Apples to apples: can differences in out-of-hospital cardiac arrest incidence and outcomes between Sweden and Ireland be explained by core Utstein variables?

Background Variation in reported incidence and outcome based on aggregated data is a persistent feature of out-of-hospital cardiac arrest (OHCA) epidemiology. Objective To investigate the extent to which patient-level analysis using core ‘Utstein’ variables explains inter-country variation between Sweden and the Republic of Ireland. Methods A retrospective cross-sectional comparative study was performed, including all Swedish and Irish OHCA cases attended by Emergency Medical Services (EMS-attended OHCA) where resuscitation was attempted from 1st January 2012 to 31st December 2014. Incidence rates per 100,000 population were adjusted for age and gender. Two subgroups were extracted: (1) Utstein - adult patients, bystander-witnessed collapse, presumed medical aetiology, initial shockable rhythm and (2) Emergency Medical Service (EMS)-witnessed events. Multivariable logistic regression analysis was used to identify predictors of survival following multiple imputations of data. Results Five thousand eight hundred eighty six Irish and 15,303 Swedish patients were included. Swedish patients were older than Irish patients (median age 71 vs. 66 years respectively). Adjusted incidence was significantly higher in Sweden compared to the Republic of Ireland (52.9 vs. 43.1 per 100,000 population per year). Proportionate survival in Sweden was greater for both subgroups and all age categories. Regression analysis of the Utstein subgroup predicted approximately 17% of variation in outcome, but there was a large unexplained ‘country effect’ for survival in favour of Sweden (OR 4.40 (95% CI 2.55–7.56)). Conclusions Using patient level data, a proportion of inter-country variation was explained, but substantial variation was not explained by the core Utstein variables. Researchers and policy makers should be aware of the potential for unmeasured differences when comparing OHCA incidence and outcomes between countries. Electronic supplementary material The online version of this article (10.1186/s13049-018-0505-2) contains supplementary material, which is available to authorized users.

The OHCA register is based in Gothenburg, the West Coast in Sweden and collaborates with the Emergency Ambulance Services (EMS) in all 21 counties. The OHCA Steering Group members are from several counties in Sweden but the there is a working group situated in Gothenburg.

Funding
OHCAR is funded by the National Ambulance Service and the Pre-Hospital Emergency Care Council The register is funded by The Swedish Association of Local Authorities and Regions

Patient and Event Variables
In Ireland, statutory Emergency Medical Services use a standardised Patient Care Report (PCR) which contains a section dedicated to data collection for OHCAR. Specially designed OHCA envelopes have also been provided to each ambulance station. In the event of OHCA, practitioners place completed PCRs in OHCA envelopes. On a monthly or fortnightly basis, envelopes are collected together with all PCRs from each station. All PCRs are scanned and stored digitally and cases in OHCA envelopes are given priority in the scanning process to facilitate OHCAR. OHCAR variables are manually entered onto an electronic database. This database is then forwarded to OHCAR together with a scanned copy of each PCR for case-by-case validation.
There is an standard web template which will be documented in connection to a treated OHCA. The EMS crew is responsible for filling in the web template. All data is collected in a database.

Time variables
Dispatch data (i.e. time variables) is collected directly from the ambulance dispatch centre by registry staff The time variables are availible from the dispatch center and from the EMS medical journal.

Outcome data
Outcome data is requested by registry staff from receiving hospitals Outcome data is availible from medical journals from EMS and in-hospital and also from Statistics Sweden. How are missing cases identified?
Cases that are not placed in OHCA envelopes are not processed through the OHCAR data collection system and must be identified separately. Missing case identification is performed on a monthly basis and repeated on an annual basis to capture delayed reports. First, a search of the digital scanning archive is performed based on the 'Chief Complaint' field in the PCR using the word 'arrest'. Reported cases are excluded from the results and then the digital scan of the PCR associated with each call found is viewed. Missed OHCAR cases are identified and captured during the viewing process. Next, emergency control data is filtered to identify all calls with an AMPDS© designation of 'ECHO' at the time of resource deployment. A further seventeen 'DELTA' codes that may signify arrest occurred are included in the filter. Reported calls are then excluded from the filtered list. PCRs on the filtered list are then viewed and remaining unreported OHCAR cases are identified.
The EMS crew makes regular retrospective observations with the aim of searching undocumented OHCA. The searching procedure is performed by a digital searching programme, and manually searching if medical journals have been documented in papers. Missed OHCA cases are identified and imported to the database. Missed OHCAs are labelled in order to identify them as retrospective data.

Table S1
Comparison of Irish and Swedish OHCA resuscitation registries (continued)

Sweden Ireland
How is data quality assured in the registry?
Data is received by OHCAR in an electronic database together with a scanned copy of each PCR. Each electronic entry is checked by OHCAR staff against the data in the PCR to ensure accuracy of manual data entry. Checked data is then forwarded to the OHCAR manager who performs a random check of cases before finally adding data to the master OHCAR database. For cases that are identified through the missing case identification process, data is extracted from the scanned PCR by OHCAR staff and manually entered onto an electronic database by OHCAR staff. The 'missing' database is then forwarded to the OHCAR manager, who validates each entry using the corresponding scanned PCR. Once validation of the missing cases is complete, they are added to the master OHCAR database.
Registry data is compared to documented data in medical journals using variables such as incidence, place OHCA occurred, treatment and survival.