Premium
Accuracy of perception and touch for detecting fever in adults: a hospital‐based study from a rural, tertiary hospital in Central India
Author(s) -
Singh Manoj,
Pai Madhukar,
Kalantri S. P.
Publication year - 2003
Publication title -
tropical medicine and international health
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.056
H-Index - 114
eISSN - 1365-3156
pISSN - 1360-2276
DOI - 10.1046/j.1365-3156.2003.01049.x
Subject(s) - medicine , cohen's kappa , kappa , pediatrics , machine learning , computer science , philosophy , linguistics
Summary objectives To evaluate the diagnostic accuracy of the patient's perception, and of the touch of patient attendants and a doctor for detecting fever. methods We enrolled patients older than 13 years who presented with history of fever to the in‐ and out‐patient departments of a rural teaching hospital. The design was a double‐blind, cross‐sectional analysis of a hospital‐based case series, independently comparing reported history of fever and touch of patient attendant and that of doctor against an established reference standard (axillary temperature > 37.5 °C). Diagnostic accuracy was measured by computing sensitivity, specificity, and likelihood ratio values. The agreement between the patient, his attendant and the doctor was assessed by kappa statistic. results We studied 462 patients of whom 274 (59.3%) were men. A total of 206 patients (44.58%) had fever. The patient's perception of fever (LR+ 1.77; 95% CI 1.52, 2.06), patient attendant's touch (LR+ 2.03; 95% CI 1.74, 2.36) and the doctor's touch (LR+ 3.08; 95% CI 2.51, 3.71) did not accurately distinguish those with and without fever. Doctors (LR− 0.20; 95% CI 0.17, 0.34) and patient attendants (LR− 0.24; 95% CI 0.14, 0.28) were more accurate in ruling out fever. The patient's perception agreed moderately with patient attendant's touch ( κ = 0.44; 95% CI 0.36, 0.53), and the doctor's assessment ( κ = 0.47; 95% CI 0.39, 0.55). There was moderate agreement between patients' attendants and the study doctor ( κ = 0.48; 95% CI 0.40, 0.56). conclusions Our findings suggest that patients, their attendants or doctors cannot accurately detect the presence of a fever without using a thermometer. Doctors should confirm a history of fever by recording temperature.