Change in physical activity after a diabetes diagnosis: opportunity for intervention

Abstract

Introduction

Moderate intensity physical activity is recommended for individuals with diabetes to control glucose and prevent diabetes-related complications. The extent to which a diabetes diagnosis motivates patients to increase physical activity is unclear. This study used data from the Women’s Health Initiative Observational Study (baseline data collected from 1993-1998) to examine change in physical activity and sedentary behavior in women who reported a diabetes diagnosis compared to women who did not report diabetes over 7 years of follow-up (up to 2005).

Methods

Participants (n=84,300) were post-menopausal women who did not report diabetes at baseline [mean age=63.49; standard deviation (SD)=7.34; mean BMI=26.98 kg/m²; SD=5.67]. Linear mixed model analyses were conducted adjusting for study year, age, race/ethnicity, BMI, education, family history of diabetes, physical functioning, pain, energy/fatigue, social functioning, depression, number of chronic diseases and vigorous exercise at age 18. Analyses were completed in August 2012.

Results

Participants who reported a diabetes diagnosis during follow-up were more likely to report increasing their total physical activity (p=0.002), walking (p<0.001) and number of physical activity episodes (p<0.001) compared to participants who did not report a diabetes diagnosis. On average, participants reporting a diabetes diagnosis reported increasing their total physical activity by 0.49 MET-hours/week, their walking by 0.033 MET-hours/week and their number of physical activity episodes by 0.19 MET-hours/week. No differences in reported sedentary behavior change were observed (p=0.48).

Conclusion

A diabetes diagnosis may prompt patients to increase physical activity. Healthcare professionals should consider how best to capitalize on this opportunity to encourage increased physical activity and maintenance.

Introduction

Paragraph Number 1

A disease diagnosis may motivate health behavior changes in smoking cessation,(28) weight loss,(14, 28) reducing fat intake,(9) completing regular check-ups and improving diet.(18) Most studies of health behavior change following disease diagnoses focus on cancer [e.g. (15, 18)]. Few studies examine behavior change following a diagnosis of type 2 diabetes despite the fact that encouraging lifestyle changes are recommended for diabetes management.(2, 17)

Paragraph Number 2

Following a type 2 diabetes diagnosis, recommendations include achieving at least 150 minutes per week of moderate intensity aerobic activity and, if indicated, resistance training twice a week.(2) Physical activity is recommended to assist with glucose control, weight loss and to prevent diabetes-related complications, but whether patients with recently diagnosed diabetes change their physical activity in response to the diagnosis is unclear. One study of individuals diagnosed with type 2 diabetes reported that a shorter time since diagnosis was associated with greater increases in physical activity over six months.(42) As participants were diagnosed prior to inclusion in the study, it is unknown whether their physical activity changed as a result of the diabetes diagnosis or whether their greater level of activity reflected a previous habit. A large, prospective study of middle-aged and older adults reported that among those with higher education levels, more were engaging in physical activity following a diabetes diagnosis, while those with lower education levels were less likely to engage in physical activity after diagnosis.(38)

Paragraph Number 3

Physical activity recommendations include encouragement to reduce sedentary behavior.(17) Evidence has accumulated on the dangers of excessive sedentary behavior, regardless of physical activity level. Sedentary behavior has been associated with type 2 diabetes,(24) obesity(20, 24, 45) and mortality,(39) but it is unknown if a diabetes diagnosis prompts changes in sedentary behavior.

Paragraph Number 4

The present study aims to expand upon previous research by using the Women’s Health Initiative data to examine the relationship between change in physical activity and sedentary behavior for women reporting a new onset of diabetes compared to those not reporting diabetes. This prospective study reduces the bias often observed in studies of post-diagnosis behavior as physical activity data was collected annually for 9 years and not based on retrospective report. We hypothesize that women who report a diabetes diagnosis during follow-up will report increased physical activity and reduced sedentary behavior, compared to women who do not report a diabetes diagnosis. Additionally, we hypothesize that higher education and a family history of diabetes will be associated with greater increases in physical activity and greater reductions in sedentary behavior in women recently diagnosed with diabetes. We will further explore whether race/ethnicity modify the relationship between diabetes diagnosis and change in physical activity and sedentary behavior.

Methods

Procedures

Paragraph Number 5

The present study is a secondary data analysis within the Women’s Health Initiative (WHI). The WHI was a national, longitudinal study that enrolled 68,132 participants into three clinical trials and enrolled 93,676 participants into an observational study (WHI-OS) at 40 clinical centers across the U.S. from 1993-1998. The present study used data from the WHI-OS participants from baseline (1993-1998) and periodic follow-up assessments up to 2005. The design of the WHI has been described elsewhere.(32) New diabetes diagnosis, physical activity and sedentary behavior were collected repeatedly over approximately 7 years of follow-up and participant retention and data collection completion rates were >95%. The WHIOS study protocol was approved by the IRBs of WHI institutions and participants provided written informed consent. The University of Massachusetts Medical School IRB granted an exemption for the current study.

Participants

Paragraph Number 6

WHI-OS eligibility criteria included: postmenopausal woman, aged 50 to 79 years, reliable/mentally competent, and expected survival and local residency for at least 3 years. Exclusion criteria included current alcoholism, drug dependency and dementia, or other conditions that would limit full participation in the study. Since the present study aimed to examine change in physical activity and sedentary behavior after a diabetes diagnosis, participants with a history of diabetes diagnosis at baseline (n=3,902) were excluded from the analysis. In addition we excluded 5,474 participants who were missing baseline information on diabetes history (n=121), self-reported diabetes diagnosis during follow-up (n=602), physical activity at baseline (n=979) or all physical activity data during follow-up (n=3,772). The analytic sample included 84,300 participants, 90% of the WHI-OS sample.

Measures

Paragraph Number 7

Diabetes diagnosis

Diabetes at baseline was defined as a self-report of “ever having received a physician diagnosis of diabetes when not pregnant.” At each annual follow-up visit (years 1-9), participants reported whether a doctor prescribed pills or insulin shots for diabetes. Although self-reported diabetes cases were not adjudicated, self-reported diabetes was found to be a valid indicator of diagnosed diabetes in the WHI, as assessed using medication and laboratory data, though undiagnosed diabetes was likely missed.(35) This method was used in other WHI studies.(34, 50)

Paragraph Number 8

Physical activity

Physical activity during the past week was measured via a self-report questionnaire that demonstrated concurrent validity and test-retest reliability.(13, 36) For baseline and years 3-8, participants indicated the number of times per week and usual duration for walking, and for activities of mild intensity (e.g., slow dancing, bowling), moderate intensity (e.g., biking outdoors, using a treadmill) and vigorous intensity (e.g., aerobics, swimming laps). Frequency and duration data were converted to weekly energy expenditure (MET-hours/week) during walking, mild intensity exercise, moderate intensity exercise and vigorous intensity exercise. Total MET-hours/week was calculated by summing the MET-hours/week for walking, mild, moderate and vigorous intensity activity. The number of physical activity episodes that were 20 minutes or longer was also collected. All of these variables were examined as outcomes to examine whether physical activity change could be attributed to a particular type or intensity of activity. Annual change in each type of physical activity was calculated using difference scores (current year minus previous year).

Paragraph Number 9

Sedentary behavior

Sedentary behavior was assessed using the item for hours of sitting time which asked participants, “During a usual day and night, about how many hours do you spend sitting? Be sure to include the time you spend sitting at work, sitting at the table eating, driving or riding in a car or bus, and sitting up watching TV or talking.” This item demonstrated adequate test-retest reliability.(36) Responses were categorical and ranged from “less than 4 hours” to “16 or more hours.” The midpoint of each category was used to compute hours of sedentary behavior and used in statistical analyses. Participants answered this item at baseline, year 3 and year 6. Thus, change data was computed using difference scores at year 3 (year 3 minus baseline hours) and at year 6 (year 6 minus year 3).

Paragraph Number 10

Covariates

Several baseline variables were included as covariates for their association with physical activity (6, 16, 22, 25, 26, 30, 31, 48, 49) or behavior change following a disease diagnosis.(5, 18, 33) Age, race/ethnicity, education, family history of diabetes and vigorous exercise at age 18 were self-reported. WHI-certified clinic staff measured height using a stadiometer and weight using a balance-beam scale to calculate BMI (kg/m²). Four subscales from the RAND Health Survey (SF-36; 52) pertaining to physical functioning, pain, energy/fatigue and social functioning were included as covariates. Depression symptoms were measured using the Center for Epidemiological Studies Depression Scale (CES D) 6-item short form.(53) Number of chronic diseases, defined by Seguin (44) was defined as number, up to 5, of the following: cardiovascular disease ever (myocardial infarction, angina, coronary artery bypass graft or percutaneous transluminal coronary angioplasty), hypertension (treated or high blood pressure defined as systolic ≥140 or diastolic ≥ 90 mmHg), congestive heart failure ever, stroke ever, cancer ever, treated diabetes, arthritis ever, falls (2 or more falls in the last 12 months was coded as a chronic disease), emphysema and hip fracture age 55 years or older. Participants with previous diabetes diagnoses were excluded from this study, thus diabetes was not a part of the number of chronic diseases calculation for the current study’s participants. Lastly, study year was included in the model to examine change in physical activity after a diabetes diagnosis from years 3 through 8.

Analytic plan

Paragraph Number 11

This study compared participants with and without a diabetes diagnosis on change in physical activity and in sedentary behavior during 6 years of follow-up assessments. Change in physical activity was computed as current reported value minus previous reported value for each individual at each scheduled follow-up clinic visit (years 3, 4, 5, 6, 7, 8). Thus, among participants who self-reported a diabetes diagnosis, change in physical activity was calculated from before, to in close proximity (i.e., within 1 year) of reporting a diabetes diagnosis. For participants with missing physical activity data at the previous timepoint, the year prior to that was selected to capture change, but only up to 1 year earlier. For instance, if a participant was missing year 4, then change at year 5 was calculated as year 5 minus year 3; however change at year 4 was still regarded as missing. Analyses were conducted with and without using the previous timepoint for missing cases to ensure consistent results. Change in sedentary behavior was computed as current reported value minus previous for each individual at each scheduled follow-up (years 3 and 6). Participants with missing sedentary activity data were omitted from these analyses (n=7,352).

Paragraph Number 12

Linear mixed models were conducted using PROC MIXED in SAS version 9.2. (SAS Institute Inc., Cary, NC) with a repeated statement for subject and specified a first-order Toeplitz covariance structure. As the response variables were difference scores between 2 timepoints, correlations among the variables within an individual were small. More complicated covariance structures gave similar results. Three separate models were run for each outcome. The first model was the unadjusted model, the second model was adjusted for baseline BMI only (to minimize missing data from other covariates) and the third model controlled for the remaining covariates. Physical activity was not included as a covariate when sedentary behavior was the outcome and sedentary behavior was not a covariate when physical activity was the outcome. Analyses were completed in August 2012.

Paragraph Number 13

For the effect modification analyses, interaction terms were created with the diabetes diagnosis variable and each of the moderator variables: education, family history of diabetes and race/ethnicity. These terms were then entered into the model that controlled for all covariates.

Preliminary analyses

Paragraph Number 14

Continuous variables were assessed for normality by examining skew and kurtosis. Physical activity and sedentary behavior change values were fairly symmetric, but ‘no change’ was reported more often than expected for a normal distribution. The CES-D and social functioning variables demonstrated significant skew and were coded as categorical variables using standard cutpoints [0.06 for the CES-D; (53) and 50 for the social functioning subscale(19)]. Since the timing of assessments is critical to this analysis, study year was compared with a variable specifying the number of days since enrollment to completion of the yearly assessments to ensure that assessments occurred in a logical sequence.

Results

Paragraph Number 15

The sample consisted of 84,300 postmenopausal women whose average age=63.5 years (SD=7.3), average BMI was 27.0 kg/m² (SD=5.7) and 15% were from ethnically/racially diverse groups. Participant characteristics, split by whether they reported diabetes during follow-up, are described in Table 1. Of the participants, 5% reported being diagnosed with diabetes during follow-up up. The number of participants who reported being diagnosed with diabetes in a given year ranged from 407 in year 8 to 1,521 in year 3 (year 3 includes diagnoses that occurred in years 1-3). Incidence ranged from 5.7 (per 1000) in year 8 to 7.8 in year 5, which is lower than population estimates (in 2010, women aged 45-64 had a crude incidence=11.5 per 1000 and women aged 65-79 had a crude incidence=11.9 per 1000).(7) Participants who were diagnosed with diabetes during follow-up had, on average, a higher BMI at baseline (p<0.001), lower educational level (p<0.001) and were less likely to be non-Hispanic White (p<0.001), compared to participants who did not report diabetes during follow-up. Participants who reported diabetes during follow-up also reported less total physical activity at baseline (e.g., difference in total physical activity was −4.1 MET-hours/week; 10.1 vs 14.2 MET-hours/week) and greater sedentary behavior (p<0.001).

Table 1.

Baseline participant characteristics in the WHI Observational Study by diabetes diagnosis status (N=84,300).

	Participants who develop diabetes in follow up
	Yes (n=4,385)	No (n=79,915)
	Mean±SD or n (%)	Mean±SD or n (%)	p-value
Age at screening (years)	63.6±7.2	63.5±7.4	0.264
Education (years)	14.1±2.9	14.9±2.6	<0.001
< High school	375(8.6)	3,269(4.1)	<0.001
High school diploma	2,595(59.8)	41,156(51.9)
College graduate	1,373(31.6)	34,866(44.0)
Race/ethnicity
American Indian or Alaskan Native	41(0.94)	268(0.34)	<0.001
Asian or Pacific Islander	158(3.6)	2,211(2.8)
Black or African-American	631(14.4)	5,247(6.6)
Hispanic/Latino	268(6.1)	2,575(3.2)
White (not of Hispanic origin)	3,213(73.5)	68,540(86.0)
Other	59(1.4)	867(1.1)
Body Mass Index (kg/m2)	31.5±6.8	26.7±5.5	<0.001
Physical Functioning Score	73.2±23.1	82.8±19.1	<0.001
Pain Score	68.5±26.0	75.4±22.9	<0.001
Energy/Fatigue Score	58.1±20.7	64.7±19.0	<0.001
Social Functioning Score	85.7±20.8	90.2±17.6	<0.001
<50	252 (5.8)	2,650 (3.3)	<0.001
≥50	4,089 (94.2)	76,639 (96.7)
CES-Da	0.06±0.16	0.04±0.13	<0.001
<0.06	3,639 (85.8)	69,900 (89.4)	<0.001
≥0.06	604 (14.2)	8,305 (10.6)
Vigorous exercise at 18 years
No	2,083 (49.7)	43,189 (56.0)	<0.001
Yes	2,112 (50.4)	33,913 (44.0)
Chronic disease history (number)	1.6±1.1	1.2±1.0	<0.001
Relative had adult diabetes
No	1,926(44.2)	53,147(66.7)	<0.001
Yes	2,200(50.4)	23,141(29.1)
Don’t know	236(5.4)	3,371(4.2)
Walking MET-hours per week	3.9±5.5	5.3±6.3	<0.001
Mild MET-hours per week	1.1±2.6	1.4±3.2	<0.001
Moderate MET-hours per week	2.7±4.8	3.4±5.5	<0.001
Vigorous MET-hours per week	2.5±6.5	4.1±8.7	<0.001
Total MET-hours per week	10.1±11.9	14.2±14.5	<0.001
Frequency of recreational physical activity per week >= 20 minutes	3.3±3.6	4.4±4.0	<0.001
Number of hours spent sitting (hours)	7.6±3.4	7.3±3.3	<0.001

^aCES-D= Center for Epidemiological Studies Depression Scale.

Paragraph Number 16

Results from mixed models revealed significant associations between reported diabetes diagnosis and physical activity change (Table 2). Participants who reported a diabetes diagnosis were more likely to report increased total weekly physical activity (β=0.61, SE=0.18, p<0.001 for crude model), walking (β=0.26, SE=0.07, p<0.001 for crude model), and physical activity episodes (β=0.23, SE=0.05, p<0.001 for crude model). Results remained significant after adjusting for BMI (all p-values<0.05) and other covariates (all p-values<0.05). Results also held when limited to those with complete data for computing changes (all p-values<0.01). On average, participants reporting a diabetes diagnosis increased their total physical activity by 0.49 MET-hours/week (SD=10.9), their walking by 0.033 MET-hours/week (SD=4.6) and their number of physical activity episodes by 0.19 MET-hours/week (SD=3.5). For participants who did not report diabetes, mean change was minimal and decreased for total physical activity (M=−0.13 MET-hours/week, SD=11.1), walking (M=−0.23 MET-hours/week; SD=4.7) and number of physical activity episodes (M=−0.041 MET-hours/week, SD=3.4). Table 3 displays yearly change in total physical activity, walking and number of physical activity episodes.

Table 2.

Change in physical activity^a for participants with reported diabetes compared to those without diabetes.

	Crude N=84,300			BMI adjusted N=83,324			Full modelc N=76,020
	Beta	SEb	p-value	Beta	SE	p-value	Beta	SE	p-value
Walking MET hours/week	0.259	0.074	<0.001	0.230	0.075	0.002	0.263	0.078	<0.001
Mild MET hours/week	0.025	0.056	0.660	0.027	0.057	0.632	0.027	0.060	0.649
Moderate MET hours/week	0.091	0.090	0.312	0.071	0.091	0.433	0.095	0.096	0.321
Vigorous MET hours/week	0.239	0.118	0.043	0.189	0.119	0.113	0.200	0.125	0.110
Total MET hours/week	0.613	0.176	<0.001	0.517	0.178	0.004	0.585	0.186	0.002
Episodes ≥20 min /week	0.234	0.054	<0.001	0.217	0.054	<0.001	0.260	0.057	<0.001

^aComputing change from earlier time point for those missing data (up to 1 year earlier, for instance if missing year 4 then change at year 5 is year 5 minus year 3)

^bSE- standard error

^cFull model included year, age, ethnicity, BMI, education (years), family history of diabetes, physical functioning (SF-36), pain (SF-36), energy/fatigue (SF-36), social functioning (cut at 50), depression (CES-D, cut at 0.06), number of chronic diseases and strenuous/hard exercise at age 18 years.

Table 3.

Physical activity change and sedentary behavior change for those reporting a diabetes diagnosis and those not, by year.

Physical Activity		No diabetes reported during year Mean change (SD)	Diabetes reported during year Mean change (SD)
Walking MET hours/week	Year 3	−0.48 (5.55)	−0.24 (5.17)
	Year 4	−0.25 (4.55)	0.41 (4.18)
	Year 5	−0.15 (4.41)	0.05 (4.16)
	Year 6	−0.09 (4.38)	0.26 (4.02)
	Year 7	−0.19 (4.29)	0.18 (3.87)
	Year 8	−0.14 (4.32)	0.19 (4.13)
Mild MET hours/week	Year 3	0.10 (3.51)	0.14 (2.96)
	Year 4	0.10 (3.55)	0.14 (2.73)
	Year 5	−0.03 (3.53)	0.13 (3.29)
	Year 6	−0.17 (3.51)	−0.22 (3.13)
	Year 7	0.24 (3.59)	0.22 (3.35)
	Year 8	−0.05 (3.64)	−0.35 (3.11)
Moderate MET hours/week	Year 3	0.12 (6.11)	0.09 (5.34)
	Year 4	−0.15 (5.65)	0.07 (5.39)
	Year 5	0.01 (5.48)	0.13 (4.62)
	Year 6	−0.14 (5.54)	−0.08 (5.49)
	Year 7	0.27 (5.58)	0.37 (4.88)
	Year 8	0.02 (5.72)	0.16 (4.93)
Vigorous MET hours/week	Year 3	−0.09 (8.12)	0.17 (6.71)
	Year 4	−0.02 (7.15)	0.43 (6.53)
	Year 5	−0.05 (7.07)	0.09 (6.31)
	Year 6	0.26 (7.28)	0.52 (6.5)
	Year 7	0.19 (7.4)	0.14 (7.46)
	Year 8	0.05 (7.46)	0.41 (6.81)
Total MET hours/week	Year 3	−0.35 (12.5)	0.17 (11.5)
	Year 4	−0.31 (10.8)	1.06 (10.9)
	Year 5	−0.22 (10.3)	0.41 ( 9.7)
	Year 6	−0.14 (10.7)	0.48 (9.3)
	Year 7	0.50 (10.9)	0.92 (10.5)
	Year 8	−0.13 (11.0)	0.40 (10.9)
Episodes ≥20 min /week	Year 3	−0.07 (3.72)	0.10 (3.52)
	Year 4	−0.09 (3.35)	0.39 (3.49)
	Year 5	−0.05 (3.25)	0.09 (3.42)
	Year 6	−0.09 (3.28)	0.18 (3.42)
	Year 7	0.12 (3.30)	0.40 (3.18)
	Year 8	−0.04 (3.34)	0.16 (3.48)
Sedentary hours/week	Year 3	−0.41 (3.03)	−0.48 (3.34)
	Year 6	−0.07 (2.92)	−0.03 (3.16)

Paragraph Number 17

Diabetes diagnosis had no association with mild (p=0.66) or moderate intensity activity change (p=0.31). Results did not change with adjustment for BMI (all p-values >0.43) or other covariates (all p-values >0.30). Diabetes diagnosis was significantly associated with vigorous intensity physical activity for the crude model (p=0.04), but the relationship abated after controlling for BMI (p=0.11) and other covariates (p=0.11). Diabetes diagnosis also had no impact on change in sedentary behavior for the crude (p=0.86), BMI adjusted (p=0.63) or full model (p=0.48; Table 4).

Table 4.

Change in sedentary hours for participants with reported diabetes compared to those with no diabetes.

Sedentary hours/week	Beta	SEa	p-value
Crude (N=76,948)	−0.010	0.056	0.862
BMI adjusted (N=76,069)	−0.027	0.057	0.633
Full modelb (N=69,763)	−0.043	0.06	0.478

^aSE – standard error

^bFull model containing year, age, ethnicity, BMI, education (years), family history of diabetes, physical functioning (SF-36), pain (SF-36), energy/fatigue (SF-36), social functioning (cut at 50), depression (CES-D, cut at 0.06), number of chronic diseases and strenuous/hard exercise at age 18 years.

Paragraph Number 18

Neither education, family history of disease nor race/ethnicity modified the relationship between diabetes diagnosis and change in any of the physical activity outcomes (all p-values>0.14).

Discussion

Paragraph Number 19

Post-menopausal women who reported a diabetes diagnosis reported significant increases in their physical activity compared to women who did not report a diabetes diagnosis. The increase in physical activity appeared to be driven by a small increase in reported walking and physical activity episodes. Women who reported a diabetes diagnosis reported an increase of 0.5 MET-hours/week in total physical activity. While an increase of this magnitude represents only 6% of the recommended amount of physical activity (i.e., 7.5 MET-hours/week),(2) any increase in recreational physical activity, particularly in those at greater risk for comorbid disease, may confer benefits. A pooled analysis of 6 prospective cohort studies demonstrated that adults who reported 0.1 - 3.74 MET-hours/week had a 32% lower mortality risk and an average of 1.8 years longer life expectancy compared to adults who reported no recreational physical activity.(37) The slight increase observed, coupled with research demonstrating that physical activity tends to decline with age,(51) justifies a need for future research to investigate how a diabetes diagnosis may encourage physical activity and how public health practitioners and clinicians might capitalize on this opportunity to promote activity.

Paragraph Number 20

Results suggest that a diabetes diagnosis may provide some motivation for increasing physical activity in a population that is often quite sedentary.(4) The diabetes diagnosis may have highlighted participants’ worsening health, prompting them to increase their physical activity, without further intervention or guidance, in an attempt to assuage health concerns. They also may have received counseling from a health professional on increasing physical activity.(47)

Paragraph Number 21

Prior to diagnosis, people with type 2 diabetes often have warning signs or receive a ‘pre-diabetes’ diagnosis that could heighten their awareness for the necessity of behavior change. Since physical activity may help prevent diabetes, research should examine whether classifying a patient as pre-diabetic or using an impaired glucose tolerance diagnosis can similarly motivate increased physical activity. A qualitative study of individuals with impaired glucose tolerance who participated in a diabetes prevention lifestyle intervention trial and maintained improvements in lifestyle behaviors, mentioned that their impaired glucose tolerance diagnosis was a strong motivating factor for behavior change.(41) The best strategies for facilitating behavior change during these instances require further study.

Paragraph Number 22

Though the participants who developed diabetes reported increased physical activity, the average amount was still lower than the physical activity of participants who never developed diabetes and the amount recommended for health benefits. The fact that those who developed diabetes continued to lag behind individuals who did not develop diabetes, coupled with the evidence that individuals with diabetes are more likely to relapse to a sedentary lifestyle after beginning an exercise program,(29) raises the question as to how to further increase and maintain physical activity in individuals with diabetes. Access to exercise facilities, social support and preventing pain and obesity are key factors for facilitating physical activity maintenance in individuals with diabetes.(43, 54) The extent to which technology, such as continuous blood glucose monitors and virtual reality, encourages physical activity maintenance appears promising, but requires further study.(1, 10)

Paragraph Number 23

No differences were observed in reported sedentary behavior. Increased physical activity may not necessarily prompt reductions in sedentary behavior.(23) Some individuals compensate for increased physical activity by increasing their sedentary behavior.(27) Participants may also have had little awareness of the importance of reducing sedentary behavior. Although recommendations to reduce sedentary behavior are included in the physical activity recommendations, the extent to which advice to reduce sedentary behavior is provided is unknown. A study of counseling by primary care physicians found that only 10% of patients received counseling to reduce sitting, compared to 53% of patients who received physical activity counseling.(46) Research is just beginning to emerge on the factors that facilitate decreases in sedentary behavior [e.g., point-of-choice prompting software(11)], which could assist individuals with decreasing sedentary activity following a diabetes diagnosis.

Paragraph Number 24

The lack of a moderating effect of education on physical activity change conflicts with the Newsom and colleagues study(38) in which individuals with a higher education level reported increased physical activity following a diabetes diagnosis. That study used data from the Health and Retirement Study, which includes a probability sampling of U.S. adults aged 50 and older with oversampling of Blacks, Hispanics and Florida residents.(21) The WHI-OS sample is a more highly educated sample than the US population. This difference and the inclusion of males in the Health and Retirement study sample may have accounted for the discrepant results.

Paragraph Number 25

The main limitation of this study is that diabetes was self-reported treated diabetes, and was only captured as having occurred in the past year as the WHI does not contain the date of diabetes diagnosis. Thus, participants may have been diagnosed as long as 12 months ago or as little as 1 day before they completed the follow-up assessment. The use of self-reported physical activity and sedentary behavior is not ideal given the possibility of inaccurate reporting. The lack of objective physical activity data limits our ability to validate participants’ self-report. Additionally, sedentary behavior was only measured at baseline, year 3 and year 6, which creates a larger time span between diabetes diagnosis and the post-diagnosis sedentary behavior measurement. The WHI sample also reported little vigorous intensity physical activity,(12) which may have limited the possibility of finding change over time in vigorous intensity physical activity. Participants were less likely to report diabetes than postmenopausal women in the US, which suggests that participants may not be representative of postmenopausal women. Additionally, results may not generalize to men or premenopausal women. Strengths of the study include the large, longitudinal, national, and well-characterized racially and ethnically diverse group of postmenopausal women, the use of multiple measures of physical activity and an examination of sedentary behavior.

Paragraph Number 26

A disease diagnosis may motivate health behavior change and prompt increases in physical activity. Health professionals should be cognizant of this opportunity to discuss healthy behavior change upon diagnosis, as not all patients with diabetes report receiving encouragement from their physicians on increasing physical activity(3, 40) and opportunities to motivate behavior change can be overlooked in primary care settings.(8) Providing physical activity recommendations, relevant resources and evidence-based strategies such as tracking physical activity, scheduling physical activity and eliciting social support for activity, as well as increasing contact and support for patients who have been newly diagnosed with diabetes, may facilitate greater increases in physical activity and support physical activity maintenance.