Agreement between self‐reported and objectively assessed physical activity among out‐of‐hospital cardiac arrest survivors

Low level of physical activity is a risk factor for new cardiac events in out‐of‐hospital cardiac arrest (OHCA) survivors. Physical activity can be assessed by self‐reporting or objectively by accelerometery.


| INTRODUCTION
The World Health Organization and the European Society of Cardiology recommend that all adults should undertake regular physical activity, at moderate intensity ≥150 min/week or vigorous intensity ≥75 min/week, or a combination of both, to decrease allcause and cardiovascular mortality and morbidity (Bull et al., 2020;Chastin et al., 2015;Visseren et al., 2021).The European Association of Preventive Cardiology encourages survivors of cardiac arrest caused by a myocardial infarction to engage in physical activity to prevent new cardiac events.Supervised exercised-based cardiac rehabilitation is recommended with ≥3 days/week at moderate to high intensity ≥20 min/session during the initial months after the cardiac event (Ambrosetti et al., 2020).However, many cardiac arrest survivors do not have access to postcardiac arrest rehabilitation (Gräsner et al., 2021) including physical exercise training or counselling about physical activity (Ambrosetti et al., 2020).
Assessments of physical activity in clinical practise and research can include both self-reports by questionnaires and objective assessments with accelerometers (Skender et al., 2016).Previous studies recommend objective assessments of physical activity, preferably accelerometery, to get accurate information of physical activity among populations with cardiac disease (Kambic et al., 2021;Orrell et al., 2007).An accelerometer assesses the acceleration of body movements in different directions to interpret the amount and intensity of the physical activity (Migueles et al., 2017).Through various algorithms, raw accelerometer data are transformed into activity counts.Cut-off points of intensity thresholds identified from the number of activity counts per minutes (cpm) makes it possible to assess time spent in moderate and vigorous intensity physical activity per day (Arvidsson et al., 2019).
Self-reports are easy to acquire in clinical settings compared with accelerometery.The correlation between physical activities assessed with self-reports and objectively with accelerometery, however, range from weak to moderate (Skender et al., 2016), and in adults with heart disease, self-reported assessments overestimated moderate and vigorous intensity physical activity (Kambic et al., 2021).
The level of physical activity and adherence to physical activity guidelines among out-of-hospital cardiac arrest (OHCA) survivors is unknown (Ambrosetti et al., 2020;Bull et al., 2020;Visseren et al., 2021).While objective measures are often considered as reference standard for assessment of physical activity (Innerd, 2020), selfreports are often the only feasible method in many settings.In the large international Targeted Hypothermia versus Targeted Normothermia after Out-of-Hospital Cardiac Arrest (TTM2) trial, two questions for self-reported physical activity were included as a part of the cardiovascular risk assessment, based on the recommendations for primary and secondary prevention (Lilja et al., 2020).Although the correlation between self-reports and objective assessments has been investigated in many populations, there is currently limited knowledge in specific patient groups (Skender et al., 2016) including OHCA survivors.
This study aimed to investigate the agreement between selfreported and objectively assessed physical activity among OHCA survivors.Our hypothesis was that self-reported levels of physical activity would show moderate agreement with objectively assessed levels of physical activity among OHCA survivors.

| Study design
This cross-sectional substudy on detailed assessments of physical activity is part of the international multicenter Targeted Hypothermia versus Targeted Normothermia after Out-of-Hospital Cardiac Arrest (TTM2) trial, (Dankiewicz et al., 2019(Dankiewicz et al., , 2021)), where 8 of 61 TTM2sites participated.The substudy was registered at ClinicalTrials.gov,NCT03543332 and the protocol was published (Heimburg et al., 2021).A summary of modifications from the original protocol (Heimburg et al., 2021) are described in the Supporting Information.

| Participants
OHCA with a presumed cardiac or unknown cause of arrest were randomized to targeted hypothermia at 33°C or targeted normothermia with early treatment of fever (body temperature greater or equal to 37.8°C) (Dankiewicz et al., 2019(Dankiewicz et al., , 2021)).There were no differences between the two temperature groups (33°C/normothermia) in the TTM2-trial regarding primary or secondary outcomes (Dankiewicz et al., 2021).The main exclusion criteria for the TTM2-trial were unwitnessed cardiac arrest with an initial rhythm of asystole and temperature on admission <30°C, (Dankiewicz et al., 2019).Additional exclusion criteria for this substudy were >80 years of age, a Clinical Frailty Scale (Rockwood et al., 2005) score >7 before the cardiac arrest, a prearrest dementia diagnosis, inability to speak the local language well enough to answer the questions without an interpreter, active drug abuse, and wheel-chair bound (Heimburg et al., 2021).

| Procedure
Six months after OHCA, survivors in the TTM2-trial participated in a follow-up (Lilja et al., 2020).At this follow-up, selected sites in Sweden, Denmark, and the United Kingdom invited participants to this physical activity substudy that included more detailed assessments but for a subset of the TTM2 participants.For the OHCA survivors who consented to participate in the substudy, an additional follow-up meeting was scheduled within 4 weeks after the main study follow-up (approximately 7 months after the OHCA).At the additional follow-up, the survivor received an accelerometer and a questionnaire of two questions for self-reported physical activity.
Participants were instructed to wear the accelerometer for 7 consecutive days as an objective assessment of physical activity and live as usual.Information was provided both orally and in written format and included information on how to contact the study team during the week wearing the accelerometer, if necessary.This was considered important as a strategy to minimize the risk of missing data (Heimburg et al., 2021).After 7 days of accelerometer data acquisition, the OHCA survivors answered the questionnaire about physical activity for these specific 7 days.The accelerometer and the questionnaire were subsequently returned in a prepaid envelope.Participants were enroled between July 2018 and January 2021.
This substudy conforms to the principles outlined in the Declaration of Helsinki (World Medical, 2013).The TTM2-trial (EPN-reference number 2015/228) and this substudy (EPNreference number 2017/933) were approved by the Regional Ethics Committee at Lund University, Sweden and corresponding ethics committees in Denmark and the United Kingdom.All included participants gave their written and oral informed consent.

| Self-reported
The OHCA survivors answered two questions to report the number of days during the last week they reached (a) moderate intensity physical activity of ≥30 min and (b) vigorous intensity physical activity of ≥20 min.
Possible answers ranged from 0 to 7 days (Haskell et al., 2007).The questions were followed by examples of moderate and vigorous intensity physical activities.A brisk walk that leads to a moderate level of effort and noticeably accelerates the heart rate illustrates moderate intensity physical activity.Vigorous intensity physical activity is exemplified by jogging that causes rapid breathing and a substantial increase in heart rate (Haskell et al., 2007) (Supporting Information: Table A).These two questions are originally from the national quality registry after myocardial infarction in Sweden (Swedeheart-Sephia, 2019) and the questions were also used in the main TTM2-trial trial follow-up (Lilja et al., 2020).These self-reported questions have not been previously validated in OHCA survivors.

| Objective assessment
The accelerometer ActiGraph GT3X-BT (ActiGraph) was used to assess minutes in moderate and vigorous intensity physical activity per day.The ActiGraph has acceptable validity to assess physical activity among individuals with chronic diseases including coronary artery disease (Van Remoortel et al., 2012) but has not previously been used for OHCA.The accelerometer was attached to the right hip with a rubber strap.We chose the right hip as a placement.There is currently no consensus regarding the best practices for analysing wrist-worn accelerometer data for accurate assessment of minutes spent in moderate and vigorous intensity physical activity (Gao et al., 2021).
For extraction and analyses of raw data collected by the ActiGraph, the ActiLife software version 6.13.3 was used.A valid day was defined as a minimum of 10 h wear time (Arvidsson et al., 2019;Migueles et al., 2017).The cpm cut-off points for vector magnitude (VM), chosen for this study, were 2690-6166 cpm for moderate and ≥6167 for vigorous intensity physical activity (Sasaki et al., 2011) (see Table 1).
Sociodemographic and medical characteristics for the OHCA survivors were obtained from the TTM2 database.

| Statistical analysis
Descriptive statistics are presented with numbers and percentages for binary and categorical variables, and for continuous variables as mean and standard deviation (SD) when normally distributed, or median and quartiles [q25:q75] when nonnormally distributed.
To enable comparisons of self-reported moderate and vigorous intensity physical activity (Supporting Information: Table A) and objectively assessed moderate and vigorous intensity physical activity, only OHCA survivors with 7 valid days of accelerometer assessment were included in the analyses.For comparisons of the number of days the Wilcoxon signed rank test was used.
The OHCA survivors were then categorized in three ordered groups predefined as low, moderate, and high level of physical activity (Supporting Information: Table B) (Heimburg et al., 2021).To investigate the correlation and agreement between the three categories of self-reported and objectively assessed levels of physical activity Spearman's rho correlation coefficient (r s ) was used, and the agreement was investigated by Cohen's weighted kappa (k).
To further investigate the agreement between the two assessments of physical activity, cross tables were used.All analyses were conducted using SPSS version 27 (IBM Corp).

| RESULTS
A flow chart of the study inclusion is presented in Figure 1.Of eligible OHCA survivors, 106 of 183 (58%) were included in the substudy on physical activity.Of these, 87 of 106 (82%) had valid accelerometer data predefined as a minimum of 4 days wear time for 10 h a day.The final sample consists of 49 of 106 (46%) OHCA survivors with both self-reports and 7 days of valid accelerometer assessment to allow for comparisons.
There were no differences between all eligible OHCA survivors and the OHCA survivors with 7 valid days of accelerometery regarding age, sex, hospital length of stay, cognitive function, or self-reported fear of movement.Myocardial infarction was the most common cause of OHCA and one-fourth had normal ejection fraction of the heart in both groups (Table 2).In addition, there were no T A B L E 1 Settings in ActiLife software version 6.13.3 for analysing raw data from the accelerometer assessments regarding wear time validation, algorithms, and filters in out-of-hospital cardiac arrest survivors included in the physical activity substudy.| 147 significant differences in moderate or vigorous intensity physically activity days by self-reporting between all 104 substudy included OHCA survivors, the 85 with ≥4 valid accelerometer days and the 49 included survivors with 7 valid accelerometer days (Supporting Information: Table C).
For the 49 included participants, more days of moderate intensity physically active days were registered when assessed by the self-report compared with accelerometery (in median 5 days [3:7] vs. in median 3 days [0:5]; p < 0.001).The correlation between days in moderate intensity physical activity by self-reports and objectively Flowchart of inclusion and exclusion in the physical activity substudy of the Targeted hypothermia versus targeted normothermia after out-of-hospital cardiac arrest (OHCA) (TTM2) trial.
T A B L E 2 Characteristics among out-of-hospital cardiac arrest (OHCA) survivors included in the physical activity substudy and the OHCA survivors in the analysis with both self-reports, and 7 valid days of accelerometer assessment.assessed by accelerometery was sufficient (r s = 0.336, p = 0.018), and the agreement was fair (k = 0.269, 95% CI 0.119, 0.330, p = 0.001) (Table 3).The correlations are presented in more detail in Table 4.
The categorization of self-reported versus objectively assessed physical activity (Supporting Information: Table B) showed that 26% versus 65% of the OHCA survivors had low level of physical activity, 41% versus 29% had moderate level of physical activity and 33% versus 6% had high level of physical activity.The self-reports and the objective assessments showed that 24% of the OHCA survivors were categorized in the low group by both assessments, 14% in the moderate group and 4% in the group with high level of physical activity (Table 6).The correlation between the categorized selfreported and objectively assessed physical activity in OHCA survivors was sufficient (r s = 0.340, p = 0.017) and the agreement was slight (Cohen's k = 0.201, 95% CI 0.053, 0.349, p = 0.010).

| DISCUSSION
The main finding of this study was that OHCA survivors consistently reported more physically active days compared with the results of the accelerometer assessments during the 7 assessed days.Our hypothesis was rejected as the self-reported physically active days only correlated sufficiently and agreed fairly and none to slightly with objectively assessed active days.
In this study, we have used two questions for self-reported moderate and vigorous intensity physical activity in OHCA survivors.
Answers were compared with the objectively assessed moderate and vigorous physical activity analysed by cut-off points of a highly accurate diagnostic instrument, the accelerometer, as a reference standard (Innerd, 2020).Our results are similar to a systematic review about physical activity questionnaires and accelerometery that showed weak to moderate correlations (Skender et al., 2016).
The OHCA survivors consistently reported more physically active days compared with the results of the accelerometer assessments and Kambic et al. (2021) found the same results in patients with coronary artery disease.Clinicians and researchers thus need to pay caution when interpreting results from physical activity questionnaires.In our study, there were different patterns in the correlations between moderate and vigorous intensity physical activity by selfreports and those objective assessments.Few OHCA survivors were physically active in vigorous intensity.A quarter of the OHCA survivors reported low physical activity level by both assessments.
Our data suggest that self-reports could be used to identify OHCA T A B L E 3 Correlations by Spearman's rho (r s ) and agreements by Cohen's weighted kappa (k) between days of self-reported moderate and vigorous intensity physical activity and objectively assessed moderate and vigorous intensity physical activity by accelerometers in out-of-hospital cardiac arrest survivor with self-reports and 7 valid days of accelerometer assessment.survivors with low level of physical activity who could benefit from physical activity interventions.Survivors that report moderate or high levels of physical activity may need more detailed assessment to confirm if they indeed reach the self-reported physical activity level.
The current recommendation on physical activity (Bull et al., 2020;Visseren et al., 2021) is based on epidemiological associations between self-reported physical activity and health outcomes.In this study, 74% of the OHCA survivors were categorized in the groups moderate and high level of physical activity according to self-reports, and thus considered to fulfil the recommendations for primary prevention on physical activity.Only one-third of the OHCA survivors reached this recommendation according to accelerometery.
Epidemiological relationships based on objective assessment might result in differentiated recommendations for level of physical activity but are currently lacking and is not included in guidelines.A previous study by Troiano (2008) suggests that more than 150 min of moderate intensity physical activity per week assessed by selfreports and less than 150 min of moderate intensity physical activity per week assessed by an accelerometer may provide a significant health benefit.
The cpm cut-off points used for the accelerometer data were not adjusted to fit the OHCA survivors specifically, and the group was heterogeneous.More specifically, the cpm cut-off points were not age or ability adjusted and may not accurately represent the amount of moderate and vigorous intensity physical activity in all OHCA survivors.Their age ranged from 35 to 76 years old, and threequarters had mild to severely reduced cardiac ejection fraction during hospitalization.The median age of reference group (Sasaki et al., 2011) was 28 years compared with 64 years in our study.Higher age may affect the OHCA survivors' physical function (Heimburg et al., 2022).In addition, the maximal exercise capacity among OHCA survivors was found to be low in a previous study (Boyce et al., 2017).
T A B L E 4 Number of participants with number of self-reported days in 30 min in total of moderate intensity physical activity a day and number of days of objectively assessed moderate intensity physical activity 30 min in total a day by accelerometer (Sasaki et al., 2011)  T A B L E 5 Number of participants with number of self-reported days in 20 min in total of vigorous intensity physical activity a day and number of days of objectively assessed vigorous intensity physical activity 20 min in total a day by accelerometer (Sasaki et al., 2011) during the same week.This indicates that lower cpm cut-off points could be used to better reflect the OHCA survivors' moderate and vigorous intensity physical activity, since the recommendations need to match the physical fitness of the individual (Lee et al., 2003).An evaluation of the maximum exercise capacity in OHCA survivors could help identify the relative intensity and adapt the cpm cut-off points for accelerometery data (Lee et al., 2003;Siddique et al., 2020).This is an important target for future studies.We cannot exclude that the accelerometer underestimated the true level of moderate and vigorous physical activity in our population.This could also be a possible explanation why 42 of 49 had 0 days of vigorous physical activity days when assessed objectively (Table 5).In future studies, light intensity physical activity could also be of interest to analyse.
A strength of this study was that the study was performed within a large well-designed randomized controlled trial according to a specified protocol with high rates of follow-up at 6 months.The OHCA survivors received verbal as well as written information, to minimize the risk of flawed or missing data.Only two participants did not answer the two questions for selfreporting and 87 wore the accelerometer for a minimum of 4 days.Nearly all included OHCA survivors had worn the accelerometer to some degree during the week.To allow for comparisons with the self-reports, we only included the OHCA survivors with 7 valid days of accelerometery.Since the eligible survivors and those included in the analysis were similar, we assume that the results are representative.To improve the number of valid days of accelerometer assessment, we should have reinforced the importance to the OHCA survivors and their relatives of wearing the accelerometer for at least 10 h a day.
The cpm cut-off points by Sasaki (Sasaki et al., 2011) were developed using a treadmill protocol that involved walking and running in a controlled laboratory setting.The OHCA survivors in this study wore the accelerometer during a week where not all the minutes of moderate and vigorous intensity physical activity would have been accumulated by walking or running.It is unclear if and how for example weight training and biking were registered by the accelerometer (Migueles et al., 2017).A limitation is also that we do not know if the participants were engaging in a greater or lesser degree of moderate and vigorous intensity physical activity due to a reactive response from wearing the accelerometer (Prince et al., 2015).We asked specifically about the recent week and not a typical week.
Participation in exercised-based cardiac rehabilitation reduces cardiovascular mortality, recurrent cardiac events, and hospitalization (Dibben et al., 2021;Dibben et al., 2023).Despite this knowledge, we note that only 38% of the participating OHCA survivors in this study were provided exercised-based cardiac rehabilitation.
This study is to our knowledge the first study investigating physical activity using both self-reports and accelerometers as assessment of physical activity in OHCA survivors and provides useful information about how one could interpret physical activity data among OHCA survivors.More research that combines the strengths of both self-reports and objective assessments T A B L E 6 Cross-tabulation of physical activity categorized in three groups of low, moderate, and high level assessed by self-reporting and accelerometry during 7 days.
n 1 day, n 2 days, n 3 days, n 4 days, n 5 days, n 6 days, n 7 days, n Total, n during the same week.n 1 day, n 2 days, n 3 days, n 4 days, n 5 days, n 6 days, n 7 days, n Total, n