Can J Diabetes xxx (2014) 1e7
Contents lists available at ScienceDirect
Canadian Journal of Diabetes journal homepage: www.canadianjournalofdiabetes.com
Original Research
Gender Differences in the Relationship between Anxiety Symptoms and Physical Inactivity in a Community-Based Sample of Adults with Type 2 Diabetes Carla Lipscombe MA b, c, Kimberley J. Smith PhD a, b, Geneviève Gariépy MSc b, c, Norbert Schmitz PhD a, b, c, d, * a
Department of Psychiatry, McGill University, Montréal, Quebec, Canada Douglas Mental Health University Institute, Montréal, Quebec, Canada c Department of Epidemiology and Biostatistics, McGill University, Montréal, Québec, Canada d Montréal Diabetes Research Centre, Montréal, Quebec, Canada b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 September 2013 Received in revised form 18 November 2013 Accepted 6 December 2013 Available online xxx
Objective: To examine the association between physical inactivity and anxiety symptoms in a community-based sample of men and women with type 2 diabetes mellitus. Methods: Eligibility criteria included residents of Quebec, Canada aged between 40 and 75 years, having a diagnosis of type 2 diabetes (10 years), being insulin-naive and having participated in a previous telephone-based survey of diabetes treatments. Of the 2028 eligible respondents, 1953 (96.3%) provided information on anxiety symptoms and were included in this analysis. Participants were interviewed and provided information on diabetes-related clinical and sociodemographic factors. Results: A total of 27.3% of participants reported being physically inactive. The prevalence of mild to severe anxiety symptoms was 22.9%. Persons with mild anxiety symptoms and moderate to severe anxiety symptoms were 1.4 times and 1.7 times more likely to report being inactive than persons without anxiety symptoms, respectively. Subgroup analyses according to gender revealed that women who had mild anxiety symptoms were 1.5 times more likely to report being inactive compared with women who did not have anxiety symptoms, whereas men who had moderate to severe anxiety symptoms were 2.5 times more likely to be inactive than men who did not have anxiety symptoms. Conclusions: Anxiety symptoms in the mild and moderate to severe range are a relevant clinical comorbidity in persons with type 2 diabetes, and men may represent a particularly vulnerable subgroup. Future research is recommended to further assess the relationship between anxiety symptoms and diabetes-related health behaviours. Ó 2014 Canadian Diabetes Association
Keywords: anxiety exercise physical inactivity type 2 diabetes
r é s u m é Mots clés : anxiété exercice inactivité physique diabète de type 2
Objectif : Examiner le lien entre l’inactivité physique et les symptômes de l’anxiété d’un échantillon d’hommes et de femmes de la population générale ayant le diabète sucré de type 2. Méthodes : Les critères d’admissibilité comprenaient les résidents du Québec, au Canada, âgés de 40 à 75 ans, ayant un diagnostic de diabète de type 2 ( 10 ans) et étant insulino-naïfs, qui avaient participé à une précédente enquête par téléphone sur les traitements du diabète. Parmi les 2028 répondants admissibles, 1953 (96,3 %) répondants qui avaient fourni des détails sur les symptômes d’anxiété ont été inclus dans cette analyse. Les participants ont été interviewés et ont fourni des renseignements sur les facteurs cliniques et sociodémographiques liés au diabète. Résultats : Un total de 27,3 % des participants ont rapporté être inactifs physiquement. La prévalence des symptômes d’anxiété légère à grave était de 22,9 %. Les personnes qui avaient des symptômes d’anxiété légère et des symptômes d’anxiété modérée à grave étaient respectivement 1,4 fois et 1,7 fois plus susceptibles de rapporter être inactives que les personnes n’ayant pas de symptômes d’anxiété. Les analyses en sous-groupes selon le sexe révélaient que les femmes qui avaient des symptômes d’anxiété
* Address for correspondence: Norbert Schmitz, PhD, Douglas Mental Health University Institute, McGill University, 6875 LaSalle Boulevard, Montréal, Québec H4H 1R3, Canada. E-mail address: [email protected] 1499-2671/$ e see front matter Ó 2014 Canadian Diabetes Association http://dx.doi.org/10.1016/j.jcjd.2013.12.002
2
C. Lipscombe et al. / Can J Diabetes xxx (2014) 1e7
légère étaient 1,5 fois plus susceptibles de rapporter être inactives que les femmes qui n’avaient pas de symptômes d’anxiété, tandis que les hommes qui avaient des symptômes d’anxiété modérée à grave étaient 2,5 fois plus susceptibles d’être inactifs que les hommes qui n’avaient pas de symptômes d’anxiété. Conclusions : Les symptômes d’anxiété dans la fourchette légère et dans la fourchette modérée à grave représentent une comorbidité clinique pertinente chez les personnes ayant le diabète de type 2, et les hommes pourraient représenter un sous-groupe particulièrement vulnérable. D’autres recherches sont recommandées pour évaluer davantage le lien entre les symptômes d’anxiété et les comportements de santé liés au diabète. Ó 2014 Canadian Diabetes Association
Introduction
Subjects
Diabetes mellitus (DM) is a chronic, progressive medical condition affecting approximately 6.4% of the Canadian population; type 2 diabetes comprises 90% to 95% of all DM cases in Canada (1,2). Standard treatment for type 2 diabetes is focused on glucose regulation and includes a combination of pharmacological and nonpharmacological interventions (3). Regular physical activity is an important component of diabetes treatment and is associated with many health benefits including improvements in glycosylated hemoglobin levels, improved lipid profiles (4e6), reduced incidence of nonfatal cardiovascular events and increased insulin sensitivity (7e9). Physical inactivity is considered to be a key risk factor for the development of a number of diseases including type 2 diabetes (10,11). For persons living with type 2 diabetes, physical inactivity is associated with a 1.7-fold increase in the risk of allcause mortality (12). The mental health of people with diabetes has been shown to be associated with physical inactivity. For example, in a review of 12 studies, a depressed mood in persons with type 2 diabetes was associated with a 1.2- to 1.9-fold increase in the risk of being physically inactive (13). Anxiety is particularly relevant as it is a common comorbidity among people with diabetes. The prevalence of anxiety disorders in diabetes has been estimated at 14%, and 40% of persons with DM report elevated levels of subsyndromal anxiety (14). Additionally, anxiety and depression are often comorbid in persons with diabetes (15). Few studies to date have assessed the relationship between anxiety symptoms and physical activity in persons with type 2 diabetes. In a 2010 study by Khuwaja et al (16), subjects categorized as having anxiety were 1.47 times more likely to report being physically inactive (defined as participating in 10 are an indication of the probable presence of major depression (26). Using a cutoff value of 10, the PHQ-9 has a sensitivity of 88% and a specificity of 88% for major depression (27). In the current study, participants reporting a PHQ-9 score of 10 or greater were classified as having depression symptoms. Data analysis Main analyses Important sociodemographic (age, gender, marital status, education) and clinical factors (BMI, chronic conditions, diabetes complications, diabetes duration, depression symptoms) were compared according to physical activity status (inactive vs. not inactive). To test for significant differences, independent samples Student t tests were used for continuous data (age, diabetes duration) and chi-square tests for categorical data (gender, marital status, education, chronic conditions, complications, BMI, depression symptoms). To assess the bivariate association between anxiety symptom level and physical activity status, chi-square tests were performed for the overall sample. To determine the odds of inactivity associated with varying degrees of anxiety, a hierarchical logistic regression analysis was carried out. In step 1 (model 1), the unadjusted relationship between anxiety symptom level (none, mild, moderate-severe) and physical inactivity was assessed. In step 2 (model 2), sociodemographic and clinical factors associated with inactivity were entered into the model simultaneously. Step 3 (model 3) further adjusted for depression symptoms. All continuous variables met the assumption for normality and were
3
Table 1 Association of physical activity status with sociodemographic and clinical factors Variable
Inactive (n¼533)
Not inactive (n¼1420)
p value
Age (n¼1953) Gender (n¼1953) Female Male Highest level of education (n¼1928) Less than secondary school graduation Secondary school graduation Post high school Marital status (n¼1948) Married/partner Single Divorced Diabetes complications (n¼1848) None One Two or more Chronic conditions (n¼1886) None One Two or more BMI (n¼1867) Normal Overweight Obese Diabetes duration (n¼1948) Depression symptoms (n¼1948) No depression symptoms Depression symptoms
61.748.28
59.978.34
0.000 0.024
54.0 (1.4) 46.0 (1.4)
48.3 (0.8) 51.7 (0.8)
48.9 (3.0)
37.5 (1.8)
28.4 (1.1) 22.7 (2.4)
32.1 (0.7) 30.4 (1.5)
58.5 (2.0) 14.1 (1.8) 27.4 (2.2)
68.4 (1.3) 10.5 (1.1) 21.1 (1.4)
27.2 (3.9) 33.5 (0.7) 39.2 (3.6)
42.0 (2.3) 31.0 (0.4) 27.1 (2.2)
13.5 (3.0) 24.9 (2.1) 61.6 (3.5)
21.4 (1.8) 32. 0 (1.3) 46.6 (2.1)
14.7 (1.2) 35.6 (0.9) 49.7 (1.6) 4.893.05
17.6 (0.7) 39.1 (0.6) 43.3 (0.9) 4.483.12
90.2 (1.1) 9.8 (4.8)
96.6 (0.7) 3.4 (2.9)
0.000
0.000
0.000
0.000
0.043
0.009 0.000
Percentage of participants within each categorical variable according to physical inactivity status (inactive/not inactive) is shown. Categorical variables that had a significant association with inactivity status using a chi-square analysis are indicated in the category column. Continuous variables (age, diabetes duration) were assessed using independent samples t tests and are shown as mean SD. Values in parentheses are the standard deviation of the residual.
therefore presented as means and standard deviations. Data were analyzed using SPSS version 20.0 software (SPSS, Chicago, IL). Sensitivity analysis A series of sensitivity analyses were performed to assess the robustness of the observed effect. A logistic regression was carried out to test for a nonlinear trend in anxiety symptom level across inactivity status. To test for this, we included a quadratic term in the main model using the continuous anxiety score. Additionally, physical activity scores were re-categorized into either low physical activity (12 days of physical activity a month), and the main logistic regression was rerun with these new parameters. Subgroup analysis A final set of logistic regression analyses were performed to assess whether the association between anxiety symptom level and the likelihood of physical inactivity would vary as a function of gender. Separate regression analyses were conducted for female and male subjects. In model 1, the unadjusted association between anxiety and physical inactivity was assessed. Model 2 corrected for previously identified sociodemographic and clinical factors. Model 3 further corrected for depression symptoms. Results Sample descriptive statistics A total of 27.3% of the sample reported being inactive (0 days of activity a month). Women were slightly more likely to report being inactive than men (29.6% vs. 25%). The mean GAD-7 score for the
4
C. Lipscombe et al. / Can J Diabetes xxx (2014) 1e7
Table 2 Association between physical activity status and anxiety symptom level Variable
Inactive (n¼533)
Not inactive (n¼1420)
No anxiety symptoms (n¼1505) Mild anxiety symptoms (n¼280) Moderate-severe anxiety symptoms (n¼168)
68.9 (2.2) 17.6 (2.0) 13.5 (3.9)
80.1 (1.3) 13.1 (1.2) 6.8 (2.4)
The percentage of subjects with mild, moderate or severe anxiety symptoms according to their physical activity status (inactive/not inactive) is shown. The result of the chi-square analysis was significant (chi-square (2) ¼ 32.48, p¼0.00). Values in parentheses are the standard deviation of the residual.
sample was 2.76 (SD 4.16), reflecting mild anxiety symptoms. Although women evidenced slightly higher GAD-7 scores than men, both women (mean 3.43, SD 4.59) and men (mean 2.08, SD 3.55) reported scores in the range for mild anxiety symptoms. The prevalence of anxiety symptoms in the sample (mild to severe anxiety) was 22.9%. A total of 14.3% of the sample reported anxiety symptoms in the mild range whereas 8.6% of the sample exceeded the GAD-7 cutoff score of 10, indicating possible generalized anxiety disorder. Women were more likely than men to report having mild anxiety symptoms (17.8% vs. 10.9%) and moderate-severe anxiety symptoms (11.9% vs. 5.3%).
Univariate analyses Physical inactivity was significantly associated with all independent variables (see Table 1). Preliminary analyses assessing the bivariate relationship between anxiety symptom level and physical
inactivity status for the overall sample confirmed that inactivity was more frequent among persons with mild and moderate to severe anxiety symptoms and less frequent in persons without anxiety symptoms (see Table 2). Logistic regression analysis A logistic regression analysis examined the association between varying degrees of anxiety symptoms and the likelihood of being inactive (Table 3). After adjustment for various sociodemographic and clinical factors (model 2), mild anxiety symptoms were associated with a 1.4-fold increase in the likelihood of being inactive, and moderate to severe anxiety symptoms were associated with a 1.7-fold increase in the likelihood of inactivity. When depression symptoms were included in the model as a covariate (model 3, Table 3), point estimates for the mild anxiety symptom group remained the same, whereas point estimates for the moderatesevere anxiety symptom group were no longer significantly correlated with inactivity. In the final model, physical inactivity was significantly associated with poorer education, obesity, being single, having 1 or more diabetes complications, having 2 or more chronic conditions and having symptoms of depression (model 3). Subgroup analyses A final set of analyses assessed the relationship between physical inactivity and anxiety symptom level according to gender (Table 4). In adjusted models, mild anxiety symptoms were associated with an increased odds of inactivity in women only. This relationship was statistically significant in model 2 and approached significance in model 3 (p¼0.06). For men only, moderate to severe anxiety symptoms were significantly associated with an increased
Table 3 Logistic regression assessing the association between anxiety symptom level and physical inactivity Variable
Model 1, unadjusted OR (95% CI)
Model 2, adjusted OR (95% CI)
Model 3, adjusted OR (95% CI)
No anxiety symptoms Anxiety symptoms, mild Anxiety symptoms, moderate-severe Age Gender (female ¼ 1, male ¼ 0) Education Post-secondary Less than high school High school Marital status Married Single Divorced/widowed Duration of diabetes Diabetes complications None 1 >2 Chronic conditions None 1 >2 Body mass index Normal Overweight Obese Depression symptoms No depression symptoms Depression symptoms
1.0 1.71 (1.27e2.30)* 2.25 (1.56e3.24)*
1.0 1.43 1.70 1.02 1.04
1.0 1.39 1.33 1.02 1.04
(1.05e1.96)y (1.16e2.50)z (1.01e1.04) (0.82e1.31)
(1.01e1.90)y (0.85e2.10) (1.01e1.04)z (0.82e1.32)
1.0 1.51 (1.14e2.01)z 1.09 (0.81e1.49)
1.51 (1.13e2.00)z 1.12 (0.82e1.52)
1.0 1.69 (1.19e2.38)z 1.23 (0.93e1.62) 1.04 (.998e1.07)
1.69 (1.19e2.38)z 1.19 (0.89e1.58) 1.04 (1.00e1.08)
1.0 1.32 (0.99e1.75) 1.48 (1.10e1.99)z
1.33 (1.00e1.77)y 1.47 (1.09e1.97)y
1.0 1.02 (0.71e1.47) 1.54 (1.09e2.15)y
1.03 (0.71e1.47) 1.52 (1.08e2.13)y
1.0 1.19 (0.84e1.68) 1.55 (1.10e2.19)y
1.17 (0.83e1.66) 1.52 (1.08e2.15)y 1.0 1.80 (1.04e3.13)y
CI, confidence interval; OR, odds ratio. Reference group: inactivity. Model 1: unadjusted model, model chi-square (2) ¼ 26.45, p¼0.000. Model 2: model 1 adjusted for age, gender, education, marital status, chronic conditions, diabetes complications, diabetes duration, body mass index; step statistic chi-square (13) ¼ 74.54, p¼0.000; model chi-square (15) ¼ 100.99, p¼0.000; HosmerLemeshow test chi-square (8) ¼ 7.83, p¼0.45. Model 3: model 2 adjusted for depression symptoms; step statistic chi-square (1) ¼ 4.36, p¼0.037; model chi-square (16) ¼ 105.36, p¼0.000; Hosmer-Lemeshow test chi-square (8) ¼ 5.02, p¼0.755. * p
Contents lists available at ScienceDirect
Canadian Journal of Diabetes journal homepage: www.canadianjournalofdiabetes.com
Original Research
Gender Differences in the Relationship between Anxiety Symptoms and Physical Inactivity in a Community-Based Sample of Adults with Type 2 Diabetes Carla Lipscombe MA b, c, Kimberley J. Smith PhD a, b, Geneviève Gariépy MSc b, c, Norbert Schmitz PhD a, b, c, d, * a
Department of Psychiatry, McGill University, Montréal, Quebec, Canada Douglas Mental Health University Institute, Montréal, Quebec, Canada c Department of Epidemiology and Biostatistics, McGill University, Montréal, Québec, Canada d Montréal Diabetes Research Centre, Montréal, Quebec, Canada b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 September 2013 Received in revised form 18 November 2013 Accepted 6 December 2013 Available online xxx
Objective: To examine the association between physical inactivity and anxiety symptoms in a community-based sample of men and women with type 2 diabetes mellitus. Methods: Eligibility criteria included residents of Quebec, Canada aged between 40 and 75 years, having a diagnosis of type 2 diabetes (10 years), being insulin-naive and having participated in a previous telephone-based survey of diabetes treatments. Of the 2028 eligible respondents, 1953 (96.3%) provided information on anxiety symptoms and were included in this analysis. Participants were interviewed and provided information on diabetes-related clinical and sociodemographic factors. Results: A total of 27.3% of participants reported being physically inactive. The prevalence of mild to severe anxiety symptoms was 22.9%. Persons with mild anxiety symptoms and moderate to severe anxiety symptoms were 1.4 times and 1.7 times more likely to report being inactive than persons without anxiety symptoms, respectively. Subgroup analyses according to gender revealed that women who had mild anxiety symptoms were 1.5 times more likely to report being inactive compared with women who did not have anxiety symptoms, whereas men who had moderate to severe anxiety symptoms were 2.5 times more likely to be inactive than men who did not have anxiety symptoms. Conclusions: Anxiety symptoms in the mild and moderate to severe range are a relevant clinical comorbidity in persons with type 2 diabetes, and men may represent a particularly vulnerable subgroup. Future research is recommended to further assess the relationship between anxiety symptoms and diabetes-related health behaviours. Ó 2014 Canadian Diabetes Association
Keywords: anxiety exercise physical inactivity type 2 diabetes
r é s u m é Mots clés : anxiété exercice inactivité physique diabète de type 2
Objectif : Examiner le lien entre l’inactivité physique et les symptômes de l’anxiété d’un échantillon d’hommes et de femmes de la population générale ayant le diabète sucré de type 2. Méthodes : Les critères d’admissibilité comprenaient les résidents du Québec, au Canada, âgés de 40 à 75 ans, ayant un diagnostic de diabète de type 2 ( 10 ans) et étant insulino-naïfs, qui avaient participé à une précédente enquête par téléphone sur les traitements du diabète. Parmi les 2028 répondants admissibles, 1953 (96,3 %) répondants qui avaient fourni des détails sur les symptômes d’anxiété ont été inclus dans cette analyse. Les participants ont été interviewés et ont fourni des renseignements sur les facteurs cliniques et sociodémographiques liés au diabète. Résultats : Un total de 27,3 % des participants ont rapporté être inactifs physiquement. La prévalence des symptômes d’anxiété légère à grave était de 22,9 %. Les personnes qui avaient des symptômes d’anxiété légère et des symptômes d’anxiété modérée à grave étaient respectivement 1,4 fois et 1,7 fois plus susceptibles de rapporter être inactives que les personnes n’ayant pas de symptômes d’anxiété. Les analyses en sous-groupes selon le sexe révélaient que les femmes qui avaient des symptômes d’anxiété
* Address for correspondence: Norbert Schmitz, PhD, Douglas Mental Health University Institute, McGill University, 6875 LaSalle Boulevard, Montréal, Québec H4H 1R3, Canada. E-mail address: [email protected] 1499-2671/$ e see front matter Ó 2014 Canadian Diabetes Association http://dx.doi.org/10.1016/j.jcjd.2013.12.002
2
C. Lipscombe et al. / Can J Diabetes xxx (2014) 1e7
légère étaient 1,5 fois plus susceptibles de rapporter être inactives que les femmes qui n’avaient pas de symptômes d’anxiété, tandis que les hommes qui avaient des symptômes d’anxiété modérée à grave étaient 2,5 fois plus susceptibles d’être inactifs que les hommes qui n’avaient pas de symptômes d’anxiété. Conclusions : Les symptômes d’anxiété dans la fourchette légère et dans la fourchette modérée à grave représentent une comorbidité clinique pertinente chez les personnes ayant le diabète de type 2, et les hommes pourraient représenter un sous-groupe particulièrement vulnérable. D’autres recherches sont recommandées pour évaluer davantage le lien entre les symptômes d’anxiété et les comportements de santé liés au diabète. Ó 2014 Canadian Diabetes Association
Introduction
Subjects
Diabetes mellitus (DM) is a chronic, progressive medical condition affecting approximately 6.4% of the Canadian population; type 2 diabetes comprises 90% to 95% of all DM cases in Canada (1,2). Standard treatment for type 2 diabetes is focused on glucose regulation and includes a combination of pharmacological and nonpharmacological interventions (3). Regular physical activity is an important component of diabetes treatment and is associated with many health benefits including improvements in glycosylated hemoglobin levels, improved lipid profiles (4e6), reduced incidence of nonfatal cardiovascular events and increased insulin sensitivity (7e9). Physical inactivity is considered to be a key risk factor for the development of a number of diseases including type 2 diabetes (10,11). For persons living with type 2 diabetes, physical inactivity is associated with a 1.7-fold increase in the risk of allcause mortality (12). The mental health of people with diabetes has been shown to be associated with physical inactivity. For example, in a review of 12 studies, a depressed mood in persons with type 2 diabetes was associated with a 1.2- to 1.9-fold increase in the risk of being physically inactive (13). Anxiety is particularly relevant as it is a common comorbidity among people with diabetes. The prevalence of anxiety disorders in diabetes has been estimated at 14%, and 40% of persons with DM report elevated levels of subsyndromal anxiety (14). Additionally, anxiety and depression are often comorbid in persons with diabetes (15). Few studies to date have assessed the relationship between anxiety symptoms and physical activity in persons with type 2 diabetes. In a 2010 study by Khuwaja et al (16), subjects categorized as having anxiety were 1.47 times more likely to report being physically inactive (defined as participating in 10 are an indication of the probable presence of major depression (26). Using a cutoff value of 10, the PHQ-9 has a sensitivity of 88% and a specificity of 88% for major depression (27). In the current study, participants reporting a PHQ-9 score of 10 or greater were classified as having depression symptoms. Data analysis Main analyses Important sociodemographic (age, gender, marital status, education) and clinical factors (BMI, chronic conditions, diabetes complications, diabetes duration, depression symptoms) were compared according to physical activity status (inactive vs. not inactive). To test for significant differences, independent samples Student t tests were used for continuous data (age, diabetes duration) and chi-square tests for categorical data (gender, marital status, education, chronic conditions, complications, BMI, depression symptoms). To assess the bivariate association between anxiety symptom level and physical activity status, chi-square tests were performed for the overall sample. To determine the odds of inactivity associated with varying degrees of anxiety, a hierarchical logistic regression analysis was carried out. In step 1 (model 1), the unadjusted relationship between anxiety symptom level (none, mild, moderate-severe) and physical inactivity was assessed. In step 2 (model 2), sociodemographic and clinical factors associated with inactivity were entered into the model simultaneously. Step 3 (model 3) further adjusted for depression symptoms. All continuous variables met the assumption for normality and were
3
Table 1 Association of physical activity status with sociodemographic and clinical factors Variable
Inactive (n¼533)
Not inactive (n¼1420)
p value
Age (n¼1953) Gender (n¼1953) Female Male Highest level of education (n¼1928) Less than secondary school graduation Secondary school graduation Post high school Marital status (n¼1948) Married/partner Single Divorced Diabetes complications (n¼1848) None One Two or more Chronic conditions (n¼1886) None One Two or more BMI (n¼1867) Normal Overweight Obese Diabetes duration (n¼1948) Depression symptoms (n¼1948) No depression symptoms Depression symptoms
61.748.28
59.978.34
0.000 0.024
54.0 (1.4) 46.0 (1.4)
48.3 (0.8) 51.7 (0.8)
48.9 (3.0)
37.5 (1.8)
28.4 (1.1) 22.7 (2.4)
32.1 (0.7) 30.4 (1.5)
58.5 (2.0) 14.1 (1.8) 27.4 (2.2)
68.4 (1.3) 10.5 (1.1) 21.1 (1.4)
27.2 (3.9) 33.5 (0.7) 39.2 (3.6)
42.0 (2.3) 31.0 (0.4) 27.1 (2.2)
13.5 (3.0) 24.9 (2.1) 61.6 (3.5)
21.4 (1.8) 32. 0 (1.3) 46.6 (2.1)
14.7 (1.2) 35.6 (0.9) 49.7 (1.6) 4.893.05
17.6 (0.7) 39.1 (0.6) 43.3 (0.9) 4.483.12
90.2 (1.1) 9.8 (4.8)
96.6 (0.7) 3.4 (2.9)
0.000
0.000
0.000
0.000
0.043
0.009 0.000
Percentage of participants within each categorical variable according to physical inactivity status (inactive/not inactive) is shown. Categorical variables that had a significant association with inactivity status using a chi-square analysis are indicated in the category column. Continuous variables (age, diabetes duration) were assessed using independent samples t tests and are shown as mean SD. Values in parentheses are the standard deviation of the residual.
therefore presented as means and standard deviations. Data were analyzed using SPSS version 20.0 software (SPSS, Chicago, IL). Sensitivity analysis A series of sensitivity analyses were performed to assess the robustness of the observed effect. A logistic regression was carried out to test for a nonlinear trend in anxiety symptom level across inactivity status. To test for this, we included a quadratic term in the main model using the continuous anxiety score. Additionally, physical activity scores were re-categorized into either low physical activity (12 days of physical activity a month), and the main logistic regression was rerun with these new parameters. Subgroup analysis A final set of logistic regression analyses were performed to assess whether the association between anxiety symptom level and the likelihood of physical inactivity would vary as a function of gender. Separate regression analyses were conducted for female and male subjects. In model 1, the unadjusted association between anxiety and physical inactivity was assessed. Model 2 corrected for previously identified sociodemographic and clinical factors. Model 3 further corrected for depression symptoms. Results Sample descriptive statistics A total of 27.3% of the sample reported being inactive (0 days of activity a month). Women were slightly more likely to report being inactive than men (29.6% vs. 25%). The mean GAD-7 score for the
4
C. Lipscombe et al. / Can J Diabetes xxx (2014) 1e7
Table 2 Association between physical activity status and anxiety symptom level Variable
Inactive (n¼533)
Not inactive (n¼1420)
No anxiety symptoms (n¼1505) Mild anxiety symptoms (n¼280) Moderate-severe anxiety symptoms (n¼168)
68.9 (2.2) 17.6 (2.0) 13.5 (3.9)
80.1 (1.3) 13.1 (1.2) 6.8 (2.4)
The percentage of subjects with mild, moderate or severe anxiety symptoms according to their physical activity status (inactive/not inactive) is shown. The result of the chi-square analysis was significant (chi-square (2) ¼ 32.48, p¼0.00). Values in parentheses are the standard deviation of the residual.
sample was 2.76 (SD 4.16), reflecting mild anxiety symptoms. Although women evidenced slightly higher GAD-7 scores than men, both women (mean 3.43, SD 4.59) and men (mean 2.08, SD 3.55) reported scores in the range for mild anxiety symptoms. The prevalence of anxiety symptoms in the sample (mild to severe anxiety) was 22.9%. A total of 14.3% of the sample reported anxiety symptoms in the mild range whereas 8.6% of the sample exceeded the GAD-7 cutoff score of 10, indicating possible generalized anxiety disorder. Women were more likely than men to report having mild anxiety symptoms (17.8% vs. 10.9%) and moderate-severe anxiety symptoms (11.9% vs. 5.3%).
Univariate analyses Physical inactivity was significantly associated with all independent variables (see Table 1). Preliminary analyses assessing the bivariate relationship between anxiety symptom level and physical
inactivity status for the overall sample confirmed that inactivity was more frequent among persons with mild and moderate to severe anxiety symptoms and less frequent in persons without anxiety symptoms (see Table 2). Logistic regression analysis A logistic regression analysis examined the association between varying degrees of anxiety symptoms and the likelihood of being inactive (Table 3). After adjustment for various sociodemographic and clinical factors (model 2), mild anxiety symptoms were associated with a 1.4-fold increase in the likelihood of being inactive, and moderate to severe anxiety symptoms were associated with a 1.7-fold increase in the likelihood of inactivity. When depression symptoms were included in the model as a covariate (model 3, Table 3), point estimates for the mild anxiety symptom group remained the same, whereas point estimates for the moderatesevere anxiety symptom group were no longer significantly correlated with inactivity. In the final model, physical inactivity was significantly associated with poorer education, obesity, being single, having 1 or more diabetes complications, having 2 or more chronic conditions and having symptoms of depression (model 3). Subgroup analyses A final set of analyses assessed the relationship between physical inactivity and anxiety symptom level according to gender (Table 4). In adjusted models, mild anxiety symptoms were associated with an increased odds of inactivity in women only. This relationship was statistically significant in model 2 and approached significance in model 3 (p¼0.06). For men only, moderate to severe anxiety symptoms were significantly associated with an increased
Table 3 Logistic regression assessing the association between anxiety symptom level and physical inactivity Variable
Model 1, unadjusted OR (95% CI)
Model 2, adjusted OR (95% CI)
Model 3, adjusted OR (95% CI)
No anxiety symptoms Anxiety symptoms, mild Anxiety symptoms, moderate-severe Age Gender (female ¼ 1, male ¼ 0) Education Post-secondary Less than high school High school Marital status Married Single Divorced/widowed Duration of diabetes Diabetes complications None 1 >2 Chronic conditions None 1 >2 Body mass index Normal Overweight Obese Depression symptoms No depression symptoms Depression symptoms
1.0 1.71 (1.27e2.30)* 2.25 (1.56e3.24)*
1.0 1.43 1.70 1.02 1.04
1.0 1.39 1.33 1.02 1.04
(1.05e1.96)y (1.16e2.50)z (1.01e1.04) (0.82e1.31)
(1.01e1.90)y (0.85e2.10) (1.01e1.04)z (0.82e1.32)
1.0 1.51 (1.14e2.01)z 1.09 (0.81e1.49)
1.51 (1.13e2.00)z 1.12 (0.82e1.52)
1.0 1.69 (1.19e2.38)z 1.23 (0.93e1.62) 1.04 (.998e1.07)
1.69 (1.19e2.38)z 1.19 (0.89e1.58) 1.04 (1.00e1.08)
1.0 1.32 (0.99e1.75) 1.48 (1.10e1.99)z
1.33 (1.00e1.77)y 1.47 (1.09e1.97)y
1.0 1.02 (0.71e1.47) 1.54 (1.09e2.15)y
1.03 (0.71e1.47) 1.52 (1.08e2.13)y
1.0 1.19 (0.84e1.68) 1.55 (1.10e2.19)y
1.17 (0.83e1.66) 1.52 (1.08e2.15)y 1.0 1.80 (1.04e3.13)y
CI, confidence interval; OR, odds ratio. Reference group: inactivity. Model 1: unadjusted model, model chi-square (2) ¼ 26.45, p¼0.000. Model 2: model 1 adjusted for age, gender, education, marital status, chronic conditions, diabetes complications, diabetes duration, body mass index; step statistic chi-square (13) ¼ 74.54, p¼0.000; model chi-square (15) ¼ 100.99, p¼0.000; HosmerLemeshow test chi-square (8) ¼ 7.83, p¼0.45. Model 3: model 2 adjusted for depression symptoms; step statistic chi-square (1) ¼ 4.36, p¼0.037; model chi-square (16) ¼ 105.36, p¼0.000; Hosmer-Lemeshow test chi-square (8) ¼ 5.02, p¼0.755. * p
