Comparative assessment of automated and manual white blood cell counts at Kenyatta National Hospital Haematology Labaratory

Abstract

Background: Assessment of White Blood Cell values is essential in patient management. However many methods currently in use tend to vary, with most centres using different types of automated machines and a few that have no electrical installations may still have to rely on the manual methods. The manual methods are also retained for purposes of calibration of new machines and to occasionally tie-break the results obtained by different methods. It is therefore rational to remain acquainted with and retain the capability of manual methods while being conventional with automated methods

Objective: To compare the WBC values by automated and manual methods

Study Design: A cross-sectional study of automated and manual white blood cell values from March to May 2008.

Setting: Kenyatta National Hospital Haematology department and wards

Measurable Values and Study Outcomes: Age, sex, diagnosis, WBC total counts, WBC differential counts, Hb, PCV, MCV, platelet counts, NRBC, and ESR.

Materials and methods: Blood specimens from the wards sent to the Haematology laboratory for complete (total) blood counts and those which met the inclusion criteria were processed. The white blood cell counts were estimated using cell counters (CD3200 and CD1300), PBF and Neubauer Chamber. ESR was also done by Westergren method.

In addition, the patients were traced to the wards for the purposes of obtaining desired clinical information from the patient notes and sample for making the PBF using Native blood. The data was entered into a proforma questionnaire. Comparison of the results by the automated and manual methods was undertaken and data analysis was done using standard statistical tests.

Results: The manual and automated methods had coefficient of variations below 5% for total WBC using control blood. The manual methods showed high accuracy in estimation of WBC compared to automated methods with a correlation coefficient, r2>0.9. Manual method with EDT A anti-coagulated blood and native blood were comparable for WBC counts. The TWBC and differentials highly correlated (R2>O.91) and the inter-manual (EM and NM) variations were not significant in the mean estimates of these cell populations (P>O.05).

The age, sex, Hb, PCV, MCV did not affect the WBC counts, while RBC, platelets and ESR (in females) were found to affect the WBC counts. Manual methods showed a higher sensitivity (85% and 96%), specificity (82% and 67%) and accuracy (42% and 47%) for detection of immature WBC and NRBC respectively when compared to CD3200 sensitivity (83% and 80%), specificity (53% and 33%) and accuracy (35% and 39%) respectively.

Observation: Manual methods produced similar results with automated methods for estimation of total WBC, lymphocytes and neutrophils and can be used to compliment automated machines or used in place of automated machine in case of machine downtime with similar accuracy expected. Therefore, manual differential count can be performed when necessary on Native blood. CD3200when compared to manual methods (EM and NM), showed less accuracy in flagging immature white cells and nucleated red cells.