PT - JOURNAL ARTICLE AU - Xiao Wu AU - Vivek B Kalra AU - David Durand AU - Ajay Malhotra TI - Utility analysis of management strategies for suspected subarachnoid haemorrhage in patients with thunderclap headache with negative CT result AID - 10.1136/emermed-2015-204634 DP - 2016 Jan 01 TA - Emergency Medicine Journal PG - 30--36 VI - 33 IP - 1 4099 - http://emj.bmj.com/content/33/1/30.short 4100 - http://emj.bmj.com/content/33/1/30.full SO - Emerg Med J2016 Jan 01; 33 AB - Purpose To determine the most effective follow-up strategy for evaluation of patients with thunderclap headache and negative initial non-contrast CT for acute subarachnoid haemorrhage (SAH).Materials and methods Institutional review was performed to assess the frequency of CT angiography (CTA) in screening patients with negative non-contrast CT. A comparative effectiveness analysis based on decision tree modelling was subsequently performed to assess three different strategies—no follow-up, lumbar puncture (LP) and CTA. The clinical probabilities and utilities from literature were used to design the decision tree. Base-case scenario utility calculations, sensitivity analyses and probabilistic Monte Carlo simulation were performed.Results Institutional review of recent data in the last two years demonstrates frequent use of CTA in patients with thunderclap headache with limited utility. The decision tree analysis shows CT with LP follow-up to be the most effective strategy with the highest expected utility of 0.79926 quality-adjusted life-year (QALY) compared with 0.79875 QALY for no follow-up and 0.79869 QALY for CTA follow-up. Monte Carlo simulation showed LP was the best strategy in 86.4% of all iterations. Sensitivity analyses demonstrate that CT without follow-up is the best strategy only when the sensitivity of CT is very high (99.6%) or the pre-test probability of SAH in a patients with thunderclap headache with negative initial CT is low (1.6%).Conclusions CT with no follow-up was shown to be the best strategy when the pre-test probability of SAH is low (<1.6%) or the sensitivity of initial non-contrast CT for blood is high (>99.6%). Otherwise, LP should be the preferred strategy for follow-up.