Validity and reliability of methods to microscopically detect and quantify malaria parasitaemia.

ABSTRACT

OBJECTIVES: The recommended microscopy method by WHO to quantify malaria parasitaemia yields inaccurate results when individual leucocyte (WBC) counts deviate from 8000 leucocytes/µl. A method avoiding WBC count assumptions is the Lambaréné method (LAMBA). Thus, this study compared validity and reliability of the LAMBA and the WHO method. METHODS: Three methods for counting parasitaemia were applied in parallel in a blinded assessment the LAMBA, the WHO method using a standard factor of 8000 leucocytes/µl ['simple WHO method' (sWHO)] and the WHO method using measured WBC counts ['accurate WHO method' (aWHO)]. Validity was assessed by comparing LAMBA and sWHO to the gold standard measurement of aWHO. Reliability was ascertained by computation of intraclass correlation coefficients (ICCs). RESULTS: 787 malaria-positive thick smears were analysed. Parasitaemia as determined by LAMBA and sWHO increasingly deviated from aWHO the more patients' WBCs diverged from 8000/µl. Equations of linear regression models assessing method deviation in percent from gold standard as function of WBC count were y = -0.00608x (95% CI -0.00693 to -0.00524) + 47.8 for LAMBA and y = -0.0125x (95% CI -0.01253 to -0.01247) + 100.1 for sWHO. Comparison of regression slopes showed that the deviation was twice as high for sWHO as for LAMBA (P < 0.001). ICCs were excellent (>90%) for both methods. CONCLUSIONS: The LAMBA has higher validity than the sWHO and may therefore be preferable in resource-limited settings without access to routine WBC-evaluation.