Incidental and Planned Exercise Questionnaire for Seniors

Incidental and Planned Exercise Questionnairefor Seniors: Validity and Responsiveness

DAFNA MEROM1,3, KIM DELBAERE2, ROBERT CUMMING3, ALEX VOUKELATOS4, CHRIS RISSEL3,5,HIDDE P. VAN DER PLOEG3,6, and STEPHEN R. LORD2

1School of Science and Health, University of Western Sydney, Campbelltown, AUSTRALIA; 2Neuroscience Research Australia(NeuRA), University of NSW, Sydney, AUSTRALIA; 3School of Public Health, University of Sydney, Sydney, AUSTRALIA;4Health Promotion Service, Clinical Support Division (Western), Sydney, AUSTRALIA; 5Centre for Health Advancement, NSWHealth Department, Liverpool, AUSTRALIA; and 6Department of Public and Occupational Health, EMGO Institute for Healthand Care Research, VU University Medical Centre, Amsterdam, THE NETHERLANDS

ABSTRACT

MEROM, D., K. DELBAERE, R. CUMMING, A. VOUKELATOS, C. RISSEL, H. P. VAN DER PLOEG, and S. R. LORD. Incidental

and Planned Exercise Questionnaire for Seniors: Validity and Responsiveness. Med. Sci. Sports Exerc., Vol. 46, No. 5, pp. 947–954,

2014. Purpose: This study assessed the criterion validity and responsiveness of the Incidental and Planned Exercise Questionnaire

(IPEQ) specifically developed for aging research. Methods: The sample comprised 315 community-dwelling inactive older adults (mean

age = 73.2 yr) who participated in a trial investigating the effect of a walking program on falls. At baseline, a subsample (n = 177) also wore

an accelerometer (ActiGraph GT1M) for 7 d, and 126 of those had four valid days or more of accelerometer data. Validity coefficients

(Spearman Q) were calculated between accelerometer counts per minute, average steps per day, average moderate- to vigorous-intensity

physical activity (MVPA) using two cut points (760 and 1041 counts per minute), and IPEQ-MVPA (the sum of physical activity

hours per week derived from nine IPEQ items excluding the indoor chores question). Responsiveness was the mean IPEQ change

in the intervention group divided by the SD of the mean change in control group. Results: The correlation coefficients between

IPEQ-MVPA and counts per minute per day and steps per day were 0.33 and 0.31, respectively. The coefficients with accelerometer

MVPA at low and medium cut points were 0.29 and 0.33, respectively, and 0.26 and 0.35 for walking. The Bland–Altman plot showed

increased errors with higher values of IPEQ-MVPA. IPEQ incidental physical activity questions and IPEQ as a whole were significantly

correlated with accelerometer light intensity (100–760 counts per minute, Q = 0.29 and 0.23). The 12-month responsiveness of

total IPEQ was 0.30 and 0.44 for the ‘‘planned walking’’ item. Conclusions: IPEQ provides a practical and valid measure of MVPA in

surveillance and intervention studies. Incidental activities are mostly of a light-intensity nature, and their low ambulatory component

may explain the lower correlation between IPEQ as a whole and accelerometer measures. Key Words: ACCELEROMETER, AGED,

PHYSICAL ACTIVITY, SELF-REPORT, WALKING

An increasing number of older adults are facing thechallenges of aging successfully, that is, main-taining a low risk of disease and disease-related

disabilities and high cognitive and physical functional ca-pacities and engagement in social and productive activities(8,30). Regular physical activity (PA) and exercise may dis-criminate between adults who successfully age from those

who do not (13,23,39). PA can attenuate the physiologicaldeclines associated with aging and muscular disuse, preventthe onset of chronic diseases even if commenced late in life,prevent or mitigate functional limitations, and improve cog-nitive functioning and mental health (13,15). The valid andreliable methods for monitoring PA levels in older age andthe success/failure of promotional efforts are highly importantto public health and individuals.

Self-reported PA questionnaires have a prominent role inresearch and practice settings (1) and remain the mostpractical method for assessing PA in population samples andinterventions that require large sample sizes (31). Literaturereviews of the suitability and quality of PA questionnairesfor older people highlight that very few questionnaires havebeen developed to quantify light-intensity functional dailyPA, which is the main source of energy expenditure in olderpeople (14,22,31,37). Furthermore, the most prominentform of exercise/PA for older adults is walking, with half ofthe active older population engaged exclusively in walking(26). However, to date, only one study has assessed the

Address for correspondence: Dafna Merom, Associate Professor, School ofScience and Health, Campbelltown Campus, Locked bag 1797, PenrithNSW 2751, Australia; E-mail: [email protected] for publication July 2013.Accepted for publication October 2013.Supplemental digital content is available for this article. Direct URL cita-tions appear in the printed text and are provided in the HTML and PDFversions of this article on the journal_s Web site (www.acsm-msse.org).

0195-9131/14/4605-0947/0MEDICINE & SCIENCE IN SPORTS & EXERCISE�Copyright � 2014 by the American College of Sports Medicine

DOI: 10.1249/MSS.0000000000000196

947

EPIDEM

IOLO

GY

Copyright © 2014 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.

http://www.acsm-msse.org

measurement properties of walking questions in an olderpopulation (16).

Past validation studies have had certain methodologicallimitations (e.g., small samples, recruitment of volunteers),and few validation studies have used objective measures ofPA (e.g., motion sensors). Previous studies have also pro-vided limited information on measurement properties acrossspecific groups (e.g., functional ability) (14,22). Finally,only one questionnaire designed for older adults has beenevaluated in relation to sensitivity to intervention effect (14).

It has also been suggested that existing surveillance ques-tionnaires may perform differentially for older adults becauseof age-related cognitive declines, imposing greater cognitiveburden on older people compared with younger adults (10).This was corroborated recently by cognitive interviewingwhile older adults attempt to respond to two widely usedsurveillance instruments (the International Physical ActivityQuestionnaire and the Active Australia Questionnaire). Re-spondents were not clear of what activities fitted within thescope of generic questions, struggled estimating frequencyand amount in open-ended questions, and had particular dif-ficulties in assessing PA episodes in the quanta of 10 minperiods (20,21).

To circumvent some of the above limitations, the Com-munity Healthy Activities Model Program for Seniors(CHAMPS) and the Yale Physical Activity Survey (YPAS)were developed with lists of specified activities and easyresponses format. However, their use in surveillance orlarge-scale interventions is not practical as both take 15 or20 min to complete. A third questionnaire developed forolder people (the Physical Activity Scale for the Elderly[PASE]) has been extensively used in epidemiological re-search with older adults due to its practicality (it takes 5 minto complete). However, this scale has limitations in thatdaily chores are assessed in a ‘‘yes/no’’ format, walking forerrands or exercise cannot be distinguished, and a total PAscore cannot be translated to any meaningful threshold,such as compliance with public health recommendations.Therefore, there is a need to develop and properly validateself-reported questionnaires specifically for older adults.

The Incidental and Planned Exercise Questionnaire (IPEQ)was designed specifically for use in aging research. TheIPEQ takes 5 min to complete and has excellent reliabilityand construct validity (8). IPEQ can discriminate betweenPA levels across groups that differ by fall risk factors (8). Assuch, it is a promising tool for older adult PA surveillance andlarge-scale intervention studies. However, the criterion va-lidity of the IPEQ using motion sensors and its responsive-ness to change is yet to be determined.

This study was conducted alongside a community-basedrandomized controlled trial designed to test the effect of awalking program on rate of falls and aimed to examine (i)the criterion validity of IPEQ walking, light- and moderate-intensity PA questions against two accelerometer cut pointssuggested for brisk walking, and lifestyle PA (7,25) and (ii)the responsiveness of IPEQ total and subcomponent scores.

The study methodology follows the new checklist forassessing PA internal and external data validity (18).

METHODS

Participants in the study comprised community-dwellingolder adults (Q65 yr) who enrolled in a randomized trial thatevaluated the effect of a home-based self-paced walkingprogram on falls and its associated risk factors. A detailedprotocol of the study methodology has been published else-where (ACTRN12610000380099) (36). In short, participantswere recruited through advertisements and the NSW electoralroll from August 2009 to August 2011. Potential partici-pants first had a short telephone screening interview todetermine eligibility. To be eligible, participants neededto be physically inactive (exercising less than three timesand/or G120 minIwkj1), without medical issues that pre-cluded regular exercise, and without cognitive impairment,defined as score of G7 on the Short Portable Mental StatusQuestionnaire (28).

The sample for the present validity study comprised asubsample of participants in the RCT, who after completingthe baseline telephone screening interview agreed also tohave physical performance tests in their homes and under-take activity monitoring while wearing an accelerometer(n = 177). The sample for the responsiveness aspect of thepresent study comprised all those who enrolled in the studyand completed both baseline and follow-up telephone in-terviews (n = 315). The study was approved by the ResearchEthics Review Committee of the Sydney South West AreaHealth Service–Eastern Zone (X08-0279 and HREC/08/RPAH/477), and written informed consent was obtained fromeach participant.

Procedures. After telephone screening, eligible partic-ipants completed a 20-min telephone interview, which in-cluded the IPEQ. Participants who agreed to have homevisits were also mailed an ActiGraph GT1M accelerometer(ActiGraph LLC, Fort Walton Beach, GFL) with written il-lustrated instructions on how to wear it on the right mid-axillary line of the hip using the belt provided. Participantswere asked to wear the accelerometer for seven consecutivedays and to only remove it when showering, when perform-ing water-based activities, and when sleeping. Approximately8–10 d after mailing, a trained research assistant visitedparticipants’ homes to collect the accelerometers and to con-duct physical measurements, including height and weightmeasurements, the Short Physical Performance Battery (SPPB)(17), and a test of quadriceps strength. After the comple-tion of the home visit, participants were randomized intoexercise and control groups using sealed opaque envelopes.At 12 months from randomization, participants who remainedin the study were recontacted by telephone for a follow-upinterview, which was identical to that conducted at baseline.Participants in the home visit subsample were also assessedat 12 months.

http://www.acsm-msse.org948 Official Journal of the American College of Sports Medicine

EPIDEM

IOLO

GY


Intervention. Participants in the intervention group re-ceived a self-managed, progressive walking program. Thisconsisted of a sequential mailing of three manuals, a pe-dometer for self-monitoring (Yamax S270), and seven tele-phone coaching sessions during the 48 wk of the intervention,at weeks 1, 3, 6, 12, 16, 24, and 36. The control group re-ceived health information for the duration of the study.Control group participants were requested not to change theirPA habits during the study, after which they also receivedthe walking kit.

Measures. The IPEQ consists of 10 questions on PApertaining to an average week in the past 3 months. Thisversion of the IPEQ was chosen because of its higher test–retest repeatability (ICC = 0.84) compared with the IPEQpast week version (ICC = 0.77) (8). The frequency and thesession duration of both planned exercise and plannedwalking were measured in a first set of six questions. Thesewere followed by four questions on incidental PA: theweekly frequency and the daily duration of walks to get toplaces, a question on the average time spent daily doingchores outside the home, and the average time undertak-ing indoor chores. All answers were close-coded in definedintervals of duration and frequency, except the responsesto planned exercise other than walking, which used anopen-ended format (see Supplemental Digital Content 1,http://links.lww.com/MSS/A338, Incidental and Planned Phy-sical Activity Questionnaire [IPAQ]). Participants’ responseswere recorded on hard copy during a telephone interview andwere manually entered into an Access database and checked forentry errors by research assistants. Average weekly time spenton each IPEQ question was calculated by multiplication of theweekly frequency by daily duration using the median point foreach response category (i.e., one to two times a week equiv-alent to 1.5 or 15–30 min was 0.375 of an hour, which is22.5 min) (8). Four IPEQ indicators were examined: (i) ascore for moderate- to vigorous-intensity PA (IPEQ-MVPA)comprising the sum of hours across all questions, excludingthe indoor chores question (according to the PA compen-dium, an absolute MET value of 3 will be assigned to theseIPEQ questions (2), which is the accepted lower cut point formoderate-intensity (27), although for adults age 65 yr andabove, the relative intensity of these activities can be muchabove this threshold, in particular for those who are unfit); (ii)a total walking score comprising the summation of hoursspent ‘‘walking for exercise’’ and ‘‘walking for errands’’; (iii)an incidental PA score comprising walking for errands andoutdoor and indoor chores; and (iv) a total IPEQ score com-prising the sum of all questions.

Accelerometer data were recorded in 10-s intervals andaggregated into 1-min epochs. Accelerometry data reductionwas performed with an Excel Macro program developedby Trost. Zero counts for more than 60 min were consid-ered as nonwear time. Days where the participant worethe accelerometer for at least 10 h were considered a validday. Outliers (i.e., epochs with more than 20,000 counts perminute) were excluded from the data, and if a day had 10 or

more outliers, data for this the day were excluded from furtheranalyses. To be included in the validity analyses, participantshad to have at least four valid days of data.

There is no consensus on acceptable cut points to delin-eate PA intensity (25), and very few calibration studies in-volved adults 65 yr or older (34). We chose two cut pointsthat were calibrated to detect different PA behaviors. Amedium cut point of 1041 counts per minute was based ona calibration study in older adults by Copland and Dale (7).This medium level of PA correlated highly (r = 0.60) withoxygen consumption while walking on a treadmill at apace of 3.2 kmIhj1 and corresponded to 3.7 METs. A lowercut point of 760 counts per minute was based on calibrationstudies in free-living conditions reviewed by Matthews (25).This low cut point covers a wide range of daily activities ofmoderate intensity such as shopping for groceries, mopping,and vacuum cleaning as well as leisure time activities suchas light calisthenics, playing with children and gardening(25). Light intensity ranged between 100 counts per min-ute, which is currently the widely accepted cut point forsedentary activity and each of these cut points (i.e., 760 or1041). We hypothesized that incidental PA would be bettercaptured by light-intensity activity using the low cut point(G760 counts per minute). The incidental questions refer tooutdoor and indoor chores such as ironing, laundry, andcooking, many of which are classified at an absolute METrange of 1.5–2.9 (2).

Accelerometer indicators were as follows: 1) averagedaily counts per minute per day, 2) average steps per day, 3)average MVPA minutes per day, 4) average MVPA 10-minbout per day (allowing two episodes 2 min each to beinterrupted), and 5) average light-intensity activity minutesper day (100 to G760 counts per minute and 100 to G 1041counts per minute).

IPEQ responsiveness was measured on 315 participantswho remained in the study and completed the 12-monthfollow-up interview, which included the IPEQ. We com-pared the responsiveness of IPEQ-MVPA scores to the re-sponsiveness of accelerometer measures of MVPA (1041counts per minute) and light-intensity PA (100 to G1041counts per minute) in 85 people who completed the validitystudy and had valid accelerometer data at 12 months.

Statistical analysis. SAS 9.3 software was used foranalysis. Accelerometer person averages were calculated asthe sum over the valid days divided by the number of validdays. Descriptive statistics (mean, SD, and frequencies)were calculated to describe the demographic, health-relatedcharacteristics, and PA levels of participants. Because of theskewed distribution of IPEQ data, Spearman correlationcoefficients (r) were used to assess associations betweenIPEQ weekly hours and other measures. Bland–Altmanplots were produced after excluding one outlier for IPEQ-MVPA (4). For the plot, the week scores from the IPEQwere converted to day scores to enable a direct comparisonwith the accelerometer minutes per day. Responsiveness tochange was computed using the Tuley et al. (35) index,

VALIDITY AND RESPONSIVENESS OF THE IPEQ Medicine & Science in Sports & Exercised 949

EPIDEM

IOLO

GY


http://links.lww.com/MSS/A338

which is computed from paired t-tests: R = Dintervention / Rcontrol),where D is the mean change in the intervention group and R

is the SD of the mean change in the control group. This indexquantifies the minimum clinically important difference andis less affected by baseline skewness than other methods. Wealso computed effect size using Cohen’s formula to comparethe IPEQ to the CHAMPS questionnaire, which is the onlyquestionnaire to date that was tested for its responsiveness inthis age-group (33). Classifications of effect sizes accordingto Cohen were low (r e 0.2), moderate (r = 0.2–0.5), or high(90.5) (6).

RESULTS

Table 1 presents the descriptive characteristics of theparticipants who received home visits (n = 177), the subset ofthese who had at least four valid days of data and formedthe validity sample (n = 126), the responsiveness studysample (n = 315), and the comparative New South Wales(NSW) population for people age 65 yr and older. The va-lidity sample was not different to the home visit sample butwas significantly different (P G 0.05) from the responsive-ness sample by sex, living arrangement, and proportion of

retirees. Compared with the NSW data, both study samplesoverrepresented people of younger age (G75), women, se-niors with higher education, seniors who lived alone, andthose still working. The samples underrepresented peopleborn in Australia but had similar proportions of seniors whospoke English (5). The study samples had similar fall ratesand proportions of older adults who rated their health asexcellent or good to those reported for NSW but under-represented those at the lowest self-rated health category(fair or below) (5).

At baseline, participants reported, on average, 20 h perweek on all light-, moderate-, and vigorous-intensity PA(Table 1). Of this time, 7.3 h per week were devoted toat least moderate intensity (i.e., mean = 62 minIdj1), whichwas higher than the accelerometer MVPA estimate of43 min at a cut point of 1041 counts per minute and lowerthan the accelerometer MVPA mean of 85 min at a cut pointof 760 counts per minute. Participants reported engagingin 2.1 h per week of planned exercise, including walkingfor exercise (data not shown). The mean daily steps were5267 and mean counts per minute were 174.

Criterion validity. The correlation coefficients of totalIPEQ PA with accelerometer counts per minute were low

TABLE 1. Sociodemographics, health, physiological function, and PA levels of the samples of the validity and responsiveness studies.

Home Visit Sample, n = 177 Validity Sample n = 126 Responsiveness Sample, n = 315 NSW Older (Q65) Population

Age 65–74 (%) 63.1 66.9 64.6 54.0Males (%) 30.2 32.5 27.0 45.5Education (%)

Less than high school 22.4 21.1 19.9 50.4High school/technical 37.9 39.8 43.3 22.5University 39.7 39.1 36.9 15.8

Australia born (%) 65.3 62.0 66.4 72.8Speak English at home (%) 87.2 87.2 89.9 91.1Living alone (%) 37.9 37.4 42.0 33.3Retired (%) 76.4 76.4 82.7 91.4Self-rated health

Very good/excellent 44.1 47.2 47.9 47.7Good 45.8 44.8 40.6 30.9Fair/poor 10.1 8.0 11.4 21.4

Falls history0 77.4 74.8 76.5 74.41 11.9 12.8 12.4 15.2+ 10.7 12.8 11.1 9.9

Perceived walking abilityG15 min 14.7 13.6 12.4 —15–30 min 24.9 21.6 24.8930 min 60.4 64.8 62.9

BMI measured (%) a

Normal weight (G25) 22.7 21.6 NA 44.0Overweight (25–29) 43.7 47.2 36.1Obese (Q30) 33.5 31.2 19.8

Gait speed (mIsj1) 0.98 (0.29) 0.96 (0.26) NA —Timed 5 chair rises (s) 13.0 (4.5) 12.6 (4.2) —IPEQ scores (hIwkj1) NA

Total PA 22.1 (15.2) 20.9 (13.3) 21.9 (13.8) Not askedWalking (planed and travel) 3.8 (5.4) 3.7 (4.6) 3.4 (4.1)Outdoor PA 2.8 (4.5) 2.4 (3.5) 3.1 (5.3)Indoor PA 14.5 (10.9) 13.7 (10.1) 14.4 (10.2)MVPA 7.7 (8.0) 7.3 (7.0) 7.4 (7.1)

Accelerometer (cut point 1041)Counts (counts per minute) 174 (83.6)Steps (steps per day) 5266 (2526)MVPA (minIdj1) 43 (27.4)

aBased on the self-reported height and weight.MVPA, moderate- to vigorous-intensity PA.


EPIDEM

IOLO

GY


(0.17; Table 2), but were higher for IPEQ-MVPA (Q = 0.33).Copland’s cut point was more strongly correlated (Q = Q30)than Matthew’s cut point with regard to IPEQ-MVPA andwalking. On the other hand, Matthews’s cut point was morestrongly correlated with incidental PA and total IPEQ thanCopland’s cut point. Coefficients measured by accelerome-ter MVPA minutes per day in 10-min bouts were slightlyhigher than those without restriction on bout length for thelow cut point.

The Bland–Altman plot (Fig. 1) for IPEQ-MVPA, ex-pressed as minutes per day, and accelerometer mean MVminutes per day (cut point 1040 counts per minute) showsvalues above and below their combined mean but with aclear pattern of increased errors with higher values of IPEQ-MVPA. The mean differences between the two instrumentswere +15.5 min, and the limits of agreement were wide(105 to j74 min). The mean difference for IPEQ walkingquestions (data not shown) wasj12min below accelerometerMVPA and the limits of agreement large (56 to j80).

There were significant variations in the correlation co-efficients across population groups (Table 3). Coefficientswere better for younger, female, overweight, and obeseparticipants, those with better gait speed (90.95 mIsj1), andthose who performed better in the repeated sit-and-stand test(G12 s). The coefficients between IPEQ-MVPA and the ac-celerometer low cut point were lower than for the mediumcut points across all subgroups.

Responsiveness. Table 4 presents the responsivenessindex for IPEQ measures. Responsiveness ranged from thelowest 0.13 for incidental walking or other planned exerciseto the highest 0.45 for walking for exercise. The Cohen’seffect size for intervention target (data not shown) was 0.49,which is moderate (6). The total responsiveness score forIPEQ was 0.31, lower for total IPEQ-MVPA (R = 0.34). The12-month accelerometer measurement on a subsample par-tially supports changes observed in the IPEQ measures:minutes spent in MVPA (although not specific to walking) andin light-intensity increased in both groups. The accelerometer

TABLE 2. Validity coefficients (Spearman’s rho) for selected Items in IPEQ and all IPEQ questions against ActiGraph GT1M by two cutoffs to delineate moderate intensity.

IPEQ

Accelerometer MVPA Walking Incidental All PA

Counts per minute 0.33** 0.31** 0.15 0.17Steps per day 0.31** 0.28** 0.07 0.11Copland’s cutoff point for moderate-intensity PA (1040 counts per minute)

MVPA (minIdj1) 0.33** 0.35** 0.04 0.09MVPA 10-min bouts (minIdj1) 0.33** 0.36** j0.01 0.09Light (minIdj1) 0.12 0.03 0.14 0.11

Matthew’s cutoff point for moderate-intensity PA (760 counts per minute)MVPA (minIdj1) 0.26* 0.29** 0.13 0.15MVPA 10-min bouts (minIdj1) 0.30** 0.33** 0.16 0.13Light (minIdj1) j0.02 j0.07 0.29** 0.23**

*P G 0.05 **P G 0.01.MVPA, planned exercise/PA, walking for exercise, walking for errands, and outdoor chores; All PA, all IPEQ questions which is all of the above plus indoor chores.

FIGURE 1—Bland–Altman plot comparing IPEQ moderate- to vigorous-intensity PA (IPEQ-MVPA) with accelerometer MVPA (cut point 1041counts per minute).


EPIDEM

IOLO

GY


responsiveness coefficients for MVPA and light-intensityactivity were higher (R = 0.65 and R = 0.78, respectively)than for IPEQ measures, probably due to smaller measure-ment errors.

DISCUSSION

To the best of our knowledge, this is the first study thatpresents validity data for low-intensity incidental PA at anyage, and it is the only second study of an older adult PAquestionnaire that has examined responsiveness to changeas a result of an intervention (33). The IPEQ was selected asthe study PA measurement due to its excellent repeatability,its short format (5-min interview), the quantification of low-intensity daily activities (i.e., removing floor effect), and theseparate measurement of planned walking (intervention fo-cus) from other PA. No other older adults’ PA questionnairemet all of these criteria. IPEQ correlation coefficients withMVPA, after excluding the indoor chores question, were atthe upper range of the validity coefficients (0.2–0.4) reportedfor PA questionnaires (14). As hypothesized, the positive and

significant correlation between incidental PA questions withlight-intensity activity at the low cut point and the lack ofcorrelation with MVPA suggests that incidental PA is pri-marily of a light-intensity nature. IPEQ mean estimates wereabove or below accelerometer mean MVPA, depending onthe cut point used. However, the limits of agreement arewithin older population averages for time spent in moderate-intensity minutes per day as recently demonstrated (12). Thecorrelation coefficients were better for women and youngerold (65–74), which could be due to the small numbers ofmen and older old in the sample. IPEQ also performed betterfor older people with better functional status, in line withfindings from other validation studies (19) Finally, IPEQwas found to be sensitive to detect changes in PA over time,and as expected, the highest responsiveness score was forplanned walking.

Certain limitations of this study must be acknowledged:the nonrandom selection of people; the timing of acceler-ometer measurement, which may not represent a usual weekover the last 3 months; the relative small sample of 12-monthaccelerometer measurement of responsiveness; and the

TABLE 3. Correlation coefficients (Spearman’s rho) between IPEQ moderate-to-vigorous (MV) questions (with and without outdoor chores), walking questions and accelerometer countsper minute per day and MV minutes per day by population groups.

Accelerometer Counts per Minuteper Day, IPEQ-MVPAa

Accelerometer MV Minutesper Day, IPEQ-MVPAb

Accelerometer MV Minutesper Day, IPEQ-MVPAc

Accelerometer MV Minutesper Day, IPEQ-Walking

Age65–74 (n = 84) 0.27* 0.27* 0.25** 0.36**Q75 (n = 42) 0.22 0.14 0.13 0.18

SexMen (n = 40) 0.22 0.27 0.20 0.28Women (n = 86) 0.39** 0.36** 0.29** 0.38**

Body mass indexe25 (n = 27) 0.19 0.19 0.08 0.12925 (n = 97) 0.38** 0.38** 0.34** 0.33**

Gait speed (median split)G0.95 mIsj1 (n = 66) 0.34** 0.27* 0.29* 0.33**Q0.95 mIsj1 (n = 60) 0.19** 0.16** 0.12 0.35**

Timed chair stand (median split)e12.0 s (n = 54) 0.30** 0.27** 0.25* 0.34*912.0 s (n = 72) 0.34* 0.37* 0.27* 0.32*

aMVPA = planned exercise, walking for exercise, walking for errands, and outdoor chores.bMV = accelerometer moderate to vigorous cut point set at 1040 counts per minute per day.cMV minute per day = accelerometer moderate to vigorous, cut point set at 760 counts per minute per day.*P G 0.05 **P G 0.01.

TABLE 4. Responsiveness to change of the Incidental and Planned Exercise Questionnaire (IPEQ) in RCT that promoted planned walking and in a subsample with accelerometer valid daysat baseline and 12-month follow-up.

Intervention (n = 190) Control (n = 196) Responsiveness index

IPEQ (n = 315) Mean Change Hours per Week (SD) Mean Change Hours per Week (SD) R (95% CI)

Walking for exercise 1.42 (1.95) 0.27 (3.13) 0.45 (0.32; 0.59)Walking for errands 0.62 (3.78) 0.60 (4.69) 0.13 (0.01; 0.26)All walking 2.05 (4.57) 0.87 (5.73) 0.36 (0.21; 0.50)Other planned exercise 0.26 (1.74) 0.55 (2.02) 0.13 (0.0; 0.26)IPEQ-MVPAa 2.46 (8.01) 1.50 (7.23) 0.34 (0.16; 0.52)Outdoor chores 1.17 (8.48) 0.64 (7.05) 0.17 (j0.01; 0.34)Indoor chores 1.24 (7.58) j0.18 (8.01) 0.15 (0.01; 0.30)Incidental PA 2.63 (13.45) 0.71 (13.02) 0.20 (0.02; 0.38)Total PA 4.36 (13.92) 1.57 (14.08) 0.31 (0.16; 0.46)Accelerometer (n = 85) N = 42 N = 43MVb (minIdj1) 13.3 (82.90) 15.9 (20.50) 0.65 (j0.63; 1.93)Light (minIdj1) 42.6 (67.09) 65.7 (55.80) 0.78 (0.35; 1.18)

aMVPA = planned exercise, walking for exercise, walking for errands, and outdoor chores.bMV = moderate-to-vigorous intensity.


EPIDEM

IOLO

GY


exclusion of ‘‘active’’ individuals. The latter may underesti-mate the IPEQ coefficients that had been included in thispopulation, and the correlation coefficients for MVPA wouldlikely have been higher. For example, using Mini-Log, thevalidity coefficients of PASE, YPAS, and CHAMPS in asample of 51 inactive older adults were 0.38, 0.17, and 0.20,respectively, but much higher for the 36 community activevolunteers (0.55, 0.55, and 0.36) (19). The desirable validitycoefficients for PA questionnaires compared with accelerome-ters have been suggested to be Q0.50, but no self-administeredquestionnaires used in older adults have achieved this goal(14). Two validity studies of the interview-administeredPASE questionnaire have neared this level; one studyreported a correlation coefficient of 0.49 (38), and the otherstudy reported a coefficient of 0.43 (9). These data, how-ever, were garnered from small samples (n = 20 and n = 52)of healthy active volunteers; hence, the generalizability ofthe PASE questionnaire to the older general population isstill questionable.

Self-reported moderate-intensity and walking questionshave consistently shown lower repeatability and validityscores than vigorous intensity (40). This is likely due to lowsalience (i.e., cognitively less distinctive) of daily activities(3,31). Further, the low ambulatory components of light-intensity activities may lower the coefficients of the inci-dental PA questions despite applying a low cut point. It ispossible that a 24-h diary would provide a better means fordetermining criterion validity for incidental PA or otherquestionnaires that emphasize low-intensity PA. IPEQwalking questions had higher correlations with accelerome-ter MVPA minutes per day than with steps counts, contra-dicting what is usually reported (25). One explanation couldbe that the types of planned exercise reported by the validitysample comprise complex movements such as yoga, tai chi,dance, Pilates, muscle strengthening, bowls, and balance ex-ercise, which are not characterized by high-frequency gaitmovements. The CHAMPS questionnaire, which was vali-dated against pedometer step counts, had a higher correla-tion with walking frequency (rho = 0.57) than did the IPEQand a coefficient similar to IPEQ for duration of walking(rho = 0.40) (16).

Misunderstanding the wording of questions or responseformat due to impaired cognition (3,11) may have alsocontributed to the lower coefficients for IPEQ as a whole.The definition of indoor chores as ‘‘standing on your legsdoing tasks such as housework, and self-care’’ may promptparticipants to ignore dynamic light activities such as sweep-ing, mopping and vacuum cleaning. A second source of erroris the choice of response categories to walking questions,which may introduce misclassification. The first responsecategory is ‘‘1–2 days a week’’ and the second categoryis ‘‘3–6 days a week.’’ A person who walked for exercise5 dIwkj1 will be classified the same as one who walks3 dIwkj1 (median = 4.5 d). It is possible that equal intervalswould result in further improved validity of the walkingquestions and should be tested.

To date, CHAMPS is the only older adult questionnairethat has been evaluated in relation to response to a multi-faceted intervention using Cohen’s d effect size, which was0.38 for moderate intensity expressed as kilocalories perweek (14). Using the Tuley et al. index, CHAMPS respon-siveness in an Australian context after a 4-month telephone-based lifestyle intervention was 0.24 for walking and 0.32for all MVPA items (29). In our study, we noted increasedparticipation by control group participants in planned exer-cise by 0.42 h (or 25 min). The increases in PA levels inthe control group may be attributed to the Hawthorn effectwhereby participants modify their behavior in response tobeing in the study (32). Although social desirability mayhave exaggerated the changes in self-reported planned walk-ing or other exercise for the control and intervention partici-pants, changes were also detected in activities that were notpromoted (i.e., incidental PA). Further, the responsivenessto change in IPEQ-MVPA (approximately 0.34) and inciden-tal PA were corroborated by objectively measured MVPA andlight-intensity PA. This suggests that IPEQ is sensitive tochanges in PA over time with and without intervention.

Among the strengths of this study are the use of largesample, the inclusion of a representative population of olderpeople of different ethnic backgrounds, and the examina-tion of measurement properties according to physical abili-ties and covariates. Further, the study provides coefficientsfor incidental PA, which is the main source of activity forolder people and presents two validity measures. Thesefindings add to the previously established stability andcontent validity of IPEQ (8) and demonstrate that IPEQ hasnow undergone four out of the five recommended validityassessments (24).

CONCLUSION

IPEQ is a practical PA measure for surveillance and in-tervention. The scale’s validity is satisfactory for MVPA,particularly for walking. IPEQ incidental PA is cut point de-pendent and has satisfactory correlation with light intensity,considering the difficulty to recall activities that are moreroutine and the low ambulatory component of indoor chores.Future research should determine whether small changes tothe IPEQ response format improve its criterion validity, assessIPEQ’s validity in a sample that includes also sufficiently ac-tive older people, and validate IPEQ incidental questionsagainst 24-h diary entries.

This study was funded by the NSW Health, Health Promotion Re-search Demonstration Grant Scheme. The authors thank CameronHicks, Joanna Jaques, and Mayna Ratanapongleka for collecting dataand the general day-to-day operation of the study.

DM, SL, and KD received salary funding from the Australian Na-tional Health and Medical Research Council.

The authors declare no conflict of interests.The results of the present study do not constitute endorsement

by the American College of Sports Medicine.


EPIDEM

IOLO

GY


REFERENCES

1. Ainsworth BE, Caspersen CJ, Matthews CE,Masse LC, Baranowski T,Zhu W. Recommendations to improve the accuracy of estimates ofphysical activity derived from self report. J Phys Act Health. 2012;9:S76–84.

2. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium ofphysical activities: an update of activity codes and MET intensi-ties. Med Sci Sports Exerc. 2000;32(9 Suppl):S498–516.

3. Baranowski T. Validity and reliability of self-report measures ofphysical activity: an information porocesing perspective. Res QExerc Sport. 1988;59(4):314–27.

4. Bland MJ, Altman DG. Statistical methods for assessing agree-ment between two methods of clinical measurement. Lancet.1986;1(8476):307–10.

5. Centre for Health Advancement and Centre for Epidemiology andResearch. New South Wales Fall Prevention Baseline Survey:2009 Report. Sydney: New South Wales Department of Health;2010. Available from: New South Wales Department of Health.

6. Cohen J. Statistical Power Analysis for Behavioral Sciences.Hillsdale (NJ): Lawrence Earlbaum associates; 1988.

7. Copland JL, Dale EW. Accelerometer assessment of physical ac-tivity in active healthy older adults. J Aging Phys Act. 2009;17:17–30.

8. Dalbaere K, Hauer K, Lord S. Evaluation of the Incidental andPlanned Activity Questionnaire (IPAQ) for older people. Br JSports Med. 2010;44:1029–34.

9. Dinger MK, Oman RF, Taylor EL, Yesely SK, Abie J. Stabilityand convergent validity of the Physical Activity Sacle for the El-derly (PASE). J Sports Med Phys Fitness. 2004;44(2):186–92.

10. DiPietro L. Physical activity in aging: changes in patterns andtheir relationship to health and function. J Gerontol. 2001;56A(Special Issue II):13–22.

11. Durante R, Ainsworth BE. The recall of physical activity: usingcognitive model of the question-answering process.Med Sci SportsExerc. 1996;28(10):1282–91.

12. Evenson KR, Buchner DM, Morland KB. Objective measurementof physical activity and sedentary behavior among US adults aged60 years or older. Prev Chronic Dis. 2012;9(11):010109.

13. Fiatarone-Singh MA. Exercise comes of age: rationale and rec-ommendations for a geriatric exercise prescription. J Gerontol.2002;57A(5):M262–82.

14. Forsen L, Loland NW, Vuillemin A, et al. Self-administered phys-ical activity questionnaires for elderly. Sports Med. 2010;40(7):601–23.

15. Garber CE, Blissner B, Deschenes MR, et al. Quantity and qualityof exercise for developing and maintaining cardiorespiratory,musculoskeletal, and neuromotor fitnss in apparently healthyadults: guidance for prescribing exercise. Med Sci Sports Exerc.2011(7);1334–59.

16. Giles K, Marshall AL. Repeatability and accuracy of CHAMPS asa measure of physical activity in a community sample of olderAustralian adults. J Phys Act Health. 2009;6(2):221–9.

17. Guralnik JM, Simonsick EM, Ferrucci L, et al. A short physicalperformance battery assessing lower extremity function: associa-tion with self-reported disability and prediction of mortality andnursing home admission. J Gerontol. 1994;49(2):M85–94.

18. Hagstromer M, Ainsworth BE, Kwak L, Bowles H. A checklist forevaluation the methodology qaulity of validation studies on self-report instruments for physical activity and sedentary behavior.J Phys Act Health. 2012;9:S29–36.

19. Harada ND, Chiu V, King AC, Stewart AL. An evaluation of threeself-report physical activity instruments for older adults. Med SciSports Exerc. 2001;33(6):962–70.

20. Heesch KC, Uffelen van JGZ, Hill RL, Brown WJ. What do IPAQQuestions mean for older adults. Lessons from cognitive in-terviews. Int J Behav Nutr Phys Act. 2010;7:35.

21. Heesch KC, van Uffelen JGZ, Brown W. How do older adultsrespond to Active Australia physical activity questions? Lessonsfrom cognitive interviews. J Aging Phys Act. 2014;22(1):74–86.

22. Jorstand-Stein EC, Hauer K, Becker C, et al. Suitability of physicalactivity questionnaires for older adults in fall-prevention trials: asystematic review. J Aging Phys Act. 2005;13:461–81.

23. King AC, Guralnik JM. Maximizing potential of an aging popu-lation. JAMA. 2010;304(17):1944–5.

24. Masse LC, Niet JE. Source of validity evidence needed with self-report measures of physical activity. J Phys Act Health. 2012;9(1 Suppl):S44–55.

25. Matthews CE. Calibration of accelerometer output for adults.Med Sci Sports Exerc. 2005;37(11 Suppl):S512–22.

26. Merom D, Cosgrove C, Venugopal K, Bauman A. How diversewas the leisure time physical activity of older Australian over thepast decade. J Sci Med Sport. 2012;15(3):213–9.

27. Nelson ME, Rejeski WJ, Blair SN, et al. Physical activity andpublic health in older adults: recommendation from the AmericanCollege of Sports Medicine and the American Heart Association.Med Sci Sports Exerc. 2007;39(8):1435–45.

28. Pfeiffer E. A short portable mental status questionnaire for theassessment of organic brain deficit in elderly patients. J Am GeriatrSoc. 1975;23:433–41.

29. Reeves MM, Marshall AL, Owen N, Winkler EAH, Eakin EG.Measuring physical activity change in broad-reach interventiontrials. J Phys Act Health. 2010;7:194–203.

30. Rowe JW, Kahn RL. Successful aging. Gerontologist. 1997;37(4):433–40.

31. Shephard RJ. Limits to the measurement of habitual physical ac-tivity by questionnaires. Br J Sports Med. 2002;37:197–206.

32. Sluijs EMF, Poppel MNM, Twisk JWR, Mechelen W. Physicalactivity measurements affected participants’ behavior in a randomisedcontrolled trial. J Clin Epidemiol. 2006;59:404–11.

33. Stewart AL, Mills A, King AC, Haskell WL, Gills D, Ritter PL.CHAMPS Physical activity questionnaire for older adults: out-come for interventions. Med Sci Sports Exerc. 2001;33:1126–41.

34. Strath SJ, Pfeiffer KA, Whitt-Glover MC. Accelerometer use withchildren, older adults, and adults with functional limitations. MedSci Sports Exerc. 2012;44:S77–85.

35. Tuley MR, Mulrow CD, McMaHan AC. Estimating and testing anindex of responsiveness and the relationship of the index to power.J Clin Epidemiol. 1991;44:417–21.

36. Voukelatos A, Merom D, Rissel C, Sherington C, Watson W,Waller K. The effect of walking on falls in older people: the EasySteps to Health randomised controlled trial study protocol BMCpublic health. 2011;11:888.

37. Washburn RA. Assessment of physical activity in older adults. ResQ Exerc Sport. 2000;71(2):79–88.

38. Washburn RA, Flicker JL. Physical Activity Scale for the El-derly (PASE): the relationship with activity measured by aportable accelerometer. J Sports Med Phys Fitness. 1999;39:336–40.

39. Wojtek J, Chodzko-Zajko W, Proctor DN, et al. Exercise andphysical activity for older adults. Position Stand. Med Sci SportsExerc. 2009;41(7):1510–30.

40. Yore MM, Ham SA, Ainsworth BE, et al. Reliability and validityof the instrument used in BRFSS to assess physical activity. MedSci Sports Exerc. 2007;39(8):1267–74.


EPIDEM

IOLO

GY


Incidental and Planned Exercise Questionnaire for Seniors

Documents

Transcript of Incidental and Planned Exercise Questionnaire for Seniors