Journal of Aging and Physical Activity, 2011, 19, 347-372
© 2011 Human Kinetics, Inc.
Assessment of Physical Activity
in Older People With and Without
Cognitive Impairment
Klaus Hauer, Stephen R. Lord, Ulrich Lindemann,
Sarah E. Lamb, Kamiar Aminian, and Michael Schwenk
The purpose of this study was to validate a new interview-administered physical
activity questionnaire (Assessment of Physical Activity in Frail Older People;
APAFOP) in older people with and without cognitive impairment. The authors
assessed feasibility, validity, and test–retest reliability in 168 people (n = 78 with,
n = 88 without cognitive impairment). Concurrent validity was assessed against an
inertia-based motion sensor and an established questionnaire. Sensitivity to change
was tested in an ongoing study in patients with mild to moderate dementia (n = 81).
Assessment of physical activity by the APAFOP and the motion sensor correlated
well in the total sample (TS; p = .705), as well as in the subsamples with cognitive
impairment (CI; p = .585) and without CI (p = .787). Excellent feasibility with an
acceptance rate of 100%, test–retest reliability (intraclass correlation coefficients
ranging from .973 (TS) to .975 (CI) to .966 (no CI), and sensitivity to change
(effect sizes: 0.35–1.47) were found in both subsamples.
Keywords: activity questionnaire, dementia, validity, reliability, responsiveness,
aged
A recent systematic review of questionnaires to measure physical activity in
older people (Jorstad-Stein et al., 2005) concluded that although some measures
had specific strengths, none were satisfactory for use in older frail or impaired
people. Major limitations of these questionnaires derive from the fact that only a
few have been developed specifically for older adults. Even those that were were
developed for and validated in relatively “young older” people—those age 60–70
years (Stewart et al., 2001; Voorrips, Ravelli, Dongelmans, Deurenberg, & van
Staveren, 1991). None were specifically developed for and validated in the “old
old” or people with cognitive impairment.
Hauer and Schwenk are with the Bethanien-Krankenhaus-Geriatric Center, University of Heidelberg,
Heidelberg, Germany. Lord is with the Prince of Wales Medical Research Institute, University of New
South Wales, Sydney, Australia. Lindemann is with the Dept. of Clinical Gerontology, Robert Bosch
Krankenhaus, Stuttgart, Germany. Lamb is with the Warwick Clinical Trials Unit Health Sciences
Research Institute, University of Warwick, Warwick, UK. Aminian is with the Laboratory for Movement
Analysis and Measurement, Lausanne Federal Poytechnical School, Lausanne, Switzerland.
347
348 Hauer et al.
The assessment of physical activity in very old people has specific challenges
with regard to the accurate measurement of activity patterns. In contrast to sports
activities in younger people, such activities are often characterized by short activities of daily living (ADLs) or nonexercise activity thermogenesis activities, which
are not easily remembered and therefore often not documented in assessment of
physical activity. However, such low levels of physical activity represent the most
prevalent form of activity even in the general population (Shephard, 2003; Washburn, Heath, & Jackson, 2000) and make a significant contribution to total energy
expenditure (Blair, Kohl, & Barlow, 1993; Masse et al., 1998). Nonexercise activity thermogenesis activities determine total caloric expenditure substantially and
exceed the impact of sports activities in sedentary people (Levine et al., 2005). In
frail older people or very sedentary people, health benefits may result from very low
levels of activity that are unlikely to induce breathlessness, sweating, or increase
aerobic fitness (Blair, Cheng, & Holder, 2001; Wannamethee & Shaper, 2001). It
seems that documentation of multiple sporadic activities that are typical of ADLs
is especially relevant in frail older people (Hardman, 2001).
A major challenge regarding the oldest old is the documentation of such lowlevel and fragmented physical activity. Many questionnaires suffer from floor effects
(Tudor-Locke & Myers, 2001), and even well-established questionnaires do not
measure activities less intense than brisk walking or activities with a duration less
than 10 min (Blair et al., 1985), yet these are the typical activity patterns in old age.
Accuracy of physical activity documentation depends on recall of activities,
and not every type of activity is recalled with the same accuracy. Previous studies
have reported better accuracy in the recall of high-intensity exercise than more
common, lower intensity activities (Dipietro, Caspersen, Ostfeld, & Nadel, 1993;
Baranowski, 1988). Structured exercise activities are usually performed in a stable,
scheduled time frame and are therefore easy to recall (Caspersen, 1989). In contrast,
brief episodes or low-intensity activity as in ADLs are harder to remember. Other
factors influencing recall and accuracy of physical activity include assessment
factors such as form of administration, duration of assessment period, relevant and
meaningful questionnaire items, and interview structure and strategies (Baranowski,
1988; Blair, Dowda, & Pate, 1991; Durante & Ainsworth, 1996; Shephard, 2003;
Washburn, Smith, Jettre, & Janney, 1993).
Patients with cognitive impairment show illness-related symptoms such as
memory impairment, loss of orientation in time and locus, impairment in semantic
performance, decreased ability for mathematical calculation, and impaired selfperception (American Psychiatric Association, 1994). Such deficits may also occur
in the course of age-related decline and substantially restrict the accuracy of physical
activity reports, resulting in the exclusion of the old or cognitively impaired people
from most previous studies.
To our knowledge no specific questionnaire has been developed for frail,
impaired older people that addresses these limitations in a comprehensive approach.
We therefore developed an interview-administered physical activity questionnaire
(Assessment of Physical Activity in Frail, Older People [APAFOP]) specifically
designed for frail older people with and without mild to moderate cognitive impairment. In this article we present data on its concurrent validity, test–retest reliability,
feasibility, and sensitivity to change.
Physical Activity Questionnaire in the Elderly 349
Method
Participants
Participants were recruited from a geriatric rehabilitation ward and a communitydwelling population after screening for the following inclusion criteria: age >65
years, no severe psychological or somatic disease, ability to walk 10 m without
using a walking aid, and provision of written, informed consent. In people with
cognitive impairment with a legal representative, an additional written, informed
consent was obtained. Participants with scores on the Mini Mental State Examination (MMSE; Folstein, Folstein, & McHugh, 1975) below 24 were assigned to
the cognitively impaired group, and those with scores of 24–30 to the cognitively
intact group. The study was performed according to the Helsinki declaration, and
approval by the ethics committee of the local university was obtained. Sensitivity
to change of the questionnaire was determined in an ongoing study in patients with
mild to moderate dementia using inclusion criteria similar to those in the original
sample just described.
Descriptive Measures
Age, gender, education (years of educational and professional training), a singleitem question on fear of falling (Maki, Holiday, & Topper, 1991), the Barthel Index
on ADLs (Mahoney & Barthel, 1965), five chair stands (Guralnik et al., 1996), the
SF-12 (Gandek et al., 1998), the MMSE, and number of falls in the previous year
were documented by standardized patient interview or testing.
Development of the APAFOP
The questionnaire was developed in a European research network (Prevention of
Falls Network Europe; ProFaNE; http://www.profane.eu.org). In a systematic review
of existing questionnaires to assess physical activity used in fall-prevention trials, no
questionnaires were identified that met the criteria for adequately documenting the
types of physical activity typical for frail older people (Jorstad-Stein et al., 2005).
As a first step, the then unpublished questionnaire by Delbaere, Hauer, and Lord
(the Incidental and Planned Activity Questionnaire; 2009) was used as an initial
methodological approach to cover the main physical activity domains relevant to
older people. This questionnaire was substantially modified to meet the needs of
the target group. In contrast to previous questionnaires that have used an extended
list of single activities, the newly developed APAFOP focuses on items such as
walking (including walking periods for less than 3 min), standing and time on
feet indoors and outdoors (including mixed activities such as standing or walking
periods of less than 3 min), sitting, and lying, which have different intensity ratings
and better represent typical forms of activity in old age. The focus on such ADLrelated activities prevents floor effects in the otherwise sedentary target group. To
prevent ceiling effects in high-functioning people and allow for documentation of
physical training in intervention studies, sports activities were also included. The
questionnaire items and ratings are provided in Appendix A.
350 Hauer et al.
Assessing Total Activity and Energy Expenditure
To capture total activity we used an established calculation method for physical
activity (Schuler, Richardson, Ochoa, & Wang, 2001): multiplication of MET levels
by the duration of an activity per time unit (24 hr). As an option for energy estimation
not used in this study, MET per time equivalents allows for an approximation of total
individual energy expenditure when individual body weight is included (kcal/day).
Intensity Rating
We used adjusted intensity codes as suggested by Ainsworth et al. (2000) and modified by Stewart et al. (2001) and adjusted MET values for our target population.
Readjustment was necessary because the study participants were almost a decade
older than the original target population and included frail people with multiple
morbidities following the same strategy for modification as the previously mentioned expert group. A comprehensive list of activities and related METs used in
this study, including MET values by Ainsworth et al. and Stewart et al., on which
our modification of MET values is based, is presented for comparison in the interviewer’s manual (see http://www.profane.eu.org/).
Methods to Facilitate Accurate Reporting and Diminish
Recall Failure
Self-administered questionnaires and long retrospective assessment periods increase
recall problems and decrease the accuracy of reports. We therefore developed a
highly structured interview-based questionnaire to support recall and restricted the
assessment period to 24 hr.
Theoretically based approaches to foster memory have been suggested in different research areas (Baranowski, 1988; Durante & Ainsworth, 1996), but divergent
concepts have so far not been integrated into one comprehensive method to assess
physical activity. In the development of the APAFOP we applied a number of
strategies to foster recall aimed specifically at our target population with memory
impairments. Methods used in this study included comprehensive verbal information and feedback from participants before and in the course of the assessment, to
eliminate miscomprehension or ambiguities that are substantial sources of errors
in the question-answering process (Durante & Ainsworth, 1996); reassurance and
generation of an informal conversational approach to prevent fear of failure in
comprehension and recall; use of specific time frames as “anchors” to segment
the recall period to clue and structure memory (Baranowski, 1988; Baranowski,
Sworkin, & Cieslik, 1984); addressing typical or habitual activities to clue recall
and improve the completeness of reports (Baranowski, 1988); and rehearsal of
activities reported by summarizing reported activities at the end of the interview
(Souchay, Moulin, Isinarini, & Conway, 2008).
Documentation, Interview Process, and Physical Activity
Measurements
Specific activity, duration, and intensity were documented using the assessment
form (see Appendix B). Examples of MET-intensity scores for different forms of
Physical Activity Questionnaire in the Elderly 351
activities are listed to support intensity rating and allow comparison with comparable ratings (see Appendix B). The interview process and the use of assessment
forms have been summarized in a comprehensive interviewer’s manual that will
be published at an open-access Web site (http://www.profane.eu.org/) to enable
training and comprehensive instruction for users.
Patients were interviewed at their homes without participation of proxies or
caregivers, and the physical activity assessment was performed in both the impaired
and nonimpaired groups in a standardized manner. Descriptive data were documented for most of the participants (n = 123) during previous hospital stays. For
those participants who had been recruited in the community, descriptive data were
obtained during their interview at home. Interviewers were trained intensively with
the test manual and underwent repeated, supervised test runs to be able to perform
the standardized interview assessments. Measurements that covered a 24 hr-period
were initiated between 8 and 10 a.m. and continued throughout the day and night
on random days of the week except for Sundays. Measurements with wear times
of less than 23 hr were excluded from analysis.
Assessment of Measurement Properties:
Acceptability and Feasibility
To address issues of acceptability and feasibility we examined response rates for
completion of the interview, missing data, and completion time for documentation.
Concurrent Validity
We used two independent measures of direct validation strategies: an objective,
technologically advanced motion-capture system (Physilog) and the established
Physical Activity Questionnaire for the Elderly (PAQE; Voorrips et al., 1991).
Objective Measurement Using the Physilog System. The Physilog system
(BioAGM, CH) is a small (95 × 60 × 22 mm), light (122-g), long-term recording
system containing inertial sensors (two accelerometers and one gyroscope) with
software developed to identify postural positions and movements such as walking,
standing, sitting, or lying every second during a measuring period of up to 48 hr. The
analysis algorithm is described elsewhere in detail and has been validated in different
groups of subjects including elderly adults. It has proven to be sensitive (87–99%)
and specific (87–99.7%) for postural positions and detection of walking in different
samples of older adults and patient groups (Najafi et al., 2003; Paraschiv-Ionescu,
Buchser, Rutschmann, Najafi, & Aminian, 2004) and Parkinson’s patients (Salarian,
Russmann, Vingerhoets, Burkhard, & Aminian, 2007). The technical features of
the Physilog allowed a direct comparison of patient reports by the APAFOP for
single items, as well as total scores (e.g., single activities such as walking could be
directly compared with the APAFOP item “walking”). To evaluate the overall daily
activity as measured by the Physilog, we developed an activity score based on the
Physilog activity assessment. Activities such as lying or sitting were weighed by a
factor of 1, standing by a factor of 1.5, and walking by a factor of 2 as an approach
to adjust for typical intensities in this old-age cohort and to allow comparison with
the intensity-weighted total APAFOP score.
352 Hauer et al.
PAQE. The PAQE has good construct validity and retest reliability (Voorrips et
al., 1991) and has been rated highly in a previous systematic review (Jorstad-Stein
et al., 2005). Although it is one of the few questionnaires especially developed for
older people and shares some common features with the APAFOP (application
as interview and inclusion of ADL activities), the PAQE has some substantial
differences from the APAFOP. It was developed for and validated in the “young
old.” ADLs and household activities are solely assessed as the frequency of different
activities in a predefined scoring system not including duration and intensity, it
emphasizes intensity in non-ADL activities, it has no specific interview techniques
to prevent recall problems, and the assessment period is 1 year. To better compare
results of the two questionnaires and adjust for memory limitations of the study
group we restricted the assessment period of the PAQE to 1 week. The PAQE
questionnaire was administered within 3 days of Physilog and APAFOP assessments
to ensure that a comparable time period was covered by the three measurements.
Test–Retest Reliability
Test–retest reliability was assessed in 30 people recruited with subsamples of 15
people for each group according to cognitive status. The questionnaire was administered twice with an interval of 24 hr between by the same interviewer to exclude
interobserver variability.
Sensitivity to Change
For obtaining sensitivity-to-change data we used a subsample of 81 people with
mild to moderate dementia from a large ongoing randomized, controlled intervention trial. The intervention group performed high-intensity, progressive strength
(70–80% of maximal individual performance) and progressive functional training
for 4 hr/week. The control group performed a low-intensity stretching/ball-game
exercise program while seated for 2 hr/week. Both sessions were group-based,
supervised by an experienced trainer, and strictly standardized with respect to
duration and intensity. The total activity (including habitual and training activity)
was documented by the questionnaire (APAFOP) and the objective measurement
(Physilog system) at baseline before intervention and at the end of the intervention
period including a day of training. Interobserver variability was excluded because
all interviews were performed by the same interviewer.
Analytic Strategy
Group comparisons between subgroups for descriptive variables were made by
t test for continuous or chi-square test for dichotomous variables as appropriate.
Test-retest-reliability analysis was evaluated using intraclass correlations (ICCs)
using a one-way random-effect model (single measure) or, alternatively, a twoway mixed-effects model (single measure), as indicated in tables. For correlations,
Spearman’s rank correlation and Pearson’s correlation coefficient were used as
appropriate. Correlations were considered low (r < .2), moderate (r = .2–.5), or
high (r > .5) according to the recommendations of Cohen (1988).
Physical Activity Questionnaire in the Elderly 353
In addition, a Bland–Altman plot for visualization of study results was composed (Bland & Altman, 1986). Effect sizes for change within groups were calculated as mean changes divided by pooled standard deviation at baseline (Kazis,
Anderson, & Meenan, 1989). Effect sizes represent established methods for reporting the magnitude of change. They are expressed in units of variability without unit
of measurement (Cohen, 1988, 1992). Effect sizes of .20 are considered small, .50
medium, and above .80 large (Cohen, 1992; Katz, Larson, Phillips, Fossel, & Liang,
1992; Kazis et al., 1989). We tested group differences in correlation coefficients by
means of the standard procedure of applying Fisher’s z transformation and testing
the critical ratio of the difference of the z scores as normally distributed (Altman
& Gardner, 2000; Howell, 2007). Statistical analysis was performed with SPSS
for Windows, Version 16.0.
Results
The total sample included 168 people living at home (n = 152, 90%) and in seniors’
homes (n = 16, 10%) comprising old and functionally limited people, with more
than half of the participants reporting one or more falls during the previous 12
months. Data for the settings (community or institutionalized) were pooled and
then stratified according to cognitive impairment to yield three samples for analysis:
the whole sample (unstratified), a sample comprising people with an MMSE <23
(20.2 ± 2.5), and a sample comprising people with an MMSE ≥24 (26.1 ± 1.9).
The subgroups did not differ with respect to motor performance, history of falls,
depressive symptoms, or gender. The cognitively impaired group was on average
2.5 years older (Table 1).
Table 1 Sample Characteristics
Variable
Age, M (SD)
Men, n (%)
MMSE score, M (SD)
Five-chair-rise, s, M (SD)
Depression (SF12 Item 6/7),
n (%)
Fear of falling, n (%)
One or more falls during previous 12 months, n (%)
Total group,
N = 166
81.92 (6,80)
39 (23.5)
23.29 (3.64)
18.28 (8.98)
Cognitively
impaired,
n = 78
83.26 (6.05)
21 (26.9)
20.24 (2.49)
18.30 (8.73)
Cognitively
intact,
n = 88
80,74 (7.22)
18 (20.5)
26.06 (1.87)
18.26 (9.26)
Differences
between
subgroups
p = .017
p = .327
p < .001
p = .979
58 (34.94)
30 (18.1)
32 (41.03)
13 (16.7)
26 (29.55)
17 (19.3)
p = .132
p = .635
95 (57.2)
47 (60.3)
48 (54.5)
p = .499
Note. MMSE = Mini Mental State Examination; SF = short form. Comparisons were made by t tests for continuous
or chi-square test for dichotomous variables. Answer categories of fear of falling (no vs. some to very much) and
SF12 Item 6/7 (depression: no vs. mostly to always) have been dichotomized.
354 Hauer et al.
Feasibility and Acceptability
We excluded 4 of 82 people (5%) with cognitive impairment because in these
people recollection of physical activity events was not possible because of advanced
memory impairment, disorientation on time and locus, or confabulation. No
participant objected to the assessment procedure, and data documentation was
comprehensive, with no missing responses for any questionnaire item in either
subgroup. Physilog measurements were refused because of unwillingness to wear
the device in at least one out of a series of four measurements in 19% of the people
assessed. Three Physilog measurements could not be analyzed because of technical
problems. Interviewer reports in a subsample of interviews indicated a mean time
of 24 ± 5 min to complete the interview.
Association of Activity Scores by APAFOP With Objective
Measurement (Physilog)
The results of the APAFOP correlated well with the results of the objective Physilog
measurement. Associations for the total activity scores, time on feet, and summarized scores for inactivity (including sitting and lying) and activity (including
walking, standing, and, for the APAFOP, sports activities) were more pronounced
than with other activities. However, all except one correlation (sitting for the cognitively impaired: ρ = .172) showed significant associations (range ρ = .230–.787).
Results for people without cognitive impairment showed higher associations for
most activity-related parameters (including total score activity, walking, total activity) but not for subscores associated with inactivity (including sitting, lying, total
inactivity; see Table 2). The scores are based on duration and intensity of activities.
To separately document the association of duration of activities, including standing, time on feet, and walking, a Bland–Altman plot is provided in Figure 1. The
mean association was –.306, indicating a moderate influence of activity status (low
vs. high) on association between measures. Activity durations as measured by the
Physilog and APAFOP correlated highly with each other (ρ = .691).
Association Between the APAFOP, PAQE, and Physilog
In a subsample of 108 participants including all patients of the intervention sample
for which all measurements were available (cognitively impaired, n = 65; cognitively intact, n = 43), the APAFOP, Physilog, and PAQE results were compared.
The subsample of 108 people did not differ from the total sample (N = 168). The
APAFOP showed moderate to good association with the PAQE. Compared with
the APAFOP, the PAQE had a weaker association with the objective measurement
(Physilog), especially in the impaired group. The APAFOP showed excellent
correlation to the Physilog in all study groups. In all correlations performed, the
cognitively intact people tended to higher associations between assessment tools
(see Table 3).
Test–Retest Reliability of the APAFOP
In a subsample of 30 participants (15 cognitively intact and 15 cognitively impaired)
consecutively recruited in the intervention sample the APAFOP showed excellent
355
(p < .001)
ρ = .715
(p < .001)
ρ = .688
(p < .001)
ρ = .705
(p < .001)
ρ = .511
(p < .001)
ρ = .623
(p < .001)
ρ = .230
(p = .003)
ρ = .422
Correlation
coefficient
.597–.760
.630–.782
.287–.539
.080–.369
.519–.707
.388–.614
.618–.773
95% CI
ρ = .311
(p = .006)
ρ = .615
(p < .001)
ρ = .600
(p < .001)
ρ = .585
(p < .001)
ρ = .351
(p = .002)
ρ = . 623
(p < .001)
ρ = .172
(p = .132)
Correlation
coefficient
.433–.724
.452–.735
–.053 to
.379
.093–.480
.462–.741
.137–.530
.413–.712
95% CI
Cognitively Impaired,
n = 78
ρ = .469
(p < .001)
ρ = . 784
(p < .001)
ρ = . 753
(p < .001)
ρ = .787
(p < .001)
ρ = .600
(p < .001)
ρ = .624
(p < .001)
ρ = .271
(p = .011)
Correlation
coefficient
.643–.830
p = .070
p = .032
p = .237
.285–.616
.685–.852
p = .510
p = .991
p = .039
p = .013
Differences between
subgroups
.064–.453
.474–.736
.444–.718
.690–.854
95% CI
Cognitively Intact,
n = 88
Note. Spearman’s rho (ρ) is given for correlation between variables. Activity summarizes walking and time on feet (in Physilog measurements/all specified activities in APAFOP);
Inactivity summarizes sitting and lying. Correlation coefficients were tested for statistical differences between subgroups.
Inactivity
Activity
Lying
Sitting
Time on feet
Walking
Total activity
Position/Activity
Total Group,
N = 166
Table 2 Correlations Between the Assessment of Physical Activity in Frail Older People (APAFOP) and Physilog Scores
356 Hauer et al.
Figure 1 — Bland–Altman plot for comparison of Physilog and APAFOP measurements
(Physilog + APAFOP/2 measurements for x axis and difference of scores (Physilog –
APAFOP) for y axis), including mean deviation, mean, and limits of agreement. Regression lines are given for subsamples according to cognitive status. Open circle = cognitively
impaired (n = 78); closed circle = cognitively intact (n = 88); dashed line = trend line
cognitively impaired; solid line = trend line cognitively intact.
reproducibility for the total sample (intraclass correlation [ICC] .973), as well as
for the subsamples of cognitively impaired (ICC .975) and cognitively intact people
(ICC .966) for the total score. Detailed analysis also showed good to excellent
test–retest reliability for single activities for the total sample (ICC range .907–.973),
as well as for the subgroups of cognitively impaired (ICC range .905–.998) and
cognitively intact people (ICC range .800–.991). Group means of activities assessed
at test and retest showed high agreement in all groups. Test–retest reliability tended
to be higher in the subsample of cognitively impaired people. The activity level of
the cognitively intact people was higher than that of the impaired people (Table 4).
Sensitivity to Change
The APAFOP adequately reproduced the changes in physical activity during the
3-month intervention in both study groups. High effect sizes were evident for the
high-intensity training group when training periods were compared with the habitual
sedentary lifestyle, indicating the substantial increase in physical activity induced
by training. Moderate effect sizes were documented for the less intense training.
The technical measurement was less responsive (Table 5). Patients with cognitive
impairment showed higher responsiveness in questionnaire measurements (Table 6).
357
r = .650 (p < .001)
r = .695 (p < .001)
Physilog–APAFOP
PAQE–APAFOP
.529–.749
.580–.780
.298–.598
95% CI
r = .601 (p < .001)
r = .544 (p < .001)
r = .307 (p = .013)
Correlation
coefficient
.415–.735
.358–.700
.065–.510
95% CI
Cognitively Impaired (n = 65)
r = .794 (p < .001)
r = .755 (p < .001)
r = .644 (p < .001)
Correlation
coefficient
.642–.881
.582–.858
.420–.789
95% CI
Cognitively Intact (n = 43)
.053
.062
.025
p
Note. APAFOP = Assessment of Physical Activity in Frail Older People; PAQE = Physical Activity Questionnaire for the Elderly. Presented are Pearson’s correlation coefficient
(r) and p values for significance for comparison of total scores of assessment instruments. p Values are also given for differences of correlations between subgroups (right column).
r = .462 (p < .001)
Physilog–PAQE
Assessment
Correlation
coefficient
Total Group (N = 108)
Table 3 Correlations Between Physilog, APAFOP, and PAQE
Table 4 Test–Retest Reliability for Single Activities
Item
Total activity, S
Walking, S
Total Sample
(N = 30)
M (SD)
ICC
Cognitively Impaired
(n = 15)
M (SD)
ICC
Cognitively Intact
(n = 15)
M (SD)
ICC
28.57 (2.93
.973
27.49 (2.34)
.975
29.66 (3.11)
.966
28.49 (2.87)
.972
27.48 (2.51)
.974
29.49 (2.93)
.965
1.44 (1.44)
.944
1.27 (1.30)
.966
1.62 (1.60)
.932
1.38 (1.34)
.944
1.18 (1.21)
.966
1.59 (1.47)
.927
Walking, RT
0.62 (0.56)
.919
0.59 (0.59)
.958
0.64 (0.54)
.878
0.60 (0.52)
.917
0.55 (0.54)
.958
0.64 (0.50)
.870
Time on feet OD, S
1.62 (2.47)
.946
0.89 (1.58)
.973
2.36 (3.00)
.933
1.69 (2.32)
.945
0.91 (1.58)
.971
2.47 (2.71)
.929
Time on feet OD, RT
0.69 (0.94)
.933
0.37 (0.58)
.949
1.00 (1.13)
.920
0.73 (0.95)
.932
0.39 (0.57)
.945
1.08 (1.14)
.917
Time on feet ID, S
6.34 (3.28)
.925
4.58 (1.99)
.916
8.10 (3.42)
.890
6.28 (3.18)
.923
4.61 (2.30)
.911
7.95 (3.11)
.884
Time on feet ID, RT
3.97 (1.76)
.907
2.96 (1.20)
.905
4.98 (1.96)
.846
3.94 (1.79)
.904
2.96 (1.30)
.899
4.92 (1.71)
.836
Sitting, S
8.56 (2.73)
.844
8.55 (3.36)
.950
8.58 (2.18)
.651
8.55 (2.43)
.839
8.77 (3.20)
.898
8.33 (1.41)
.657
Sitting, RT
Lying, S
Lying, RT
Sport activity, S
Sport activity, RT
8.56 (2.73)
.910
8.55 (3.36)
.950
8.58 (2.18)
.800
8.55 (2.43)
.907
8.77 (3.20)
.949
8.33 (1.41)
.796
9.96 (2.91)
.967
11.24 (3.37)
.981
8.67 (1.64)
.954
9.93 (2.91)
.963
11.08 (3.50)
.963
8.78 (1.57)
.954
9.96 (2.91)
.967
11.24 (3.37)
.981
8.67 (1.64)
.875
9.93 (2.91)
.966
11.08 (3.50)
.980
8.78 (1.57)
.869
0.65 (1.66)
.997
0.97 (2.09)
.998
0.33 (1.04)
.991
0.65 (1.65)
.985
0.93 (2.09)
.999
0.38 (1.07)
.946
0.22 (0.55)
.997
0.33 (0.70)
.998
0.11 (0.35)
.991
0.22 (0.55)
.996
0.32 (0.69)
.998
0.13 (0.36)
.991
Note. S = scores (Duration × Intensity); RT = real times measured (hr); OD = outdoors; ID = indoors. Presented
are means for test (first line) and retest (second line) and ICCs for the total sample and subsamples according to
cognitive status. ICC first line: one-way random-effects model single measure (Shrout & Fleiss, 1979, Model 1.1);
ICC second line: two-way mixed-effects model single measure (Shrout & Fleiss, 1979, Model 3.1).
358
Physical Activity Questionnaire in the Elderly 359
Table 5 Standardized Effect Sizes for Sensitivity to Change
Assessment
APAFOP T2_T vs. T1
APAFOP T2_T vs. T2_N
APAFOP T2_T vs. T3
Physilog T2_T vs. T1
Physilog T2_T vs. T2_N
Physilog T2_T vs. T3
Total group
(N = 81)
0.90
0.86
0.84
0.09
0.21
0.24
High-intensity
training (n = 37)
1.47
1.47
1.35
0.11
0.30
0.28
Low-intensity
training (n = 44)
0.40
0.35
0.40
0.10
0.15
0.22
Note. APAFOP = Assessment of Physical Activity in Frail Older People; T2_T = measurement at the
end of intervention (active day of training); T1 = baseline measurement before training started; T2_N
= measurement at the end of intervention (habitual activity, day without training); T3 = measurement at
follow-up after training had stopped. Presented are standardized effect sizes (difference between mean
scores at assessments, divided by the standard deviation of baseline scores, Kazis et al., 1989) for the
APAFOP and Physilog for 24-hr recordings during an intervention study for different training regimens.
Table 6 Standardized Effect Sizes for Change During Intervention
Total group
(N = 98)
MMSE scores <24
(n = 60)
MMSE scores ≥24
(n = 38)
APAFOP
1.01
1.11
0.84
Physilog
0.16
0.23
0.03
Parameter
Note. MMSE = Mini Mental State Examination; APAFOP = Assessment of Physical Activity in
Frail Older People. Presented are standardized effect sizes for the total group (including high- and
low-intensity training groups) and subgroups of participants according to cognitive impairment for
comparison between sedentary baseline und active training period (training day).
Discussion
The APAFOP is the first questionnaire to assess physical activity developed and
validated for use in frail older people with and without cognitive impairment. The
questionnaire is feasible, valid, reliable, and responsive and allows documentation
of approximate energy expenditure (MET-related activity scoring per time unit) for
a wide variety of activities ranging from ADL-related activities to sports activities.
Feasibility and Acceptability
Acceptability involves the physical and emotional burden of the respondent, and
feasibility focuses on the demands on those who administer a measure (Fitzpatrick,
Davey, Buxton, & Jones, 1998). As assessed in this study feasibility and acceptability were high for the APAFOP; all the cognitively intact participants and most
of the cognitively impaired could be interviewed, and physical activity could be
documented adequately. Only a small number of more severely impaired people
(5%) were excluded because of advanced disorientation in locus and time, severely
360 Hauer et al.
impaired memory, or confabulation, indicating that such assessments are not feasible
for those in advanced stages of cognitive impairment. Patients who were excluded
ranged in the lower scores of the MMSE (<20), but not all the patients with lower
scores had severe problems with the assessment. Insufficient discrimination of the
relatively small group of these patients may relate to the limitations of the MMSE,
which was developed as a screening tool for cognitive impairment, not as a detailed
diagnostic instrument for performance in specific cognitive subdomains related to
abilities specific for interview assessment.
Despite the exclusion of such severely impaired participants we found that the
assessment of specific activity patterns in older people and most people with mild
to moderate cognitive impairment was feasible. With the interview-based administration method we ascertained that data documentation was comprehensive and
no questionnaire items were left out as previously reported for self-administered
questionnaires in old or cognitively impaired people (Hauer et al., 2010). The time
to complete the interview was sufficient for a detailed questionnaire such as the
APAFOP and is comparable to the data that are available for other questionnaires
with completion times ranging from 15 min to 90 min (Sallis et al., 1985; Stewart
et al., 2001). Willingness to participate in the technical measurement was somewhat
lower than with the questionnaire assessment, although the Physilog measurement
used an individually customized, easy-to-wear support system and participants had
been contacted at home for installation and deinstallation of the system. Reasons
for discontinuing the technical measurement were “felt uncomfortable,” “could not
sleep or have a shower with device,” “forgot to put it on again,” and “forgot what
the measurement was about” and were partly related to the impaired cognitive
status of the participants.
Validity
We used two independent direct validation strategies: an objective, technologically
advanced motion-capture system (Physilog) and the PAQE based on the participant’s
physical activity ratings. Results obtained by direct measurements carry the most
weight because they measure the same construct (direct validity; McDowell &
Newell, 1996). Previous studies used different measurements as external standards
for validation (Jorstad-Stein et al., 2005). Depending on the relationship between
measurements and quality of methods used, reported correlations have been low
to moderate and have partly been limited by the indirect nature of associations,
for example, when fitness-related markers were related to physical activity measurements (Dipietro et al., 1993; Jorstad-Stein et al., 2005). Accelerometer-based
motion sensors, as used in this study, provide valid data on physical activity as
objective external measurements by detecting activity such as walking, sitting,
standing, and lying (Aminian et al., 1999; Bonnefoy et al., 2001; Bussmann, van
de Laar, Neeleman, & Stam, 1998; Grant, Dall, Mitchell, & Granat, 2008; Levine,
Melanson, Westerterp, & Hill, 2001).
In the current study, results of the APAFOP correlated very well with the
objective results derived from the motion sensors. As would be expected the
associations tended to be higher in people without cognitive impairment, because
cognitive deficits may have limited recall of activities. However, based on the
type of method to support recall, such as the structured face-to-face interview,
Physical Activity Questionnaire in the Elderly 361
short assessment period, and so on, the validity of reports in people with cognitive
impairment was comparable to results of other successful validation studies for
questionnaires targeting unimpaired older people (Dipietro et al., 1993; Harada,
Chiu, King, & Stewart, 2001; Jorstad-Stein et al., 2005; Stel et al., 2004). The total
activity scores and summarized scores for inactivity (including sitting and lying)
and activity (including walking, standing, and sports activities) all showed high
associations to the objective measurement (Physilog), confirming the excellent
recording of overall physical activity by the APAFOP. Comparison of single activities (e.g., sitting) between the Physilog system and single questionnaire items was
somewhat limited by technical issues. Differentiation between lying and sitting is
difficult because many older people rest in armchairs while seated in an almost
lying position or may lift their upper body in bed because of breathing problems
when lying. A motion-capturing system such as the Physilog, based on algorithms
to determine relative positions of the body, is therefore limited when defining a
position. However, when both sitting and lying were combined in an inactivity score,
associations between technical measurements and APAFOP results were excellent.
When comparing single questionnaire items such as walking (defined as walking
periods reported by participants) or time on feet (which includes a combination of
walking, standing, and other leisure sports activities) with the Physilog items walking or standing summarizing all walking and all standing measured objectively, it
is obvious that associations will be limited. However, the correlations were good to
excellent for single or combined activities and compared favorably with previous
validation studies (Dipietro et al., 1993; Jorstad-Stein et al., 2005; Stel et al., 2004).
Comparison of the APAFOP and PAQE
We chose the PAQE, a questionnaire developed for and validated in older people
and rated highly in a systematic review of questionnaires for use with older people
(Jorstad-Stein et al., 2005) to compare with the APAFOP. The PAQE showed
moderate (cognitively impaired) to good (cognitively intact) correlations with the
APAFOP. Compared with objective measurements (Physilog) it showed weaker
associations, especially in the cognitively impaired group (see Table 3). The modification of the original PAQE with a shortened recall period of 1 week instead of
1 year as used in this study may have altered the results in favor of the PAQE,
because many of the participants had not been able to give estimates of activities
for a time period of 1 year as suggested for the original PAQE version. This shorter
recall period also ensured that the 24-hr Physilog measurement covered a current
activity level comparable to that documented by the APAFOP and Physilog. The
superior results of the APAFOP may derive from the short recall period and the
specific interview techniques used in the APAFOP, supporting the accuracy of
reports, especially in people with cognitive impairment.
Test–Retest Reliability of the APAFOP
As reported in the systematic review, previous test–retest reliability studies on
questionnaires used to measure physical activity in older people have shown differing results depending on the setting and method, with only some questionnaires
presenting solid evidence of retest stability (Jorstad-Stein et al., 2005). In previous
362 Hauer et al.
studies test–retest reliability was measured by repeated assessment of physical
activity, with time intervals ranging from 2–3 weeks (Voorrips et al., 1991) to 6–12
months (Stel et al., 2004; Stewart et al., 2001) between assessments, which reduces
comparability of results. Such a methodological approach implies that the test
parameter (physical activity) is stable over time. This may be problematic because
methodological aspects (unreliable documentation from the method of investigation vs. unreliable, modified test criteria) are mixed because physical activity may
change for many reasons. High variability (Stel et al., 2004) and long time periods
between assessments have been discussed as reasons for low test–retest reliability
in established physical activity questionnaires, resulting in lower bound estimates
of reliability (Harada et al., 2001; Stel et al., 2004; Stewart et al., 2001). In some
questionnaires the risk of documenting such individual variance has been limited
by asking the participant to give a mean of average activities over time ranging
from 1 week (Stewart et al., 2001) to 1 year (Voorrips et al., 1991). However, such
complex questions (e.g., “What was the average walking time per week last year?”
“At what intensity?”) presuppose an intact memory, orientation in time and locus,
ability for mathematical calculation, and intact self-perception besides advanced
interviewer skills. Frail older people, especially those with cognitive impairment,
may be seriously overtaxed by such an approach.
In the current study we used a different methodological approach based on the
APAFOP’s short recall period of 24 hr. Participants were interviewed twice for the
same recall period (24 hr) within 2 days. The recall period was thereby identical,
so results were not influenced by individual variance of physical activity as in other
reliability studies using longer assessment periods. Results of the current study
therefore relate exclusively to physical activity reports and not to varying activity
levels. In this approach the APAFOP showed excellent test–retest reliability for
both study groups in total scores, as well as single activities. The trend for slightly
higher test–retest reliability in the subsample of cognitively impaired people may
have been the result of lower but stable activity levels that are remembered more
easily than more frequent and more variable activities in the cognitively intact
participants. High reliability was also achieved by the specific strategies to improve
recall of individual physical activity as developed for the APAFOP, which is particularly relevant for older people with cognitive impairments.
Sensitivity to Change
Sensitivity to change is a mandatory biometrical property of questionnaires to
detect changes in physical activity over time or effects of interventions in randomized controlled trials. As reported in the systematic review on questionnaires used
in geriatrics research (Jorstad-Stein et al., 2005), sensitivity to change has rarely
been addressed. For questionnaires that have reported responsiveness, results were
derived from very different observation periods ranging from 2 weeks to 1 year,
limiting direct comparisons between results. In some but not all studies presenting responsiveness data, a standardized intervention targeted to influence the
physical activity in the study group was performed during the observation period
(Jorstad-Stein et al., 2005). The interventions differed from each other, further
limiting comparability of the results. Effect sizes reported in these studies were
small, ranging from .03 to .30, with only one questionnaire presenting detailed
and moderate to good sensitivity-to-change data with effects size ranging from
Physical Activity Questionnaire in the Elderly 363
.38 to .64 (Stewart et al., 2001). Sensitivity-to-change measurements suffer from
a methodological problem: Do insufficient effect sizes represent a methodological
fault of the questionnaire, or do insufficient results represent insufficient change of
physical activity over time? In the case of intervention studies to increase physical
activity, this might relate to an insufficient effect of the study intervention. Such
methodological pitfalls have also been identified in previous studies, leading to the
request that “ideally the outcome measure would focus primarily on the types of
physical activities that are being targeted for change in the intervention to improve
sensitivity” (Stewart et al., 2001, p. 1128).
To prevent methodological inconsistency in this study we examined the effect
sizes of two strictly standardized training regimens to control for a standardized
change in physical activity over time. The APAFOP adequately assessed the differences between study groups and the change from primarily sedentary life style
to comparatively high physical activity induced by two different training regimens
during a standardized intervention. It also showed moderate to large responsiveness
in both subsamples with respect to cognitive status. The slightly higher effect sizes
in impaired people may be attributable to the lower overall activity at baseline and
the higher increase induced by training. The good responsiveness of the APAFOP
exceeded that of the objective external standard. Despite rapid technological
development, even advanced motion-analytic systems are not able to detect and
classify all physical activities performed by humans. Because the system used in
this study (Physilog) classifies posture positions rather than their intensity, activities such as bicycling or strength training while seated on training machines were
classified as sitting position without considering the activity of lower limbs. As a
consequence these training activities were estimated with lower energy expenditure. Such methodological pitfalls have been reported before in comparable studies indicating specific limitations of objective measurements (Westerterp, 1999).
Because the training intervention of the trial included bicycling on a stationary bike
and strength training while seated, the substantial change of activity induced by
this intervention was not adequately recorded by the Physilog system. However,
such limitations will not be relevant in most other interventions or daily activities
that are not performed while seated. In these cases intervention effects will be
well assessed by the Physilog or other comparable devices documenting walking,
running, and standing as physical activity categories. Although riding a bicycle or
performing strength training on machines may still not be a very frequent activity
for most old, frail, and multimorbid people in everyday life, results point to the
limitations, as well as the potential, of both technical and interview-based assessment of physical activity.
Limitations
Some limitations of the APAFOP should be noted. Normative values for the metabolic performance of the oldest old and impaired people are lacking, so the scoring
system of the APAFOP is based on an MET- or time-unit-based concept modified
for use in very old people according to an established scoring system suggested
by Ainsworth et al. (2000). Although an approximation of caloric expenditure is
feasible for the APAFOP, we did not make use of this option because we only have
exact caloric values for younger people.
364 Hauer et al.
The short assessment period of the APAFOP, which promotes better recall of
activities, may not fully cover day-to-day variability in physical activity, although
physical activity variance seems to decrease with age and limited functional status.
For a younger population, data from pedometers suggest a minimum of 3 days of
recording to counterbalance day-to-day variability in physical activity (TudorLocke et al., 2005). Another study with people over 60 years old (some were still
employed) using pedometers and devices with one accelerometer suggested 2 days
of recording (Rowe, Kemble, Robinson, & Mahar, 2007). Nonetheless, there is very
limited information on the variability of physical activity patterns in older retired
persons using well-validated movement sensors. When examined on a group level a
low day-to-day variability could be detected in a recent study in older, communitydwelling people including high-functioning participants (Nicolai et al., 2010). To
our knowledge no data on older adults with cognitive impairment and advanced
frailty have been published. The current study used such a sedentary sample. A
24-hr monitoring may therefore prove to be sufficient to document habitual physical
activity because of the low day-to-day variability in frail, impaired, sedentary people.
For comparative validation we used the PAQE questionnaire. To allow an
assessment period comparative to the APAFOP and feasibility of the assessment in
patients with dementia, we reduced the assessment period of the PAQE to 1 week.
The assessment of activities during 1 week is already part of the original PAQE
assessment, complemented for sport and leisure activities by year-round activity.
For the modified PAQE, results of validation are therefore limited.
As with any other subjective reports by questionnaires, assessment of physical activity depends on preserved cognitive performance. Supported by specific
methodological approaches, the APAFOP showed good to excellent measurement
qualities even in the subgroup of people with mild to moderate cognitive impairment. However, assessment by questionnaire may not be feasible in more severely
impaired people.
Because of the rapid aging of populations worldwide, programs promoting
health and related research will increasingly focus on the old and oldest-old. Physical
activity has been identified as a high-impact lifestyle factor to improve health even
in frail, multimorbid, and impaired people. The APAFOP closes a methodological
gap and demonstrates good to excellent biometric qualities to adequately assess
physical activity in frail older people with and without cognitive impairment. It
may be useful in future studies to further examine cognitive and psychological
processes influencing physical activity assessment to further improve the accuracy
of documentation.
Acknowledgments
This research was supported in parts by the Dietmar Hopp Stiftung, the Robert Bosch
Stiftung, the Landesstiftung Baden-Württemberg, and the Landesgraduiertenstiftung BadenWürttemberg. We thank Ruth Hofrichter and Laura Coll-Planas for support in data collection,
Anna Tremmel and Anna Czempik for support in patient management, and Clemens Becker
for support in developing the questionnaire.
Physical Activity Questionnaire in the Elderly 365
References
Ainsworth, B.E., Haskell, W.L., Leon, A.S., Jacobs, D.R., Montoye, H.J., Sallis, J.F., &
Paffenbarger, R.S., Jr. (1993). Compendium of physical activities: Classification of
energy costs of human physical activities. Medicine and Science in Sports and Exercise, 25(1), 71–80.
Ainsworth, B.E., Haskell, W.L., Whitt, M.C., Irwin, M.L., Swartz, A.M., Strath, S.J., . . .
Leon, A.S. (2000). Compendium of physical activities: An update of activity codes
and MET intensities. Medicine and Science in Sports and Exercise, 32(9), S498–S504.
Altman, D.G., & Gardner, M.J. (2000). Regression and correlation. In D.G. Altman, D.
Machin, T.N. Bryant, & M.J. Gardner (Eds.), Statistics with confidence (2nd ed.).
London: BMJ Books.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental
disorders. DSM4 (4th ed.). Washington, DC: Author.
Aminian, K., Robert, Ph., Buchser, E.E., Rutschmann, B., Hayoz, D., & Depairon, M. (1999).
Physical activity monitoring based on accelerometry: Validation and comparison with
video observation. Medical & Biological Engineering & Computing, 37(3), 304–308.
Baranowski, T. (1988). Validity and reliability of self reports measures of physical activity:
An information processing perspective. Research Quarterly for Exercise and Sport,
59(4), 314–327.
Baranowski, T., Sworkin, R.J., & Cieslik, C.J. (1984). Validity of children’s self report of
aerobic activity. Research Quarterly for Exercise and Sport, 55, 309–317.
Blair, S.N., Cheng, Y., & Holder, J.S. (2001). Is physical activity or physical fitness more
important in defining health benefits? Medicine and Science in Sports and Exercise,
33(6, Suppl.) S379–S399.
Blair, S.N., Dowda, M., & Pate, R.R. (1991). Reliability in long-term recall of participation
in physical activity by middle-aged men and women. American Journal of Epidemiology, 133(3), 266–275.
Blair, S.N., Haskell, W.L., Ho, P., Paffenbarger, R.S., Jr., Vranizan, K.M., Farquhar, J.W., &
Wood, P.D. (1985). Assessment of habitual physical activity by a seven-day recall in
a community survey and controlled experiments. American Journal of Epidemiology,
122(5), 794–804.
Blair, S.N., Kohl, H.W., & Barlow, C.E. (1993). Physical activity, physical fitness, and
all-cause mortality in women. Do women need to be active? Journal of the American
College of Nutrition, 12(4), 368–371.
Bland, J.M., & Altman, D.G. (1986). Statistical methods for assessing agreement between
two methods of clinical measurement. Lancet, 1(8476), 307–310.
Bonnefoy, M., Normand, S., Pachiaudi, C., Lacour, J.R., Laville, M., & Kostka, T. (2001).
Simultaneous validation of ten physical activity questionnaires in older med: A doubly
labeled water study. Journal of the American Geriatrics Society, 49(1), 28–35.
Bussmann, J.B.J., van de Laar, Y.M., Neeleman, M.P., & Stam, H.J. (1998). Ambulatory
accelerometry to quantify motor behavior in patients after failed back surgery: A validation study. Pain, 74(2-3), 153–161.
Caspersen, C.J. (1989). Physical activity epidemiology: Concepts, methods, and applications
to exercise science. Exercise and Sport Sciences Reviews, 17, 423–473.
Cohen, J. (1988). Statistical power analysis for the behavioural sciences (2nd ed.). Hillsdale,
NJ: Lawrence Erlbaum.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159.
Delbaere, K., Hauer, K., & Lord, S.R. (2009). Evaluation of the Incidental and Planned
Activity Questionnaire (IPAQ) for older people. British Journal of Sports Medicine.
366 Hauer et al.
Dipietro, L., Caspersen, C.J., Ostfeld, A.M., & Nadel, E.R. (1993). A survey for assessing
physical activity among older adults. Medicine and Science in Sports and Exercise,
25(5), 628–642.
Durante, R., & Ainsworth, B. (1996). The recall of physical activity: Using a cognitive
model of question-answering process. Medicine and Science in Sports and Exercise,
28(10), 1282–1291.
Fitzpatrick, R., Davey, C., Buxton, M.J., & Jones, D.R. (1998). Evaluating patient-based
outcome measures for use in clinical trials. Health Technology Assessment, 2(14), I–IV,
1–74.
Folstein, M.F., Folstein, S.E., & McHugh, P.R. (1975). “Mini-Mental State”: A practical
method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12(3), 189–198.
Gandek, B., Ware, J.E., Jr., Aaronson, N.K., Alonso, J., Apolone, G., Bjorner, J., . . . Sullivan, M. (1998). Tests of data quality, scaling assumptions, and reliability of the SF-36
in eleven countries: Results from the IQOLA Project. International Quality of Life
Assessment. Journal of Clinical Epidemiology, 51(11), 1149–1158.
Grant, P.M., Dall, P.M., Mitchell, S.L., & Granat, M.H. (2008). Activity-monitor accuracy
in measuring step number and cadence in community-dwelling older adults. Journal
of Aging and Physical Activity, 16(2), 201–214.
Guralnik, J.M., Simonsick, E.M., Ferrucci, L., Glynn, R.J., Berkman, L.F., Blazer, D.G., . . .
Wallace, R.B. (1996). A short physical performance battery assessing lower extremity function: Association with self-reported disability and prediction of mortality and
nursing home admission. Journal of Gerontology, 49(2), M85–M94.
Harada, N.D., Chiu, V., King, A.C., & Stewart, A.L. (2001). An evaluation of three selfreport physical activity instruments for older adults. Medicine and Science in Sports
and Exercise, 33(6), 962–970.
Hardman, A.E. (2001). Issues of fractionalisation of exercise (shorts vs long bouts). Medicine
and Science in Sports and Exercise, 33(6, Suppl.) 421–428.
Hauer, K., Yardley, L., Beyer, N., Kempen, G., Dias, N., Campbell, M., . . . Todd, C. (2010).
Validation of the Falls Efficacy Scale (FES) and Falls Efficacy Scale International (FESI) in geriatric patients with and without cognitive impairment: Results of self-report
and interview- based questionnaires. Gerontology, 56(2), 190–199.
Howell, D.C. (2007). Statistical methods for psychology (6th ed., pp. 35–38). Belmont,
CA: Duxbury Press.
Jorstad-Stein, E.C., Hauer, K., Becker, C., Bonnefoy, M., Nakash, R.A., Skelton, D.A., &
Lamb, S.E. (2005). Suitability of physical activity questionnaires for older adults in
fall-prevention trials: A systematic review. Journal of Aging and Physical Activity,
13(4), 461–481.
Katz, J.N., Larson, M.G., Phillips, C.B., Fossel, A.H., & Liang, M.H. (1992). Comparative
measurement sensitivity of short and longer health status instruments. Medical Care,
30(10), 917–925.
Kazis, L.E., Anderson, J.J., & Meenan, R.F. (1989). Effect size for interpreting changes in
health status. Medical Care, 27(3), 178–189.
Levine, J., Melanson, E.L., Westerterp, K.R., & Hill, O.J. (2001). Measurement of the
components of nonexercise activity thermogenesis. American Journal of Physiology.
Endocrinology and Metabolism, 281(4), 670–675.
Levine, J.A., Lanningham-Foster, L.M., McCrady, S.K., Krizan, A.C., Olson, L.R., Kane,
P.H., . . . Clark, M.M. (2005). Inter-individual variation in posture allocation: Possible
role in human obesity. Science, 307(5709), 584–586.
Mahoney, F.J., & Barthel, D.W. (1965). Functional evaluation: The Barthel-Index. Maryland
Medical Journal, 14, 61–65.
Maki, B.E., Holiday, P.J., & Topper, A.K. (1991). Fear of falling and postural performance
in the elderly. Journal of Gerontology, 46(4), M123–M131.
Physical Activity Questionnaire in the Elderly 367
Masse, L.C., Ainsworth, B.E., Torteleo, S., Leon, S., Fulton, J.E., Henderson, K.A., & Mayo,
K. (1998). Measuring physical activity in midlife, older and minority women: Issues
from an expert panel. Journal of Women’s Health, 7(1), 57–67.
McDowell, I., & Newell, C. (1996). Measuring health: A guide to rating scales and questionnaires (2nd ed., pp. 35–38). New York: Oxford University Press.
Najafi, B., Aminian, K., Paraschiv-Ionescu, A., Loew, F., Büla, C., & Robert, P. (2003).
Ambulatory system for human motion analysis using a kinematic sensor: Monitoring
of daily physical activity in the elderly. IEEE Transactions on Bio-Medical Engineering, 50(6), 711–723.
Nicolai, S., Benzinger, P., Skelton, D.A., Aminian, K., Becker, C., & Lindemann, U. (2010).
Day-to-day variability of physical activity of older persons living in the community.
Journal of Aging and Physical Activity, 18(1), 75–86.
Paraschiv-Ionescu, A., Buchser, E.E., Rutschmann, B., Najafi, B., & Aminian, K. (2004).
Ambulatory system for the quantitative and qualitative analysis of gait and posture
in chronic pain patients treated with spinal cord stimulation. Gait & Posture, 20(2),
113–125.
Rowe, D.A., Kemble, C.D., Robinson, T.S., & Mahar, M.T. (2007). Daily walking in older
adults: Day-to-day variability and criterion-referenced validity of total daily step counts.
Journal of Physical Activity and Health, 4(4), 434–446.
Salarian, A., Russmann, H., Vingerhoets, F., Burkhard, P.R., & Aminian, K. (2007). Ambulatory monitoring of physical activities in patients with Parkinson’s disease. IEEE
Transactions on Bio-Medical Engineering, 54(12), 2296–2299.
Sallis, J.F., Haskell, W.L., Wood, P.D., Fortman, S.P., Rogers, T., Blair, S.N., & Paffenbarger,
R.S., Jr. (1985). Physical activity assessment methodology in the five city project.
American Journal of Epidemiology, 121(1), 91–106.
Schuler, P.B., Richardson, M.T., Ochoa, P., & Wang, M.Q. (2001). Accuracy and repeatability of the Yale Physical Activity Survey in assessing physical activity of older adults.
Perceptual and Motor Skills, 93(1), 163–177.
Shephard, R.J. (2003). Limits to the measurement of habitual physical activity by questionnaires. British Journal of Sports Medicine, 37(3), 197–206.
Shrout, P.E., & Fleiss, J.L. (1979). Intraclass correlation: Uses in assessing rater reliability.
Psychological Bulletin, 86, 420–428.
Souchay, C., Moulin, C.J., Isinarini, M., & Conway, M.A. (2008). Rehearsal strategy use in
Alzheimer’s disease. Cognitive Neuropsychology, 25(6), 783–797.
Stel, V.S., Smit, J.H., Pluijm, M.F., Visser, M., Deeg, D.J., & Lips, P. (2004). Comparison of
the LASA Physical Activity Questionnaire with a 7-day diary and pedometer. Journal
of Clinical Epidemiology, 57(3), 252–258.
Stewart, A.L., Mills, K.M., King, A.C., Haskell, W.L., Gillis, D., & Ritte, P.L. (2001).
CHAMPS Physical Activity Questionnaire for Older Adults: Outcomes for interventions. Medicine and Science in Sports and Exercise, 33(7), 1126–1141.
Tudor-Locke, C., Burkett, L., Reis, J.P., Ainsworth, B.E., Macera, C.A., & Wilson, D.K.
(2005). How many days of pedometer monitoring predict weekly physical activity in
adults? Preventive Medicine, 40(3), 293–298.
Tudor-Locke, C.E., & Myers, A.M. (2001). Challenges and opportunities for measuring
physical activity in sedentary adults. Sports Medicine), 31(2), 91–100.
Voorrips, L.E., Ravelli, A.C., Dongelmans, P.C., Deurenberg, P., & van Staveren, W.A. (1991).
A physical activity questionnaire for the elderly. Medicine and Science in Sports and
Exercise, 23(8), 974–979.
Wannamethee, S.G., & Shaper, A.G. (2001). Physical activity in the prevention of cardiovascular disease: An epidemiological perspective. Sports Medicine), 31(2), 101–114.
Washburn, R.A., Heath, G.W., & Jackson, A.W. (2000). Reliability and validity issues concerning large scale surveillance of physical activity. Research Quarterly for Exercise
and Sport, 71(2, Suppl.) 104–113.
368 Hauer et al.
Washburn, R.A., Smith, K.W., Jettre, A.M., & Janney, C.A. (1993). The Physical Activity
Scale for the Elderly (PASE): Development and evaluation. Journal of Clinical Epidemiology, 46(2), 153–162.
Westerterp, K.R. (1999). Physical activity assessment with accelerometers. International
Journal of Obesity, 23(3, Suppl.) 45–49.
Appendix A: Questionnaire Items and Intensity
Rating of the APAFOP
Specific questionnaire items have been developed and an intensity rating has been
established for the APAFOP. Appendix Table 1 gives additional information and
examples for the questionnaire items, and Appendix Table 2 on the intensity rating
of the APAFOP compared with an established questionnaire in which the rating
has been based.
Intensities of activities are rated in the APAFOP according to a MET-based
scoring system. MET values of different intensity levels for examples of frequent
activities are given in Appendix Table 2. MET values are derived from an established coding system by Ainsworth et al. (1993, 2000) that has been modified for
use in older populations (CHAMPS; Stewart et al., 2001). We adjusted those MET
values for the older and frailer target population of the APAFOP. The original MET
levels suggested by Ainsworth et al. and Stewart et al. are integrated in Appendix
Table 2 for comparison.
Age-adjusted MET values are given for the APAFOP and for original values
given by the CHAMPS questionnaire (Stewart et al., 2001) for comparison.
For categories such as lying or indoor activity, which were not presented in the
CHAMPS, MET values derived from the original compendium by Ainsworth et
al. (1993, 2000) were documented.
Appendix Table 1 Types of Activity as Examples for Questionnaire Items
Item
1. Walking
(longer than
3 min)
2. Outdoor
activity
3. Indoor
activity
4. Sitting
5. Lying
6. Sports
activity
Related activity, clues for interrogation
Walking, hiking, walking the dog, shopping,
going to doctor, walking to run errands, going
to the cemetery, going to social events, walking
around in the garden; Walking in the house
in case of bad weather or restriction to home,
walking passages or corridors in institutions
(seniors’ homes).
Gardening, construction, or cleaning activity
around the house (e.g., sweep the yard), time
spent in shops with mixed activity.
Mixed household activities (walking or standing),
all activities while standing or walking, such
as talking to a neighbor, standing at a window,
going to the toilet, fetching something from the
basement, doing the laundry, personal hygiene,
care for dependent proxy.
Sitting at home (while watching TV, while
preparing meals, reading newspaper, knitting);
sitting in a restaurant or theater, during
transportation (car, bus), or outdoors (garden).
Lying in bed at night, taking a nap during the
day, lying on a sofa while watching TV or
reading a book.
Home training (stretching exercise, home trainer,
unsupervised fitness exercise). Exercise/sport
clubs, training sessions in seniors’ homes.
Inclusion of specific activities
Short walking activities mixed
with other activities (indoors,
e.g., walking to the toilet,
walking around the kitchen)
relates to Item 3; mixed outdoor
activity relates to Item 2.
Walking to shops relates to
Item 1.
Short walking periods relate
to this item.
Difference between sitting and
lying depends on the postural
position or definition.
369
370
aFrom
Nonstrenuous activity: around the home (e.g., watering the garden, strolling in
garden) or away from home (mixed activity while shopping).
Moderate activity: Cleaning windows, sweeping the yard or street, shopping while
carrying a basket, increased percentage of walking during mixed activities.
Strenuous activities: Chopping wood, lawn mowing, more strenuous gardening
activities.
Nonstrenuous activities: household activities: (washing dishes, cooking, watering flowers, personal hygiene, making a telephone call while standing, talking to a
neighbor, during mixed activities focus on standing.
Moderate activities: doing the laundry, storing shopping goods, vacuuming, mixed
activities focus on walking.
Strenuous activities: cleaning floor, climbing stairs, carrying heavy loads, intensive
care or support for dependent proxies.
Activities while sitting: reading, watching TV, making a telephone call, having a
conversation, sitting during transport, etc.
Activities while lying: reading, watching TV, sleeping.
Nonstrenuous activity: stretching exercise while seated or lying, nonstrenuous
mixed activity.
Moderate activity: moderate endurance home training, moderate exercise while
standing or walking, mixed activities.
Strenuous activity: progressive resistance training; challenging progressive functional training; fall-prevention exercise; more strenuous endurance exercise such as
bicycling, swimming, walking/jogging; other challenging sports exercise.
Walking uphill or briskly in nonfrail people.
Item-related activities according to different intensity levels
Habitual walking (doing errands, walking leisurely).
Brisk walking (walking as an exercise).
the original compendium by Ainsworth et al. (1993, 2000).
5. Lying
6. Sports activity
4. Sitting
3. Indoor activity
2. Outdoor activity
Item/ Activity
1. Walking
1
1
2
3
4
2.5
3
4–5
3
3.5–7.5a
1–1.8a
0.9–1.0a
2
2.0–3.0a
4
4.0
1.5
3
3.0
1.8–2.5a
2
4
METs
modified
for APAFOP
2
3
2.25
6
Original METs according to
Ainsworth et al. (1993, 2000)
& Stewart et al. (2001)
2.5
3.5
Appendix Table 2 Intensity Scoring on the Assessment of Physical Activity in Frail Older People (APAFOP) for Different
Types of Activity
Appendix Figure A — Assessment form: Activities of daily living daily protocol. Note.
MET = metabolic equivalent.
371
372 Hauer et al.
Appendix B: APAFOP Assessment Form
The documentation form allows an assessment of type of activity and intensity
during the interrogation period of 24 hr for subscores and total scores.
Appendix Figure B — Assessment form: Sport activity and activity type. Note. MET =
metabolic equivalent.