Research Quarterly for Exercise and Sport

ISSN: 0270-1367 (Print) 2168-3824 (Online) Journal homepage: https://shapeamerica.tandfonline.com/loi/

urqe20

An Accurate VO2max Nonexercise Regression

Model for 18–65-Year-Old Adults

Danielle I. Bradshaw , James D. George , Annette Hyde , Michael J. LaMonte ,

Pat R. Vehrs , Ronald L. Hager & Frank G. Yanowitz

To cite this article: Danielle I. Bradshaw , James D. George , Annette Hyde , Michael J. LaMonte ,
Pat R. Vehrs , Ronald L. Hager & Frank G. Yanowitz (2005) An Accurate VO2max Nonexercise
Regression Model for 18–65-Year-Old Adults, Research Quarterly for Exercise and Sport, 76:4,
426-432, DOI: 10.1080/02701367.2005.10599315

To link to this article: https://doi.org/10.1080/02701367.2005.10599315

Published online: 23 Jan 2013.

Submit your article to this journal

Article views: 1316

Citing articles: 42 View citing articles

Full Terms & Conditions of access and use can be found at

https://shapeamerica.tandfonline.com/action/journalInformation?journalCode=urqe20
 Physiology
Bradshaw, George, Hyde, LaMonte, Vehrs, Hager, and Yanowitz

Research Quarterly for Exercise and Sport

©2005 by the American Alliance for Health,
Physical Education, Recreation and Dance
Vol. 76, No. 4, pp. 426–432

An Accurate VO2max Nonexercise Regression Model

for 18–65-Year-Old Adults
Danielle I. Bradshaw, James D. George, Annette Hyde, Michael J. LaMonte, Pat R. Vehrs,
Ronald L. Hager, and Frank G. Yanowitz

The purpose of this study was to develop a regression equation to predict maximal oxygen uptake (VO2max) based on nonexercise
(N-EX) data. All participants (N = 100), ages 18–65 years, successfully completed a maximal graded exercise test (GXT) to
assess VO2max (M = 39.96 mL·kg-¹·min-¹, SD = 9.54). The N-EX data collected just before the maximal GXT included the
participant’s age; gender; body mass index (BMI); perceived functional ability (PFA) to walk, jog, or run given distances; and
current physical activity (PA-R) level. Multiple linear regression generated the following N-EX prediction equation (R = .93,
SEE = 3.45 mL·kg-¹·min-¹, % SEE = 8.62): VO2max (mL·kg-¹·min-¹) = 48.0730 + (6.1779 x gender; women = 0, men = 1) –
(0.2463 x age) – (0.6186 x BMI) + (0.7115 x PFA) + (0.6709 x PA-R). Cross validation using PRESS (predicted residual
sum of squares) statistics revealed minimal shrinkage (Rp = .91 and SEEp = 3.63 mL·kg-¹·min-¹); thus, this model should yield
acceptable accuracy when applied to an independent sample of adults (ages 18–-65 years) with a similar cardiorespiratory
fitness level. Based on standardized β-weights, the PFA variable (0.41) was the most effective at predicting VO2max followed by
age (-0.34), gender (0.33), BMI (-0.27), and PA-R (0.16). This study provides a N-EX regression model that yields relatively
accurate results and is a convenient way to predict VO2max in adult men and women.

Key words: cardiorespiratory fitness, exercise testing, maxi- component of health and physical fitness (ACSM). The
mum oxygen uptake assessment of CRF is valuable when educating individu-
als about their overall fitness status, developing exercise

C ardiorespiratory fitness (CRF) is the ability to per-

form dynamic, moderate- to high-intensity exercise
using the large muscle groups for long periods of time
(American College of Sports Medicine [ACSM], 2000).
CRF depends on the respiratory, cardiovascular, and
skeletal muscle systems and, therefore, is an important
Submitted: January 14, 2004
Accepted: December 4, 2004
At the time of this study, Danielle I. Bradshaw and Annette Hyde
were with the Department of Exercise Sciences at Brigham
Young University. James D. George, Pat R. Vehrs, and Ronald L.
Hager, are with the Department of Exercise Sciences at Brigham
Young University. Michael J. LaMonte is with the Epidemiology
Division at The Cooper Institute. Frank G. Yanowitz is with the
School of Medicine at the University of Utah and The Fitness
Institute at LDS Hospital, Salt Lake City, UT.
programs, and stratifying cardiovascular risk (ACSM).
The standard test for determining CRF is the mea-
surement of maximal oxygen uptake (VO2max) when
performing a maximal graded exercise test (GXT; ACSM,
2000). VO2max is the most accurate way to assess CRF,
however, the test requires costly equipment, space to house
the equipment, and trained personnel to administer the
test. In addition, maximal GXTs are unappealing to
some individuals, because the test requires strenuous
exercise to the point of volitional exhaustion. Because
of this, some older or higher risk individuals should not
perform the test without medical supervision (ACSM).
Due to the possible drawbacks of maximal GXTs and
the direct measurement of VO2max, submaximal exercise
tests are available and provide an acceptable prediction
of CRF and VO2max. Submaximal exercise tests use pre-
diction variables, such as age, gender, body mass, exercise
pace, and exercise heart rate to predict VO2max (ACSM,

426 RQES: December 2005

Bradshaw.pmd 426 11/9/2006, 5:49 PM


2000). Although not as accurate as maximal GXTs,
submaximal exercise tests are easier to perform, require
less time and effort to complete, and can be administered
at lower costs and reduced risk (ACSM). These tests use a
variety of exercise modes, including cycle ergometry, step-
ping and walking, and jogging or running on a treadmill
or track (ACSM, 2000; Bruce, Kusumi, & Hosmer, 1973;
Cooper, 1968; Kline, Porcari, Hintermeister, Freedson, &
Ward, 1987; Astrand & Ryhming, 1954).
Other methods that do not require exercise are also
available to predict VO2max (ACSM, 2000). Nonexercise
(N-EX) regression equations provide a convenient esti-
mate of CRF without needing to perform a maximal or
submaximal exercise test. This approach is inexpensive,
time efficient, and realistic for large groups. To date, N-
EX predictor variables include age, gender, body mass
index (BMI), percentage of body fat, physical activity rat-
ing (PA-R; Heil, Freedson, Ahlquist, Price, & Rippe, 1995;
Jackson et al., 1990), and perceived functional ability (PFA;
George, Stone, & Burkett, 1997). The PFA includes
simple questions that ask individuals to rate their ability
to exercise at a comfortable pace for 1 and 3 miles (1.6
and 4.8 km). Studies show N-EX equations are relatively
accurate and a quick and easy way to predict VO2max
(George et al., 1997; Heil et al., 1995; Jackson et al., 1990).
To date, PA-R has been validated in a large sample of 18–
70-year-old men and women, but PFA has only been vali-
dated in a sample of college-aged men and women. Thus,
there is a need to further evaluate PFA as a predictor vari-
able in estimating CRF with a broader age range.
Method
Participants
One hundred individuals (50 women and 50 men),
ages 18–65 years, participated in this study. Participants
were recruited from the Y-Be-Fit Wellness Program at
Brigham Young University (BYU) and employees from
the LDS Hospital in Salt Lake City, UT. Participants were
classified as either low or moderate risk according to
ACSM guidelines for exercise testing. Moderate risk
individuals were tested at the LDS Hospital under phy-
sician supervision to comply with ACSM recommenda-
tions (ACSM, 2000). All participants completed a
Physical Activity Readiness Questionnaire and signed
an informed consent document that was approved by the
BYU Institutional Review Board for Human Subjects.
Procedures
A stadiometer and balance beam scale were used to
measure and record participants’ height and body mass,
RQES: December 2005
Bradshaw.pmd 427
Bradshaw, George, Hyde, LaMonte, Vehrs, Hager, and Yanowitz
respectively, while they wore lightweight clothing and no
shoes. All participants completed a maximal treadmill
GXT, and, prior to testing, all received instructions about
the maximal GXT involving the protocol, electronic heart
rate monitor, head gear, mouth piece, nose clip, and rat-
ing of perceived exertion (RPE) scale (Borg, 1982).
Maximal Treadmill Graded Exercise Test
Participants performed a maximal GXT using a
modified version of the Arizona State University (ASU)
maximal protocol (George, 1996), with only a slight
warm-up change that required participants to complete
up to three 4-min stages and reach 75% of maximum
heart rate (HRmax). In the first stage, participants
walked at a self-selected brisk pace, 3–4 mph (4.8–6.4
km/h). If they progressed to the second stage, they in-
creased their speed to a self-selected jogging pace be-
tween 4 and 6 mph(4.8 and 6.4 km/h). If participants
did not achieve 75% of HRmax by the end of the sec-
ond stage, they continued to the third stage where the
speed increased to 6 mph (9.6 km/h) or greater. Fol-
lowing the warm-up, participants rested for approxi-
mately 5 min, while the test administrator explained the
test procedures for the maximal GXT and fit partici-
pants with a mouthpiece and nose clip. The ending
speed of the warm-up served as the necessary pace for
the maximal GXT. Thereafter, the grade was increased
1.5% each minute, while pace remained the same, and
continued until participants reached volitional fatigue
and were no longer able to continue.
During the maximal GXT, metabolic gases were
collected using the TrueMax 2400 metabolic measure-
ment system (Consentius Technologies, Sandy, UT). An
on-line computer system computed, averaged, and
printed the VO2 and respiratory exchange ratio (RER)
every 15 s. The participants’ exercise heart rate (HR)
and RPE score were recorded at the end of each stage.
VO2max was considered valid when participants met at
least two of the following three criteria (ACSM, 2000;
Howley, Bassett, & Welch, 1995):
1. Maximal heart rate within 15 beats of age predicted
maximal heart rate.
2. Respiratory exchange ratio equal to or greater than
1.10.
3. No increase in VO2 despite an increase in work load.
Participants who failed to meet at least two of these cri-
teria were dropped from the study.
Nonexercise Questionnaire
Prior to exercise testing, participants completed the
PFA (George et al., 1997) and a modified PA-R nonexer-
427
11/9/2006, 5:49 PM
 Bradshaw, George, Hyde, LaMonte, Vehrs, Hager, and Yanowitz

cise questionnaire (George et al., 1997; Jackson et al., Results

1990). The PFA includes two questions that ascertain how
fast participants feel they can cover a 1- and 3-mile (1.6
and 4.8 km) distance at a comfortable pace. The
participant’s sum total of both 13-point questions is
counted as the PFA score (range: 2–26). The original
PA-R questionnaire has individuals rate their activity
level over the previous month using a seven-point scale.
However, Kolkhorst and Dolgener (1994) noted that an
extended time reference might represent participants’
overall average physical activity level more accurately.
Thus, the PA-R questionnaire was modified with a
longer 6-month time reference and an expanded 10-
point scale.
Statistics
Standard multiple linear regression was used to
create a VO2max regression model using age, gender,
BMI, PFA, and PA-R as predictor variables. The validity
of the VO2max equation was evaluated based on the
correlation coefficient and the standard error of esti-
mate (SEE). Predicted residual sum of squares (PRESS)
statistics (Holiday, Ballard, & McKeown, 1995) were also
computed to estimate the degree of shrinkage one could
expect when the VO2max prediction equation is used
across similar, but independent, samples. Our data were
also input into the Jackson et al. (1990) N-EX equation
to cross validate its accuracy in predicting VO2max. The
statistical significance level was set at p < .05.
One hundred nine participants (ages 18–65 years)
volunteered for this study; however, 9 were dropped from
the analysis, because they did not satisfy the VO2max crite-
ria. Descriptive statistics for the total sample and specific
age groups are presented in Table 1. All participants
achieved a valid VO2max during the maximum GXT
with the average (± SD) VO2max equal to 40.0 ± 9.5
mL·kg-¹·min-¹ (range = 20.2–61.1 mL·kg-¹·min-¹). Corre-
sponding maximum RERmax values (1.19 ± 0.05),
HRmax (184.4 ± 12.3 beats·min-¹), and maximum RPE
scores (19.2 ± 0.8) were indicative of maximum effort.
Participants’ body mass, height, and BMI ranged from
48.08 to 115.94 kg, 1.47 to 1.92 m, and 18.79 to 38.93
kg·m-2, respectively. Participants’ PFA and PA-R scores
ranged from 2 to 26 and 1 to 10, respectively. Descrip-
tive statistics for the male and female samples are also
presented in Table 2.
Multiple linear regression generated the following
N-EX prediction equation: (R = .93; SEE = 3.45 mL·kg-¹
·min-¹): VO2max (mL·kg-¹·min-¹) = 48.073 + (6.178 x gen-
der; women = 0, men = 1) – (0.246 x age) – (0.619 x BMI)
+ (0.712 x PFA) + (0.671 x PA-R). The predictor vari-
ables in this model explain 87% (R2 = .87) of the VO2max
variance. Each independent variable was statistically
significant (p < .05) in predicting VO2max. Standardized
β-weights show the PFA variable (0.41) to be the most ef-
fective at predicting VO2max and the PA-R variable to be
the least effective (0.16; see Table 3). The cross-validation

Table 1. Descriptive statistics for exercise and nonexercise data

Variable Total Ages 18–29 years Ages 30–39 years Ages 40–49 years Ages 50–65 years
(N = 100) (n = 30) (n = 23) (n = 26) (n = 21)
M SD M SD M SD M SD M SD

Age (years) 37.5 13.2 22.1 1.9 32.2 2.9 45.2 3.2 55.7 4.9
Women 50 15 13 13 9
Men 50 15 10 13 12
Height (m) 1.71 0.09 1.74 0.10 1.72 0.10 1.70 0.09 1.70 0.06
Body mass (kg) 74.7 15.2 72.4 14.5 74.9 13.9 75.2 15.8 77.2 17.3
Body mass index (kg·m-2) 25.22 4.17 23.85 3.41 25.15 3.37 25.86 4.68 26.43 4.92
Maximal treadmill test
VO2max (L·min-¹) 2.96 0.82 3.25 0.73 3.28 0.85 2.78 0.83 2.41 0.59
VO2max (mL·kg-¹·min-¹) 40.0 9.5 45.2 7.6 43.9 8.9 37.3 9.0 31.5 5.9
HRmax (beats·min-¹) 184.4 12.3 194.7 5.4 189.3 8.2 178.7 9.2 171.4 10.8
RPEmax (15-point scale) 19.2 0.8 19.5 0.7 19.3 0.6 18.9 0.7 19.9 0.9
RERmax (VCO2 ÷ VO2) 1.19 0.05 1.19 0.05 1.16 0.04 1.20 0.05 1.18 0.04
Nonexercise questions
PFA (2–26 point scale) 14.5 5.5 15.9 4.9 17.1 5.1 14.2 5.1 10.1 4.8
PA-R (10-point scale) 4.9 2.3 4.6 1.7 5.6 2.6 5.5 2.4 4.0 2.2

Note. M = mean; SD = standard deviation; VO2max = maximal oxygen uptake; HRmax = maximal heart rate; RPEmax = maximal rating of
perceived exertion; RERmax = maximal respiratory exchange ratio; PFA = perceived functional ability; PA-R = physical activity rating.

428 RQES: December 2005

Bradshaw.pmd 428 11/9/2006, 5:49 PM

 Bradshaw, George, Hyde, LaMonte, Vehrs, Hager, and Yanowitz

PRESS statistics (Rp = .91 and SEEp = 3.63 mL·kg-¹·min-¹)
show minimal shrinkage in predictive accuracy. Figure
1 provides a scatter plot of predicted versus observed
VO2max values for the N-EX regression model. Based
on a gender-specific analysis of the N-EX prediction
equation, the male sample yielded a slightly more accu-
rate VO2max prediction (R = .94, SEE = 2.89 mL· kg-¹·min-¹,
n = 50) than the female sample (R = .90, SEE = 3.92 mL·
kg-¹·min-¹, n = 50).
According to PFA scores, 25 participants were in the
walking category (PFA = 7.1 ± 2.2, measured VO2max =
29.51 ± 5.21 mL·kg-¹·min-¹), 53 were in the jogging cat-
egory (PFA = 14.9 ± 2.3, measured VO2max = 40.42 ± 6.25
mL·kg-¹·min-¹), and 22 were in the running category
(PFA = 21.7 ± 2.1, measured VO2max = 50.71 ± 5.77
mL·kg-¹·min-¹), with each group significantly different
(p < .05; see Figure 2).
Discussion
The N-EX regression model developed in this study
predicted VO2max well (R = .93; SEE = 3.45 mL·kg-¹·min-¹)
in this sample of men and women, ages 18–65 years. Other
age-generalized N-EX models created by Jackson et al.
Table 2. Descriptive statistics for the total, female, and male
participants
Variable Total Women Men
(N = 100) (n = 50) (n = 50)
M SD M SD M SD
(1990; R = .78, SEE = 5.70 mL·kg-¹·min-¹) and Heil et al.
(1995; R = .88, SEE = 4.90 mL·kg-¹·min-¹) are also relatively
accurate but less precise. The results of this study are simi-
lar to those reported by George et al. (1997; R = .85 and
SEE = 3.44 mL·kg-¹·min-¹) who was the first to include the
PFA variable. Surprisingly, the best maximal exercise (EX)
models, such as the 1.5-mile (2.4 km) run developed by
Cooper (1968; R = .90), the Bruce treadmill test (1973; R
=.93, SEE = 3.13 mL·kg-¹·min-¹), and the ASU treadmill test
(George, 1996; R = .95, SEE = 2.13 mL·kg-¹·min-¹) are only
slightly more accurate than the present N-EX model. Sev-
eral submaximal EX models are less accurate, such as the
Rockport Fitness Walking Test (Liu, Plowman, & Looney,
1992; R = .88 and SEE = 5.00 mL·kg-¹·min-¹), the 20-m
shuttle test (R = .72 and SEE = 5.26 mL·kg-¹·min-¹), and the
Ebbeling, Ward, Puleo, Widrick, and Rippe (1991) tread-
mill walking test (R = .91, SEE = 4.84 mL·kg-¹·min-¹).
The β-weight analyses show that the PFA variable
explains most of the variance in predicting VO2max.
Dropping PFA out of the model lowers the multiple
correlation coefficient (R) from .93 to .89 and increases
the SEE from 3.45 to 4.20 mL·kg-¹·min-¹. A possible rea-
son for the effectiveness of the PFA variable is that it
mimics the role of an EX variable and helps to classify
participants as walkers, joggers, or runners based on their
response to how fast they feel they can cover a 1- (1.6
km) and 3-mile (4.8 km) distance at a comfortable pace.
In a similar way, EX models, such as the 1.5-mile (2.4
km) run, separate participants into categories based on
how fast they complete the distance (Cooper, 1968).
As a cross-validation comparison, we entered our
data (age, gender, BMI, and PA-R scores) into the Jack-
son et al. (1990) N-EX regression equation. The esti-
mated VO2max results were relatively accurate (r = .89,

Age (years) 37.5 13.2 37.2 13.8 37.8 12.8

Height (m) 1.72 0.09 1.65 0.07 1.78 0.06 Table 3. Nonexercise maximal oxygen uptake regression
Body mass (kg) 74.71 15.19 66.40 12.39 83.02 13.10 equation (N = 100)
BMI (kg·m-2) 25.22 4.17 24.27 3.85 26.16 4.29
PFA 14.5 5.5 12.9 5.3 16.1 5.3 Variable β β-weight
PA-R 4.9 2.3 4.9 2.3 5.0 2.3
Maximal treadmill values Intercept 48.073
VO2max (L·min-¹) 2.96 0.82 2.35 0.49 3.56 0.63 Gender (0 = women, 1 = men) 6.178 0.326
VO2max Age -0.246 -0.341
(mL·kg-¹·min-¹) 39.96 9.54 36.32 8.91 43.59 8.80 Body mass index (kg·m-2) -0.619 -0.27
HRmax Perceived functional ability 0.712 0.412
(beats·min-¹) 184.4 12.3 182.0 13.7 186.8 10.4 Physical activity rating 0.671 0.162
RPE R 0.93
(15-point scale) 19.2 0.8 19.3 0.7 19.1 0.8 SEE (mL· kg-¹·min-¹) 3.44
RER (VCO2 ÷ VO2) 1.19 0.05 1.18 0.05 1.19 0.05 R PRESS 0.91
SEEPRESS (mL·kg-¹·min-¹) 3.63
Note. M = mean; SD = standard deviation; PFA = perceived % SEE (% of VO2max) 8.62
functional ability; VO2max = maximal oxygen uptake; HRmax =
maximal heart rate; BMI = body mass index; PA-R = physical Note. VO2max = maximal oxygen uptake; β-weight = standard
activity rating; RPE = rating of perceived exertion; RER = multiple regression coefficients; RPRESS = (1-(PRESS/SStotal))-1/2;
respiratory exchange ratio. SEEPRESS = (PRESS/n)-1/2.

RQES: December 2005 429

Bradshaw.pmd 429 11/9/2006, 5:49 PM

 Bradshaw, George, Hyde, LaMonte, Vehrs, Hager, and Yanowitz

SEE = 4.42 mL·kg-¹·min-¹) and similar to the current N-EX mL·kg-¹·min-¹, measured VO2max = 39.95 mL·kg-¹·min-¹;
model, with the PFA variable not included (R = .89, SEE = p < .0001). These small differences may result from
4.20 mL·kg-¹·min-¹). The Jackson model slightly underes- sample differences or differences in response to the
timated measured VO2max (predicted VO2max = 38.06 original versus modified PA-R question.

70

R = .93
60 SEE = 3.45 mL·kg -1 ·min -1

50
Predicted VO2 ma x (mL·kg -1 ·min -1 )

40

30

20

female
10 male

0
0 10 20 30 40 50 60 70
Measured VO2 max (mL·kg -1 ·min -1 )

Figure 1. Predicted maximal oxygen uptake (VO2max) versus measured VO2max scatter plot.

6
0
Measured VO2max (mL?kg-¹?min-¹)

5
0

4
0

3
0

2
0

10

0
Walkers (n = 25) Joggers (n = 53) Runners (n = 22)
PFA = 7.1 ± 2.2 PFA = 14.9 ± 2.3 PFA = 21.7 ± 2.1

Figure 2. Measured maximal oxygen uptake (VO2max) versus perceived functional ability (PFA; mean ± standard deviation). Signifi-
cantly different (p < .05) between groups.

430 RQES: December 2005

Bradshaw.pmd 430 11/9/2006, 5:49 PM


Dustman-Allen et al. (2003) reported contrasting
results in a cross validation of the N-EX regression equa-
tion developed by George et al. (1997). Dustman-Allen
et al. (2003) assessed 66 participants who were similar
in age (18–29 years) to the validation sample. The study
reported r and SEE values of .53 and 5.3 mL·kg-¹·min-¹,
respectively, for women and .59 and 4.4 mL·kg-¹·min-¹,
respectively, for men. The measured and predicted
VO2max values were similar for women, (measured
VO 2max = 40.8 ± 6.2 mL·kg -¹·min -¹ and predicted
VO2max = 39.9 ± 4.8 mL·kg-¹·min-¹) but dissimilar for
men (measured VO2max = 50.4 ± 5.4 mL·kg-¹·min-¹ and
predicted VO2max = 47.2 ± 5.0 mL·kg-¹·min-¹). It is diffi-
cult to explain why the results of the Dustman-Allen et
al. (2003) study were so different from the present study
and the original George et al. (1997) study. In addition,
based on our findings we may assume that if Dustman-
Allen et al. (2003) had not included the PFA data in
predicting VO2max, their results would have been worse.
Future studies need to clarify the accuracy of the PFA
variable across various samples.
It is interesting to note that the 30–39-year-old group
reported a higher mean PFA (15.9 vs. 17.1) and PA-R scores
(4.6 vs. 5.6) than the 18–29-year-old group; in addition the
40–49-year-old group reported a higher PA-R score (4.6
vs. 5.5) than the 18–29-year-old group (see Table 1). How-
ever, the mean VO2max values decreased progressively
across the four age groups as the participants increased
in age. One possible reason the PFA and VO2max scores
do not perfectly parallel one another across the four
groups is that the PFA questions are based on endurance
performance not fully explained by VO2max. For example,
it is commonly accepted that endurance performance is
mostly related to three variables: VO2max, lactate thresh-
old, and exercise economy (Bassett & Howley, 2000). If
more participants in the 30–39-year-old group had a higher
lactate threshold than the 18–29-year-old group, it may ex-
plain why they reported a higher PFA score yet exhibited
a lower VO2max. With respect to the PA-R variable, it may
be that the two older groups were simply more physically
active than younger group. Because one’s physical activity
does not fully explain all the VO2max variance, it is con-
ceivable that an older age group could be more physically
active yet have a lower VO2max than a younger age group.
Finally, the subjective nature of the N-EX questionnaire
could partially explain some of these discrepancies in the
PFA, PA-R, and VO2max data; however, these small mean
differences between age groups do not appear to affect
the overall accuracy of the model.
Although the N-EX model developed in this study
presents excellent results, there are obvious limitations
when using nonexercise data. For instance, individuals
who are unfamiliar with prolonged aerobic exercise or
unfamiliar with the pace it would take to cover 1 and 3
miles (1.6 and 4.8 km) may report inaccurate PFA scores.
RQES: December 2005
Bradshaw.pmd 431
Bradshaw, George, Hyde, LaMonte, Vehrs, Hager, and Yanowitz
Individuals who dislike jogging or running may under-
estimate their PFA score, because they prefer to walk.
Former athletes may overestimate their PFA score, be-
cause they believe they still have the ability to run at a
fast pace. The PA-R score is subjective as well and de-
pends on participants’ ability to distinguish between
light, moderate, and vigorous activity. To be accurate,
participants must correctly recall their physical activity
level over the previous 6 months. Finally, it is possible to
measure body mass and height (and calculate BMI
scores) and not have it self-reported, which may help to
decrease some possible error. However, whether body
mass and height are measured or self-reported probably
has little bearing on the VO2max prediction as reported
by Jackson et al. (1990) and George et al. (1997).
There are various advantages and disadvantages to
EX and N-EX VO2max regression models. An important
feature of EX models is the actual exercise test. As the
test progresses from a low to high exercise intensity, the
participant’s HR and RPE response can be used to de-
velop a safe and effective aerobic exercise prescription.
In addition, EX models yield relatively accurate VO2max
predictions. However, EX tests take time to complete,
often require expensive equipment, sometimes require
the individual to exercise to volitional fatigue, and may
not facilitate testing large groups of people at the same
time. Despite these possible drawbacks, EX models are
still preferred over N-EX models, when the participant
can safely perform an exercise test. However, N-EX
models serve as a valuable alternative to EX tests. Such
models predict CRF through questionnaire-based data
and are administered without expensive administrative
costs and time restraints. N-EX models are the most prac-
tical method for assessing VO2max in large populations
and may be useful in large epidemiological studies. N-
EX models may also be used in university wellness and
physical education classes to teach students the relation-
ship between VO2max and PFA. Finally, because of their
subjective nature, these models are not suitable for set-
tings in which CRF or the person’s functional capacity
must be measured (e.g., when qualifying for the mili-
tary, civil service, or police academy).
It is also noteworthy to mention that the mean
VO2max of the volunteer sample in this study is relatively
high when compared to population norms. Based on
normative data published by ACSM, the mean VO2max
scores for the 20–29-year-olds, 30–39-year-olds, 40–49-
years-olds, and 50–65-year-olds are at about the 90th,
80th, 70th, and 50th percentiles, respectively (ACSM,
2000). In addition, the male sample yielded a slightly
more accurate VO2max prediction than the female
sample, as noted in the results section. These unique
sample characteristics may influence the generalizability
of the VO2max prediction model developed in this study
(see Table 3) when it is applied to other healthy adults.
431
11/9/2006, 5:49 PM
 Bradshaw, George, Hyde, LaMonte, Vehrs, Hager, and Yanowitz
In conclusion, the new N-EX prediction model
developed in this study provides an accurate and con-
venient way to predict VO2max in healthy adult men and
women. The accuracy of this N-EX model is similar to
many of the popular N-EX models. The results show that
the PFA variable significantly improves the ability of the
regression model to predict VO2max accurately. Future
studies are needed to evaluate the accuracy and
generalizability of the PFA variable and N-EX regression
model in a variety of samples.
References
American College of Sports Medicine. (2000). ACSM’s guide-
lines for exercise testing and prescription (6th ed.). Philadel-
phia: Lippincott Williams & Wilkins.
Astrand, P.-O., & Ryhming, I. (1954). A nomogram for calcu-
lation of aerobic capacity (physical fitness) from pulse
rate during submaximal work. Journal of Applied Physiol-
ogy, 7, 218–221.
Bassett, D. R., & Howley, E. T. (2000). Limiting factors for
maximum oxygen uptake and determinants of endur-
ance performance. Medicine & Science in Sports & Exer-
cise, 32, 70–84.
Borg, G. A. V. (1982). Psychophysical bases of perceived exer-
tion. Medicine & Science in Sports & Exercise, 14, 377–381.
Bruce, R. A., Kusumi, F., & Hosmer, D. (1973). Maximal oxy-
gen intake and nomographic assessment of functional
aerobic impairment in cardiovascular disease. American
Heart Journal, 85, 546–562.
Cooper, K. H. (1968). A means of assessing maximal oxygen
intake. Journal of American Medical Association, 203, 135–138.
Dustman-Allen, K. M., & Plowman, S. A. (2003). Validation of
a non-exercise regression equation for the prediction
of maximal aerobic capacity. Medicine & Science in Sports
& Exercise, 35, S310.
Ebbeling, C. B., Ward, A., Puleo, E. M., Widrick, J., & Rippe,
J. M. (1991). Development of a single-stage submaximal
treadmill walking test. Medicine & Science in Sports & Ex-
ercise, 23, 966–973.
432
Bradshaw.pmd 432
George, J. D. (1996). Alternative approach to maximal exer-
cise testing and VO2max prediction in college students.
Research Quarterly for Exercise and Sport, 67, 452–457.
George, J. D., Stone, W. J., & Burkett, L. N. (1997). Non-exer-
cise VO2max estimation for physically active college stu-
dents. Medicine & Science in Sports & Exercise, 29, 415–423.
Heil, D. P., Freedson, P. S., Ahlquist, L. E., Price, J., & Rippe,
J. M. (1995). Nonexercise regression models to estimate
peak oxygen consumption. Medicine & Science in Sports
& Exercise, 27, 599–606.
Holiday, D. B., Ballard, J. E., & McKeown, B. C. (1995). PRESS-
related statistics: Regression tools for cross-validation and
case diagnostics. Medicine & Science in Sports & Exercise,
27, 612–620.
Howley, E. T., Bassett, D. R., Jr., & Welch, H. G. (1995). Crite-
ria for maximal oxygen uptake: Review and commentary.
Medicine & Science in Sports & Exercise, 27, 1293–1301.
Jackson, A. S., Blair, S. N., Mahar, M. T., Wier, L. T., Rossand,
R. M., & Stuteville, J. E. (1990). Prediction of functional
aerobic capacity without exercise testing. Medicine &
Science in Sports & Exercise, 22(6), 863–870.
Kline, G. M., Porcari, J. P., Hintermeister, R., Freedson, P. S.,
Ward, A., McCarron, R. F., et al. (1987). Estimation of
VO2max from a one-mile track walk, gender, age, and body
weight. Medicine & Science in Sports & Exercise, 19, 252–259.
Kolkhorst, F. W., & Dolgener F. A. (1994). Nonexercise model
fails to predict aerobic capacity in college students with high
VO2peak. Research Quarterly for Exercise and Sport, 65, 78–83.
Liu, N. Y.-S., Plowman S. A., & Looney M. A. (1992). The reli-
ability and validity of the 20-meter shuttle test in Ameri-
can students 12 to 15 years old. Research Quarterly for Exercise
and Sport, 64, 360–365.
Authors’ Note
Please address all correspondence concerning this ar-
ticle to James D. George, Department of Exercise Sci-
ences, Brigham Young University, 228-A Smith Field
House, Provo, Utah 84602.
E-mail: jim_george@byu.edu
RQES: December 2005
11/9/2006, 5:49 PM