Am Fam Physician. 2005;71(2):370
There are a variety of approaches to testing for group A streptococcal disease (i.e., “strep throat”), and recommendations vary by organization, depending on whether the affected population comprises children or adults. For example, the Infectious Diseases Society of America recommends treatment only in patients who have a confirmed positive culture or rapid strep test result. Other groups recommend using only rapid tests and treating patients with positive results, without culture confirmation of negative results. In a large study, there were no differences in complication rates using this strategy compared with a culture-based strategy. McIsaac and colleagues prospectively tested six strategies to determine which cost the least, missed the fewest cases of strep throat, and used the least amount of unnecessary antibiotics.
Canadian patients three to 69 years of age who presented with acute sore throat were enrolled. All patients were swabbed twice, and the swabs were sent for culture or rapid strep testing, depending on which one of the following six strategies was used: (1) culture all patients and treat those with positive results; (2) perform rapid strep testing on all children and treat those with positive results; obtain cultures in children with negative rapid test results and treat those with positive results; and perform rapid strep testing on all adults and treat only those with positive results without further confirmation of negative results; (3) treat children per strategy 2; perform rapid testing on all adults with a Centor score of 2 or 3 and treat those with positive test results; and empirically treat all adults with Centor scores of 4 or more; (4) treat children per strategy 2 and empirically treat adults with a Centor score of 3 or 4; (5) culture all children and adults with a Centor score of 2 or 3 and treat those with positive results; and empirically treat children and adults with a Centor score of 4 or more; and (6) perform rapid testing on all children and adults and treat those with positive results without further confirmation of negative results.
In this population, the overall prevalence of strep throat was 29 percent, which is similar to the reported prevalence; children had higher rates of strep throat than adults. All strategies except number 6 had a sensitivity greater than 90 percent overall, but sensitivities varied when broken down for adults and children. All strategies except for number 4 had a specificity greater than 90 percent. The highest rate of antibiotic prescriptions was associated with strategy 4 (strategy 5 in children alone). The highest rate of unnecessary antibiotic prescriptions was associated with strategies 4 and 5 (18.9 and 4.8 percent, respectively). Strategy 5 required the least number of tests per person and generally fewer follow-up telephone calls for positive cultures.
Although an all-culture strategy has the best sensitivity and specificity, it is not the most practical. Strategy 5 would identify all cases of strep throat with less testing but with more unnecessary antibiotic use than other strategies except number 4. Because the amount of unnecessary antibiotic use associated with strategy 5 is still less than that associated with current practices, the authors suggest that strategy 5—performing a throat culture on everyone with a modified Centor score of 2 or 3 and empirically treating those with a higher score—represents the best compromise when choosing among the six strategies.