Protocol Summary:
Background:
Acute headache is a common emergency department (ED) problem accounting for 4.5% of all ED visits.
The most important emergent diagnosis in this patient population is non-traumatic subarachnoid hemorrhage (SAH),
a type of hemorrhagic stroke which has a 50% mortality rate but only accounts for 1% of ED visits for headache.
Current strategies using computed tomography (CT) of the patient’s head and/or lumbar puncture (LP)
(collects spinal fluid via the patient’s back) to investigate patients with headache to rule out SAH are inefficient and
lead to increased ED length of stay and increased health care costs.
The ED use of CT has risen substantially in Canada leading to considerable inefficiency,
with over 95% of scans for “rule out SAH” being negative for SAH. LP is recommended when CT is negative,
as CT can miss 2-10% of SAHs. LPs are invasive, painful procedures with potential patient complications
(infection, post LP headache).
Additional costs occur when patients from rural/remote areas are transferred for investigation.
Despite the inefficiency of current investigations, up to 40% of patients with SAH are
initially misdiagnosed resulting in poorer outcomes.
In phase I, we prospectively enrolled 1808 patients from six tertiary care hospital EDs with 115 positive SAH cases.
This represented 69% of all eligible patient visits during the study period.
From this cohort we derived several highly sensitive clinical decision rules to direct investigations to rule out SAH.
We then selected the three best performing rules with retrospective sensitivity of 100% (95% CI: 97-100%) for SAH,
which reduce investigation rates to as low as 63% from the current rate of 83%.
Following prospective validation, the final clinical decision rule will allow physicians to be more selective
in ordering investigations with fewer missed SAH and fewer complications.
We are currently awaiting funding decisions from CIHR for phase II.
Objectives:
The goal of phase II is to prospectively assess the accuracy of the three best performing clinical
decision rules derived in phase I in a new set of alert neurologically intact ED patients with
acute headache without altering patient care.
The specific objectives are to assess prospectively:
i) the accuracy or classification performance of each proposed clinical decision rule to rule out SAH;
ii) the reliability of each clinical decision rule;
iii) the clinical sensibility, i.e. physicians’ accuracy, comfort and ease of use with applying each rule;
iv) the potential of the each rule to reduce investigations (i.e. head CT and lumbar puncture);
v) the potential for refinement of the rules (improve specificity);
vi) the potential savings with widespread implementation of the most efficient rule in a preliminary economic evaluation.
Methods:
This four year prospective cohort study will be conducted in the EDs of 10 university affiliated Canadian hospitals.
Data management and analysis will be conducted at the clinical epidemiology unit of the Ottawa Hospital Research Institute.
In this validation study, patients will be managed according to standard clinical practice and
not according to any of the proposed clinical decision rules.
The study will enrol consecutive alert patients over 15 years of age with acute non-traumatic headache.
Emergency physicians will assess and classify each patient for the need for investigation according
to each proposed rule to rule out SAH prior to CT or LP.
A subset of patients will be examined independently by two physicians.
The classification performance of each rule for predicting SAH will be assessed with 95% confidence intervals
for sensitivity, specificity, negative predictive value, positive predictive value and likelihood ratios,
which will allow users to estimate the probability of the outcomes in patients,
given a negative status for the final “Canadian SAH Rule”.
Reliability will be assessed by the kappa coefficient.
We will attempt to refine (i.e. improve specificity without loss of sensitivity) the rules by
recursive partitioning analysis techniques.
The preliminary economic analysis will determine the potential incremental cost savings from
a societal perspective that could be attributed to use of the rule.
The estimated sample size is 2700 patients (including 170 SAH cases).
Importance:
This prospective validation study is an essential step in the development of an accurate, reliable, and
clinically sensible decision rule and must be undertaken before any rule can be recommended for actual patient care.
The final SAH Rule will permit physicians to standardize care of alert neurologically intact
non-traumatic acute headache patients and be much more selective in performing CT and LP.
The best rule will be subsequently implemented to lower health care costs,
decrease morbidity of unnecessary LPs and decrease the mortality of missed SAHs.
The impact of this rule should be similar to the Ottawa Ankle Rules that have been readily adopted
by physicians and have significantly reduced health care costs.
Ottawa Team
Partners