Interobserver agreement of pallor in detection of anemia

Introduction: The diagnosis and management of anemia largely depends on clinical assessment of pallor. There are varied opinions regarding the subjective variation in the assessment of pallor. Objective: The objective is to study the interobserver agreement of pallor between two independent observers. Methodology: This observational study done in a medical college hospital on 50 children in the age group of 6 months to 5 years. Pallor was assessed in four sites namely conjunctiva, tongue, nail bed, and palm. When pallor was present, it was graded as definitely present and borderline. Second examiner independently assessed the pallor at all four sites within an hour. Statistical analysis was done by percentage and kappa value. Results: Out of 50 children aged between 6 months to 5 years, 30 were male and 20 were female. Eleven were infants of 6 months to 12 months, 19 children were between 1 and 3 years and 20 children were 3–5 years of age group. Twenty-eight children had anemia with hemoglobin <11 g/dl. Kappa value for conjunctival and nailbed assessment of pallor were 0.484 and 0.441, respectively, indicating good agreement. The value was 0.644 and 0.73 for tongue pallor and palmar pallor respectively suggesting very good agreement. Conclusion: Interobserver agreement was found to be good in the evaluation of pallor. Palmar and tongue pallor had better agreement than conjunctival and nailbed pallor in our study. By assessing in all four sites, for pallor interobserver variability can be minimized.