Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature

HLC 2293 No. of Pages 18

REVIEW

Heart, Lung and Circulation (2017) xx, 1–18 1443-9506/04/$36.00 http://dx.doi.org/10.1016/j.hlc.2017.01.005

Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature Muaddi Alharbi, RN, BN, MN *, Adrian Bauman, MBBS, MPH, PhD, FAFPHM, Lis Neubeck, RN, PhD, Robyn Gallagher, RN, MN, PhD Charles Perkins Centre, University of Sydney, Sydney, NSW, Australia Received 9 June 2016; received in revised form 27 December 2016; accepted 8 January 2017; online published-ahead-of-print xxx

Background

Assessment of physical activity (PA) for cardiac rehabilitation (CR) participants is critical to monitor changes. However, the validity and reliability of PA measures to assess PA throughout the day, not only during exercise training, is poorly investigated.

Aim

To establish a reliable and valid measure to assess overall PA in CR participants.

Methods

A narrative literature review was performed based on a systematic search of EMBASE, CINAHL, MEDLINE and PubMed databases. Eight studies comparing two or more PA measures with at least one direct measure met the inclusion criteria.

Results

Methodological designs were heterogeneous. Correlations and levels of agreement between self-reported measures and direct measures were weak to moderate, while the correlations between direct measures were high. Of the direct measures, the SenseWear armband had the highest validity, and the PA diary and MobilePAL questionnaires performed better than other self-reported PA measures.

Conclusion

Direct measures were more valid and reliable than self-reported measures. No recommendation for a definitive PA measure was made due to lack of strong evidentiary support for one PA measure over another. There is a need for accurate measures of overall PA in evaluating current and changing PA levels following CR.

Keywords

Cardiac rehabilitation  Physical activity assessment  Questionnaire  Activity monitor  Validity  Reliability

Introduction Cardiovascular disease (CVD) is a major cause of death and disability [1]. However, both cardiac and total mortality may be reduced by cardiac rehabilitation (CR) [2] particularly when supervised, structured exercise is included. The benefits of exercise include stabilisation or reversal of the atherosclerotic process and psychological well-being [3,4], particularly when exercise achieves recommended levels

over the whole day, not only during supervised exercise training at CR but also at home [5]. However, non-adherence to PA recommendations remains a major concern in the CR population [6], and methods are needed to accurately quantify overall PA, not just in supervised CR sessions. Accurate quantification of PA in CR is crucial [7]. Accuracy is important to monitor trajectories of PA, assess the effectiveness of interventions, examine dose-response relationships, and define which PA dimensions (i.e. frequency,

*Corresponding author at: Office 2W12/Level 2/Building 17, Charles Perkins Centre University of Sydney, NSW, Australia, 2006. Mobile: +61 4 202 31 229., Email: [email protected] © 2017 Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) and the Cardiac Society of Australia and New Zealand (CSANZ). Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

HLC 2293 No. of Pages 18

2

M. Alharbi et al.

duration and intensity) are important for specific health outcomes [7]. Nonetheless, measuring PA in the CR setting is challenging because there are substantial variations in the population including age, diagnoses, disease severity, and stage of recovery [7]. In general, self-reported and direct measures are proposed in the literature on measurement of PA and their strengths, weaknesses, reliability and validity are comprehensively discussed in a number of reviews [7–9]. Self-reported measures that assess overall PA are the most frequently used approaches in CR due to their practicality and cost-effectiveness [7]. However, most self-reported PA measures for cardiac patients have great variability, low validity and reliability, and are typically suitable for epidemiologic studies rather than CR settings [7]. Direct measures of PA are likely to be superior to indirect measures in minimising over- or under-reporting. Of direct measures, accelerometry technologies have distinct benefits in continuously measuring activities of daily living, metabolic expenditures (METs), and step counts [10,11]. Use of such measures enables clinicians to monitor the progress of the patients’ activity levels remotely (i.e. outside of CR settings) and intervene in a timely way. For instance, step counts and active minutes tracked per day could be used to evaluate if the patient attains the CR daily PA recommendation (10,000 steps/day or 30 minutes or more of moderate to vigorous physical activity [MVPA]) [10]. To date, there is a substantial body of literature related to the validity and reliability of PA measures in healthy people [8,9,12]. There are, however, far fewer validation studies for people with existing CHD in CR settings [13,14]. Hence, achieving a precise measurement of overall PA in cardiac patients during and following rehabilitation remains a significant clinical and public health issue [7]. The aim of this study is to establish a reliable and valid measure to assess overall PA in CR participants by performing a narrative literature review that compares two or more PA measures with at least one direct measure.

participants who were commencing or had completed a CR program were eligible. Studies that compared two or more PA measures with at least one direct measure were also eligible. Finally, the primary outcomes were validity and/or reliability on self-reported and/or direct measures of PA that examined the PA construct, including measures such as frequency, intensity and duration of activity rather than physical function or fitness in a CR setting.

Screening of Search Findings The initial search output was 461 studies (summarised [3_TD$IF]in Figure 1). Where there was uncertainty regarding eligibility, the full text was evaluated and a decision made following discussion between two team members (MA and RG). After assessment of the abstract of each study and a hand search of the reference list, 13 potentially relevant studies were identified and were subject to a full-text review. Of the 13 studies, 5 did not meet the inclusion criteria and were excluded from this review.

Quality Appraisal The Kowalski et al. (2012) [8] checklist was used to assess the quality of the reviewed studies (see Appendix A). It consists of 21 items (nine quality of reporting criteria, three external validity criteria, and nine internal validity criteria) with a maximum score of 22 points. Overall, total scores of the reviewed studies ranged from 10 to 18, with a mean total score of 14.3. The external validity ratings of most studies were moderate with a range of items from one to two and a mean external validity score of 1.6. Similarly, the internal validity ratings of most studies were moderate with a range of items from two to seven and a mean internal validity score of 4.8.

Results

Methods

The main characteristics of the eight studies of PA measures in CR are summarised in Table 1, validity outcomes are synthesised in Table 2, and reliability outcomes in Table 3.

Search Strategy

Participants and Setting

A search strategy was developed in consultation with the health librarian. The following electronic databases were searched: EMBASE; CINAHL; MEDLINE; and PubMed. A search of Google Scholar and a hand search of the reference lists in the selected studies were also performed to identify further relevant studies. The key search terms included: (1) ‘‘physical activity”, or ‘‘exercise”; (2) ‘‘cardiac rehabilitation” or ‘‘secondary prevention”; (3) ‘‘survey”, ‘‘measure”, ‘‘instrument”, ‘‘questionnaire”, ‘‘monitor” or ‘‘track”; (4) ‘‘validity”, or ‘‘development”. Eligible for review were primary research studies published in English that reported evidence for validity and/ or reliability on self-reported and/or direct measures of PA among CR participants including randomised controlled trials and validation studies. Only studies including adult

In total, the eight studies involved 397 participants, although sample sizes varied, ranging from 9 [15] to 73 [16]. The mean age ranged from 57.6 [17] to 72.1 [18] years and mean BMI ranged from 27 [14] to 32.8 [11] kg/m2[34_TD$IF], with the minority of participants, female (21.5%). Cardiac diagnoses of participants were diverse, but the predominant diagnosis was myocardial infarction. Most of the reviewed studies indicated a low PA population, with participants having an average 7134  2808 steps/day [15]; spending 95  76 [18] or 119.5 [16] minutes of MVPA/week; a minority (only 3%) achieving the recommended PA levels of 30 minutes of MVPA/day [13]; and reaching a PA energy expenditure of 1.69  0.1 METs/minutes [19]. Recruitment of participants in relation to CR enrolment varied: before commencing CR [18]; current enrolments [11,13,15,16,19]; CR completed [17]; and

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

HLC 2293 No. of Pages 18 Measuring Overall Physical Activity for Cardiac Rehabilitation Participants

3

[3_TD$IF]Figure 1 Flow chart of study.

one study recruited participants who were either currently enrolled in a CR program or who had previously benefited from CR [14]. The CR programs which participants attended also varied considerably in exercise content, number of sessions, length of program, and stage of recovery. The study designs and settings varied. Four studies utilised criterion validity design [11,13,16,18], two studies utilised concurrent validity design [15,17], and two studies utilised both criterion and concurrent validity designs [14,19]. All were conducted in Western countries.

Physical Activity Measures Measurement of PA varied between studies and included direct measures, existing PA questionnaires, newly developed PA questionnaires and PA diaries. In total, the eight studies used 10 direct measures of which the accelerometer was the most common (n = 8); an indirect calorimetry (n = 1), and a pedometer (n = 1) were also used. All studies compared results to another PA measure concurrently, but the study by Frederix et al. (2011) [15] used a two period crossover design between the pedometer and the accelerometer for four weeks each. Activity monitor wearing time was during all waking hours in all reviewed studies. A valid day was defined as wearing the accelerometer for at least 10 hours [18,19], 11 hours [13,16], and 12 hours [14] of valid time, which was defined as activity records longer than 60 min of 0 counts [13,16,19]. All studies measured PA across at least seven consecutive wearing days (to include weekend days and weekdays) except for the study by Hertzog et al. (2007) [18] which measured PA across two periods of three consecutive wearing days including one weekend day. The

accelerometers were placed on the waist [13,16,18,19] (n = 4) and one study each for the upper arm [11], the chest [14], and the wrist [15]. There were eight self-reported PA measures across the reviewed studies, including the Health Survey for England (HSE) [13], International Physical Activity Questionnaire (IPAQ, long form) [19], and Dijon Physical Activity Questionnaire (DPAQ) [17]. Newly developed PA questionnaires (n = 3) included Mobile phone physical activity level questionnaire (MobilePAL) [19], Total Activity Measure (TAM) [16], and Swiss Physical Activity Questionnaire (SWISSPAQ) [14]. The PA diaries (n = 2) used were hand written records of time spent in PA. The recording period ranged from three [18] to seven days [14]. There was wide variability in the duration of the validation test among the reviewed studies, ranging from eight minutes [11] to four weeks [15]. Over half of the included studies were of seven days’ duration [13,14,16,17,19], however, the study by Hertzog et al. (2007) [18] was of three days’ duration. Four studies included assessment of reliability with the time between assessments ranging from seven days [16,19] to three months [18]. The length of data collection to achieve acceptable levels of reliability for measures of PA ranged from two [18,19] to four days [18].

Validity Results All eight studies used appropriate criterion measures to assess the validity of nine PA measures (Table 2). One study [13] investigated the sensitivity and specificity of the selfreported measure against the accelerometer in classifying participants who did light, MVPA. In general, the results

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

4 HLC 2293 No. of Pages 18

Study &

Sample

Direct PA measures

Self-reported PA measures

design

Bahler et al.

Size: n = 48 CR

ACTIHEART

Existing questionnaires

Newly developed questionnaires

PA diaries

N/A

Swiss Physical Activity Questionnaire

Physical Activity Diary

[14] Location:

(SWISSPAQ) Female: 23%

Switzerland

Administration mode: Self-reported

Manufacturer: Cambridge

Administration mode: Self-reported

Neurotechnology Ltd., Papworth, UK Age: 60  8

Type: Uniaxial

No. of items: 5

Time: NR

BMI: 27  4

Size and weight:

Time frame: A typical week in the

Relative/perceived intensity: Borg

0.7  3.3 cm (8 g)

previous 2 months.

rating of perceived exertion scale

Placement: Chest

Derived measures: Frequency, duration,

Modification or reference of original

and intensity of activities performed in a

version: NR

typical week Epoch length: 15 s to 1 min

Modification or reference of original version: Based on a German PA Questionnaire and developed on the basis of personal and telephone interviews, using qualitative and quantitative approaches

Memory capacity: 21 days PA measurements: Steps, active minutes, and estimates of EE Cole et al.

Size: n = 24 CR

[11]

Indirect calorimetry (IC)

N/A

N/A

N/A

provided information on EE of PA.

Location:

Female: 20.8%

SenseWearTM (SW) armband

Age: 62  8.1

Manufacturer: BodyMediaTM,

USA Pittsburgh, PA, USA BMI: 32.8  6.5

Type: Biaxial Size and weight: 8.5  5.3  2 cm (85 g) Placement: Upper arm (triceps) Epoch length: Collects data at a rate of 32 times/second Memory capacity: 28 days PA measurements: Steps, active minutes, and estimates of EE

M. Alharbi et al.

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 1 Characteristics of included studies.

Sample

Direct PA measures

Self-reported PA measures

design

Frederix

Size: n = 9 CR

Polar FA20 activity watch

Female: 44.4%

Manufacturer: Polar FA20

Existing questionnaires

Newly developed questionnaires

PA diaries

N/A

N/A

N/A

Dijon Physical Activity

N/A

N/A

et al. [15] Location: Belgium

activity watch Age: 63  9.0

Type: Uniaxial

BMI: 28.2  5.0

Size and weight: 12.5  10  7.5 cm (45 g) Placement: Wrist Epoch length: Not reported on manufacturer’s website Memory capacity: 9 weeks PA measures: Steps, active minutes, and estimates of EE A pedometer (Omron 720IT) Manufacturer: Omron 720IT, Omron Healthcare, Inc Type: Biaxial Size and weight: 4.7  7.3  1.6 cm (241 g) Placement: Waist Epoch length: Not reported on manufacturer’s website Memory capacity: 41 days PA measurements: Steps, active minutes, and estimates of EE

Guiraud

Size: n = 70 CR

MyWellness Key accelerometer

et al. [17] Location:

Questionnaire (DPAQ) Female:17%

France Age:

Manufacturer: Technogym,

Administration mode: Self-

Gambettola, Italy

reported

Type: Uniaxial

No. of items: 9

Size and weight:

Time frame: From a typical

8.5  2.0  0.7 cm (18.7 g)

week to more than 9 months

Placement: Waist

Derived measures: Overall

57.6  11.6 BMI: 27.5  3.8

assessment of PA, usual daily activities, sports and leisure activities, and rest time.

HLC 2293 No. of Pages 18

Study &

Measuring Overall Physical Activity for Cardiac Rehabilitation Participants

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 1. (continued).

5

6

Sample

Direct PA measures

HLC 2293 No. of Pages 18

Study &

Self-reported PA measures

design Existing questionnaires

Newly developed questionnaires

PA diaries

N/A

N/A

Physical Activity Diary

Epoch length: Not reported on manufacturer’s website Memory capacity: 7 days PA measurements: Steps, active minutes, and estimates of EE Hertzog

Size: n = 64 CR

RT3 triaxial accelerometer

et al. [18] Location:

Female: 14%

Administration mode: Self-reported

Age: 72.1  5.4

Time: on an hourly basis

BMI: 28.1  5.0

Relative/perceived intensity: Light,

USA

moderate, hard, and very hard Modification or reference of original version: Adapted from 7Day Activity Interview, and Bouchard and colleagues (1983) diary format Orrell et al.

Size: n = 72 CR

RT3 triaxial accelerometer

[13] Location:

Health Survey for England

N/A

N/A

Total Activity Measure (TAM). Two

N/A

(HSE) Female: 19.4%

UK Age:

Manufacturer: Stayhealthy Inc.

Administration mode:

Monrovia, CA.

Interview

Type: Triaxial

No. of items: 22

Size and weight:

Time frame: Past month

65.9  7.45 BMI: NR

7.1  5.6  2.8 cm (56.2 g) Placement: Hip or waist

Derived measures: Overall activity level

Epoch length: 1 s to 1 min Memory capacity: 7 days PA measurements: Steps, active minutes, and estimates of EE Orrell et al.

Size: n = 73 CR

[16]

RT3 triaxial accelerometer

N/A

versions: TAM1 reports the average time spent at each activity level and TAM2 reports the total time spent at each activity level

Location:

Female: 20.5%

Administration mode: Self-reported

Age: Male

No. of items: 6

UK (66.2  7.74)

M. Alharbi et al.

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 1. (continued).

Sample

Direct PA measures

Self-reported PA measures

design Existing questionnaires Female

Newly developed questionnaires

PA diaries

Time frame: A typical week

(65.9  5.96) BMI: NR

Derived measures: Total activity score, strenuous PA, moderate PA, and mild PA Modification or reference of original version: Developed from the Godin and Shepherd Leisure Time Exercise Questionnaire (GSLEQ)

Pfaeffli

Size: n = 19 CR

Actigraph GT1M

International Physical Activity

Mobile phone Physical Activity Level

Questionnaire (IPAQ)

Questionnaire (MobilePAL)

Manufacturer: Actigraph LLC,

Administration mode: Self-

Administration mode: Self-reported;

Pensacola, FL.

reported

answered via mobile phone

Type: Biaxial

No. of items: 27

No. of items: 2

Size and weight:

Time frame: Past 7 days

Time frame: Daily for a total of 7days

Derived measures: Domain

Derived measures: PA level values

specific scores for walking,

obtained for work/daytime activities &

moderate and vigorous

PA level from energy expended during

intensity

leisure/evening activities

et al. [19] Location:

Female: 13%

New Zealand Age: 59  10 years BMI: NR

3.8  3.7  1.8 cm (27 g) Placement: Hip or waist

Epoch length: 1 s to several

Modification or reference of original

mins

version: Adapted the original Java–based

N/A

mobile phone questionnaire Memory capacity: 1 year PA measurements: Steps, active minutes, and estimates of EE CR, cardiac rehabilitation; PA, physical activity; NR, not reported; BMI, body mass index; EE, energy expenditure; N/A, not applicable.

HLC 2293 No. of Pages 18

Study &

Measuring Overall Physical Activity for Cardiac Rehabilitation Participants

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 1. (continued).

7

8 HLC 2293 No. of Pages 18

Studies examining validity of physical activity measures in cardiac rehabilitation based on units or outcomes measured.

Study

Criterion measure

Test measure (s)

Time

Units

Validity analysis

Main findings

7 days

PA classifications*:

Sensitivity & specificity

Sensitivity (0.35) and specificity (0.92) for

A- light <1 episode of MVPA/

Kappa

Sensitivity (0.40) and specificity (0.56) for

1- PA classifications of light, moderate and vigorous Orrell et al.

RT3 accelerometer

[13]

Health Survey for England (HSE)

light PA 30 min/week

moderate PA

B- moderate = 1 to 5 episodes of

Sensitivity (1.0) and specificity (0.76) for

MVPA/30 min/week

vigorous PA

C- vigorous >5 episodes of MVPA/

HSE misclassified 63%

30 min/week Overall sensitivity (k = 0.04, 95% CI,

0.31 to

0.40) Agreement (k0.08, 95% CI,

0.20 to 0.36)

2- Duration of active minutes and steps counts Orrell et al.

RT3 accelerometer

[16]

Total Activity

7 days

Minutes/week

Pearson correlation

A- total >3 METs

Post-hoc analyses with

TAM2_ Min –accelerometer_Min: (range,

Student Newman–Keuls test

r = 0.25–0.32)

Measure (TAM).

TAM1_ Min –accelerometer_Min: (range, r = 0.05–0.27)

Two versions: TAM1 & TAM 2

Method bias: Time spent in activity as

B- low >3–<5 METs

measured on the accelerometer (119.5 min) was lower than time that measured on the TAM1 (381.2 min) and the TAM2 (359.1 min) C- moderate >5–<9 METs D- strenuous >9 METs Hertzog

RT3 accelerometer

et al. [18] Pfaeffli

Physical Activity

3 days

Minutes of MVPA

Pearson correlation

PA diary_Min –RT3_activity counts: (range,

7 days

Minutes/day of light, moderate and

Pearson correlation

IPAQ_Min

Average daily PA levels

Method bias: Time spent in activity as

Diary GT1M accelerometer

et al. [19]

Mobile phone

r = 0.42–0.46) accelerometer_Min: r = 0.41

vigorous PA

Physical Activity Level Questionnaire (MobilePAL) International Physical Activity

measured on the accelerometer (302 min)

Questionnaire

was higher than time measured on the IPAQ (149 min)

(IPAQ) Frederix et al. [15]

Polar activity watch

Omron

4 weeks:

pedometer

two period

Steps/day

Pearson correlation

Pedometer_steps

accelerometer_steps:

(r = 0.28)

cross-over T-test

Method bias: No significant difference between the pedometer and the accelerometer for recording daily steps (7636  3348 vs. 7134  2808, respectively)

M. Alharbi et al.

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 2

Criterion measure

Test measure (s)

Time

Units

Validity analysis

Main findings

Total Activity

7 days

MET/min/week

Pearson correlation

TAM1_METs –accelerometer_METs: (range,

A- total >3 METs

Post-hoc analyses with

3- Estimates of energy expenditure: A- Metabolic equivalents Orrell et al.

RT3 accelerometer

[16]

Measure (TAM).

r = 0.25–0.27)

Two versions: TAM1 & TAM 2

B- low >3–<5 METs

TAM2_METs –accelerometer_METs: (range,

Student–Newman–Keuls test

r = 0.03–0.32)

Bland–Altman plot

Method bias: The accelerometer activity score (mean = 563.4 METs/min) was lower than the TAM1 activity score (mean = 1693.1 METs/min) and the TAM2 (mean = 1515.1 METs/min).

C- moderate >5–<9 METs

Bland–Altman plot for TAM1 & TAM2 at time 2 (METs/min): The total activity scores for both versions were higher than the accelerometer and as the activity scores increased, the difference in the activity scores also increased, especially for the TAM1

D- strenuous >9 METs Pfaeffli

GT1M accelerometer

et al. [19]

Mobile phone

7 days

METs/min/day

Pearson correlation

IPAQ_METs –accelerometer _METs: r = 0.62

Average daily PA levels

Method bias: EE as measured in

Physical Activity Level Questionnaire (MobilePAL) International Physical Activity

accelerometer (1.69 METs/min) was lower

Questionnaire

than EE that measured on the IPAQ (531 METs/min)

(IPAQ) Bahler et al.

ACTIHEART

Swiss Physical

[14]

accelerometer

Activity

7 days

METs/hours/day

Pearson correlation

SWISSPAQ _METs

accelerometer_ METs:

r = 0.41

Questionnaire (SWISSPAQ) Physical Activity

T-test

SWISSPAQ_METs –PA diary_ METs: r = 0.41

Diary Bland-Altman plot

Method bias: The SWISSPAQ (6.14  4.46) estimated higher EE compared to ECGaccelerometry (5.08  4.34) by 1.05  4.79 METs/hours/day Bland–Altman plot: the mean difference between SWISSPAQ 1 and ACTIHEART was high with a mean bias of 1.05 METs-hours.

HLC 2293 No. of Pages 18

Study

Measuring Overall Physical Activity for Cardiac Rehabilitation Participants

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 2. (continued).

9

Study

Criterion measure

Test measure (s)

Time

Units

Validity analysis

Main findings 95% limits of agreement:

0.34 to 2.44

METs-hours B- Energy expenditure Cole et al.

Indirect calorimetry

SenseWearTM

[11]

(IC)

(SW) armband

8 minutes

Kilojoules

Pearson correlation

A- Results for IRS version 2.2

Analysis of variance

SenseWear_EE – indirect calorimetry_EE:

(ANOVA)

(range, r = 0.67–0.90) for arm and rowing

with three software (IRS version 2.2, IRS version 4.0 and preliminary cardiac software) during four forms of exercise 1treadmill 2- arm ergometer

ergometry, treadmill, and stepper 3- recumbent step

Bland-Altman plot

Method bias: SenseWear (version 2.2) underestimated EE (kilojoules) compared to indirect calorimetry during treadmill (214  64 vs. 184  43) and rowing activities (201  38 vs. 133  47), but there were no significant results between-method differences during the other two activities

4- rower

Bland–Altman plot: SenseWear (version 2.2)

ergometer

underestimated EE compared to indirect calorimetry for both for rower and treadmill. 95% limits of agreement were  123.73 for rowing, and  77.75 for treadmill B- Results for IRS version 4.0 SenseWear_EE – indirect calorimetry_EE: (range, r = 0.54–0.85) for arm and rowing ergometry, treadmill, and stepper Method bias: No significant difference for EE (kilojoules) estimates by SenseWear (version 4.0) and indirect calorimetry for any

161  40, respectively), recumbent stepping (203  36 Vs. 186  54, respectively), and rowing (185  54 vs. 193  68, respectively)

M. Alharbi et al.

activities: Treadmill (189  34 vs. 192  58, respectively), arm ergometry (170  44 vs.

HLC 2293 No. of Pages 18

10

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 2. (continued).

Criterion measure

Test measure (s)

Time

Units

Validity analysis

Main findings Bland–Altman plot: SenseWear (version 4.0) underestimated EE compared to indirect calorimetry for both for stepping and arm ergometry. 95% limits of agreement were  46.5 for arm ergometry, and  94.5 for recumbent stepping C- Results for preliminary cardiac software SenseWear_EE – indirect calorimetry_EE: (range, r = 0.78–0.90) for arm and rowing ergometry, treadmill, and stepper Method bias: No significant difference for EE estimates by SenseWear (cardiac software) and indirect calorimetry for any activities: Treadmill (192  58 Vs. 192  44, respectively), arm ergometry (161  36 vs. 161  40, respectively), recumbent stepping (187  48 vs. 186  54, respectively), and rowing (193  63 vs. 193  68, respectively) Bland–Altman plot: SenseWear (cardiac software) showed good agreement with indirect calorimetry for estimating EE for all the four activities and more tightly clustered than of either of the 2 versions (version 2.2 and version 4.0) of the software. 95% limits of agreement were  35 for arm ergometry,  63 for recumbent stepping,  74 for rowing, and  73 for treadmill

Frederix

Polar activity watch

et al. [15]

Omron

4 weeks:

pedometer

two period

Kcal/day

Pearson correlation

Pedometer_Kacl – accelerometer_Kcal: (r = 0.15)

cross-over T-test

Method bias: No significant difference between the pedometer and the accelerometer for recording daily calories burned (243  119 vs. 237  118, respectively)

Guiraud

MyWellness Key

Dijon Physical

et al. [17]

accelerometer

Activity

7 days

Kcal/week

Spearman correlation

DPAQ_ Kcal – accelerometer_Kcal: (range, Rho = 0.37–0.40

Questionnaire (DPAQ) 3 days

1 Total energy expenditure

Pearson correlation

11

RT3 accelerometer

HLC 2293 No. of Pages 18

Study

Measuring Overall Physical Activity for Cardiac Rehabilitation Participants

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 2. (continued).

Study

Criterion measure

Test measure (s)

Hertzog

Physical Activity

et al. [18]

Diary

Time

Units

Validity analysis

Main findings PA diary_TEE – RT3_TEE: (range, r = 0.57– 0.77)

Kcal/day

T-test

PA diary_Kcal – RT3_Kcal: (range, r = 0.46– 0.49)

Bland–Altman plot

Method bias: at 3 weeks, the RT3 (2122.0  422.0) estimated higher TEE compared to the PA diary (2036.0  457.7) by 86.0  301.1 kilocalories. At 3 months, the RT3 (2507.2  681.6) estimated higher TEE compared to the PA diary (2249.6  561.0) by 257.6  589.0 kilocalories. Bland–Altman plot: Showed the RT3 estimates to be, on average, higher than those from the PA diary, especially for participants with the lowest overall EE.

4- Physical activity level Pfaeffli et al. [19]

GT1M accelerometer

Mobile phone

7 days

1 Daily PA level from MobilePAL

Physical Activity

against accelerometer (daily: counts

Level

per minute, active minutes, and

Questionnaire

METs)

Pearson correlation

MobilePAL– accelerometer_METs: (r = 0.45)

Regression analyses

MobilePAL– accelerometer: (range, r = 0.30–

(MobilePAL) International

2 Daily PA level from Mobile PAL

Physical Activity

against IPAQ (daily: counts per

Questionnaire

minute, active minutes, and METs)

0.45; beta = 0.42)

(IPAQ) Bland–Altman plots

MobilePAL–IPAQ: (range, r = 0.48–0.49; beta = 0.51) Bland–Altman plots of energy expenditure obtained using accelerometer_METs and MobilePAL: Good agreement between the methods with a mean bias of 0.08 METs. 95% limits of agreement:

0.29 to 0.14 METs

Physical activity classifications*: Non weight bearing and water-based activities excluded; HSE, Health Survey for England; PA, physical activity; MVPA, Moderate to vigorous physical activity; TAM, Total Activity tionnaire; MobilePAL , Mobile Phone Physical Activity Level Questionnaire; IC, Indirect calorimetry; DPAQ, Dijon Physical Activity Questionnaire

M. Alharbi et al.

Measure; EE, Energy Expenditure; Kcal, Kilo/calories; METs, Metabolic Equivalents, TEE, Total Energy Expenditure; SWISSPAQ, Swiss Physical Activity Questionnaire; IPAQ, International Physical Activity Ques-

HLC 2293 No. of Pages 18

12

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 2. (continued).

PA measure(s)

Time

Units

Reliability analysis

Main findings

Bahler et al. [14]

Swiss Physical

3 weeks

METs/hours/day

Pearson correlation

SWISSPAQ 1 and 2: r = 0.62

Activity Questionnaire (SWISSPAQ) Bland–Altman plot

Bland–Altman plot: The mean difference between SWISSPAQ 1 and 2 was minimal with a mean bias of 0.06 METs-hours. 95% limits of agreement:

Orrell et al. [16]

Total Activity

7 days

Measure (TAM). Two versions:

1.26 to1.37 METs-hours

METs/min/week

Intra-class correlation

Test–retest repeatability ICC:

A- total >3 METs

coefficients (ICCs) Bland–Altman plot

TAM1 (range, ICC = 0.53–0.82)

TAM1 & TAM 2 B- low >3–<5 METs

TAM2 (range, ICC = 0.54–0.85)

C- moderate >5–<9

Bland–Altman plot: Inspection of the Bland–

METs

Altman plots indicate good agreement and that the reproducibility of the TAM2 is more consistent than the TAM1

Hertzog et al.

RT3

2, 3 and 4 days at each

D- strenuous >9 METs Number of days

[18]

accelerometer

time point (3 weeks, 6

required to estimate:

Intra-class correlation coefficients (ICCs)

Based on 2 to 4 day of data collection:

weeks, and 3 months) Physical

Total energy

Activity Diary

expenditure

RT3 (range, ICC = 0.90–0.98) for TEE

Kcal/day

RT3 (range, ICC = 0.82–0.93) for kcal/day

Minutes of MVPA

RT3 (range, ICC = 0.79–0.92) for minutes of MVPA PA diary (range, ICC = 0.82–0.95) for TEE PA diary (range, ICC = 0.69–0.92) for kcal/day PA diary (range, ICC = 0.67–0.91) for minutes of MVPA

Pfaeffli et al.

Mobile phone

[19]

Physical

differences in MobilePAL between any 2 days

Activity Level

were observed, except for days 6 and 7 (adjusted

Questionnaire (MobilePAL)

P = 0.04)

7 days

Daily PA level

Tukey–Kramer test

After Tukey–Kramer adjustment, there was no

PA, physical activity; MVPA, Moderate to vigorous physical activity; TAM, Total Activity Measure; Kcal, Kilo/calories; METs, Metabolic Equivalents, TEE, Total Energy Expenditure; SWISSPAQ, Swiss Physical Activity

13

Questionnaire; MobilePAL, Mobile Phone Physical Activity Level Questionnaire; ICC, Intra-class correlation coefficients

HLC 2293 No. of Pages 18

Study

Measuring Overall Physical Activity for Cardiac Rehabilitation Participants

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

Table 3 Studies examining reliability (test-retest & repeatability) of physical activity in cardiac rehabilitation.

HLC 2293 No. of Pages 18

14

showed that HSE misclassified 63% of PA levels for participants, had very poor overall sensitivity (k = 0.04, 95% CI, 0.31–0.40), and very poor agreement with the accelerometer (k = 0.08, 95% CI, 0.20–0.36) [13]. The validity of PA measures in assessing active minutes of PA [16,18,19] or step counts [15] were examined. The three studies assessing minutes of PA investigated the validity of self-reported measures in a comparison with the accelerometer to measure time spent in daily [18,19] or weekly [16] average for active minutes of light, moderate and/or vigorous PA. The one study [15] assessing step counts investigated the validity of the pedometer in a comparison with the accelerometer to measure daily steps. In all three studies investigating active minutes, moderate correlations were noticed: between the TAM2 and accelerometer (range, r = 0.25–0.32) [16], between the IPAQ and accelerometer (range, r = 0.41) [19], and between the PA diary and accelerometer (range, r = 0.42–0.46) [18]. The results demonstrated that self-reported measures overestimated time spent in PA compared to the accelerometer [16,19]. Frederix et al. (2011) [15] demonstrated a modest correlation (r = 0.28) between the pedometer and accelerometer, but no statistically significant difference was found between the pedometer and the accelerometer for recording daily step counts (7636  3348 vs. 7134  2808, respectively). Three studies [14,16,19] examined the validity of selfreported measures compared to the accelerometer in measuring METs. The correlation between self-reported measures and the accelerometer was low for the TAM2 (range, r = 0.03–0.32) [14,16], but was better for both the SWISSPAQ (r = 0.41) [14,16] and the IPAQ (r = 0.62) [19]. In all three studies [14,16,19] the self-reported measures estimated higher METs compared to the accelerometer. In addition, Bland-Altman plots for METs showed the mean difference between self-reported measures were higher than for the accelerometer and as the amount of METs increased the difference in METs also increased [14,16]. Only Bahler et al. (2013) [14] examined the validity of self-reported measures against the PA diary in measuring METs and the correlation between the two measures was weak (r = 0.14). The validity of diverse PA measures in measuring active energy expenditure (AEE; energy burned over a period of time due to PA during periods of non-rest) [11,15,17,18] and total energy expenditure (TEE; energy burned by the person over time, determined by the sum of three components: basal energy expenditure, diet induced thermogenesis, and PA) [18] was also examined. An indirect calorimetry measure against accelerometer [11], an accelerometer against a pedometer [15], an accelerometer against PA diary [18], and an accelerometer against self-reported PA questionnaire [17] were all used in the validation studies to estimate EE. Cole et al. (2004) [11] validated the SenseWear (SW) armband with three types of software (IRS version 2.2, IRS version 4.0, and preliminary cardiac software) during four forms of exercise: treadmill, arm ergometer, recumbent step, and rower ergometer. A high correlation was found between EE estimates by the SW cardiac software and indirect calorimetry

M. Alharbi et al.

(IC) (range, r = 0.78–0.90) for all of the four activities. There was also no significant difference in EE estimates by the SW cardiac software and IC for any of the four activities. BlandAltman plots also showed the SW cardiac software had good agreement with IC in estimating EE for all four activities and were more tightly clustered than the other two types of SW software (IRS version 2.2 and IRS version 4.0). Furthermore, Frederix et al. (2011) [15] demonstrated a weak correlation (r = 0.15) between the pedometer and accelerometer, but no statistically significant difference between the pedometer and the accelerometer for recording daily calories burned (243  119 vs. 237  118, respectively). Similarly, a moderate correlation was found between the DPAQ and accelerometer (range, Rho = 0.37–0.40) [17]. Hertzog et al. (2007) [18] revealed a better correlation between PA diary and accelerometer in estimating total EE (range, r = 0.57–0.77), but a weak correlation between the two measures in estimating AEE (range, r = 0.46–0.49). Hertzog et al. (2007) [18] exhibited the accelerometer slightly estimated higher TEE compared to the PA diary and Bland-Altman plots demonstrated the accelerometer estimates to be, on average, higher than those from the PA diary, especially for participants with the lowest overall EE. One study [19] investigated the validity of MobilePAL questionnaire against the accelerometer and IPAQ in measuring daily PA level and found moderate correlations between MobilePAL and accelerometer (range, r = 0.30– 0.45) as well as MobilePAL and IPAQ (range, r = 0.48– 0.49). Bland-Altman plots of EE obtained using accelerometer METs and MobilePAL showed a good agreement between the methods with a mean bias of 0.08 METs. Ninety-five per cent limits of agreement ranged from 0.29 to 0.14 METs.

Reliability Results In total, four studies [14,16,18,19] assessed the reliability of PA measures (Table 3). Two studies showed TAM2 and SWISSPAQ had overall acceptable reliability in measuring METs and Bland-Altman plots for the reproducibility of TAM2 and SWISSPAQ exhibited a good level of agreement and the mean difference was minimal [14,16]. However, Pfaeffli et al. (2013) [19] stated no differences in MobilePAL when measuring daily PA levels between any two days, except for days 6 and 7. Hertzog et al. (2007) [18] demonstrated that, based on two to four days of data collection, the PA diary and the accelerometer showed overall an acceptable to high reliability coefficient in measuring energy expenditure and minutes of MVPA, but the reliability coefficients were higher for the accelerometer than for the PA diary.

Discussion Overall, direct measures were substantially better than selfreported PA measures in terms of validity and reliability. Physical activity diary and MobilePAL do better than other self-reported measures. Direct measures also had good agreement when the comparison was made with the criterion

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

HLC 2293 No. of Pages 18 Measuring Overall Physical Activity for Cardiac Rehabilitation Participants

measures. Moreover, utilising a specific software version for CR enhanced the validity of the PA measure. Self-reported questionnaires had overall poor sensitivity, very poor agreement, and their validity coefficients were considerably lower than validity coefficients of direct measures. Most selfreported PA questionnaires exhibited weak correlation, with all estimating higher measured PA parameters compared to the accelerometer. Unlike self-reported PA questionnaires, the PA diary had slightly higher validity coefficients and did not overestimate the measured PA parameters when compared to the accelerometer. Self-reported questionnaires showed overall acceptable reliability, but reliability coefficients for the PA diary and the accelerometer ranged from acceptable to strong. However, the reliability coefficients were even higher for the accelerometer than for the PA diary. Most of the reviewed studies used METs and EE units as the derived PA measure. Notably, the average METs in cardiac patients is significantly lower than the commonly accepted value of 3.5 mL O2.kg(-1).min(-1) [20]. Most of the reviewed self-reported measures also used the absolute intensity of effort for specific PA. However, with increasing age there is a physiologic decline in maximum oxygen uptake, maximum achievable heart rate, and skeletal muscle mass, all of which may influence the perception of effort [21]. It is expected, therefore, that elderly cardiac participants will most likely perceive exercise intensity differently to fitter and younger participants as a result of age-related changes in their physical functions [7]. As patients recover from their cardiac event, the perception and accuracy of physical function improves [22], so the perception of intensity may differ during recovery [7]. Importantly, one of the principal objectives of a CR program is to increase patient PA levels by promoting the recommended minutes of MVPA and step counts rather than METs. Therefore, future PA measures in CR should target these PA parameters as the derived measure of PA as patients become more familiar with these units while attending CR. Of note, our results demonstrated the SenseWear armband has the highest accuracy among the reviewed PA measures. This finding has been reported in a population diagnosed with chronic obstructive, pulmonary disease [23]. Interestingly, similar PA monitors such as Fitbit, a pedometer, or PA tracker are increasingly popular in the market for the public to monitor their progress towards PA goals. The growing popularity of these wearable activity trackers has also extended to the clinical setting and their data will increasingly be understood in terms of establishing disease-specific clinical measures and assessments [24]. Thus, when selecting an accelerometer, researchers and clinicians should consider its validity and reliability, cost, features such as size, and practicality to access PA data. Importantly, wearable PA monitoring devices have become standard direct methods for assessing PA [25] and have shown the usefulness of accelerometry in tracking activities of daily living, minutes of MVPA, and METs [10], as well as motivating individuals to increase their daily PA level [26–28]. These tracking devices may be ideal in this population to help monitor PA and

15

encourage sustained changes in PA [28–30]. Despite this potential, there remains limited evidence on the validity of these devices and the capacity to use them to assess or motivate PA in CR patients. In this review, only two studies investigated the validity of such devices in measuring step counts and EE only. Therefore, future studies are required to validate wearable PA monitoring devices and to examine their feasibility for both CR patients and clinical staff in CR. This is because many current PA monitoring devices provide useful information for clinical professionals to track if the patient is achieving the daily recommended PA level as well as to track the overall time spent in a day being active. There are a number of methodological limitations of the reviewed studies that may threaten the external and internal validity of the findings. There is wide variability in the duration of the studies [11,15]. Some of the PA questionnaires reviewed relied on a recall period of greater than one week [13,14,17], when the length of recall should be limited to the previous week in self-reported PA because recall is often problematic for CABG patients [7]. Moreover, the studies reported the criterion measure and test measures were assessed non-concurrently [15] or the length of recall for self-reported measures [13,14,17] was different to the criterion measures, both of which may impact the validity of the study. The recruitment of participants from different phases in the CR program (but mostly phase III), along with gender bias (participants were predominately male), may have influenced the results in the reviewed studies. In addition, different models of PA monitoring devices with a wide range of sensors were placed in different locations on the body, all of which may impact the validity of the PA measures. For instance, triaxial sensors provide a more comprehensive assessment of body movements than uniaxial sensors [31]. Notably, while different models for data collection and analysis of accelerometers have been proposed [32– 34], there is no universally accepted definition [35,36]. The inconsistencies that then exist between data analysis methods among the reviewed studies may lead to large differences in outcomes such as in MVPA. Accordingly, researchers should consider utilising the cut-points specifically established for CHD populations, for example, the recently developed coronary artery disease cut-points by Prince et al. (2015) [36]. Considering the limitations identified in all of the studies selected for review, it may be argued that generalisability is somewhat weakened in terms of their evidence for determining a reliable and valid standard measure for PA assessment in CR settings. As such, an ideal study should consider all of the elements discussed to support and enhance its validity and reliability. Our findings showed direct PA measures appear to perform substantially better than self-reported measures in terms of validity and reliability. The PA diary and MobilePAL do better than other self-reported measures. Nevertheless, the results presented in this review differ to those provided in the review of self-reported measures in cardiac patients by Le Grande et al. (2008) [7] in one key aspect. This review refrains from making a recommendation for a

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

HLC 2293 No. of Pages 18

16

M. Alharbi et al.

informative studies from non-English speaking countries were missed.

definitive PA measure due to diverse PA measures used across the reviewed studies and the lack of strong evidentiary support for one PA measure over another. Le Grande et al. (2008) [7] nominated five self-reported measures to assess low-intensity PA for cardiac patients. However, this review is consistent with other literature [7,12] as it highlights the wide variations in self-reported measures and that such measures are mostly suitable to epidemiological studies rather than CR settings [7]. This review is also consistent with other reviews in that it demonstrates that the correlations between self-reported and direct measures are generally low to moderate and generally overestimate the measured PA constructs compared to the direct measures [12]; whereas the correlation between direct and other direct measures was generally high and agreements between these two were strong [8]. The findings presented in this review must be interpreted with several limitations in mind. The nature of a narrative literature review is highly subjective in the determination of which studies to include, the way the studies are analysed, and the conclusions drawn. Furthermore, only a limited number of databases were searched and the study was limited to English language, thus making it likely that

Conclusion Overall, the reviewed studies show low validity of selfreported measures as the strength of correlations were only weak to moderate with a poor level of agreement, and were more likely to overestimate the measured PA parameters compared to direct measures. Direct measures are far more precise than self-reported measures in evaluating current and changing overall PA in and following CR. A review of the methodology of these studies reveals several factors that may explain the discrepant findings. These include, but are not limited to, the issue of methodological design such as underpowered and small sample size, differences in patient populations and heterogeneity often found among the accelerometer models and sensors. These elements may affect the measured outcomes and may explain why there is inconsistency among the findings produced by these studies. Therefore, a recommendation is not made in this review for a specific PA measure in CR.

Appendix A. Quality appraisal checklist of the reviewed studies Author.

Reporting

EV

IV

Total

Year

Bahler et al. 2013

Score 1

2

3

4

5

6

7

8

9

1

2

3

1

2

3

4

5

6

7

8

9

(0–22)

1

0

1

1

1

1

1

1

1

0

1

1

0

0

1

0

0

1

1

1

2

16 13 11

Cole et al. 2004

1

1

1

1

n/a

1

0

1

1

0

0

n/a

0

0

1

1

0

0

1

1

0

Frederix et al. 2011

1

1

1

0

n/a

1

1

1

0

0

1

1

0

0

0

0

0

0

1

1

0

Guiraud et al. 2012

1

1

1

0

1

1

0

1

0

0

0

1

0

0

1

0

0

0

1

1

0

10 18 14

Hertzog et al. 2007

1

1

1

1

1

1

1

1

1

0

1

1

0

0

1

1

1

1

1

1

1

Orrell et al. (a) 2007

1

1

1

1

1

1

1

0

1

0

1

1

1

0

1

0

1

0

1

0

0

Orrell et al. (b) 2007

1

1

1

1

1

1

1

1

1

0

0

1

1

0

1

1

1

1

1

0

1

Pfaeffli et al. 2013

1

1

1

1

1

1

0

1

1

0

1

1

0

0

1

1

0

1

1

1

0

17 15

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

HLC 2293 No. of Pages 18 Measuring Overall Physical Activity for Cardiac Rehabilitation Participants

EV, External Validity; IV, Internal Validity; 1 is Yes and 0 is No for all questions except IV question 9, where 2 is Yes,1 is Partially and 0 is No. When not applicable (n/a) to study, question was given 1 point. Reporting 1. Is the hypothesis/aim/objective clearly described? 2. Are the operational definitions of the main physical activity constructs to be validated clearly described in the Introduction or Method section. 3. Are the characteristics of the participants to be included in the study clearly described? 4. Are the distributions of principle cofounders clearly described? 5. For studies evaluating an existing measure has the original source been cited? For studies evaluating a newly developed PA measure or a modified version of an existing measure, has the original source been cited and the modifications described. 6. Are the methods of administration and/or data reduction for the examined measure and the reference measure clearly described? 7. Have the characteristics of the participants with missing, incomplete, and/or invalid data been described? 8. Does the study provide information about the variability in the data for the main physical activity constructs? 9. Have limits of agreement and/or confidence interval been reported for the main analysis? Reporting Score: ___/9 External Validity. 1. Were the individuals asked to participate in the study representative of the entire population from which they were recruited? 2. Were the participants who were enrolled in the study representative of the entire population from which they were recruited? 3. Was the examined measure administration (e.g., researcher-participant contact, survey mode etc.) representative of the procedures applied under epidemiological, or behavioural research constraints? External Validity Score: ___/3 Internal Validity. 1. Was an attempt made to minimise altered physical activity behaviour by the participant in response to awareness and burden of measurement? 2. Was an attempt made to blind research staff to the activity levels or characteristics of the participants to prevent leading responses to the examined measure? 3. Does the reference measure assess the physical activity construct(s) of interest with greater accuracy than the examined measure, and are errors in the reference method uncorrelated with errors in the examined measure? 4. Did the examined measure and the reference measure assess physical activity in the same time frame?

17

5. Was compliance with the measurement protocol acceptable? 6. Was reproducibility of the main physical activity constructs reported for the examined measure? 7. Were statistical tests used appropriate to assess validity for the main physical activity constructs between the examined measure and the reference measure? 8. If any of the results of the study were based on ”data dredging” was this made clear? 9. Did the study have sufficient sample size to assess agreement? Internal Validity Score: ____/10

References [1] Australian Heart Foundation. Data and statistics; 2011, https://www. heartfoundation.org.au/about-us/what-we-do/ heart-disease-in-australia. accessed 20 April 2015).. [2] Taylor RS, Dalal H, Jolly K, Moxham T, Zawada A. Home-based versus centre-based cardiac rehabilitation. Cochrane Database Syst Rev 2015;1. http://dx.doi.org/10.1002/14651858.cd007130.pub3. [3] Anderson L, Oldridge N, Thompson DR, Zwisler AD, Rees K, Martin N, et al. Exercise-based cardiac rehabilitation for coronary heart disease: Cochrane systematic review and meta-analysis. J Am Coll Cardiol 2016; 67:1–12. [4] Anderson L, Thompson DR, Oldridge N, Zwisler AD, Rees K, Martin N, et al. Exercise-based cardiac rehabilitation for coronary heart disease. Cochrane Database Syst Rev 2016;1. http://dx.doi.org/10.1002/ 14651858.cd001800.pub3. [5] Pogosova N, Saner H, Pedersen SS, Cupples ME, McGee H, Ho¨fer S, et al. Psychosocial aspects in cardiac rehabilitation: From theory to practice. A position paper from the Cardiac Rehabilitation Section of the European Association of Cardiovascular Prevention and Rehabilitation of the European Society of Cardiology. Eur J Prev Cardiol 2014;22:1290–306. [6] Kotseva K, Wood D, De Backer G, De Bacquer D, Pyo¨ra¨la¨ K, Keil U. EUROASPIRE III: A survey on the lifestyle, risk factors and use of cardioprotective drug therapies in coronary patients from 22 European countries. Eur J Cardiovasc Prev Rehabil 2009;16:121–37. [7] Le Grande MR, Elliott PC, Worcester MU, Murphy BM, Goble AJ. An evaluation of self-report physical activity instruments used in studies involving cardiac patients. J Cardiopulm Rehabil Prev 2008;28:358–69. [8] Kowalski K, Rhodes R, Naylor P-J, Tuokko H, MacDonald S. Direct and indirect measurement of physical activity in older adults: A systematic review of the literature. Int J Behav Nutr Phys Act 2012;9:148–60. [9] Helmerhorst HJ, Brage S, Warren J, Besson H, Ekelund U. A systematic review of reliability and objective criterion-related validity of physical activity questionnaires. Int J Behav Nutr Phys Act 2012;9:103–57. [10] Bidargaddi NP, Sarela A. Activity and heart rate-based measures for outpatient cardiac rehabilitation. Methods Inf Med 2008;47:208–16. [11] Cole PJ, Lemura LM, Klinger TA, Strohecker K, McConnell TR. Measuring energy expenditure in cardiac patients using the Body Media(trademark) Armband versus indirect calorimetry: A validation study. J Sports Med Phys Fitness 2004;44:262–71. [12] Prince SA, Adamo KB, Hamel ME, Hardt J, Connor Gorber S, Tremblay M. A comparison of direct versus self-report measures for assessing physical activity in adults: A systematic review. Int J Behav Nutr Phys Act 2008;5:56–80. [13] Orrell A, Doherty P, Coulton S, Miles J, Stamatakis E, Lewin R. Failure to validate the Health Survey for England physical activity module in a cardiac population. Health Policy 2007;84:262–8. [14] Bahler C, Bjarnason-Wehrens B, Schmid JP, Saner H. SWISSPAQ: Validation of a new physical activity questionnaire in cardiac rehabilitation patients. Swiss Med Wkly 2013;143:w13752. [15] Frederix I, Dendale P, Berger J, Vandereyt F, Everts S, Hansen D. Comparison of two motion sensors for use in cardiac telerehabilitation. J Telemed Telecare 2011;17:231–4. [16] Orrell A, Doherty P, Miles J, Lewin R. Development and validation of a very brief questionnaire measure of physical activity in adults with coronary heart disease. Eur J Cardiovasc Prev Rehabil 2007;14:615–23.

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005

HLC 2293 No. of Pages 18

18

[17] Guiraud T, Granger R, Bousquet M, Gremeaux V. Validity of a questionnaire to assess the physical activity level in coronary artery disease patients. Ann Phys Rehabil Med 2012;55:317–9. [18] Hertzog MA, Nieveen JL, Zimmerman LM, Barnason SA, Schulz PM, Miller CL, et al. Longitudinal field comparison of the RT3 and an activity diary with cardiac patients. J Nurs Meas 2007;15:105–20. [19] Pfaeffli L, Maddison R, Jiang Y, Dalleck L, Lof M. Measuring physical activity in a cardiac rehabilitation population using a smartphone-based questionnaire. J Med Internet Res 2013;15:61–78. [20] Savage PD, Toth MJ, Ades PA. A re-examination of the metabolic equivalent concept in individuals with coronary heart disease. J Cardiopulm Rehabil Prev 2007;27:143–8. [21] Hollmann W, Stru¨der HK, Tagarakis CV, King G. Physical activity and the elderly. Eur J Cardiovasc Prev Rehabil 2007;14:730–9. [22] Worcester MUC, Murphy BM, Elliott PC, Grande MR, Higgins RO, Goble AJ, et al. Trajectories of recovery of quality of life in women after an acute cardiac event. Br J Health Psychol 2007;12:1–15. [23] Van Remoortel H, Raste Y, Louvaris Z, Giavedoni S, Burtin C, Langer D, et al. Validity of six activity monitors in chronic obstructive pulmonary disease: A comparison with indirect calorimetry. PLoS One 2012;7: 391–8. [24] Chiauzzi E, Rodarte C, DasMahapatra P. Patient-centered activity monitoring in the self-management of chronic health conditions. BMC Med 2015;13:77–82. [25] Bassett Jr DR, Rowlands AV, Trost SG. Calibration and validation of wearable monitors. Med Sci Sports Exerc 2012;44:32–8. [26] Tudor-Locke C. Taking steps toward increased physical activity: Using pedometers to measure and motivate. Pres Council Phys Fit & Sport Res Dig 2002;3:1–8. [27] Gallagher R, Kirkness A, Armari E, Davidson PM. Weight management issues and strategies for people with high cardiovascular risk undertaking

M. Alharbi et al.

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

an Australian weight loss program: A focus group study. Nurs Health Sci 2012;14:18–24. Gallagher R, Kirkness A, Armari E, Davidson PM. Participants’ perspectives of a multi-component, group-based weight loss programme supplement for cardiac rehabilitation: A qualitative study. Int J Nurs Pract 2012;18:28–35. Annette Dean MC. Using pedometer step-count goals to promote physical activity in cardiac rehabilitation: A feasibility study of a controlled trial. Int J Phys Med Rehabil 2013;1:157–62. Izawa KP, Watanabe S, Omiya K, Hirano Y, Oka K, Osada N, et al. Effect of the self-monitoring approach on exercise maintenance during cardiac rehabilitation: A randomized, controlled trial. Am J Phys Med Rehabil 2005;84:313–21. Butte NF, Ekelund U, Westerterp KR. Assessing physical activity using wearable monitors: Measures of physical activity. Med Sci Sports Exerc 2012;44:5–12. Catellier DJ, Hannan PJ, Murray DM, Addy CL, Conway TL, Yang S, et al. Imputation of missing data when measuring physical activity by accelerometry. Med Sci Sports Exerc 2005;37:555–62. Hart TL, Swartz AM, Cashin SE, Strath SJ. How many days of monitoring predict physical activity and sedentary behaviour in older adults. Int J Behav Nutr Phys Act 2011;8:1–7. Herrmann SD, Barreira TV, Kang M, Ainsworth BE. How many hours are enough? Accelerometer wear time may provide bias in daily activity estimates. J Phys Act Health 2012;10:742–9. Hagstro¨mer M, Oja P, Sjo¨stro¨m M. Physical activity and inactivity in an adult population assessed by accelerometry. Med Sci Sports Exerc 2007;39:1502–8. Prince SA, Reed JL, Mark AE, Blanchard CM, Grace SL, Reid RD. A comparison of accelerometer cut-points among individuals with coronary artery disease. PLoS One 2015;10:1–12.

Please cite this article in press as: Alharbi M, et al. Measuring Overall Physical Activity for Cardiac Rehabilitation Participants: A Review of the Literature. Heart, Lung and Circulation (2017), http://dx.doi.org/10.1016/j.hlc.2017.01.005