The CARE-SORE THROAT Study
a multinational e-mail/web-based study
Copyright ©2002, CARE-SORE THROAT

The United States National Ambulatory Medical Care Survey: 1999 Summary estimates over 15 million yearly patient visits to for sore throats. This represents the second most common reason for symptomatic visits. While patients visit physicians primarily because of discomfort, physicians have focused on the management of pharyngitis for several reasons. Approximately 10-20% of sore throats are due to group A beta hemolytic streptococci. Untreated this organism can cause both suppurative and non-suppurative complications. Treatment decreases rates of tonsillar abscess as well as rates of acute rheumatic fever. Untreated group A beta hemolytic streptococcal pharyngitis is contagious both within households as well as in schools and work places. Twenty-four hours of treatment markedly decreases the public health spread of this infection. Recent data demonstrate that the duration of discomfort for group A beta hemolytic streptococcal phyarngitis decreases by as much as two days with immediate treatment. ¹

Other organisms can cause pharyngitis. Of the beta hemolytic streptococci, group C represents the second most important cause. Few other bacteria frequently cause pharyngitis; however, viral infections are a very common cause.

In order to decrease unnecessary antibiotic use, we would like to quickly and appropriately make decisions about patients with acute pharyngitis. Two recent systematic reviews of this topic have been published,^2,3 one in the Rational Clinical Examination Series and one in the Clinical Practice Guideline Series by the American College of Physicians-American Society of Internal Medicine. The rational clinical examination series examined a number of articles and looked at the likelihood of strep throat infections and the presence of individual clinical findings. They also looked at and recommended several clinical prediction rules as having both good accuracy and good validation.

Their analysis suggests that well validated clinical prediction rules can help physicians make informed use of both rapid antigen tests and throat cultures.

The ACP-ASIM guideline was published in the Annals of Internal Medicine in 2001. They recommended the use of the prediction rule developed by Centor in 1981. Neither the ACP guideline nor the Rational Clinical Exam article considered the impact of telephone screening. In today's primary care environment we often need triage procedures for telephone screening of symptoms. One of the clinical predication rules mentioned in the rational clinical examination review did look at a possible telephone prediction rule. However, this rule had not been independently validated in populations other than in the original paper.

Finally, unpublished data from Centor and colleagues suggested that ordinal clinical variables might add accuracy to the analysis of individual clinical signs and symptoms. The proposed telephone-screening rule used such data. There are no good prospective analyses of dichotomous (yes/no signs or symptoms) vs. ordinal graded symptoms (absent, mild, moderate or severe).

1. In patients with acute sore throats what are the test characteristics of the clinical examination. Candidate variables include historical variables-cough, fever history, rhinitis, difficulty swallowing, presumed exposure and age. Physical findings include: temperature, tonsillar or pharyngeal exudates, anterior cervical adenopathy, posterior cervical adenopathy, and tonsillar swelling.
2. What is the reliability of recording these signs?
3. What is the comparison between patient recorded historical findings and physician recorded historical findings.
4. What are the test characteristics of three previously described prediction rules?
   1. The original rule described by Centor in 1981.⁴
   2. The modified rule described by McIsaac in 1998.⁵
   3. The telephone prediction rule described by Clancey in 1988.⁶

• Investigators will be invited to join using the Internet.
• Investigators should come from primary care settings.
• The Internet will be used to enroll in the study and to complete data collection
• Each investigator group will be responsible for obtaining local ethics approval and proof must be forwarded to the studying coordinating center on request.
• Investigator completing the clinical exam should be blind to the results of any rapid strep testing and to the throat culture.
• Each investigator should enroll at least 20 patients.
• If possible a second investigator should repeat the clinical examination (blind to the results of the first examination) within 30 minutes of the initial assessment.
• The 2 investigators performing the clinical exam should not change throughout the study.

Patients will be those presenting with acute (< 3 days) onset of sore throat to a primary care site (this will include acute care clinics, emergency rooms, and student health sites and primary care offices).

Exclusions:
• Patients < 4 years of age
• Patients unable to give consent
• Patients treated with antibiotics within previous week

The exam should be performed at the initial visit. The following signs and symptoms should be evaluated in an absent mild, moderate, severe ranking to be determined subjectively by individual patients and physicians. Descriptions of these ordinal variables can be found in the appendix. Prior to the patient being examined by the physician, each patient will be asked to fill out a short form on the historical features of their sore throat. This will be a short piece of paper that includes the history forms found below. Physicians will be blind to the patient's completed assessment form. The time required for completion of the patient assessment form as well as the physician assessment should be recorded respectively.

• History
  • Cough
  • Rhinitis
  • Fever history (absent, present subjectively or measured)
  • Difficulty swallowing
  • History of presumed exposure to group A strep (this is a dichotomist variable)
  • Patient age
• Physical examination
  • Temperature either in degree centigrade or degree Fahrenheit
  • Exudates (tonsillar and/or pharyngeal)
  • Anterior cervical adenopathy
    • If present, is it tender (yes or no)
  • Posterior cervical adenopathy
    • If present, is it tender (yes or no)
  • Tonsillar swelling

When possible, a second independent investigator will repeat the clinical examination within 30 minutes of the initial assessment and will be blinded to the results of that initial assessment.

Each patient will have a throat culture with specific typing of beta hemolytic streptococci with specific reference to group A and, if possible, group C beta hemolytic streptococci. If available, a rapid antigen test could also be performed (record the type of rapid test and the result).

Physician participants should record their clinical diagnosis at the time of the exam. They should also record antibiotic prescription (which antibiotic and what dose). Finally, they will record what diagnostic/therapy they would have done if not participating in the study (culture or rapid test, or antibiotic use).

To be developed by the CARE group.

Sensitivity/specificity likelihood ratios and ROC curves for each of the clinical exam maneuvers will be calculated using the results of the throat culture as a reference standard. Non-parametric correlations for inner observer agreement on the clinical examination will be calculated.

The three previously described multi-variable clinical prediction rules will be calculated for each patient and compared using ROC analysis and likelihood ratios. We will also use the data to attempt to improve on the multi-variable analysis using ordinal variables.

Enrollment should continue for one year.

First analyses and drafts will be sent to participating investigators

After the investigators are satisfied with the manuscript and the analysis it will be submitted to a journal by mutual agreement of the authors and include a line stating "on behalf of the CARE-SORE THROAT Study Group"

All investigators and methodologist will be included and will be listed in descending order by the number of patients they enroll.

Proposed by Andrea Cherrington, M.D. and Robert M. Centor, M.D. of Birmingham, Alabama

* Any comments or concerns about this protocol, please email carestudy@rogers.com.