Time to Specialized Admission in Case of Severe Brain Haemorrhage.

Background:

Spontaneous (non-traumatic) subarachnoid haemorrhage (sSAH) is rare but timely detection is critical. A feared complication is rebleeding, with an overall risk around 4% in the first 24 hours. Early specialized medical, neurosurgical or neuroradiological interventions are required. A Danish study of pre-hospital delay was made in 2010. In this, a median delay of 60 minutes in a pooled population of patients having contacted general practitioners, the emergency phone number or presented directly to the emergency room. Total time from symptom onset to neurosurgical admission was 5 hours 30 minutes.

Reorganizations have since been made in the pre-hospital system to optimize correct and fast triage. A criteria-based dispatch (CBD) system, the Danish Index for Emergency Care (Danish Index), is now used in the Danish Emergency Medical Communications Centers (EMCC). The dispatcher is guided through a list of questions to ask the caller. An algorithm then suggests a prehospital response.

The Danish Index contains 37 main symptom categories, each divided into five degrees of urgency, which again are subdivided into specific symptoms. Headache is a main category with ´acute severe headache´ resulting in the most urgent response; level A.

In Denmark there is one national emergency phone number, ´112´. Depending on your location, your call will be directed to one of the five EMCCs. At the EMCC in the Capital Region of Denmark the Danish Index was introduced in May 2011. In January 2012 an additional phone line, ´1813´, was opened for non-urgent contacts out of general practice (GP) opening hours. At this non-urgent phone-line the Danish Index is used as well. By January 1, 2014, ´112´ and ´1813´ became the only out-of-hours entry points to the health care system.

Even when aided by a decision support tool telephone triage lacks the nonverbal information obtained through visual cues, vital signs and physical exam findings of a face-to-face consultation. This has been argued to lead to poor agreement between telephone triage decisions and face-to-face triage.

A systematic review concluded that 10% of telephone-triage contacts were "unsafe". This proportion was increased to 50% if including only studies using simulated high-risk patients. Likewise, inappropriate response decisions at telephone triage leads to significantly increased time to reperfusion therapy for patients with myocardial infarction, illustrating the direct negative consequence of incorrect triage. Twelve to nineteen percent of patients with sSAH have been found to be misdiagnosed on their first face-to-face contact with a health care professional. Initial misdiagnosis of sSAH has also been found to result in delayed specialized treatment and neurologic deterioration.

In addition to being rare, the clinical presentation of sSAH may vary. Some studies report as little as 40% of patients to have presented with classic textbook symptoms. Half are initially in an intact neurological state, and only 7-12% of those actually presenting with thunderclap headache have sSAH.

Hence, an electronic decision support tools such as the Danish Index may be of good use in telephone visitation of patients with sSAH, but the association between symptoms and time to neurosurgical admission remains unknown.

Primary aim:

The primary aim of this study is to describe time from first contact to the Copenhagen Emergency Medical Communications Center to admission to a neurosurgical department in patients with spontaneous subarachnoid haemorrhage. Further, to identify predictors for short time to neurosurgical department.

Secondary analyses:

• The proportion of patients with ´acute severe headache´, that are admitted to a neurosurgical department within four hours of initial contact to the EMCC.
• Sensitivity and positive predictive value (PPV) of the Danish Index category ´acute severe headache´ for sSAH.
• Sensitivity and PPV for sSAH of symptoms other than ´acute severe headache´ in the Danish Index.
• The proportion of emergency calls from sSAH-patients, that trigger a level A response, regardless of which symptom category it is triggered by.

Hypothesis:

The hypothesis is, that less severe symptoms are associated with longer delay to neurosurgical admission.

Study design:

Register-based retrospective cohort study.

Data sources:

• Existing local sSAH-database:

All patients registered with sSAH (International Classification of Disease version 10 diagnoses DI60-DI60.9) in the Danish National Patient Register, admitted to a hospital in the Capital Region of Denmark between 1 May 2011 - 31 December 2014, aged ≥18 years on admission are registered in this database. Patients diagnoses were double-validated by medical record review. Admission date and -time was recorded arrival at a neurosurgical department.

• Copenhagen Emergency Medical Service, Emergency Medical Communications Center:

At both ´112´ and ´1813´ the Danish Index is used. Data on date, time, type of incident (sickness, trauma, assault etc.) and activated response are registered.

Some patients may have been admitted via general practitioners (GP). In the EMCC database these will appear if an ambulance has been requested, but no data on symptoms are registered in these cases.

Data correctness from ´112´ and ´1813´ are expected to be high. The same is expected for data on ambulance requests from GPs. Data may not be complete for ´112´ and ´1813´ due to missing Central Person Register numbers (CPR-numbers), but the extend is unknown.

Linking data sources:

Anyone born in, or immigrated to, Denmark are given a unique identification number in the Central Person Register; a CPR-number. CPR-numbers are used as identification throughout the public sector and in all health care registers. This allows data from multiple sources to be linked on an individual level.

Variables:

Based on the CPR-numbers from the local sSAH-database, emergency phone calls to the EMCC (both ´112´ and ´1813´) leading up to admission are identified. From 1 May 2011 data is extracted from calls to ´112´ and from January 30, 2012 data from ´1813 ´ is extracted. The inclusion period ends on 31 December 2014.

The following variables are collected from calls to the Copenhagen EMCC:

• Date and time of phone call.
• Main complaint/ the complaint that determined the initiated response.
• Any supplementary/ secondary complaints registered.
• Activated response.
• Number of calls within the last four weeks.
• Date, time and primary complaint of previous calls.
• For ambulance requests from GP´s: Date and time. All these variables are routinely collected from each call. To determine the sensitivity and the PPV of symptoms the total number of callers with the respective primary complaint within the study period is extracted. No data on individuals are extracted for this.

Missing data:

If more than 10% of data are missing a strategy for imputation will be made.

Statistical analyses:

Normally distributed data will be reported as means with 95% confidence intervals, skewed data as median values with range and inter-quartile-range (IQR); proportions are reported with 95% confidence intervals. Comparisons are made using chi-square, Mann-Whitney's or Fisher´s test as appropriate.

For the primary analysis, time from EMCC-contact to neurosurgical admission will be presented as median (range, IQR). The following variables will be tested as predictors in a logistic regression model with time >/\<4hrs as the dependent variable: symptoms (classical acute headache yes/no), age (years), gender (m/f), person who is calling in (patient vs bystander).

The sensitivity for sSAH of ´acute severe headache´ will be reported as a percentage with the numerator being sSAH-positive registered at the EMCC with ´acute severe headache´ as the primary symptom, and the denominator being all sSAH-positive with a contact to the EMCC.

The PPV of ´acute severe headache´ in the Danish Index is calculated as the number of sSAH-patients with this registered complaint in the numerator and the total number of contacts from sSAH-patients in the denominator.

The sensitivity for sSAH of symptoms other than ´acute severe headache´ will be calculated for each symptom registered as the primary complaint in a patient. Numerators being the number of sSAH-patients presenting with the respective primary complaint and the denominator being all sSAH-positive patients contacting the EMCC with the respective complaint within the study period.

The PPV for sSAH of other symptoms than ´acute severe headache´ is calculated as the number of sSAH-patients with the respective complaint in the numerator and all patients contacting the EMCC with the complaint within the study period in the denominator.

The proportion of emergency calls from sSAH-patients, that trigger a level A response, regardless of which symptom category it is triggered by, is calculated as the number of sSAH-patients triggering an A-response over the total number of sSAH-patients that contacted the EMCC.