- Email: [email protected]
Pedometer-determined ambulatory activity in individuals with type 2 diabetes
Diabetes Research and Clinical Practice 55 (2002) 191– 199 www.elsevier.com/locate/diabres
Pedometer-determined ambulatory activity in individuals with type 2 diabetes Catrine E. Tudor-Locke a,b,*, Rhonda C. Bell c, Anita M. Myers a,d, Stewart B. Harris e, Nicola Lauzon a, N. Wilson Rodger f a
The Center for Acti6ity and Ageing, The Uni6ersity of Western Ontario and St. Joseph’s Health Center, London, Ont., Canada b The Pre6ention Research Center, The Uni6ersity of South Carolina, Columbia, SC 29208, USA c Agricultural, Food and Nutritional Science, The Uni6ersity of Alberta, Edmonton, Alta., Canada d Department of Health Studies and Gerontology, The Uni6ersity of Waterloo, Waterloo, Ont., Canada e Faculty of Medicine and Dentistry, The Uni6ersity of Western Ontario, London, Ont., Canada f The Lawson Diabetes Center, The Uni6ersity of Western Ontario, St. Joseph’s Health Center, London, Ont., Canada. Received 13 October 2000; received in revised form 7 July 2001; accepted 25 July 2001
Abstract This cross-sectional study presents the first normative data on pedometer-determined ambulatory activity, defined as steps/day, in 160 (98 males, 62 females; age =52.49 5.3 years; BMI = 32.39 5.7) free-living individuals with type 2 diabetes. Participants took 6662 93077 steps per day, less than that reported in nondiabetic samples and more than that reported for samples living with more restrictive chronic conditions including claudication, joint replacement, chronic obstructive lung disease, and chronic heart failure. Steps/day and BMI were inversely and significantly correlated (r= −0.27, PB0.01). Further, there was a significant difference between BMI categories (from normal weight to obesity class III) with regard to steps/day (F =2.96, P B0.05). The difference was most apparent between the highest obesity classes (II and III) and normal weight categories. This data is useful for sample comparison purposes. In addition the standard deviation or variance estimates can be used to calculate samples sizes for intervention efforts. Objective quantification of ambulatory activity via simple and inexpensive pedometers permits researchers and practitioners to easily screen for level of activity along a continuum. This study opens the door for future research and clinical applications including identifying threshold values related to important health outcomes and evaluating incremental change due to various interventions in this population. Published by Elsevier Science Ireland Ltd. Keywords: Walking; Exercise; Physical activity
* Corresponding author. Present address: CLAS Bldg., Room 160, Department of Exercise and Wellness, Arizona State University East, 7001 E. Williams Field Rd., Mesa, AZ, 85212-0180, USA. Tel.: + 1-480-727-1944. E-mail address: [email protected] (C.E. TudorLocke).
1. Introduction Despite the well-known multiple benefits of a physically active lifestyle, individuals with type 2 diabetes are generally described as sedentary. The
0168-8227/02/$ - see front matter. Published by Elsevier Science Ireland Ltd. PII: S 0 1 6 8 - 8 2 2 7 ( 0 1 ) 0 0 3 1 7 - 5
192
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
empirical evidence to support this characterization is meager, however. For instance, a review of survey data from the 1990 United States National Health Interview Survey found that 66% of individuals with diabetes (both type 1 and type 2) reported no participation in regular leisure-time physical activity, compared with 59% of nondiabetics, although this difference was negated when activity limitations were considered [1]. These distinct populations did not differ on mean number of bouts of exercise, average number of minutes per bout of exercise, or mean total weekly hours of exercise [1]. Differences were noted, however, in activity choices. People with diabetes were more likely to engage in walking and less likely to engage in more vigorous activities (e.g. jogging, dancing, tennis, bicycling, skiing, weight-lifting, etc.) than were people without diabetes [1]. However, walking is considered the least reliably recalled physical activity [2– 4]. In addition, most, if not all, self-report instruments suffer from restrictive floor effects that limit researchers to describing the percentage of individuals who fail to register on the preferred instrument [5]. For example, Hays and Clark [6] recently reported that almost 55% of a sample of 260 individuals with type 2 diabetes reported zero minutes of weekly physical activity in response to questions about walking behaviors. Surely, these individuals are doing some level of physical activity that apparently continues to escape our best efforts at quantification through self-report questionnaires. If we are to document the physical activity patterns of this population, as well as investigate how changes in habitual behaviors impact on important health outcomes, then it is imperative that we accurately quantify walking behaviors. The use of simple and inexpensive pedometers to quantify physical activity, specifically ambulatory activity, is a recent development [5,7– 11]. Compared with accelerometers, pedometers are less expensive and more practical for both research and clinical applications [5]. In addition to being used to quantify habitual walking activity, pedometers may be useful motivational tools to encourage individuals with type 2 diabetes to increase walking behaviors [12,13]. Researchers and clinicians focused on type 2 diabetes require
normative data to assist with interpretation of pedometer data. In addition users want to know how pedometer output (steps/day) is related to health indicators. The purpose of this paper is to present, as a starting point, descriptive data from the first pedometer-assessed ambulatory activity database of individuals with type 2 diabetes. This database also permits an examination of the relationship between steps/day and body mass index or BMI (an indicator of body composition), as well as self-reported frequency of exercise, perceived health status, and perceived readiness (i.e. stages of change) for exercise.
2. Materials and methods
2.1. Recruitment A manual search of patient files at the Lawson Diabetes Centre (LDC) in Southwestern Ontario was used to generate a sampling pool of potential subjects. The LDC offers patient-centered programs in all aspects of diabetes, as well as for a variety of health care trainees and professionals. The LDC is staffed by a team of Certified Diabetes Educators, dietitians, and nurses, some having experience in the Diabetes Control and Complications Trial. There is close cooperation with family practices and specialized medical communities, and with other community resources e.g. London InterCommunity Health Centre. Eligibility criteria consisted of: a minimum of 3 months post-diagnosis with type 2 diabetes (defined according to Canadian Diabetes Association criteria [14]); management via oral hypoglycemic medication or diet only (no insulin); aged 40–60 years; and, no diseases or conditions (e.g. eye disease or heart conditions) for which physical activity is contraindicated and, therefore, likely low. The rationale for this last criteria was that we were interested in the physical activity of those individuals with type 2 diabetes who were not restricted due to health reasons, and in fact likely received education to increase their activity. Potential subjects were contacted by telephone or directly during individual counseling sessions with education center staff. Height and weight data,
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
collected during routine visits to the study center using standardized methods, was extracted from patient files. BMI was computed as height in kilograms/weight in meters squared.
2.2. Procedures Eligible volunteers met with research staff for a 10 min instruction session at the study center. At this time subjects provided informed consent, and were given a background questionnaire to complete at home as well as detailed instruction on using the pedometer (Yamax Digiwalker Model SW-200, Accusplit, CA) for self-monitoring purposes. Prior to use, pedometers were calibrated using a method described previously [7]. Briefly, pedometers were evaluated for accuracy during walking trials against steps counted using a handtally counter. Only instruments that were exactly correct were used for monitoring purposes, only two of 90 pedometers pre-tested in this manner proved untrustworthy. The background questionnaire solicited demographic information, reported exercise frequency, perceived health, and stages of change for physical activity. The exact wording for the reported exercise frequency [15] and perceived health status questions are shown in the Appendix. A Canadian version [15] (replacing the term exercise with physical activity) of the stages of change six stage scale [16] was used to classify readiness to change. Briefly, the stages are: precontemplation (no intent to change); contemplation (considering change); preparation (making small changes); action (actively engaged in changing behavior); and maintenance (continued successful behavior change) [17]. A sixth stage, relapse (discontinued behavior), included in the original scale [16], was also assessed in the version used here [15]. Subjects were instructed to wear the pedometer clipped to their clothing at the waist, on either side, centered over the foot. If the pedometer could not be held in a vertical plane (e.g. due to excessive abdominal adiposity), subjects were instructed to wear the pedometer at the back of the waistband, mid-line of the leg. Pedometer-determined ambulatory activity was measured on 3 consecutive days (including one weekend day and
193
two weekdays). Pilot testing using this 3 day monitoring frame on five subjects at two time points separated by one month elicited an intraclass correlation coefficient of 0.91. Subjects were instructed to set the pedometer to zero at the beginning of the first day of self-monitoring and to seal it with a sticker. Upon completion of the 3 days of measurement, participants mailed back the sealed pedometer and the background questionnaire in a pre-addressed and stamped padded envelope. Using a similar process described by Sequira et al. [18], the average number of steps recorded on the pedometers during mail transit was determined during pilot testing. Ten pedometers were reset to zero, placed in padded mailing envelopes, and mailed to the research center from different locations within the study catchment area. An average of 803 steps were accumulated during mail transit. This value was used to correct values recorded for the present study. The pedometer reading was immediately recorded upon its receipt at the study center, corrected for mail transit, and divided by 3 days of assessment to arrive at the variable steps/day used for analysis.
2.3. Data treatment and statistical analysis Pedometer-determined ambulatory activity data was treated as both a continuous variable (steps/ day) and as a categorical variable defined according to tertiles identified by the 25 and 75th percentiles. Recommended categories for BMI distribution [19] were used to represent normal weight (BMI B 25), overweight (BMI 35.0–29.0), obesity Class I (BMI 30.0–29.9), obesity Class II (BMI 35.0–39.9), and obesity Class III (BMI] 40). No individuals in this study were classified as underweight (BMIB 18.5). Age, BMI, and steps/day data were normally distributed. Descriptive data are expressed as means 9 S.D. (95% CI). Independent t-tests were used to compare mean steps/day between genders. Means of steps/day across categories of reported frequency of exercise and perceived health were compared with one way analysis of variance (ANOVA). Pearson-product moment correlation coefficients were used to examine the relationship between steps/day and BMI as continuous vari-
194
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
ables, after checking for evidence of collinearity. Regression analysis was used to predict a line of best fit for those data. ANOVA was used to examine, (1) whether steps/day varied significantly among BMI categories; and (2) whether BMI varied significantly among tertiles of steps/day. If the overall, or omnibus, ANOVA was significant, Student –Newman –Keuls tests were employed to compare significant differences between groups. Data were analyzed using SPSS Version 8.0 Statistical Software; P B0.05 was considered to be statistically significant. Stages of change for exercise behavior was compressed to three stages with respect to reported behavior, irregardless of intention to change. Individuals in precontemplation, contemplation, or relapse were classified as similarly inactive (e.g. I am not physically active, I do not exercise, etc.); individuals in preparation were classified as irregularly active (e.g. I am physically active once in a while, I exercise once in a while but not regularly, etc.); while those in action or maintenance were classified as regularly active (e.g. I am currently physically active, I participate in regular physical activity, I exercise regularly, etc.). Stages of change have been previously collapsed into this three stage activity categorization and validated according to self-reported participation in vigorous leisure time activity [20].
not be contacted due to wrong phone numbers (115 or 29.0%) or inability to reach patient by telephone after four to five repeated attempts (290 or 71.0%). Of those contacted, 225 (65.4%) agreed to participate in the project. The remainder refused (119 or 34.6%) citing no time (14 or 11.8%), not interested (75 or 63.0%), reporting exclusion criteria not documented in patient charts (8 or 6.7%), transportation issues (3 or 2.5%), moving out of town (5 or 4.2%), and other medical reasons (14 or 11.8%). One hundred and eighty-four (82%) pedometers were returned to the study center. The remaining pedometers were not returned, lost, or broken during mail-back. Data from 24 subjects were deleted prior to analysis due to missing data from patient files (e.g. weight or height, n= 23) or incomplete background questionnaires (n = 1). The final sample (N= 160) consisted of 98 (61.3%) males and 62 (38.8%) females. Education level was reported as public school (1 or 0.6%), some secondary school (37 or 23.1%), high school diploma (42 or 26.3%), some post-secondary (36 or 22.5%), and college or university graduate (40 or 25.0%). Descriptive data for the total sample by gender are presented in Table 1. There were no significant differences between males and females for age, BMI, or steps/day. Mean duration of diabetes was 1.99 2.2 (CI 1.6–2.2) years. Eighty-six (53.8%) were treated with diet only. There was a non-significant trend (P = 0.09) for higher steps/ day in individuals treated with diet only compared with individuals also taking oral hypoglycemic medication (70439 3009 vs. 62189 3116). Seventy-two subjects (45%) reported having been diagnosed with high blood pressure, 56 (35%) with high cholesterol, and 8 (5%) with heart troubles. Perceived health status was reported as
3. Results Seven hundred and forty eligible subjects were identified through manual search of patient files at the study center. Research staff attempted contact with potential subjects by telephone. Three hundred and ninety-six (53.5%) eligible subjects could Table 1 Descriptive characteristics for total sample and each gender group
Age BMI Steps/day
Total Sample (n= 160)
Males (n = 98)
Females (n = 62)
52.4 95.3 (51.6–53.3) 32.39 5.7 (31.4–33.2) 66629 3077 (6181–7142)
51.8 9 5.4 (50.7–52.9) 31.89 5.2 (30.8–32.9) 6859 9 3086 (6240–7477)
53.4 95.1 (52.1–54.7) 33.1 9 6.4 (31.4–34.7) 6350 9 3062 (5572–7128)
Note: values are means 9 S.D. (95% CI). No significant differences between males and females for age, BMI, or steps/day.
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
195
Table 2 Comparison of variables between categories of perceived health status Variable
Age BMI Steps/day
Category of perceived health status Poor (n =7)
Fair (n=39)
Good (n =105)
Excellent (n =9)
53.1 93.2 (50.2–56.1) 33.39 3.8 (29.7–36.8) 5499 91781 (3852–7146)
51.3 95.6 (49.5–53.1) 32.7 95.6 (30.9–34.5) 61699 2713 (5290–7049)
52.6 9 5.4 (51.5–53.6) 32.5 95.9 (31.3–33.6) 6740 9 3275 (6107–7374)
55.2 9 4.3 (51.9–58.5) 27.7 94.2 (24.5–30.9) 8781 91985 (7256–10307)
Note: values are means 9S.D. (95%CI). No significant differences between categories for age, BMI, or steps/day. Table 3 Comparison of variables between categories of exercise frequency Variable
Age BMI Steps/day
Category of exercise frequency (number of times per week) Rarely or never (n = 56)
Once or twice (n = 54)
At least three times (n = 50)
52.2 9 5.1 (50.9–53.6) 32.09 3.9 (31.0–33.1) 63049 2971 (5509–7100)
52.7 95.3 (51.2–54.1) 32.4 95.7 (30.8–33.9) 6594 9 3046 (5763–7456)
52.4 9 5.7 (50.8–54.0) 32.5 9 7.3 (30.4–34.6) 7135 93226 (6218–8052)
Note: values are means 9S.D. (95%CI). No significant differences between categories for age, BMI, or steps/day.
poor (7 or 4.4%), fair (39 or 24.4%), good (105 or 65.6%), or excellent (9 or 5.6%). Reported frequency of exercise, over a typical 7-day period, sufficiently prolonged and intense to cause sweating and rapid heart beat, was reported as rarely or never (56 or 35.0%), normally once or twice (54 or 33.8%), and at least three times (50 or 31.3%). There were no significant differences between perceived health (Table 2) or reported exercise frequency (Table 3) categories for age or BMI. Although mean steps/day increased with both increased perceived health (Table 2) and higher reported frequency of exercise (Table 3), the differences were not statistically significant (F = 2.16, P= 0.095; and F= 0.98, P =0.38, respectively) due to a noticeable overlap in 95% confidence intervals for the mean. According to the collapsed 3-stage activity categorization based on stages of change, 41 (25.6%) of the sample were inactive (5 precontemplation, 25 contemplation, 11 relapse) and took an average of 59139 2962 (CI 4977– 6848) steps/day. Fifty-four (33.8%) were irregularly active (preparation), and averaged 66679 2890 (CI 5,878– 7456) steps/day. Fifty-eight (36.3%) were regularly active (26 action, 32 maintenance) with an aver-
age of 72799 3300 (CI 6411– 8147). As in the case of perceived health and reported exercise frequency, stages of change for exercise, compressed to three stages to reflect reported behavior, did not differ significantly across stages for steps/day (F= 2.38, P= 0.096), despite increasing mean steps/day. Pedometer-determined tertiles of activity were defined as low (below the 25th percentile of distribution, or 5 4716 steps/day) moderate (between the 25 and 75th percentiles of distribution, or 4717–8644 steps/day) and high activity (above the 75th percentile of distribution or ] 8645 steps/ day). Table 4 displays the comparison of age and BMI for the low, moderate, and high categories of pedometer-determined ambulatory activity (steps/ day). There was no significant difference in age across pedometer-determined ambulatory activity groups. BMI was significantly different across pedometer-determined ambulatory activity tertiles (F= 6.85, PB 0.01). Individuals falling in the high activity category had a significantly lower BMI than those in either the moderate or low categories; the latter groups did not differ significantly.
196
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
Analyzed as continuous variables, steps/day and BMI were inversely and significantly correlated (r = − 0.27, P B0.01). Fig. 1 presents a visual depiction of the line of best fit against the distribution based on regression analysis. Table 5 presents the descriptive data for age and steps/day classified according to BMI category. Age did not differ significantly across categories. There was a significant difference between BMI categories with regard to steps/day (F = 2.96, P B 0.05). The difference was most apparent between the highest obesity classes (II and III) and normal weight
categories, although only eight individuals were classified as normal weight in this sample.
4. Discussion This is the first study to describe ambulatory activity, measured using a simple and inexpensive pedometer, in free-living individuals with type 2 diabetes. The pedometer values recorded herein (66629 3080 steps/day) are slightly lower in comparison to two other similarly-aged samples with-
Table 4 Comparison of variables between categories of pedometer-determined ambulatory activity Variable
Age BMI
Categories of pedometer-determined ambulatory activity Low (n= 40)54716 steps/day
Moderate (n = 80)4717–8644 steps/day
High (n = 40)]8645 steps/day
53.1 95.5a (51.4–54.9) 34.3 95.4a (32.6–36.0)
51.8 9 5.2a (50.6–53.0) 32.695.7a (31.3–33.8)
53.0 95.3a (51.3–54.7) 29.8 95.4b (28.1–31.5)
Note: values are means 9S.D. (95% CI). Categories of Digiwalker pedometer-assessed physical activity determined using percentiles of distribution: low 525th, \25th moderateB75th, high ]75th. Means with different letter superscripts are significantly different PB0.01.
Fig. 1. Regression of average steps/day versus body mass index (BMI) in individuals with type 2 diabetes.
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
197
Table 5 Comparison of variables between categories of BMI-defined body composition for total sample Variable
Age Steps/day
Categories of BMI-defined body composition Normal Weight (n =8)
Overweight (n =56)
Obesity Class I (n = 55)
Obesity Class II (n = 27)
Obesity Class III (n =14)
55.6 95.4a (51.1–60.1) 8607 9 3508a (5674-11539)
52.7 9 5.2a (51.3–54.1) 7423 9 2691a,b (6702-8144)
52.3 95.1a (51.0–53.7) 6309 93,083a,b (5475-7142)
52.3 9 5.6a (50.1–54.5) 5621 93223b (4346–6896)
50.1 9 5.9a (46.7–53.5) 5899 93175b (4066–7732)
Note: values are means 9 S.D. (95% CI). Categories of BMI are defined in kg/m2 as Normal Weight 18.5–24.9, Overweight 25.0–29.9, Obesity Class I 30.0–34.9, Obesity class II 35.0–39.9, Obesity Class III]40. No subjects were classified as underweight (BMIB18.5). Means with different letter superscripts are significantly different PB0.01.
out reported diabetes. Tudor-Locke et al. [21] reported values of 737093080 steps/day in a bi-ethnic sample of 109 adults, mean age 44.99 15.8 years. McClung et al. [22] reported 77819 2807 steps/day in a subsample of 58 healthy adults aged 22–82 years. The values for individuals with type 2 diabetes in the present study are, however, higher than those reported for individuals with claudication (approximately 4100– 5300 steps/day) [23], joint replacements (approximately 3500– 6000 steps/day) [22,24], chronic obstructive pulmonary disease (approximately 3800 steps/day) [25], and chronic heart failure (approximately 3500– 3700 steps/day) [26–28]. The values in the present study of free-living individuals also exceed those reported by Yamanouchi et al. [12] for hospital-based individuals with type 2 diabetes serving as control subjects (approximately 4500 steps/day) in a pedometer-based intervention study. Interestingly, the treatment group in that study averaged 19200 steps/day over 6–8 weeks. The inverse relationship found between pedometer-determined steps/day and BMI is consistently reported in other nondiabetic samples of adults [21,22] and children [29]. Although McClung et al. [22] did not report a correlation coefficient, the value we computed (r= − 0.27) is on the order of that reported by Tudor-Locke et al. [21] (r = − 0.30) and Rowlands et al. [29] (r = −0.42). A simple question querying perceived health status failed to differentiate physical activity quantified as steps/day. Similarly, querying about frequency of exercise participation in the manner
employed failed to identify individual’s true activity levels. Although the simple and inexpensive pedometer is admittedly not the best measure of overall physical activity, it appears to be useful for objectively describing daily incidental and intentional ambulatory activity. Although pedometers are not sensitive to intensity of activity, we know that individuals with type 2 diabetes do not typically participate in vigorous activities and that the most common activity choice is walking, or ambulatory activity [1]. Although it has been suggested that a combination of pedometer and questionnaire methods may be able to minimize collective weaknesses [30], that postulation has yet to be explored empirically. The 3-stage activity categorization has been previously used to infer level of physical activity from reported intention. It has been validated only against participation in vigorous leisure time physical activity behavior [20] using a commonly employed self-reported measure of physical activity, namely, the 7-day physical activity recall questionnaire [31]. A difference was not evident between categories for participation in moderate intensity activities in that study [20]. Using an objective and direct measure of ambulatory activity (i.e. pedometers) we found no difference in steps/day between stages of change categories. Our data suggests that stages of change is not a good indicator of current physical activity, at least in individuals with type 2 diabetes, and should, therefore, not be used as a physical activity screening tool or as an outcome measure of change due to intervention.
198
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
The cross-sectional design of this study limits the ability to make firm causal conclusions. Further, the study sample is a convenience sample of volunteers recruited according to strict eligibility criteria (e.g. no insulin, no physical activity limitations/contraindications) and, therefore limits generalizability of the findings. We do not know how, or if, individuals who did not participate (e.g. not eligible or not interested) differ from the sample studied. The 3-day monitoring frame may also be considered a study design weakness; it has been previously reported that 5– 6 days of pedometer data are required to accurately describe physical activity with less than 5% error [32]. That study was conducted with young males purposefully recruited to represent a wide range of physical activity levels. Relevant to the present study, Sieminski et al. [33] reported an intraclass correlation statistic of 0.86 for pedometer-determined ambulatory activity averaged over 2 days (determined 1 week apart) in older, free-living claudication patients. Similarly, we found evidence of high repeatability during pilot work using the 3-day pedometer assessment protocol as described earlier. It appears that a shorter monitoring frame is sufficient in samples that exhibit relatively stable behavior (i.e. individuals living with chronic conditions, including type 2 diabetes). In summary, this study presents the first normative data on pedometer-determined ambulatory activity in individuals with type 2 diabetes. This data is useful for sample comparison purposes. In addition the standard deviation or variance estimates can be used to calculate samples sizes for intervention efforts. A primary finding was a distinct and consistent inverse relationship between steps/day and BMI, an indicator of body composition. Another lesson learned was that questions about frequency of exercise, perceived health, or stages of change were not able to capture objective differences in ambulatory activity in this population. Objective quantification of ambulatory activity via simple and inexpensive pedometers permits researchers and practitioners to easily screen for level of activity along a continuum. This study opens the door for future research and clinical applications including identifying threshold values
related to important health outcomes and evaluating incremental change due to various interventions in this population.
Acknowledgements We would like to acknowledge the financial support of the Canadian Diabetes Association. We would also like to acknowledge the assistance of the staff and educators at the Lawson Diabetes Center, St. Joseph’s Health Center, London, Ontario, Canada.
Appendix A Over a typical 7-day period (1 week), how many times do you engage in physical activity that is sufficiently prolonged and intense to cause sweating and a rapid heart beat? (Check one). At least three times. Normally once or twice. Rarely or never. In general, how would you describe your current state of health? Excellent. Good. Fair. Poor.
References [1] E.S. Ford, W.H. Herman, Leisure-time physical activity patterns in the U.S. diabetic population. Findings from the National Health Interview Survey – Health Promotion and Disease Prevention Supplement, Diabetes Care 18 (1995) (1990) 27 – 33. [2] A.M. Kriska, W.C. Knowler, R.E. LaPorte, et al., Development of questionnaire to examine relationship of physical activity and diabetes in Pima Indians, Diabetes Care 13 (1990) 401 – 411. [3] B.E. Ainsworth, A.S. Leon, M.T. Richardson, D.R. Jacobs, R.S. Paffenbarger, Accuracy of the college alumnus physical activity questionnaire, J. Clin. Epidemiol. 46 (1993) 1403 – 1411. [4] M.T. Richardson, A.S. Leon, D.R. Jacobs, B.E. Ainsworth, R. Serfass, Comprehensive evaluation of the
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Minnesota Leisure Time Physical Activity Questionnaire, J. Clin. Epidemiol. 47 (1994) 271 –281. C.E. Tudor-Locke, A.M. Myers, Challenges and opportunities for measuring physical activity in sedentary adults, Sports Med. 31 (2001) 91 –100. L.M. Hays, D.O. Clark, Correlates of physical activity in a sample of older adults with type 2 diabetes, Diabetes Care 22 (1999) 706. C.E. Tudor-Locke, A.M. Myers, Methodological considerations for researchers and practitioners using pedometers to measure physical (ambulatory) activity, Res. Q Exerc. Sport 72 (2001) 1 –12. A.V. Rowlands, R.G. Eston, E.K. Ingledew, Measurement of physical activity in children with particular reference to the use of heart rate and pedometry, Sports Med. 24 (1997) 258 – 272. G.J. Welk, C.B. Corbin, D. Dale, Measurement issues in the assessment of physical activity in children, Res. Q Exerc. Sport 71 (2000) 59 –73. P.S. Freedson, K. Miller, Objective monitoring of physical activity using motion sensors and heart rate, Res. Q Exerc. Sport 71 (2000) 21 –29. D.R. Bassett, Validity and reliability issues in objective monitoring of physical activity, Res. Q Exerc. Sport 71 (2000) 30 – 36. K. Yamanouchi, T. Takashi, K. Chikada, et al., Daily walking combined with diet therapy is a useful means for obese NIDDM patients not only to reduce body weight but also to improve insulin sensitivity, Diabetes Care 18 (1995) 775 – 778. C. Tudor-Locke, A.M. Myers, N.W. Rodger, Formative evaluation of The First Step Program: A practical intervention to increase daily physical activity, Can. J. Diabetes Care 24 (2000) 34 –38. S. Meltzer, L. Leiter, D. Daneman, et al., Clinical practice guidelines for the management of diabetes mellitus in Canada. Canadian Diabetes Association, CMAJ 159 (1998) S1 – 29. Canadian Society for Exercise Physiology, The Canadian Physical Activity, Fitness and Lifestyle Appraisal. Ottawa: The Canadian Society for Exercise Physiology; 1998. B.H. Marcus, V.C. Selby, R.S. Niaura, J.S. Rossi, Selfefficacy and the stages of exercise behavior change, Res. Q Exerc. Sport 63 (1992) 60 –66. B.H. Marcus, L.R. Simkin, The transtheoretical model: applications to exercise behavior, Med. Sci. Sports Exerc. 26 (1994) 1400 – 1404. M.M. Sequeira, M. Rickenbach, V. Wietlisbach, B. Tullen, Y. Schutz, Physical activity assessment using a pedometer and its comparison with a questionnaire in a large population survey, Am. J. Epidemiol. 142 (1995) 989 – 999. National Institutes of Health, Clinical guidelines on the identification, evaluation, and treatment of overweight
[20] [21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
199
and obesity in adults —The evidence report., Obesity Res. 6 (1998) 51S – 209S. B.H. Marcus, L.R. Simkin, The stages of exercise behavior, J. Sports Med. Phys. Fitness 33 (1993) 83 – 88. C. Tudor-Locke, B.E. Ainsworth, M.C. Whitt, R.W. Thompson, C.L. Addy, D.A. Jones, The relationship between pedometer-determined ambulatory activity and body composition variables, Int. J. Obes. 25 (2001) 1571 – 1578. C.D. McClung, C.A. Zahiri, J.K. Higa, H.C. Amstutz, T.P. Schmalzried, Relationship between Body Mass Index and activity in hip or knee arthorplasty patients, J. Ortho. Res. 18 (2000) 35 – 39. A.W. Gardner, D.J. Sieminski, L.A. Killewich, The effect of cigarette smoking on free-living daily physical activity in older claudication patients, Angiology 48 (1997) 947 – 955. T.P. Schmalzried, E.S. Szuszczewicz, M.R. Northfield, et al., Quantitative assessment of walking activity after total hip or knee replacement, J. Bone Joint Surg. 80-A (1998) 54 – 59. B. Schonhofer, P. Ardes, M. Geibel, D. Kohler, P.W. Jones, Evaluation of a movement detector to measure daily activity in patients with chronic lung disease, Eur. Respir. J. 10 (1997) 2814 – 2819. D.J. Hoodless, K. Stainer, N. Savic, P. Batin, M. Hawkins, A.J. Cowley, Reduced customary activity in chronic heart failure: Assessment with a new shoemounted pedometer, Int. J. Cardiol. 43 (1994) 39 – 42. J.T. Walsh, A. Charlesworth, R. Andrews, M. Hawkins, A.J. Cowley, Relation of daily activity levels in patients with chronic heart failure to long-term prognosis, Am. J. Cardiol. 79 (1997) 1364 – 1369. A.J. Cowley, L. Fullwood, K. Stainer, J.R. Hampton, Exercise tolerance in patients with heart failure--how should it be measured?, Eur. Heart J. 12 (1991) 50 – 54. A.V. Rowlands, R.G. Eston, D.K. Ingledew, Relationship between activity levels, aerobic fitness, and body fat in 8to 10-year-old children, J. App. Physiol. 86 (1999) 1428 – 1435. A. Kriska, Physical activity and the prevention of type 2 diabetes mellitus: How much for how long?, Sports Med. 29 (2000) 147 – 151. S.N. Blair, W.L. Haskell, P. Ho, et al., Assessment of habitual physical activity by a seven-day recall in a community survey and controlled experiments, Am. J. Epidemiol. 122 (1985) 794 – 804. R. Washburn, M.K. Chin, H.J. Montoye, Accuracy of pedometer in walking and running, Res. Q Exerc. Sport 51 (1980) 695 – 702. D.J. Sieminski, L.L. Cowell, P.S. Montgomery, S.B. Pillai, A.W. Gardner, Physical activity monitoring in patients with peripheral arterial occlusive disease, J. Cardpulm. Rehabil. 17 (1997) 43 – 47.
Pedometer-determined ambulatory activity in individuals with type 2 diabetes Catrine E. Tudor-Locke a,b,*, Rhonda C. Bell c, Anita M. Myers a,d, Stewart B. Harris e, Nicola Lauzon a, N. Wilson Rodger f a
The Center for Acti6ity and Ageing, The Uni6ersity of Western Ontario and St. Joseph’s Health Center, London, Ont., Canada b The Pre6ention Research Center, The Uni6ersity of South Carolina, Columbia, SC 29208, USA c Agricultural, Food and Nutritional Science, The Uni6ersity of Alberta, Edmonton, Alta., Canada d Department of Health Studies and Gerontology, The Uni6ersity of Waterloo, Waterloo, Ont., Canada e Faculty of Medicine and Dentistry, The Uni6ersity of Western Ontario, London, Ont., Canada f The Lawson Diabetes Center, The Uni6ersity of Western Ontario, St. Joseph’s Health Center, London, Ont., Canada. Received 13 October 2000; received in revised form 7 July 2001; accepted 25 July 2001
Abstract This cross-sectional study presents the first normative data on pedometer-determined ambulatory activity, defined as steps/day, in 160 (98 males, 62 females; age =52.49 5.3 years; BMI = 32.39 5.7) free-living individuals with type 2 diabetes. Participants took 6662 93077 steps per day, less than that reported in nondiabetic samples and more than that reported for samples living with more restrictive chronic conditions including claudication, joint replacement, chronic obstructive lung disease, and chronic heart failure. Steps/day and BMI were inversely and significantly correlated (r= −0.27, PB0.01). Further, there was a significant difference between BMI categories (from normal weight to obesity class III) with regard to steps/day (F =2.96, P B0.05). The difference was most apparent between the highest obesity classes (II and III) and normal weight categories. This data is useful for sample comparison purposes. In addition the standard deviation or variance estimates can be used to calculate samples sizes for intervention efforts. Objective quantification of ambulatory activity via simple and inexpensive pedometers permits researchers and practitioners to easily screen for level of activity along a continuum. This study opens the door for future research and clinical applications including identifying threshold values related to important health outcomes and evaluating incremental change due to various interventions in this population. Published by Elsevier Science Ireland Ltd. Keywords: Walking; Exercise; Physical activity
* Corresponding author. Present address: CLAS Bldg., Room 160, Department of Exercise and Wellness, Arizona State University East, 7001 E. Williams Field Rd., Mesa, AZ, 85212-0180, USA. Tel.: + 1-480-727-1944. E-mail address: [email protected] (C.E. TudorLocke).
1. Introduction Despite the well-known multiple benefits of a physically active lifestyle, individuals with type 2 diabetes are generally described as sedentary. The
0168-8227/02/$ - see front matter. Published by Elsevier Science Ireland Ltd. PII: S 0 1 6 8 - 8 2 2 7 ( 0 1 ) 0 0 3 1 7 - 5
192
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
empirical evidence to support this characterization is meager, however. For instance, a review of survey data from the 1990 United States National Health Interview Survey found that 66% of individuals with diabetes (both type 1 and type 2) reported no participation in regular leisure-time physical activity, compared with 59% of nondiabetics, although this difference was negated when activity limitations were considered [1]. These distinct populations did not differ on mean number of bouts of exercise, average number of minutes per bout of exercise, or mean total weekly hours of exercise [1]. Differences were noted, however, in activity choices. People with diabetes were more likely to engage in walking and less likely to engage in more vigorous activities (e.g. jogging, dancing, tennis, bicycling, skiing, weight-lifting, etc.) than were people without diabetes [1]. However, walking is considered the least reliably recalled physical activity [2– 4]. In addition, most, if not all, self-report instruments suffer from restrictive floor effects that limit researchers to describing the percentage of individuals who fail to register on the preferred instrument [5]. For example, Hays and Clark [6] recently reported that almost 55% of a sample of 260 individuals with type 2 diabetes reported zero minutes of weekly physical activity in response to questions about walking behaviors. Surely, these individuals are doing some level of physical activity that apparently continues to escape our best efforts at quantification through self-report questionnaires. If we are to document the physical activity patterns of this population, as well as investigate how changes in habitual behaviors impact on important health outcomes, then it is imperative that we accurately quantify walking behaviors. The use of simple and inexpensive pedometers to quantify physical activity, specifically ambulatory activity, is a recent development [5,7– 11]. Compared with accelerometers, pedometers are less expensive and more practical for both research and clinical applications [5]. In addition to being used to quantify habitual walking activity, pedometers may be useful motivational tools to encourage individuals with type 2 diabetes to increase walking behaviors [12,13]. Researchers and clinicians focused on type 2 diabetes require
normative data to assist with interpretation of pedometer data. In addition users want to know how pedometer output (steps/day) is related to health indicators. The purpose of this paper is to present, as a starting point, descriptive data from the first pedometer-assessed ambulatory activity database of individuals with type 2 diabetes. This database also permits an examination of the relationship between steps/day and body mass index or BMI (an indicator of body composition), as well as self-reported frequency of exercise, perceived health status, and perceived readiness (i.e. stages of change) for exercise.
2. Materials and methods
2.1. Recruitment A manual search of patient files at the Lawson Diabetes Centre (LDC) in Southwestern Ontario was used to generate a sampling pool of potential subjects. The LDC offers patient-centered programs in all aspects of diabetes, as well as for a variety of health care trainees and professionals. The LDC is staffed by a team of Certified Diabetes Educators, dietitians, and nurses, some having experience in the Diabetes Control and Complications Trial. There is close cooperation with family practices and specialized medical communities, and with other community resources e.g. London InterCommunity Health Centre. Eligibility criteria consisted of: a minimum of 3 months post-diagnosis with type 2 diabetes (defined according to Canadian Diabetes Association criteria [14]); management via oral hypoglycemic medication or diet only (no insulin); aged 40–60 years; and, no diseases or conditions (e.g. eye disease or heart conditions) for which physical activity is contraindicated and, therefore, likely low. The rationale for this last criteria was that we were interested in the physical activity of those individuals with type 2 diabetes who were not restricted due to health reasons, and in fact likely received education to increase their activity. Potential subjects were contacted by telephone or directly during individual counseling sessions with education center staff. Height and weight data,
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
collected during routine visits to the study center using standardized methods, was extracted from patient files. BMI was computed as height in kilograms/weight in meters squared.
2.2. Procedures Eligible volunteers met with research staff for a 10 min instruction session at the study center. At this time subjects provided informed consent, and were given a background questionnaire to complete at home as well as detailed instruction on using the pedometer (Yamax Digiwalker Model SW-200, Accusplit, CA) for self-monitoring purposes. Prior to use, pedometers were calibrated using a method described previously [7]. Briefly, pedometers were evaluated for accuracy during walking trials against steps counted using a handtally counter. Only instruments that were exactly correct were used for monitoring purposes, only two of 90 pedometers pre-tested in this manner proved untrustworthy. The background questionnaire solicited demographic information, reported exercise frequency, perceived health, and stages of change for physical activity. The exact wording for the reported exercise frequency [15] and perceived health status questions are shown in the Appendix. A Canadian version [15] (replacing the term exercise with physical activity) of the stages of change six stage scale [16] was used to classify readiness to change. Briefly, the stages are: precontemplation (no intent to change); contemplation (considering change); preparation (making small changes); action (actively engaged in changing behavior); and maintenance (continued successful behavior change) [17]. A sixth stage, relapse (discontinued behavior), included in the original scale [16], was also assessed in the version used here [15]. Subjects were instructed to wear the pedometer clipped to their clothing at the waist, on either side, centered over the foot. If the pedometer could not be held in a vertical plane (e.g. due to excessive abdominal adiposity), subjects were instructed to wear the pedometer at the back of the waistband, mid-line of the leg. Pedometer-determined ambulatory activity was measured on 3 consecutive days (including one weekend day and
193
two weekdays). Pilot testing using this 3 day monitoring frame on five subjects at two time points separated by one month elicited an intraclass correlation coefficient of 0.91. Subjects were instructed to set the pedometer to zero at the beginning of the first day of self-monitoring and to seal it with a sticker. Upon completion of the 3 days of measurement, participants mailed back the sealed pedometer and the background questionnaire in a pre-addressed and stamped padded envelope. Using a similar process described by Sequira et al. [18], the average number of steps recorded on the pedometers during mail transit was determined during pilot testing. Ten pedometers were reset to zero, placed in padded mailing envelopes, and mailed to the research center from different locations within the study catchment area. An average of 803 steps were accumulated during mail transit. This value was used to correct values recorded for the present study. The pedometer reading was immediately recorded upon its receipt at the study center, corrected for mail transit, and divided by 3 days of assessment to arrive at the variable steps/day used for analysis.
2.3. Data treatment and statistical analysis Pedometer-determined ambulatory activity data was treated as both a continuous variable (steps/ day) and as a categorical variable defined according to tertiles identified by the 25 and 75th percentiles. Recommended categories for BMI distribution [19] were used to represent normal weight (BMI B 25), overweight (BMI 35.0–29.0), obesity Class I (BMI 30.0–29.9), obesity Class II (BMI 35.0–39.9), and obesity Class III (BMI] 40). No individuals in this study were classified as underweight (BMIB 18.5). Age, BMI, and steps/day data were normally distributed. Descriptive data are expressed as means 9 S.D. (95% CI). Independent t-tests were used to compare mean steps/day between genders. Means of steps/day across categories of reported frequency of exercise and perceived health were compared with one way analysis of variance (ANOVA). Pearson-product moment correlation coefficients were used to examine the relationship between steps/day and BMI as continuous vari-
194
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
ables, after checking for evidence of collinearity. Regression analysis was used to predict a line of best fit for those data. ANOVA was used to examine, (1) whether steps/day varied significantly among BMI categories; and (2) whether BMI varied significantly among tertiles of steps/day. If the overall, or omnibus, ANOVA was significant, Student –Newman –Keuls tests were employed to compare significant differences between groups. Data were analyzed using SPSS Version 8.0 Statistical Software; P B0.05 was considered to be statistically significant. Stages of change for exercise behavior was compressed to three stages with respect to reported behavior, irregardless of intention to change. Individuals in precontemplation, contemplation, or relapse were classified as similarly inactive (e.g. I am not physically active, I do not exercise, etc.); individuals in preparation were classified as irregularly active (e.g. I am physically active once in a while, I exercise once in a while but not regularly, etc.); while those in action or maintenance were classified as regularly active (e.g. I am currently physically active, I participate in regular physical activity, I exercise regularly, etc.). Stages of change have been previously collapsed into this three stage activity categorization and validated according to self-reported participation in vigorous leisure time activity [20].
not be contacted due to wrong phone numbers (115 or 29.0%) or inability to reach patient by telephone after four to five repeated attempts (290 or 71.0%). Of those contacted, 225 (65.4%) agreed to participate in the project. The remainder refused (119 or 34.6%) citing no time (14 or 11.8%), not interested (75 or 63.0%), reporting exclusion criteria not documented in patient charts (8 or 6.7%), transportation issues (3 or 2.5%), moving out of town (5 or 4.2%), and other medical reasons (14 or 11.8%). One hundred and eighty-four (82%) pedometers were returned to the study center. The remaining pedometers were not returned, lost, or broken during mail-back. Data from 24 subjects were deleted prior to analysis due to missing data from patient files (e.g. weight or height, n= 23) or incomplete background questionnaires (n = 1). The final sample (N= 160) consisted of 98 (61.3%) males and 62 (38.8%) females. Education level was reported as public school (1 or 0.6%), some secondary school (37 or 23.1%), high school diploma (42 or 26.3%), some post-secondary (36 or 22.5%), and college or university graduate (40 or 25.0%). Descriptive data for the total sample by gender are presented in Table 1. There were no significant differences between males and females for age, BMI, or steps/day. Mean duration of diabetes was 1.99 2.2 (CI 1.6–2.2) years. Eighty-six (53.8%) were treated with diet only. There was a non-significant trend (P = 0.09) for higher steps/ day in individuals treated with diet only compared with individuals also taking oral hypoglycemic medication (70439 3009 vs. 62189 3116). Seventy-two subjects (45%) reported having been diagnosed with high blood pressure, 56 (35%) with high cholesterol, and 8 (5%) with heart troubles. Perceived health status was reported as
3. Results Seven hundred and forty eligible subjects were identified through manual search of patient files at the study center. Research staff attempted contact with potential subjects by telephone. Three hundred and ninety-six (53.5%) eligible subjects could Table 1 Descriptive characteristics for total sample and each gender group
Age BMI Steps/day
Total Sample (n= 160)
Males (n = 98)
Females (n = 62)
52.4 95.3 (51.6–53.3) 32.39 5.7 (31.4–33.2) 66629 3077 (6181–7142)
51.8 9 5.4 (50.7–52.9) 31.89 5.2 (30.8–32.9) 6859 9 3086 (6240–7477)
53.4 95.1 (52.1–54.7) 33.1 9 6.4 (31.4–34.7) 6350 9 3062 (5572–7128)
Note: values are means 9 S.D. (95% CI). No significant differences between males and females for age, BMI, or steps/day.
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
195
Table 2 Comparison of variables between categories of perceived health status Variable
Age BMI Steps/day
Category of perceived health status Poor (n =7)
Fair (n=39)
Good (n =105)
Excellent (n =9)
53.1 93.2 (50.2–56.1) 33.39 3.8 (29.7–36.8) 5499 91781 (3852–7146)
51.3 95.6 (49.5–53.1) 32.7 95.6 (30.9–34.5) 61699 2713 (5290–7049)
52.6 9 5.4 (51.5–53.6) 32.5 95.9 (31.3–33.6) 6740 9 3275 (6107–7374)
55.2 9 4.3 (51.9–58.5) 27.7 94.2 (24.5–30.9) 8781 91985 (7256–10307)
Note: values are means 9S.D. (95%CI). No significant differences between categories for age, BMI, or steps/day. Table 3 Comparison of variables between categories of exercise frequency Variable
Age BMI Steps/day
Category of exercise frequency (number of times per week) Rarely or never (n = 56)
Once or twice (n = 54)
At least three times (n = 50)
52.2 9 5.1 (50.9–53.6) 32.09 3.9 (31.0–33.1) 63049 2971 (5509–7100)
52.7 95.3 (51.2–54.1) 32.4 95.7 (30.8–33.9) 6594 9 3046 (5763–7456)
52.4 9 5.7 (50.8–54.0) 32.5 9 7.3 (30.4–34.6) 7135 93226 (6218–8052)
Note: values are means 9S.D. (95%CI). No significant differences between categories for age, BMI, or steps/day.
poor (7 or 4.4%), fair (39 or 24.4%), good (105 or 65.6%), or excellent (9 or 5.6%). Reported frequency of exercise, over a typical 7-day period, sufficiently prolonged and intense to cause sweating and rapid heart beat, was reported as rarely or never (56 or 35.0%), normally once or twice (54 or 33.8%), and at least three times (50 or 31.3%). There were no significant differences between perceived health (Table 2) or reported exercise frequency (Table 3) categories for age or BMI. Although mean steps/day increased with both increased perceived health (Table 2) and higher reported frequency of exercise (Table 3), the differences were not statistically significant (F = 2.16, P= 0.095; and F= 0.98, P =0.38, respectively) due to a noticeable overlap in 95% confidence intervals for the mean. According to the collapsed 3-stage activity categorization based on stages of change, 41 (25.6%) of the sample were inactive (5 precontemplation, 25 contemplation, 11 relapse) and took an average of 59139 2962 (CI 4977– 6848) steps/day. Fifty-four (33.8%) were irregularly active (preparation), and averaged 66679 2890 (CI 5,878– 7456) steps/day. Fifty-eight (36.3%) were regularly active (26 action, 32 maintenance) with an aver-
age of 72799 3300 (CI 6411– 8147). As in the case of perceived health and reported exercise frequency, stages of change for exercise, compressed to three stages to reflect reported behavior, did not differ significantly across stages for steps/day (F= 2.38, P= 0.096), despite increasing mean steps/day. Pedometer-determined tertiles of activity were defined as low (below the 25th percentile of distribution, or 5 4716 steps/day) moderate (between the 25 and 75th percentiles of distribution, or 4717–8644 steps/day) and high activity (above the 75th percentile of distribution or ] 8645 steps/ day). Table 4 displays the comparison of age and BMI for the low, moderate, and high categories of pedometer-determined ambulatory activity (steps/ day). There was no significant difference in age across pedometer-determined ambulatory activity groups. BMI was significantly different across pedometer-determined ambulatory activity tertiles (F= 6.85, PB 0.01). Individuals falling in the high activity category had a significantly lower BMI than those in either the moderate or low categories; the latter groups did not differ significantly.
196
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
Analyzed as continuous variables, steps/day and BMI were inversely and significantly correlated (r = − 0.27, P B0.01). Fig. 1 presents a visual depiction of the line of best fit against the distribution based on regression analysis. Table 5 presents the descriptive data for age and steps/day classified according to BMI category. Age did not differ significantly across categories. There was a significant difference between BMI categories with regard to steps/day (F = 2.96, P B 0.05). The difference was most apparent between the highest obesity classes (II and III) and normal weight
categories, although only eight individuals were classified as normal weight in this sample.
4. Discussion This is the first study to describe ambulatory activity, measured using a simple and inexpensive pedometer, in free-living individuals with type 2 diabetes. The pedometer values recorded herein (66629 3080 steps/day) are slightly lower in comparison to two other similarly-aged samples with-
Table 4 Comparison of variables between categories of pedometer-determined ambulatory activity Variable
Age BMI
Categories of pedometer-determined ambulatory activity Low (n= 40)54716 steps/day
Moderate (n = 80)4717–8644 steps/day
High (n = 40)]8645 steps/day
53.1 95.5a (51.4–54.9) 34.3 95.4a (32.6–36.0)
51.8 9 5.2a (50.6–53.0) 32.695.7a (31.3–33.8)
53.0 95.3a (51.3–54.7) 29.8 95.4b (28.1–31.5)
Note: values are means 9S.D. (95% CI). Categories of Digiwalker pedometer-assessed physical activity determined using percentiles of distribution: low 525th, \25th moderateB75th, high ]75th. Means with different letter superscripts are significantly different PB0.01.
Fig. 1. Regression of average steps/day versus body mass index (BMI) in individuals with type 2 diabetes.
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
197
Table 5 Comparison of variables between categories of BMI-defined body composition for total sample Variable
Age Steps/day
Categories of BMI-defined body composition Normal Weight (n =8)
Overweight (n =56)
Obesity Class I (n = 55)
Obesity Class II (n = 27)
Obesity Class III (n =14)
55.6 95.4a (51.1–60.1) 8607 9 3508a (5674-11539)
52.7 9 5.2a (51.3–54.1) 7423 9 2691a,b (6702-8144)
52.3 95.1a (51.0–53.7) 6309 93,083a,b (5475-7142)
52.3 9 5.6a (50.1–54.5) 5621 93223b (4346–6896)
50.1 9 5.9a (46.7–53.5) 5899 93175b (4066–7732)
Note: values are means 9 S.D. (95% CI). Categories of BMI are defined in kg/m2 as Normal Weight 18.5–24.9, Overweight 25.0–29.9, Obesity Class I 30.0–34.9, Obesity class II 35.0–39.9, Obesity Class III]40. No subjects were classified as underweight (BMIB18.5). Means with different letter superscripts are significantly different PB0.01.
out reported diabetes. Tudor-Locke et al. [21] reported values of 737093080 steps/day in a bi-ethnic sample of 109 adults, mean age 44.99 15.8 years. McClung et al. [22] reported 77819 2807 steps/day in a subsample of 58 healthy adults aged 22–82 years. The values for individuals with type 2 diabetes in the present study are, however, higher than those reported for individuals with claudication (approximately 4100– 5300 steps/day) [23], joint replacements (approximately 3500– 6000 steps/day) [22,24], chronic obstructive pulmonary disease (approximately 3800 steps/day) [25], and chronic heart failure (approximately 3500– 3700 steps/day) [26–28]. The values in the present study of free-living individuals also exceed those reported by Yamanouchi et al. [12] for hospital-based individuals with type 2 diabetes serving as control subjects (approximately 4500 steps/day) in a pedometer-based intervention study. Interestingly, the treatment group in that study averaged 19200 steps/day over 6–8 weeks. The inverse relationship found between pedometer-determined steps/day and BMI is consistently reported in other nondiabetic samples of adults [21,22] and children [29]. Although McClung et al. [22] did not report a correlation coefficient, the value we computed (r= − 0.27) is on the order of that reported by Tudor-Locke et al. [21] (r = − 0.30) and Rowlands et al. [29] (r = −0.42). A simple question querying perceived health status failed to differentiate physical activity quantified as steps/day. Similarly, querying about frequency of exercise participation in the manner
employed failed to identify individual’s true activity levels. Although the simple and inexpensive pedometer is admittedly not the best measure of overall physical activity, it appears to be useful for objectively describing daily incidental and intentional ambulatory activity. Although pedometers are not sensitive to intensity of activity, we know that individuals with type 2 diabetes do not typically participate in vigorous activities and that the most common activity choice is walking, or ambulatory activity [1]. Although it has been suggested that a combination of pedometer and questionnaire methods may be able to minimize collective weaknesses [30], that postulation has yet to be explored empirically. The 3-stage activity categorization has been previously used to infer level of physical activity from reported intention. It has been validated only against participation in vigorous leisure time physical activity behavior [20] using a commonly employed self-reported measure of physical activity, namely, the 7-day physical activity recall questionnaire [31]. A difference was not evident between categories for participation in moderate intensity activities in that study [20]. Using an objective and direct measure of ambulatory activity (i.e. pedometers) we found no difference in steps/day between stages of change categories. Our data suggests that stages of change is not a good indicator of current physical activity, at least in individuals with type 2 diabetes, and should, therefore, not be used as a physical activity screening tool or as an outcome measure of change due to intervention.
198
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
The cross-sectional design of this study limits the ability to make firm causal conclusions. Further, the study sample is a convenience sample of volunteers recruited according to strict eligibility criteria (e.g. no insulin, no physical activity limitations/contraindications) and, therefore limits generalizability of the findings. We do not know how, or if, individuals who did not participate (e.g. not eligible or not interested) differ from the sample studied. The 3-day monitoring frame may also be considered a study design weakness; it has been previously reported that 5– 6 days of pedometer data are required to accurately describe physical activity with less than 5% error [32]. That study was conducted with young males purposefully recruited to represent a wide range of physical activity levels. Relevant to the present study, Sieminski et al. [33] reported an intraclass correlation statistic of 0.86 for pedometer-determined ambulatory activity averaged over 2 days (determined 1 week apart) in older, free-living claudication patients. Similarly, we found evidence of high repeatability during pilot work using the 3-day pedometer assessment protocol as described earlier. It appears that a shorter monitoring frame is sufficient in samples that exhibit relatively stable behavior (i.e. individuals living with chronic conditions, including type 2 diabetes). In summary, this study presents the first normative data on pedometer-determined ambulatory activity in individuals with type 2 diabetes. This data is useful for sample comparison purposes. In addition the standard deviation or variance estimates can be used to calculate samples sizes for intervention efforts. A primary finding was a distinct and consistent inverse relationship between steps/day and BMI, an indicator of body composition. Another lesson learned was that questions about frequency of exercise, perceived health, or stages of change were not able to capture objective differences in ambulatory activity in this population. Objective quantification of ambulatory activity via simple and inexpensive pedometers permits researchers and practitioners to easily screen for level of activity along a continuum. This study opens the door for future research and clinical applications including identifying threshold values
related to important health outcomes and evaluating incremental change due to various interventions in this population.
Acknowledgements We would like to acknowledge the financial support of the Canadian Diabetes Association. We would also like to acknowledge the assistance of the staff and educators at the Lawson Diabetes Center, St. Joseph’s Health Center, London, Ontario, Canada.
Appendix A Over a typical 7-day period (1 week), how many times do you engage in physical activity that is sufficiently prolonged and intense to cause sweating and a rapid heart beat? (Check one). At least three times. Normally once or twice. Rarely or never. In general, how would you describe your current state of health? Excellent. Good. Fair. Poor.
References [1] E.S. Ford, W.H. Herman, Leisure-time physical activity patterns in the U.S. diabetic population. Findings from the National Health Interview Survey – Health Promotion and Disease Prevention Supplement, Diabetes Care 18 (1995) (1990) 27 – 33. [2] A.M. Kriska, W.C. Knowler, R.E. LaPorte, et al., Development of questionnaire to examine relationship of physical activity and diabetes in Pima Indians, Diabetes Care 13 (1990) 401 – 411. [3] B.E. Ainsworth, A.S. Leon, M.T. Richardson, D.R. Jacobs, R.S. Paffenbarger, Accuracy of the college alumnus physical activity questionnaire, J. Clin. Epidemiol. 46 (1993) 1403 – 1411. [4] M.T. Richardson, A.S. Leon, D.R. Jacobs, B.E. Ainsworth, R. Serfass, Comprehensive evaluation of the
C.E. Tudor-Locke et al. / Diabetes Research and Clinical Practice 55 (2002) 191–199
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Minnesota Leisure Time Physical Activity Questionnaire, J. Clin. Epidemiol. 47 (1994) 271 –281. C.E. Tudor-Locke, A.M. Myers, Challenges and opportunities for measuring physical activity in sedentary adults, Sports Med. 31 (2001) 91 –100. L.M. Hays, D.O. Clark, Correlates of physical activity in a sample of older adults with type 2 diabetes, Diabetes Care 22 (1999) 706. C.E. Tudor-Locke, A.M. Myers, Methodological considerations for researchers and practitioners using pedometers to measure physical (ambulatory) activity, Res. Q Exerc. Sport 72 (2001) 1 –12. A.V. Rowlands, R.G. Eston, E.K. Ingledew, Measurement of physical activity in children with particular reference to the use of heart rate and pedometry, Sports Med. 24 (1997) 258 – 272. G.J. Welk, C.B. Corbin, D. Dale, Measurement issues in the assessment of physical activity in children, Res. Q Exerc. Sport 71 (2000) 59 –73. P.S. Freedson, K. Miller, Objective monitoring of physical activity using motion sensors and heart rate, Res. Q Exerc. Sport 71 (2000) 21 –29. D.R. Bassett, Validity and reliability issues in objective monitoring of physical activity, Res. Q Exerc. Sport 71 (2000) 30 – 36. K. Yamanouchi, T. Takashi, K. Chikada, et al., Daily walking combined with diet therapy is a useful means for obese NIDDM patients not only to reduce body weight but also to improve insulin sensitivity, Diabetes Care 18 (1995) 775 – 778. C. Tudor-Locke, A.M. Myers, N.W. Rodger, Formative evaluation of The First Step Program: A practical intervention to increase daily physical activity, Can. J. Diabetes Care 24 (2000) 34 –38. S. Meltzer, L. Leiter, D. Daneman, et al., Clinical practice guidelines for the management of diabetes mellitus in Canada. Canadian Diabetes Association, CMAJ 159 (1998) S1 – 29. Canadian Society for Exercise Physiology, The Canadian Physical Activity, Fitness and Lifestyle Appraisal. Ottawa: The Canadian Society for Exercise Physiology; 1998. B.H. Marcus, V.C. Selby, R.S. Niaura, J.S. Rossi, Selfefficacy and the stages of exercise behavior change, Res. Q Exerc. Sport 63 (1992) 60 –66. B.H. Marcus, L.R. Simkin, The transtheoretical model: applications to exercise behavior, Med. Sci. Sports Exerc. 26 (1994) 1400 – 1404. M.M. Sequeira, M. Rickenbach, V. Wietlisbach, B. Tullen, Y. Schutz, Physical activity assessment using a pedometer and its comparison with a questionnaire in a large population survey, Am. J. Epidemiol. 142 (1995) 989 – 999. National Institutes of Health, Clinical guidelines on the identification, evaluation, and treatment of overweight
[20] [21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
199
and obesity in adults —The evidence report., Obesity Res. 6 (1998) 51S – 209S. B.H. Marcus, L.R. Simkin, The stages of exercise behavior, J. Sports Med. Phys. Fitness 33 (1993) 83 – 88. C. Tudor-Locke, B.E. Ainsworth, M.C. Whitt, R.W. Thompson, C.L. Addy, D.A. Jones, The relationship between pedometer-determined ambulatory activity and body composition variables, Int. J. Obes. 25 (2001) 1571 – 1578. C.D. McClung, C.A. Zahiri, J.K. Higa, H.C. Amstutz, T.P. Schmalzried, Relationship between Body Mass Index and activity in hip or knee arthorplasty patients, J. Ortho. Res. 18 (2000) 35 – 39. A.W. Gardner, D.J. Sieminski, L.A. Killewich, The effect of cigarette smoking on free-living daily physical activity in older claudication patients, Angiology 48 (1997) 947 – 955. T.P. Schmalzried, E.S. Szuszczewicz, M.R. Northfield, et al., Quantitative assessment of walking activity after total hip or knee replacement, J. Bone Joint Surg. 80-A (1998) 54 – 59. B. Schonhofer, P. Ardes, M. Geibel, D. Kohler, P.W. Jones, Evaluation of a movement detector to measure daily activity in patients with chronic lung disease, Eur. Respir. J. 10 (1997) 2814 – 2819. D.J. Hoodless, K. Stainer, N. Savic, P. Batin, M. Hawkins, A.J. Cowley, Reduced customary activity in chronic heart failure: Assessment with a new shoemounted pedometer, Int. J. Cardiol. 43 (1994) 39 – 42. J.T. Walsh, A. Charlesworth, R. Andrews, M. Hawkins, A.J. Cowley, Relation of daily activity levels in patients with chronic heart failure to long-term prognosis, Am. J. Cardiol. 79 (1997) 1364 – 1369. A.J. Cowley, L. Fullwood, K. Stainer, J.R. Hampton, Exercise tolerance in patients with heart failure--how should it be measured?, Eur. Heart J. 12 (1991) 50 – 54. A.V. Rowlands, R.G. Eston, D.K. Ingledew, Relationship between activity levels, aerobic fitness, and body fat in 8to 10-year-old children, J. App. Physiol. 86 (1999) 1428 – 1435. A. Kriska, Physical activity and the prevention of type 2 diabetes mellitus: How much for how long?, Sports Med. 29 (2000) 147 – 151. S.N. Blair, W.L. Haskell, P. Ho, et al., Assessment of habitual physical activity by a seven-day recall in a community survey and controlled experiments, Am. J. Epidemiol. 122 (1985) 794 – 804. R. Washburn, M.K. Chin, H.J. Montoye, Accuracy of pedometer in walking and running, Res. Q Exerc. Sport 51 (1980) 695 – 702. D.J. Sieminski, L.L. Cowell, P.S. Montgomery, S.B. Pillai, A.W. Gardner, Physical activity monitoring in patients with peripheral arterial occlusive disease, J. Cardpulm. Rehabil. 17 (1997) 43 – 47.