A redesign of RBC morphology reporting.

Back in April 1993, we used a team approach to redesign red blood cell morphology reporting at our lab, Metro-McNair Clinical Laboratories, a group of private labs in British Columbia. At the time, our hematology section consisted of our central lab, five larger satellites (all in the Vancouver area) and several smaller labs throughout the province. In each facility, styles used to report RBC morphology, as well as terminologies, varied.

Some of these biases were minimized in the computerized labs, but others remained, particularly in the small, noncomputerized sites. Major factors that appeared to affect reporting preferences: 1) when and where specific technologists received their training, and 2) recent mergers and acquisitions, each of which had brought individual styles, procedures, and physician preferences.

Our central laboratory was the ultimate referral center for nonstandardized RBC morphology results. Our pathologists wholeheartedly supported our initiative to create a uniform reporting system; from a clinical perspective, they felt many reports of abnormalities failed to provide client physicians with meaningful information.

Varied morphologies

Our quality control technologist confirmed there was room for improvement in this area. As part of our on-going QC program, we send out slides (with known results) bimonthly to all technologists and collate the returned data. Summaries are sent back to all referral technologists for them to compare their results with the originals and what others have reported. Thanks to this quality system, it became quite clear there was a frequent variation in morphology results, especially with more bizarre cases (see Figure 1, p. 62).

We were in the midst of rewriting our hematology procedure manual when we decided to design a standardized procedure for our RBC morphology reporting. We determined our final methodology would include clearer definitions of various abnormal red cell types noted in specific abnormatlities. This would prove to be a tremendous help to all technologists, particularly to our recent new hires fresh out of school. In addition, we all would benefit from a grading system with clearer categories than the existing few," "moderate," and "many," which tended to be influenced by subjective biases of various staff members.

Program goals

We began the process by forming an employee team made up of people with the needed expertise. Our team consisted of our QC technologist, research and development/computer applications technologist (author Payne, who became team leader, too), central hematology supervisor, student teaching technologist (author Edora), as well as a senior technologist from one of our branches and a hematopathologist.

Our mandate was to restructure our reporting of RBC morphologies to give it a more concise, meaningful format. We also hoped to provide technologists with a more definite structure for the grading system and terminology used. Our specific stated goals were to accomplish the following:

* Review abnormal poikilocytes as well as update and agree on terminology, redefine it as needed, and match it to abnormal disease conditions. This terminology then would be ready for incorporation into our new RBC morphology procedure.

* Decrease the number of grading categories for each type of RBC abnormality. A key consideration was to eliminate use of the terms "slight" and "occasional."

* Establish reproducible written guidelines on what each category represents.

We obtained the references we needed from several sources, beginning with established hematology textbooks. We conducted a literature search but were somewhat disappointed by the small number of articles available. We also received input from our teaching technologist author Edora) and from the British Columbia Institute of Technology as to what various syllabuses contained on standard reporting. Our intent was to develop a new method that would be consistent with what was currently being taught nationally by the Canadian Society of Laboratory Technology (CSLT).

In addition, we wanted our reporting to be in line with what our clients received from other institutions. To this end, our hematology manager provided us with information from other institutions, both public and private, through our provincial hematology standards committee, some members of which had been through a similar process.

Development process

We divided various abnormalities into groups, pulled information together, and identified areas where conflicts existed and where no good references could be found. We then tried to decide how, ideally, to report each abnormality. We reviewed our grading in each section closely, attempting to eliminate many of the insignificant and vague gradings, such as "occasional," "slight," and "few." Where a 1 to +4 system existed, for instance, we decreased the number of grades to three.

To test our initial version, we used 12 slides (without results or indices) that previously had RBC morphologies performed using the old system. The morphologies were revamped using our new protocol. As a result, we could identify several additional areas that needed improvement. Again, we fine-tuned our new protocol and, as a result, have developed these final recommendations:

1. A paragraph on determining the quality of the slide has been incorporated in the RBC morphology method. When a slide is of less than "optimal quality" the reviewer must be more aware of potential artifacts or, if possible, should request a new slide.

2. Abnormalities must be confirmed on other parts of the slide (five consecutive fields have been suggested). When cell types are reported as "present," such a notation implies a "significant" or "notable" finding.

3. In most cases, we have reduced the number of grading categories for any one cell type to three. Percentages that correspond to these categories are shown in Figure 2, p. 64. With some abnormal cells, gradings need to be more sensitive; however, grading should be noted and documented specifically for each abnormal cell type.

4. The following changes in or deletions from our RBC/WBC codes have been implemented:

a. Since "occ'ls" was felt to be a catch-all category (as anyone may show at least +1), we questioned the significance of reporting it; therefore, the terminology "Occ'l atypical lymphs" has been eliminated. When 10% of the lymphocytes are atypical, we grade the slide as a "few atypical lymphs" and perform and report a mono test.

b. Plasma cells have been removed from our list since they are likely to be associated with multiple myeloma or plasma cell leukemia. It became common practice for such a comment to be made in the pathologist's interpretation and not as part of an RBC morphology.

c. While we felt comfortable with five categories of platelet comments, we developed a structured grading system as to what the categories signify (Figure 3). We also replaced the terminology "few large platelets" with "large platelets present," to be used only in cases where they are "notable" or "significant."

d. We recommend the terms anisocytosis ("aniso") and poikilocytosis ("poik") only be used if "marked." We consider the terminology "slight aniso or poik" to be overused and of ambiguous clinical significance. "Moderate aniso or poik" typically is followed with a list of what actually is seen (i.e., "+1 micro," "few ellipto," etc.). Only in cases where the morphology is very unusual does "marked poik or aniso" provide the physician with clinically useful information.

e. We've retained the reporting of "moderate or marked polychromasia" since both can be significant and meaningful. We have eliminated "slight polychromasia," however, since it represents an essentially normal condition. In our final protocol, the observation "moderate or marked polychromasia" also generates a reticulocyte count.

f. We feel "slight, moderate, and marked hypochromia" are truly useful but agree the subjective nature of reporting hypochromia must be taken into account.

g. We no longer report "slight rouleaux." Yet we feel "moderate or marked rouleaux." is clinically significant and should be reported.

h. We grade microcytes and macrocytes from +1 to +3 with the gradings corresponding to the previously specified percentages. "Occ'l" has been overused in each of these categories. We also question the clinical significance of +4 (it was so drastic nobody knew what it really meant and so it was seldom used).

i. We grade spherocytes as "few" (e.g., where they may be seen as part of a mild hemolytic process) or "many" (i.e., when hereditary sphemrocytosis is suspected).

j. We've eliminated "occ'l" when describing elliptocytes, target cells, burr cells, schistocytes, acanthocytes, and teardrop cells. These cell types should be present in significant numbers -- i.e., "few," "mod," or "many" -- or not reported at all.

k. We've taken 'helmet cells' off our routine list since it is rarely used.

1. We've eliminated reporting "fine basophilic slippling." It failed to provide the physician with meaningful information unless accomplished by moderate polychromasia and a reticulocyte count.

m. We've clarified the definition of irregularly contracted red cells. Several definitions turned up in our literature search. We prefer: A densely staining red blood cell, smaller than normal, that appears contracted; the outer margin of the cell is irregular and may appear partly concave. It may be associated with drug-induced hemolytic anemias G6PD deficiency, or chemically induced hemolytic anemias."

This definition does not include "bite cells" or larger schistocytic ("helmet") cells. These abnormalities should be reported as being present only when seen in significant numbers, when they will add to the clinically relevant picture.

n. We've deleted Heinz bodies from our codes list. They were only seen when a supravital stain was performed and were not visible on routine blood smears.

o. We've changed the comment "morphology normal" to "morphology essentially normal." This reinforces techs' efforts to report only significant findings and provides a compromise between the comments "normal in every respect" and "few." See Figure 4, p. 66, for our final reporting format.

Find outcome

Once written up, our project was presented to the pathology team for final comments and recommendations. In the interim, we underwent a further centralization process and several management changes. As part of our organizational renewal and restructuring process, we are designing an electronic differential and RBC morphology workstation. We plan to implement the uniform morphology reporting project as part of these changes.

Other minor changes to our recommendations, resulting from current publications and changes in pathologist bias, are likely to occur. But we look forward to finally using our new reporting program in the summer or fall of this year.

Ultimately, we plan to test the efficiency of our new system in standardized reporting by resubmitting old test slides as new unknowns. Once we compare the new and old results, we anticipate showing a significant decrease in technologist-to-technologist variability.

Reference

[1.] Dacie JV, Lewis SM. Practical Hematology. 5th ed. Edinburgh, London: Church Hill Livingstone; 1975: 100-101.

Suggested reading

Bell A. Morphologic evaluation of erythrocytes. In: Lolspeich-Steininger CA, Stiene-Martin EA, Koepke JA, eds. Clinical Hematology. Principles, Procedures, Correlations. Philadelphia, Pa: J.B. Lippincott 1991.

Connors DM, Wilson MK. A new approach to the reporting of red cell morphology. J Med Technol. 1986; 3: 94-97.

Fairbanks VA. Is the peripheral blood film reliable for the diagnosis of iron deficiency anemia? Am J Clin Path. 1971; 55:447-451.

Glasser L. Grading red cell morphology. Diagn Med. 1980; Chapter 3.15.

Krause JR. Red cell abnormalities in the blood smear. Disease correlations. Lab Management. 1985; Chapter 23, 29.

Napoli VM, Nichols CW, Cleck S. A semi-quantitative estimate method for reporting abnormal RBC morphology. Lab Med. 1980; 11:111-115.

Savage RA. The red cells indices: Yesterday, today, and tomorrow. Clin Lab Med 1993; 13:733-785.

Walton JR. Uniform grading of hematologic abnormalities. Am J Med Technol. 1973; 39: 517-523.