Urintrol in the second quarter and lCL Scientific's Kovatrol III in the normal/negative range in the fourth quarter. Both urinalysis materials had assayed values, which became our target ranges.

A report mailed to each POL and referee listed the lab's assay values and, the target ranges, and indicated whether or not an assay value fell within acceptable limits. Educational material accompanied the reports to POLs. Topics included basic quality control; proper specimen collection and handling; variables in specimen processing (instruments, reagents, methods, standards and controls, personnel); and result evaluation and reporting.

To better assess the data, we asked both participants and referees to list the instruments and methods used to perform the different assays. Participating POLs had a variety of chemistry analyzers-Ames Seralyzer, ElectroNucleonics Gemstar, Gifford Systems 103, Abbott Vision, Kodak DT60, Nova 1, AVL 982, and the Boehringer Mannheim Unimeter, Reflotron, and Lytetek systems.

Hematology instrumentation and methodologies included the Becton Dickinson Unopette with a hemocytometer, Coulter M430, Bio-Dynamics CellTrak 3, TOA CC-110, Danam SA-1000, Ames Seralyzer, Abbott Vision, Boehringer Mannheim Reflotron, Kodak DT60, and spun hematocrit. For urinalysis, participants used the Ames N-Multistix-SG and Bio-Dynamics Chem 9 and Chem 6L urinalysis strips.

Referee labs ran their chemistries on the Kodak Ektachem 700, Beckman Astra 8, Hitachi 705, Du Pont ACA, Instrumentation Laboratory 508, and the Technicon SMA 11 and SMAC 11. Hematology tests were done on five different counters-TOA CC800, Onho ELT-8, and the Coulter S + 11, S + IV, and T660. Urinalysis strips included the Ames NMultistix-SG-read manually and with the Ames Clinitek-and the Bio-Dynamics Chem 9.

Now for the results. As might be expected, the performance of POLs veered sharply from that of the referees (Figure 1). For example, of the Ill chemistry results submitted by physicians' office labs, 61 (55 per cent) fell more than 2 standard deviations outside the acceptable range. In contrast, only 13 per cent of the referee chemistries exceeded the 2 SD limit. Uric acid testing was particularly troublesome for the officebased testing staff-nearly threequarters of their results were more than 2 SD off the mark, compared with 18 per cent of the referee results.

Hematology results followed a similar pattern. Participating labs sent in 85 results, and 40 (47 per cent) were outside the acceptable range. Twenty-eight per cent of the referee results were also beyond acceptable limits, but that figure is misleading. As noted in Figure 1, all of the referees had questionable results for the hemoglobins and hematocrits. This led us to believe that there was a problem with the control material or that the material-designed for use with basic instrumentationwas not compatible with the sophisticated counters in the referee labs.

Office laboratories posted their best performance in urinalysis, with only 10 per cent of the 273 test results falling outside the acceptable range. Physicians' office personnel did best on protein and ketone, with no out-of-range results. They had the most trouble with occult blood-about onethird of the results were out of range. The referees kept all but 3 per cent of their urinalysis results within range.

The numbers made it clear that there was a definite need for proficiency testing in physicians' office laboratories. We decided to continue the program, and in 1988, 15 POLs signed up, some of them repeaters. There was, unfortunately, no notable improvement in proficiency testing perfor

mance.

For the 1989 PT recruitment mailing, I plan to emphasize the Clinical Laboratory Improvement Amendments of 1988 in hopes of prodding physicians to get into a proficiency testing program before it becomes mandatory.

If physicians take such testing seriously enough to follow up on poor performance, and if they make sure their personnel improve their skills through such available resources as our state society's workshops, the quality of office laboratory work will improve markedly.