Analysis of data for comorbidity and survival in out-of-hospital cardiac arrest

The data presented in this article is supplementary to the research article titled ”Comorbidity and survival in out-of-hospital cardiac arrest” (Hirlekar et al., 2018). The data contains information of how Charlson Comorbidity Index (CCI) is calculated and coded from ICD-10 codes. Multivariable logistic regression was used in the analysis of association between comorbidity and return of spontaneous circulation. We present baseline characteristics of patients found in VF/VT. All patients with non-missing data on all baseline variables are analyzed separately. We compare the baseline characteristics of patients with and without complete data set. Analysis of when comorbidity was identified in relation to outcome is also shown.


Subject area
Cardiac arrest.

More specific subject area
Epidemiology of cardiac arrest.

Type of data
Tables and figures. How data was acquired Data analysis from the National Patient Registry (NPR) and the Swedish Registry of Cardiopulmonary Resuscitation (SRCR).

Data format
Analyzed.

Experimental factors
Data was analyzed to investigate whether comorbidity is associated with outcome in out-of-hospital cardiac arrest (OHCA).

Experimental features
Nationwide retrospective and population-based cohort study of patients with bystander witnessed OHCA.

Data source location
A nationwide cohort study in Sweden.

Data accessibility
The analyzed data are presented in this article.

Value of the data
The data provides information about how ICD-10 codes were used to create the categories in Charlson Comorbidity Index (CCI).
The data provides information of association between comorbidity and return of spontaneous circulation (ROSC).
The data provides information of the baseline characteristics of patients found in VF/VT. The data provides comparison of patients with complete data and patients with missing data. The data shows association between comorbidity and survival depending on when the comorbidity condition was identified.

Data
The data contains information of how ICD-10 codes were used to create the categories in Charlson Comorbidity Index (CCI) as shown in Table 1 [2]. Baseline characteristics of patients found in VF/VT are shown in Table 2. Comparison of baseline characteristics of cases with and without complete data are shown in Table 3. The association between missingness and other baseline characteristics, CCI and survival for all patients are shown in Table 4. The relation between comorbidity and the chance of any return of spontaneous circulation (ROSC) is shown in Fig. 1 and the corresponding figure for ROSC at hospital admission is shown in Fig. 2. Association between various aspects of comorbidity and 30-day survival among all patients with complete cases on all baseline characteristics (no missing) is shown in Fig. 3. In Figs. 4-6, we present the association between comorbidity and 30-day survival in relation to the time of identification of the comorbidity condition, as follows: Patients for whom comorbidity condition were identified 3-5 years before OHCA (Fig. 4); first identified within 1 year before OHCA (Fig. 5); and comorbidity condition within 1 year before OHCA irrespective of identification 1-5 years before OHCA (Fig. 6).

Experimental design, materials, and methods
We conducted an analysis of data from the Swedish Registry for Cardiopulmonary Resuscitation (SRCR) which was collected between 2011 and 2015. We linked the data from SRCR with data from the National Patient Registry (NPR). The NPR includes data on diagnoses and surgical procedure codes from hospitals and specialist clinics [3]. We had data on health disorders during the five years Table 1 Charlson comorbidity index according to ICD-10 codes.

Disease
ICD 10 Weight point  preceding the OHCA. We measured comorbidity with CCI as shown in Table 1. If the patient had any mention of an ICD-10 code listed in the NPR which was included in the category definition of CCI, the patient would get a weight point with the maximum possible score of 29.

Study design
We performed a nationwide population-based cohort study of patients with bystander witnessed OHCA which was designed to evaluate if there were any association between comorbidity and outcome. We included all cases with bystander-witnessed OHCA who were Z18 years of age. Unwitnessed and cases only witnessed by Emergency Medical Service (EMS) were excluded. For details, see Ref. [1].

Statistical analysis
We used logistic regression and made adjustments for year of OHCA, age, sex, initial rhythm, location, bystander cardiopulmonary resuscitation (CPR), mechanical chest compression, aetiology, adrenalin treatment, intubation, anti-arrhytmics, time to CPR and EMS response time. Fisher's exact test was used to test for difference between groups regarding dichotomous variables and Mann-Whitney U test for ordered/continuous variables in the baseline characteristics. We used multiple imputation for the multivariable analysis and the missing data were assumed to be missing at random (MAR). To exclude that the missing data pattern were missing completely at random (MCAR) we compared cases with no missing data with incomplete cases (Table 3) and found several major differences. The assumption of a MAR pattern was indicated to be valid by examination of the associations between missingness of each variable with other variables (Table 4). We analysed also complete cases without multiple imputation (Fig. 3). Outcome endpoint was not imputed and thus only patients with any ROSC or ROSC at hospital admission were included in the analysis in Figs. 1 and 2.   4,612). * The comorbidities of moderate or severe liver disease and AIDS/HIV were not analyzed in the specific comorbidity conditions above, due to low prevalence (0.6% and 0.1%, respectively).

Fig. 2.
Patients with ROSC at hospital admission and relation to comorbidity (n¼ 3,690). * The comorbidities of moderate or severe liver disease and AIDS/HIV were not analyzed in the specific comorbidity conditions above, due to low prevalence (0.6% and 0.1%, respectively).  . Patients for whom comorbidity conditions were identified 3-5 years before OHCA. * The comorbidities of moderate or severe liver disease and AIDS/HIV were not analyzed in the specific comorbidity conditions above, due to low prevalence (0.2% and 0.1%, respectively). . Patients for whom comorbidity was first identified within 1 year before OHCA. * The comorbidities of moderate or severe liver disease and AIDS/HIV were not analyzed in the specific comorbidity conditions above, due to low prevalence (0.2% and 0.0%, respectively).