1
This is a peer-reviewed, author accepted manuscript of the following research article: Nashandi, H. L., Reilly, J.
J., & Janssen, X. (2020). Public health surveillance of physical activity in adolescents and adults in Namibia: a
cross-sectional validation of activity questionnaires against accelerometry. Journal of Public
Health. https://doi.org/10.1093/pubmed/fdaa067
Public Health Surveillance of Habitual Physical Activity in Adolescents and Adults in Namibia: A
Cross-sectional Validation of Activity Questionnaires against Accelerometry
Authors: Hilde L Nashandi
1*
, John J Reilly
2
, Xanne Janssen
2
1 University of Namibia, Faculty of Health Sciences, School of Nursing, Windhoek, Namibia
2 University of Strathclyde Glasgow, Physical Activity for Health Group, Glasgow, Scotland G1 1XQ
*
Address for correspondence
H.L. Nashandi, MPH, School of Nursing, Faculty of Health Sciences, University of Namibia, Private Bag
13301, 340 Mandume Ndemufayo Ave, Pionierspark, Windhoek, NAMIBIA. Email:
hildeliisa@yahoo.com
Keywords: Physical activity; Global Physical Activity Questionnaire; validity; accelerometer; obesity;
exercise; Africa
2
This is a peer-reviewed, author accepted manuscript of the following research article: Nashandi, H. L., Reilly, J.
J., & Janssen, X. (2020). Public health surveillance of physical activity in adolescents and adults in Namibia: a
cross-sectional validation of activity questionnaires against accelerometry. Journal of Public
Health. https://doi.org/10.1093/pubmed/fdaa067
ABSTRACT
Background: Monitoring population-level physical activity is crucial for examining adherence to global
guidelines and addressing obesity. This study validated self-reported moderate-to-vigorous physical
activity (MVPA) against an accurate device-based method in Namibia.
Methods: Adolescent girls (n 52, mean age 16.2y [SD 1.6]) and adult women (n 51, mean age 31.3y
[SD 4.7]) completed the PACE+/GPAQ self-report questionnaires, and were asked to wear an Actigraph
accelerometer for seven days. Validity of self-reported MVPA was assessed using rank order
correlations between self-report and accelerometry, and classification ability of the questionnaires
with Mann-Whitney tests, kappa’s, sensitivity and specificity.
Results: In the adolescents, Spearman’s rank coefficients between self-reported MVPA (days/week)
and accelerometry measured MVPA were positive but not significant (r=0.240, p=0.104). In the adults,
self-reported MVPA (minutes/day) was moderately and significantly correlated with accelerometer-
measured MVPA (r=0.396, p=0.008). In both groups there was fair agreement between accelerometry
and questionnaire-defined tertiles of MVPA (adolescents κ=0.267; p=0.010; adults κ=0.284; p=0.008),
and measured MVPA was significantly higher in the individuals self-reporting higher MVPA than those
reporting lower MVPA.
Conclusions: The PACE+ and GPAQ questionnaires have a degree of validity in adolescent girls and
adult females in Namibia, though more suitable for population than individual level measurement.
Keywords: Physical activity, Validity, Adult women, Adolescent girls
3
This is a peer-reviewed, author accepted manuscript of the following research article: Nashandi, H. L., Reilly, J.
J., & Janssen, X. (2020). Public health surveillance of physical activity in adolescents and adults in Namibia: a
cross-sectional validation of activity questionnaires against accelerometry. Journal of Public
Health. https://doi.org/10.1093/pubmed/fdaa067
4
This is a peer-reviewed, author accepted manuscript of the following research article: Nashandi, H. L., Reilly, J.
J., & Janssen, X. (2020). Public health surveillance of physical activity in adolescents and adults in Namibia: a
cross-sectional validation of activity questionnaires against accelerometry. Journal of Public
Health. https://doi.org/10.1093/pubmed/fdaa067
INTRODUCTION
There has been a “physical activity transition”¹ʾ² across Africa: levels of moderate-to-vigorous physical
activity (MVPA) are now very low ³ʾ²ʾ⁴ particularly among adolescent girls and adult women, who are
also at highest risk of obesity across Sub-Saharan Africa³ʾ⁵. In Namibia, childhood obesity prevalence
is high and probably increasing⁶. The 2013 Global School-based Student Health Survey (GSHS)
reported that in Namibia, 10.4% of females aged 13-17 were overweight or obese, and the prevalence
of insufficient physical activity among adolescents exceeded 80% ⁴. In Namibia, only 5% of women
and 12% of men age 15-49 are physically active at work, while only 16% of women and 32% percent
of men engage in non-work-related physical activity⁷.
Surveillance of habitual MVPA levels in adolescents and adults in most countries still depends on
subjective, self-report, measures. For example, the PACE+
8
is used widely among adolescents as part
of the Global Student Health Survey (https://www.cdc.gov/GSHS/), and the General Physical Activity
Questionnaire (GPAQ)
9
is used widely as part of the WHO ‘STEPS’ global NCD surveillance
(https://www.who.int/ncds/surveillance/steps/en/). However, the validity of MVPA measured by self-
report is in question-in children and adolescents few self-report measures have been subject to
validations against more accurate measures of MVPA such as accelerometry
10
. In the adolescent
population, there is a dearth of validation study evidence from low-middle income countries
11
and
some validations from high-income countries have been discouraging
12
. Out of 38 validation studies
included in a recent review, only two showed acceptable correlations (r>0.40) compared to
accelerometry
13
. In adults too, validation of many self-report questionnaires used in public health
surveillance has been rare in low-middle income countries, and in high-income countries results have
also been discouraging. For example, prevalence of meeting physical activity guidelines may be
overestimated by self-report in adults in the UK and USA
14,15
. A recent review of physical activity
measures showed the mean difference between questionnaires using a 7-day recall (like those
included in this study) and the gold standard (doubly labelled water) ranged from an underestimation
of 60% to an overestimation of 60%
16
.
Increasing habitual physical activity among children, adolescents, and adults is central to prevention
of obesity and many non-communicable diseases (NCDs)
17,18,19
. If global targets are to be met, or even
monitored, unbiased methods of public health surveillance of physical activity are required. The
5
This is a peer-reviewed, author accepted manuscript of the following research article: Nashandi, H. L., Reilly, J.
J., & Janssen, X. (2020). Public health surveillance of physical activity in adolescents and adults in Namibia: a
cross-sectional validation of activity questionnaires against accelerometry. Journal of Public
Health. https://doi.org/10.1093/pubmed/fdaa067
primary aim of the present study was therefore to validate the self-report of habitual MVPA among
adolescent girls (using the PACE+ questionnaire) and among adult women (using the GPAQ) against
an accurate device-based method (accelerometry) in Namibia.
METHODS
Study sample, recruitment and ethics
The project was approved by the Ministry of Health and Social Services Research Management
Committee. All participants provided informed written consent. The present validation study was part
of a larger study focusing on lifestyle and obesity risk among adolescent girls and adult women in
Namibia. Participants of this larger study were drawn from urban public schools in Khomas region
using stratified random sampling of schools. For the larger study a total of 207 adolescent girls’
participants had their parents give consent and 212 women provided consent and participated in the
study. A subsample of 52 adolescents and 51 adult women were randomly selected (using a random
number selector) to take part in the current study. All measures were taken at school during school
time. In addition, questionnaires regarding physical activity were completed via interviews at schools.
Body fat percentage was measured accurately via deuterium oxide dilution method, as described in
detail elsewhere
6,20
though this variable was used only for descriptive purposes in the present study.
Measurement of habitual MVPA by accelerometry
Participating (n 52) adolescents and adults (n 51) were given an Actigraph GT3X+ accelerometer
(ActiGraph, Pensacola, FL) to wear on the right hip for 7 full days, taking the device off only when
sleeping or during water based activities, as described previously
21,22
. The Actigraph monitor is an
acceptable criterion measure for evaluating questionnaire validity
9
. For feasibility reasons and to
reduce participant burden, accelerometers were distributed to participants on the same day as
questionnaire completion. Data was collected in 30Hz and transformed into 15-second epochs. To be
included in the analysis participants required at least 3 valid days (i.e., at least 10 hours of wear),
including at least one weekend day, in order to obtain an estimate of habitual MVPA
23,24,25
. Data from
the devices were downloaded and cleaned using ActiLife software version 6.12.1 (ActiGraph,
Pensacola, FL). Non-wear time was removed using a 60 minutes of consecutive zeros criteria
21
. The
ActiLife data scoring programme was used to determine daily minutes spent in sedentary time, light
PA intensity (LPA), moderate PA intensity (MPA), vigorous PA intensity (VPA), and MVPA using the
Evenson activity cut point for adolescents which is most accurate to use in this age group
26
and the