 |
 |
|
Home |
Current Issue |
Past Issues |
In the Clinic |
ACP Journal Club |
CME |
Collections |
Audio/Video |
Mobile |
Subscribe |
Tools |
Help |
ACP Online
|
Rapid Responses to:
|
|
Electronic letters published:
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif){kind=icon}
Vadiation of A SARS Prediction Rule
|
|
|||
|
Matthew H-M MA, MD,PhD Department of Emergency Medicine, National Taiwan University Hospital, Shey-Ying Chen, Wen-Chu Chiang, Chan-Ping Su.
Send rapid response to journal:
mattma{at}ha.mc.ntu.edu.tw Matthew H-M MA, et al.
|
We read with great interests the article reported by Leung et al (1) on their clinical prediction rule for triaging the severe acute respiratory syndrome (SARS) in the emergency department. During the early period of the 2003 SARS epidemics in Taiwan, we have prospectively developed a clinical decision rule from a cohort of febrile patients. (3,4). The rule consisted of a 4-item symptom score, and a 6- item clinical score (3). The prediction rule was adopted in our institution and became part of an integrated decision-making process for sorting incoming febrile patients during the epidemic. When validated by a second cohort, our rule showed a sensitivity of 90.2%, specificity of 80.1%, and area under ROC 0.89(4) We applied the Leungˇ¦s rule to our cohort of 299 febrile patients, including 79 laboratory-confirmed SARS and 220 non-SARS patients. The sensitivity, specificity, positive predictive value and negative predictive value were 98.8%, 52.0%, 43.6% and 99.1% respectively. The authors should be congratulated for a well-conducted study. However, some aspects of their prediction rule warrant further elaboration. First, the study cohort had a higher proportion of non-febrile SARS patients than elsewhere reported (2,5). Due to the retrospective design, some of these patients might not have had SARS to begin with, but rather cross-infected in the hospital settings. This might introduce some misclassification bias and threatened the validity of the study. Second, the clinical utility of the rule, if applied as suggested, may be limited in a large outbreak. Over 80% of the patients still need hospitalization after triage by the rule. Housing patients with 21% risks of SARS, the lowest quartile, in a communal isolation ward can be a dangerous practice. Triage of SARS during the epidemics depends on the prevalence of the disease in the community, and on local policies. Three dimensions of information, contact history, fever, and pulmonary infiltrate, need to be considered. Febrile patients with pulmonary infiltrates should be admitted regardless of contact history. Afebrile patients without pulmonary infiltrates could be discharged home and receive daily body temperature follow-ups. Clinical decision rules are not needed for these patients. Febrile patients without pulmonary infiltrates are the most challenging for emergency department staff, and should be the focus of a prediction rule. Whether to admit afebrile patients with pulmonary infiltrates, however, is a policy issue. Well thought-out clinical prediction rules along with sound policies would help communities to tackle future SARS epidemics or similar outbreaks. References: 1.Leung GM, Rainer TH, Lau FL, Wong IOL, Wong TW, Tong A, et al. A predictive rule for diagnosing fever acute respiratory syndrome in the emergency department. Ann Intern Med. 2004; 141:333-342. 2.Chen SY, Chiang WC, Ma MHM, Su CP, Hsu CY, Ko PCI, et al. Sequential Symptomatic Analysis in Probable SARS cases. Ann Emerg Med. 2004; 43:27- 33. 3.Chen SY, Su CP, Ma MHM, Chiang WC, Hsu CY, Ko PCI, et al. Predictive model of diagnosis probably cases of SARS in febrile patients with exposure risk. Ann Emerg Med. 2004; 43:1-5. 4.Su CP, Chiang WC, Ma MHM, Chen SY, Hsu CY, Ko PCI, et al. Validation of a novel SARS scoring system. Ann Emerg Med. 2004; 43:34-42. 5.Peiris JSM, Yuen KY, Osterhaus ADME, Stohr K. The severe respiratory syndrome. N Engl J Med. 2003; 349:2431-41. Conflict of Interest:None declared |
|||
