Contents lists available at ScienceDirect

Can J Diabetes 38 (2014) 415e422

Canadian Journal of Diabetesjournal homepage:

www.canadianjournalofdiabetes.com

Original Research

Implementation of Resources to Support Patient Physical ActivityThrough Diabetes Centres in Atlantic Canada: The Effectivenessof Toolkit-Based Physical Activity Counselling

Jonathon R. Fowles BSc, MSc, PhD a,*, Chris Shields BSc, MSc, PhD a, Brittany Barron BKinH a,Stephanie McQuaid BKin a, Peggy Dunbar BSc, MEd b

aAcadia University, Wolfville, Nova Scotia, CanadabDiabetes Care Program of Nova Scotia, Halifax, Nova Scotia, Canada

a r t i c l e i n f o

Article history:Received 10 June 2013Received in revised form12 February 2014Accepted 13 February 2014

Keywords:A1CDiabetes Centrediabetes educatorexerciseimplementationNova Scotiaphysical activityToolkittype 2 diabetes

Mots clés :A1ccentre du diabèteéducateur spécialisé en diabèteexercicemise en œuvreNouvelle-É cosseactivité physiqueboîte à outilsdiabète de type 2

* Address for correspondence: Jonathon R. FowleKinesiology, 550 Main Street, Acadia University, WoCanada.

E-mail address: jonathon.fowles@acadiau.ca

1499-2671/$ e see front matter � 2014 Canadian Diahttp://dx.doi.org/10.1016/j.jcjd.2014.02.020

a b s t r a c t

Objective: The purpose of this study was to determine the effectiveness of toolkit-based physical activitycounselling on physical activity and exercise participation of type 2 diabetes patients attending diabetescentres in Atlantic Canada.Methods: Patients with type 2 diabetes (n¼198) were recruited to a quasiexperimental study comparingthe effectiveness of counselling by persons trained to use a physical activity and exercise resourcemanual (i.e. toolkit) vs. a standard of care counselling situation. Effectiveness was assessed throughquestionnaires completed by patients, and clinical data were extracted from patient charts before and6 months after a single appointment with a diabetes educator. Primary outcome measures were patientself-reported physical activity and exercise levels, efficacy perceptions and mean glycated hemoglobin.Results: There were no significant differences in primary outcomes over time. Subanalyses of the toolkit-counselled patients revealed a significant interaction for moderate-to-vigorous physical activity (MVPA[p<0.0001]), whereby patients who were not meeting Canadian Diabetes Association guidelines forphysical activity at baseline (i.e. <150 MVPA a week; n¼44) increased physical activity (from 20�23 to120�30 minutes) and patients who were active at baseline (i.e. >150 MVPA a week; n¼22) decreasedphysical activity (from 444�32 to 161�41 minutes) at 6 months.Conclusions: A single counselling appointment using the toolkit did not elicit significant changes inphysical activity or clinical outcomes measured 6 months later when compared with standard carecondition; however, increased physical activity was observed for patients who were inactive at baseline.Repeated counselling or more intensive strategies may be required to increase patient physical activitylevels and produce clinical outcomes.

� 2014 Canadian Diabetes Association

r é s u m é

Objectif : L’objet de cette étude était de déterminer l’efficacité du counseling en matière d’activitéphysique selon la boîte à outils sur l’activité physique et la participation à l’exercice physique des patientssouffrant de diabète de type 2 fréquentant des centres du diabète du Canada atlantique.Méthodes : Les patients souffrant du diabète de type 2 (n¼198) ont été recrutés dans une étude quasiexpérimentale comparant l’efficacité du counseling par des personnes formées pour utiliser le manueldes ressources sur l’activité physique et l’exercice (c.-à-d. la Boîte à outils) vs un contexte standard decounseling en soins. L’efficacité a été évaluée au moyen de questionnaires remplis par les patients et lesdonnées cliniques ont été extraites des dossiers de patients avant et 6 mois après un seul rendez-vousavec l’éducateur spécialisé en diabète. Les principaux critères d’évaluation étaient l’activité physique etles niveaux d’exercice rapportés par le patient, les perceptions d’efficacité et l’hémoglobine glyquéemoyenne.Résultats : Aucune différence significative n’a été observée dans les résultats principaux au fil du temps.Les sous-analyses des patients ayant été conseillés par la boîte à outils ont révélé une interaction

s, BSc, MSc, PhD, School oflfville, Nova Scotia B4P 2R6,

betes Association

J.R. Fowles et al. / Can J Diabetes 38 (2014) 415e422416

significative de l’activité physique modérée à vigoureuse (APMV [p < 0,0001]) selon laquelle les patientsqui au début ne répondaient aux lignes directrices sur l’activité physique de l’Association canadienne dudiabète (c.-à-d. < 150 APMV par semaine; n¼44) ont augmenté leur activité physique (de 20 � 23 à120 � 30 minutes) et les patients qui au début étaient actifs (c.-à-d. > 150 APMV par semaine; n¼22) ontdiminué leur activité physique (de 444 � 32 à 161 � 41 minutes) après 6 mois.Conclusions : Un seul rendez-vous de counseling utilisant la boîte à outils n’a pas entraîné de change-ments significatifs dans la mesure de l’activité physique ou des résultats cliniques 6 mois plus tard parrapport aux soins standards. Cependant, l’augmentation de l’activité physique a été observée chez lespatients qui étaient inactifs au début. Des séances répétées de counseling ou d’autres stratégies in-tensives s’avèrent nécessaires pour rehausser les niveaux d’activité physique des patients et donner lieu àdes résultats cliniques.

� 2014 Canadian Diabetes Association

It is well established that regular physical activity is an effectivemodality for the prevention and management of type 2 diabetesmellitus (1,2). Although “physical activity” and “exercise” are oftenused interchangeably, these terms represent different behaviours.Physical activity is defined as any bodily movement producingenergy expenditure (2) and is often unstructured and can take placein occupational, household or leisure time. Exercise is defined as asubset of physical activity and is planned, structured and repetitiveand of an intensity and duration that leads to an improvement inphysical fitness or changes in body composition (2). Research hasclearly demonstrated that regular physical activity and exerciseproduce significant improvements in glycemic control, car-diometabolic outcomes and improved quality of life for personseither at risk of diabetes or with diabetes (1e6) in a dose-responsemanner (7e10). The Canadian Diabetes Association (CDA) guide-lines for physical activity recommend 150 minutes of moderate tovigorous aerobic exercise per week and 2 but preferably 3 sessionsof resistance exercise (2). Unfortunately, the great majority (>70%)of patients with type 2 diabetes do not meet the CDA 150 minutesguidelines for physical activity (11e14). It is clear that there is amajor challenge in clinical practice to help clients meet target levelsof physical activity (15).

A meta-analysis by Umpierre et al (16) concluded that simplyproviding physical activity advice is far less effective than struc-tured exercise in eliciting improved glycemic control in type 2diabetes patients. A more recent review and meta-analysis byVuori et al (17) concluded that physical activity promotion inprimary care is effective at increasing physical activity in a rangeof populations when done in 1 or more face-to-face sessions byvarious health professionals with tailored physical activity pre-scriptions. Vuori et al (17) stated that physical activity counsellingeffectiveness in clinical practice is improved when 1) it beginswith individual assessment of needs and barriers; 2) the messageis simple, specific and realistic; 3) it employs valid behaviourchange methods; 4) it includes clear, proximal goals, and 5) itfocuses on internal motivation and factors that support self-efficacy. The use of these strategies by trained professionals indiabetes care can improve glycemic control and diabetes out-comes (18). Supervised exercise typically targets a motivatedgroup of healthy people, is expensive to develop and maintain andresults in poor long-term adherence; therefore, effective physicalactivity counselling likely shows greater promise for wide-scaledissemination and overall effectiveness in changing patientbehaviour (17e21).

In Canada, the primary provider of ongoing diabetes lifestylesupport for diabetes patients are diabetes educators in diabetescentres. Unfortunately, diabetes educators report that they havelow knowledge and confidence in their ability to counsel patientsabout physical activity or provide appropriate exercise recom-mendations (22,23). We developed a physical activity resourcemanual (i.e. the toolkit [24]) and education program, the details ofwhich are described elsewhere (25), that were designed to support

diabetes educators in their abilities to counsel regarding physicalactivity and to make appropriate recommendations for exercise. Inbrief, the manual and program are an evidence-based, theorydriven and practically oriented approach that integrates physicalactivity recommendations and behavioural management strategiesdesigned to build self-efficacy in both providers and patients and toelicit incremental improvements in patient physical activitybehaviour. This program was effective at increasing diabetes edu-cator’s confidence regarding physical activity counselling (25) andwas adopted by the CDA as a primary resource manual for diabeteseducation (26).

Although the improvement in diabetes educator confidence fortoolkit-based physical activity counselling has been established,the impact of such counselling on patient behaviour wasunknown. Therefore, the purpose of this study was to examine theeffect of the toolkit-based counselling intervention delivered bydiabetes educators in regular clinical practice (i.e. as a portion of asingle diabetes centre visit), compared to a standard care condi-tion, on self-reported physical activity and exercise, efficacy per-ceptions and self-reported fitness and health, as well as on clinicalvariables before and 6 months after a single counsellingappointment.

This evaluation was part of an action-research program, whichinvolves the process of engaging in a change in the professionalpractice by a group or organization while concurrently conductingresearch on the process and outcomes of that change, with the goalof improving the way a specific issue or problem is addressed (27).In the context of the current study, the Diabetes Care Program ofNova Scotia “Best Practice” Committee requested the involvementof researchers at Acadia University to develop, implement andevaluate resources to assist diabetes educators to better counselpatients on physical activity and provide exercise recommenda-tions in situations where they did not have access to exerciseprofessionals or exercise facilities. There was extensive involve-ment by the Diabetes Care Program of Nova Scotia and diabetes careprofessionals in the field on all aspects of the study design,administration and evaluation. This was an evaluation of theeffectiveness of translation into practice, which is an area ofresearch that is greatly needed (15,28). It was hypothesized thatpatients who received toolkit counselling would improve theirparticipation in regular physical activity more than the standardcare group.

Methods

Design

Ethics approval for this study was obtained through the AcadiaUniversity Research Ethics Board and the respective DistrictHealth Research Ethics Boards. A quasiexperimental design wasused to compare the effectiveness of a single toolkit counselling

J.R. Fowles et al. / Can J Diabetes 38 (2014) 415e422 417

session done in Nova Scotia to a single standard of care counsel-ling session done in New Brunswick. Diabetes educators in NovaScotia were trained to deliver physical activity counselling andexercise recommendations using the Diabetes Physical Activityand Exercise Toolkit, the details of the dissemination andoutcomes are described elsewhere (24,25). Patient tools andresources (brochures, online material and resistance exercisevideo) accompanied the package to support both the patient andthe provider in physical activity messaging and providing tailoredrecommendations based on individual needs and capabilities. TheDiabetes Care Program of Nova Scotia endorsed the program andencouraged diabetes educators to incorporate the methods intoclinical practice as best possible given the demands on their time,resources and care objectives. Although the toolkit was dissemi-nated across Nova Scotia, the evaluation project was done in 5diabetes centres. Two diabetes centres in New Brunswick partic-ipated as standard care control sites, and did not receive thetoolkit or training workshops.

Between 6 months and 12 months after dissemination of thetoolkit in Nova Scotia, 205 adult patients with diagnosed type 2diabetes were recruited from diabetes centres in Nova Scotia andNew Brunswick. At each site, diabetes educators identified a con-venience sample by approaching patients (approximately 1 ofevery 5 patients making regularly scheduled visits who had notpreviously received physical activity counselling from a diabeteseducator) to determine interest in participation in the study.Inclusion criteria included age of at least 18 years, having type 2diabetes and meeting the criteria that indicated referral to a dia-betes centre as identified by the patient’s physician. Patients withphysical limitations or comorbidities were not excluded frominvitation to participate in the study, although it is possible thatsome patients may have declined participation based on these orother factors. If the patient was interested, the study was explainedand informed written consent was obtained. Patients completed aquestionnaire before their counselling visit, and diabetes educatorstranscribed data from their chart to a study data collection formafter the appointment. If blood work had not been done in theprevious 2 months, blood work was requested.

Intervention

Patients in Nova Scotia (NS) were identified as the toolkitgroup, or the intervention group; they received a single counsel-ling appointment from a diabetes educator using the theories andmethods outlined in the toolkit. The toolkit intervention providedthe options of reading toolkit handout materials, completing self-management tools and participating in physical activity or exer-cises recommended by the diabetes educator. The evaluationcomponent was respectful of the clinical situation, in that diabeteseducators were trained to perform physical activity counsellingfor use in practice but there was no requirement for them todeliver certain counselling objectives as part of the study. Like-wise, patients were exposed to different counselling environ-ments (based on what was delivered by the diabetes educator)and allowed to freely choose their engagement and physical ac-tivity participation based on their own attitudes, interests andsituation.

Comparison Group

The patients from New Brunswick (NB) were identified as thestandard care group because the diabetes educators from NewBrunswick did not receive toolkit training. The control groupreceived a single standard care counselling session, with no addi-tional materials beyond reference to Canada’s physical activityguides or the CDA recommendations for exercise.

Measures

As there is little to no research examining many of the con-structs of interest within the context of diabetes education, allmeasures were developed by the researchers for use in the currentwork, with the exception of the short International Physical ActivityQuestionnaire (IPAQ) (29). All measures used were vetted forappropriateness for use with diabetes patients by an experiencedprofessional in the diabetes care community. The informationobtained on the patient questionnaire was as follows.

DemographicsBasic demographic information on age, marital status, education

level and population of their hometown was collected.

Self-reported physical activity and exerciseThe short IPAQ was used to assess reported days and minutes of

vigorous, moderate and walking physical activity over the previous7 days (29). The weekly totals for vigorous and moderate activitywere combined to determine total weekly moderate-to-vigorousphysical activity (MVPA), and walking is presented separatelybecause the measurement of walking by the IPAQ does not discernintensity. Given that the current CDA guidelines use the termexercise, which could be interpreted differently by patients as aspecific form of physical activity, we also used customized ques-tions to ask patients to report how many days, if any, over theprevious week that had they participated in planned resistanceconditioning exercises “aimed at improving yourmuscular strength(such as using resistance bands, using dumbbell, or doing bodyweight exercises like push-ups)” and planned aerobic exercise “thatelevated your heart rate, had you breathing hard and made yousweat for at least 20minutes, and that was aimed at improving youraerobic fitness (such as jogging, rowing or swimming).”

Patient perceptionsPatients’ perceptions of their own confidence in several efficacy

subdomains were evaluated. A series of 10 questions measuredparticipants’ confidence in their ability to motivate and monitortheir physical activity levels (i.e. self-regulatory efficacy) as well asconfidence in their ability to achieve the recommended amount ofphysical activity (i.e. task efficacy). A series of 9 questions was usedto assess participants’ perceived confidence in the diabetes edu-cator’s ability to motivate and educate on physical activity as well asability to help them cope with barriers, choose appropriate activityand help schedule physical activity to encourage being more active(i.e. proxy efficacy). Levels of efficacy (identified on the question-naire as “confidence”) were rated on a scale of 0% to 100%, 0% being“not at all confident” and 100% being “completely confident.” Finalvalues were calculated by summing the values of the questionspertinent to each efficacy subdomain and calculating the mean.Note that when referring to all efficacy subdomains as a group, theterm used is efficacy perceptions.

Patient attitudes toward being physically active for at least30 minutes each day in the upcoming month was assessed by 5researcher-developed questions on a 1 to 7 scale, ranging fromharmful to beneficial, good to bad, enjoyable to unenjoyable.Patients were asked to report their perceived aerobic fitness (abilityto walk or run for longer distances), their musculoskeletal fitness(ability to lift heavier objects) and their perceived overall health ona scale from 1 to 5, with 1 being poor and 5 being excellent.

Clinical variablesAt each data collection, diabetes educators provided clinical

information regarding the patients, such as date of diagnosis,height, weight, waist circumference and blood pressure. Fastingblood glucose and levels of glycated hemoglobin (A1C), total

Table 1Participant demographics at baseline, n¼198

Variable Toolkit(n¼147)

Standard care(n¼51)

Age, years 58.9�10.8 59.1�10.0SexWomen, n 102 31Men, n 45 20

Physical activity levels, minutes/week*

Vigorous 54�140 85�173Moderate 88�160 184�273Walking 213�261 176�201150 minutes/week MVPA, % 29 47

Exercise participation, days/weekResistance 0.4�1.1 0.3�0.9Aerobic 0.5�1.5 0.7�1.7

Efficacy perceptions (of 100%) and attitudes (of 7)Task efficacy 53�28 45�26Self-regulatory efficacy 64�24 59�24Proxy efficacy 71�25 64�28Attitudes 6.1�1.0 5.7�1.2

Perceived fitness and health (of 5)Perceived aerobic fitness 2.2�1.0 2.4�1.0Perceived muscular fitness 2.4�0.9 2.4�1.0Perceived overall health 2.7�0.8 2.8�0.8

Clinical variablesBody mass index, kg/m2 33.6�8.4 32.6�10.0Waist circumference, cmy 107�20 109�21Systolic blood pressure, mm Hg 127�14 126�15Diastolic blood pressure, mm Hg 72�9.8 72�8.7Glycated hemoglobin, % 7.42�1.5 7.61�2.0Fasting blood glucose, mmol/L 8.48�3.4 8.64�3.8Triglycerides, mmol/L 2.18�2.5 1.90�1.1Total cholesterol, mmol/L 4.41�1.3 4.15�0.9High-density lipoprotein, mmol/L 1.20�0.4 1.21�0.5Low-density lipoprotein, mmol/L 2.36�1.0 2.16�0.7

MVPA, moderate-to-vigorous physical activity.Values are mean�SD unless otherwise indicated.

* Physical activity data for Toolkit, n¼140, and standard care, n¼48, because 25participants exceeded the maximum reported MVPA of 960 minutes a week.

y Waist circumference data for toolkit, n¼38, and standard care, n¼39.

J.R. Fowles et al. / Can J Diabetes 38 (2014) 415e422418

cholesterol, high-density lipoprotein, low-density lipoprotein andtriglycerides from last available blood work were also provided.Any medications being taken with regard to insulin, glucose con-trol, hypertension and lipid control were reported, as well as anyincreases or decreases in medication since the previous datacollection or dietary revisions.

Timing of follow upPatients were normally scheduled for repeat appointments

6 months after their baseline appointment; at such time, patientscompleted a follow-up questionnaire assessing the same measuresas the baseline questionnaire. The follow-up procedure and time-line was the same for the toolkit sites and standard care sites.

Data analysis

There were 3 stages of data analysis: 1) Descriptive statisticswere used on baseline and 6-month variables to observe outcomessuch as levels of physical activity and exercise, efficacy perceptions,clinical outcomes and medication usage; 2) multiple repeated-measures 2�2 multivariate analysis of variance (MANOVA) wasused to compare the changes in measures over time relative to thestandard care group, and 3) outcomes between patients whoreported meeting the CDA physical activity guidelines and patientswho did not were compared using 2�2 repeated-measures analysisof variance (ANOVA). Because of incomplete data in different partsof patient questionnaires or incomplete clinical data, the number ofcases used in each analysis varies by analysis to maximize thestatistical power for any given data set. These numbers are pre-sented in tables and text for the individual analyses.

All analyses were carried out using SPSS Statistics 20 (IBMCorporation, Armonk, NY, USA). Mean with standard deviation wasused to describe demographic and outcome variables, with meanplus standard error in figures. Differences are presented as Wilks’lambda for multivariate analyses and F-statistic for univariateanalyses with omnibus or univariate statistics, respectively. Statis-tical significance was accepted at p<0.05; however, given that thiswas action research, omnibus significance values <0.10 werereported for possible trends.

Results

Demographics

We obtained complete baseline data on a total of 198 patientswho volunteered to participate in the study (Table 1). The majority(84.4%) were Caucasian, from towns with populations less than100 000 people (86.8%) andwith an education level of high school orless (53.6%). The number of patients in the entire sample meeting150 minutes of MVPA per week was 39.9%. Only 9.6% of patientsperformed either aerobic or resistance exercise 3 times per week,and only 2.5% of patients reported meeting both the guidelines foraerobic and resistance exercise 3 times per week. To test for groupequivalency at baseline, univariate ANOVA was conducted ondemographic characteristics and primary outcome variables andrevealed no differences between the intervention group (NS, n¼147)and the standard care group (NB, n¼51) at baseline (Table 1). Giventhat the onus was on diabetes educators and the patients to scheduleand make appointments, several patients were lost because ofscheduling confusion and miscommunication regarding theirparticipation in the study vs. their follow up regarding their regulardiabetes care, which resulted in a large loss of patients for the6-month follow up (total n¼78). Furthermore, we were unable tomatch the baseline response of 17 patients to 6-month question-naires because of mistakes by patients in providing unique identi-fiers for their coded data. To err on the side of caution, we eliminated

these patients from the analysis, leaving a total of 103 patients foruse in paired analyses (NS, n¼72; NB, n¼31) (Figure 1).

Physical activity and exercise

In the analysis of physical activity data over time, removal ofoutliers (patients withMVPA>960minutes aweek, n¼8) producedcell sizes of 66 for the NS group and 29 for the NB group, for a totalof 95. In examining changes in self-reported weekly minutes ofMVPA as well as weekly minutes of walking from baseline to6 months, no significant effects were detected (omnibus signifi-cance ranged from 0.15 to 0.87). A second MANOVA on exercisebehaviour also revealed no significant effects (omnibus signifi-cance, 0.13 to 0.58) for aerobic or resistance exercise (Table 2).

Patient perceptions

A significant time effect (Wilks 0.910, p¼0.014; cell sizes, NSn¼63 and NB n¼31) was noted for perceived aerobic fitness, spe-cifically, overall perceptions of aerobic fitness improved frombaseline to 6months (p¼0.003) (Table 2). No significant main effectfor group or interactive effects were seen; however, the univariatetest of perceived aerobic fitness approached statistical significance(p¼0.053). There were also no significant main effects orinteractions for efficacy perceptions or attitudes (omnibus p rangedfrom 0.14 to 0.23) although, again, the examination of the univar-iate analyses suggested that the toolkit group maintained efficacyperceptions and attitudes, whereas the standard care group tendedto decline in attitude and efficacy perceptions over time (p¼0.07).

Figure 1. Study flow diagram. See text for detailed description. Note that actual case numbers used in individual analyses may be different from those presented in the figure, basedon incomplete data sets for certain variables. MVPA, moderate-to-vigorous physical activity; PA, physical activity.

J.R. Fowles et al. / Can J Diabetes 38 (2014) 415e422 419

Clinical variables

There was no difference in the relative use of medication orchanges in medications or diet between toolkit and standard carefor patients reportingmedication use at baseline (NS group, n¼135;NB group, n¼51; omnibus p¼0.23 to p¼0.8). For patients with filesonmedication use at 6 months (NS n¼52 and NB n¼32), therewereno differences between the toolkit group and the standard caregroup for medication use or changes in medication use over the6 months of the study. A chi-square test indicated a higher pro-portion of smokers in the standard care group than in the toolkitgroup (p¼0.013). There were no other significant effects over timeor across groups on clinical variables (omnibus p¼0.22 to p¼0.83;cell sizes, NS n¼46 and NB n¼15).

Subanalysis of physical activity behaviours for active and inactivepatients

To determine whether toolkit counselling may have been moreeffective with different groups of patients, exploratory repeated-measures MANOVA was performed to examine potential changesamong patients who received toolkit counselling and were signifi-cantly different in their physical activity at baseline, namely, active(>150 minutes MVPA per week, n¼22) or inactive (<150 MVPA,n¼44). There was a significant interaction for MVPA (p<0.0001),whereby patients whowere not meeting CDA guidelines for physicalactivity at baseline (i.e. <150 MVPA minutes a week) increasedphysical activity (from 20�23 to 120�30 min), and patients whowere active at baseline (i.e. >150 MVPA minutes a week), physicalactivity decreased (from 444�32 to 161�41 minutes) at 6 months(Figure 2). There was no significant difference in walking behaviour

across time for the toolkit subgroups (inactive, 201�260 to 133�163minutes; active, 254�267 to 303�223 minutes; p¼0.143).

Discussion

The goal of this study was to determine the effectiveness ofproviding greater support for physical activity counselling, throughimplementation of a physical activity and exercise toolkit in thecontext of regular clinical practice at diabetes centres, on patients’physical activity and exercise behaviours. In contrast to ourhypothesis, a single physical activity counselling visit offered bythose trained in use of the toolkit was not sufficient for improvingphysical activity or exercise behaviours or clinical outcomes in theoverall sample of patients; however, patients who were not activeat baseline improved their MVPA, whereas patients who werealready active at baseline showed declines in MVPA, albeit stillattaining recommended guidelines. These results may imply thatimproved support for physical activity counselling may be helpfulwith patients who are inactive andwant to becomemore active andfor whom simple recommendations for beginning physical activitymay be helpful, but enhanced support for exercise may be neededin the form of greater resources, expertise and facilities for patientswho are active to keep them exercising at a high level.

This study was part of an action research program in whichresearchers attempted to address a need identified in clinicalpractice and provide evaluation of the effectiveness of the changein practice. The need addressed in this instance was providingresources and training on physical activity counselling to diabeteseducators; however, this study demonstrates the challenges ofevaluation in this type of research, where changes in practice maybe slow and therefore the intervention is inherently variable given

Table 2Paired comparison of variables from baseline to 6 months for participants who received counselling from toolkit-trained diabetes educators and participants who receivedstandard care

Variable Toolkitbaseline

Toolkit6 months

Standard carebaseline

Standard care6 months

Physical activity, minutes/week (TK¼40, SC¼18)*

Vigorous 51�155 35�83 78�136 91�176Moderate 50�115 83�130 81�125 29�49Walking 233�275 200�198 143�187 170�244150 minutes/week MVPA, % 29 28 47 30

Exercise participation, days/week (TK¼54, SC¼29)Resistance 0.5�1.1 1.1�1.6 0.4�1.2 0.6�1.3Aerobic 0.5�1.4 0.6�1.6 0.7�1.6 0.8�1.3

Efficacy (of 100%) and attitudes (of 7) (TK¼59, SC¼29)Task efficacy 58�28 55�27 48�28 46�29Self-regulatory efficacy 66�25 65�25 63�26 59�22Proxy efficacy 72�26 70�28 72�25 63�26Attitudes 6.0�1.0 6.0�1.3 6.0�1.1 5.4�1.1

Perceived fitness and health (of 5) (TK¼63, SC¼31)Perceived aerobic fitness 2.2�1.0 2.7�1.1y 2.5�1.1 2.6�1.1y

Perceived muscular fitness 2.5�0.9 2.7�0.9 2.5�1.1 2.6�1.1Perceived overall health 2.8�0.7 2.9�0.8 2.8�0.9 2.9�0.8

Clinical variables (TK¼46, SC¼15)Body mass index, kg/m2 33.3�7.6 34.6�11.0 33.5�9.9 34.3�10.6Waist circumference, cm 104�21 102�26 109�20 111�22Systolic blood pressure, mm Hg 123�13 125�16 124�11 127�12Diastolic blood pressure, mm Hg 71�8 71�7.3 70�7.4 72�6.6Glycated hemoglobin, % 7.0�1.3 7.1�1.2 7.4�1.2 7.0�0.7Fasting blood glucose, mmol/L 8.1�2.9 7.7�1.9 7.7�2.3 7.7�1.8Triglycerides, mmol/L 1.6�0.7 1.7�0.7 2.5�1.5 2.4�1.4Total cholesterol, mmol/L 4.4�1.3 4.3�0.9 4.0�0.8 4.4�0.8High-density lipoprotein, mmol/L 1.30�0.4 1.22�0.4 1.28�0.8 1.26�0.5Low-density lipoprotein, mmol/L 2.44�1.1 2.37�0.9 1.91�0.6 2.08�0.6

MVPA, moderate-to-vigorous physical activity; SC, standard care; TK, Toolkit.Values are mean�SD. Case numbers are identified for each analysis.

* Physical activity numbers are for participants with MVPA <960 minutes a week.y Significant for time p<0.005.

J.R. Fowles et al. / Can J Diabetes 38 (2014) 415e422420

the level at which individual diabetes educators may use orimplement tools and resources during the course of the evaluation.

The actual “dose” of physical activity counselling or exerciseprescription in this study is unknown, as participation in physicalactivity counselling by the diabetes educators was completelyvoluntary based on their own confidence level, time commitmentsand clinical judgement in a given appointment, and the actualcontent covered by each diabetes educator with each patient was

Figure 2. Subanalysis of toolkit-counselled patients. Data are mean�SE. Patients whowere not meeting Canadian Diabetes Association (CDA) physical activity guidelines atbaseline (inactive [solid line], i.e. <150 minutes of moderate-to-vigorous physicalactivity [MVPA] per week, n¼44) vs. patients who were exceeding the CDA physicalactivity guidelines at baseline (active [broken line], i.e. >150 minutes MVPA per week,n¼22 ) were compared. CPG, clinical practice guidelines.

not tracked. The lack of effect to elicit overall behaviour changefrom brief physical activity counselling as a portion of a single visitis not surprising, then, given the range of adoption possible amongdiabetes educators to integrate toolkit materials into practice (25)and the range of patient desires, motivations and circumstancesto act on any recommendations made.

These results may indicate that brief delivery of physical activitycounselling is not enough to significantly influence physical activitybehaviour in patients measured 6 months after the appointment orthat more effective implementation of physical activity counsellingin practice is required for a resource of this type to be broadlyeffective. For example, we observed initially as part of this projectthat only 39% of diabetes educators reported physical activitycounselling in most sessions (22), and for patients who did receivecounselling on physical activity, diabetes educators spent only a fewminutes of a 30-minute appointment on the topic (22). We recentlyreported improvements in diabetes educator confidence for phys-ical activity counselling as a result of the toolkit training andimplementation (25); however, it is unknown what the exactimpact of this increased confidence would have on the content ofindividual appointments in this study, as that was not tracked.Overall, the potency of the intervention was not enough to elicitbehaviour change in the entire intervention group of patients.

These findings are in line with a recent meta-analysis of ran-domized controlled trials that showed that supervised exercise ismore beneficial than physical activity advice to elicit changes inphysical activity and exercise behaviour to produce clinical out-comes in type 2 diabetes patients (16). A key factor in this com-parison is that supervised exercise carries greater interaction withdiabetes care providers over time and can provide more continuousmonitoring of patient behaviour. More frequent interactions or

J.R. Fowles et al. / Can J Diabetes 38 (2014) 415e422 421

longer durations of appointments may also be required for physicalactivity counselling in diabetes practice to be more effective, as ithas been shown that 2 30-minute physical activity counsellingsessions with 4 follow-up phone calls can elicit changes in physicalactivity behaviour lasting a year and produce clinical outcomes inpatients with type 2 diabetes (18). A physical activity counsellingtrial with sedentary obese patients was effective at improving self-reported physical activity as long as 13 weeks afterward (21). Thefindings of the current study provide some evidence in this regard,through the observation that patients who were physically inactiveat baseline (i.e. with an average reported MVPA of 20 minutes,n¼44) increased their MVPA to 120 minutes after toolkit counsel-ling. Although walking behaviour did not change significantly, theimprovement in MVPA is a clinically meaningful result, given thatthe greatest difference in health outcomes typically occur in thefirst 90 minutes of MVPA per week (17,30).

The use of the short IPAQ in this study was challenging becausereporting walking behaviour as a mode without an intensity limitsdifferentiation of walking for transport, leisure or pursuit of physicalactivity guidelines in the form of “brisk walking” and, therefore,produces very large variability in reporting. Including walkingminutes intoMVPAproduces excessively highproportionsof patients(approximately 80%) exceeding the MVPA guidelines of 150 minutes(29,31) and, therefore, these variableswere handled separately in thisanalysis so that the values for totalMVPAminutes and the proportionmeeting guidelines would be in line with other researchers’ obser-vations on type 2 diabetes patients (11e14). It is likely that some ofthewalking reported by patients in this studymay have been done inpursuit of physical activityguidelines (i.e. in the formof briskwalkingatmoderate or greater intensity), so the real pattern ofMVPAmay beunderreported, as some of their MVPA would appear under thewalking variable. Likewise, there may have been some shifting ofphysical activity or exercise behaviours between variables from onetimepoint to another (i.e. displacingwalkingwith a new behaviour ofMVPA such as fast cycling), or vice versa.

It has been shown that reported changes in subjective physicalactivity are not always detected by objective means (21). Objectivemeasures of physical activity were not used in this study for prac-tical reasons, to reduce the burden on both patients and diabeteseducators in the evaluation. Nevertheless, the majority of patientsin regular clinical practice aremostly inactive at baseline, so awide-scale implementation of toolkit counselling by diabetes educatorsmay still be relevant if incremental gains can be made to helpmostly inactive patients become more active. We observed anincrease in perceived aerobic fitness that wasmostly contributed bythe toolkit counselled group (interaction, p¼0.053), which is alsorelevant, as improved fitness can result from improved MVPA.

We observed trends for maintained efficacy perceptions andattitudes in toolkit counselled patients, whereas the standard caregroup efficacy perceptions and attitudes tended to decline overtime. That may help explain the improvement in physical activityfor inactive patients, as self-efficacy and, in particular, barrier effi-cacy are strong predictors of physical activity among sedentarypersons (32) and type 2 diabetes patients (19,33). The toolkittraining was effective for increasing diabetes educator self-efficacytoward physical activity counselling, and also improved diabeteseducators’ perceptions of their patients’ abilities to perform phys-ical activity (25). This subtle change in the patient-provider inter-action to improve patient confidence in their provider (i.e. proxyefficacy), along with improved information (i.e. brochures) toovercome barriers, support planning and increase confidence toinitiate physical activity (i.e. self-management and task efficacy),may have been enough to elicit this incremental change in inactivepatients’ behaviour. Efficiency of implementation can be achievedin clinical practice when a large majority of users make small im-provements. The toolkit has now been disseminated across Canada,

with diabetes educators reporting intentions to change practice as aresult of the manual and training (34), so the potential impact ofthese improvements in physical activity counselling may be sig-nificant in the scheme of diabetes care.

It was evident from the results of this study that more needs tobe done to support more formal exercise behaviours in diabetesclinical practice. We observed that the overall proportion ofpatients meeting CDA guidelines for aerobic or resistance exercise(at least 3 times a week) in the entire sample did not exceed 10%,and the MVPA of active patients declined over time (from 444minutes to 160 minutes a week). A decline in physical activityamong control patients has been shown in other physical activitytrials involving type 2 diabetes patients (18), so it is obvious thatwithout some type of support the challenge to maintain physicalactivity for persons with diabetes is great. Even so, the “active”group was still meeting the CDA guidelines of more than 150 mi-nutes MVPA per week, which can bring significant health benefits.

It is possible that seasonal effects contributed to our observedinteraction, as we performed baseline testing in the spring and earlysummer and 6-month testing in the fall. Patients who were inactiveat baseline may have overcome barriers to become somewhat activeeven in the fall, when weather may create a challenge, whereas thehighly active patients may have had extensive physical activityoutside in goodweather thatmay have been compromised in the falland early winter. We observed maintained efficacy perceptions andattitudes among toolkit counselled patients over time, but did notdetermine relationships to maintained physical activity among dia-betes patients per se, as this was not the focus of the study. There aremany factors that may limit adoption of physical activity recom-mendations by patients (11,12,14,35) or that may limit long-termadherence, such as self-efficacy, barrier and task efficacy (33,35)and autonomous motivation (35). Further examination of thesecontributors would be fertile ground for future research.

Study limitations

There are clear challenges with performing action research inclinical practice. The intent of such work is not to show the efficacyof the intervention as is done in randomized controlled trials but,rather, to demonstrate the effectiveness of implementation whendelivered in the clinical context of care. With action research, thereis undeniable loss of follow up owing to the nature of regularclinical visits and the expectations of the study in addition to reg-ular clinical work; that influences the type of measures that can beperformed and reduces the power of analyses and, therefore, sta-tistical effects. The nature of action research also means that the“intervention” can be variable, given the breadth of engagementpossible by different professionals at different times during theprocess, and that can detract from the fidelity of the interventionitself. We acknowledge these limitations in this studydthat therewere patient losses at follow up, that the intervention was appliedin a single visit, that the actual content covered by each diabeteseducator with each patient was not tracked and that subjectivemeasures of physical activity were used. Although our interactionfor MVPA was highly significant, it is recognized that subjectivemeasures of physical activity may not be consistent with objectivemeasures. Additionally, using the IPAQ limited the assessment ofwalking activity as a mode without a descriptor of intensity, so itwas analyzed separately from MVPA, and there may have beenoverlapping or underreporting of walking behaviour to attainphysical activity guidelines.

Conclusion

It is clear that there is a gap in diabetes care to help patients withtype 2 diabetes to reach targeted levels of physical activity and

J.R. Fowles et al. / Can J Diabetes 38 (2014) 415e422422

exercise. Baseline levels of physical activity were low in the sampleof type 2 diabetes patients in Atlantic Canada; however, improve-ments were made for inactive patients by providing enhancedsupport for physical activity counselling in the context of regulardiabetes centre practice. Creative and cost-effective ways to morewidely implement strategies for regular physical activity andexercise services at diabetes centres in Canada should be exploredto address the urgent and ongoing challenges of preventing andmanaging diabetes. Future work will need to confirm the efficacy ofmore intensive but clinically manageable physical activity coun-selling in diabetes clinical practice, as was performed in this study,and the resource use and cost-effectiveness of such programs foroverall diabetes management.

Acknowledgements

Financial support was provided by the Lawson Foundation(GRT 2008-006, 2010-015), Diabetes Care Program of Nova Scotiaand the Acadia Research Fund. The development of the toolkit wasfunded in part by a development grant from the Diabetes CareProgram of Nova Scotia. The dissemination and evaluation of thetoolkit was funded through a grant from the Lawson Foundation(GRT 2012-009). The TheraBand� Academy donated resistancebands for the project.

Author Disclosures

JRF has received research support from the Lawson Foundation,the TheraBand� Academy and StepsCount/Diabeaters Inc. andhonoraria from the Canadian Diabetes Association. CAS hasreceived research support from the Lawson Foundation and hon-oraria from the Canadian Diabetes Association.

Author Contributions

JRF designed the study, researched the data, contributed to thediscussion and wrote and edited the manuscript. CAS designed thestudy, researched the data, contributed to the discussion and editedthe manuscript. BB and SM researched the data and contributed tothe discussion. KLD researched the data, contributed to thediscussion and contributed to the manuscript. PD contributed tothe discussion and edited the manuscript.

References

1. Colberg SR, Sigal RJ, Fernhall B, et al. Exercise and type 2 diabetes. TheAmerican College of Sports Medicine and the American Diabetes Association:Joint position statement. Diabetes Care 2010;33:2692e6.

2. Sigal RJ, Armstrong MJ, Colby P, et al. Physical activity and diabetes. CanJ Diabetes 2013;37(Suppl):40e4.

3. Chudyk A, Petrella RJ. Effects of exercise on cardiovascular risk factors in type 2diabetes: A meta-analysis. Diabetes Care 2011;34:1228e37.

4. Snowling NJ, Hopkins WG. Effects of different modes of exercise training onglucose control and risk factors for complications in type 2 diabetic patients: Ameta-analysis. Diabetes Care 2006;29:2518e27.

5. Gillies CL, Abrams KR, Lambert PC, et al. Pharmacologic and lifestyleinterventions to prevent or delay type 2 diabetes in people with impairedglucose tolerance: Systematic review and meta analysis. BMJ 2007;344:299.

6. Nicolucci A, Balducci S, Cardelli P, et al. Improvement of quality of life withsupervised exercise training in subjects with type 2 diabetes mellitus. ArchIntern Med 2011;171:1951e3.

7. Nicolucci A, Balducci S, Cardelli P, et al. Relationship of exercise volume toimprovements of quality of life with supervised exercise training in patientswith type 2 diabetes in a randomized controlled trial: The Italian Diabetes andExercise Study (IDES). Diabetologia 2012;55:579e88.

8. Dubé JJ, Allison KF, Rousson V, et al. Exercise dose and insulin sensitivity:Relevance for diabetes prevention. Med Sci Sports Exerc 2012;44:793e9.

9. Sigal RJ, Kenny GP, Boule NG, et al. Effects of aerobic training, resistancetraining, or both on glycemic control in type 2 diabetes: A randomized trial.Ann Intern Med 2007;147:357e69.

10. Umpierre D, Ribeiro PA, Schaan BD, Ribeiro JP. Volume of supervised exercisetraining impacts glycaemic control in patients with type 2 diabetes: A sys-tematic review with meta-regression analysis. Diabetologia 2013;56:242e51.

11. Forbes CC, Plotnikoff RC, Courneya KS, Boulé NG. Physical activity preferencesand type 2 diabetes: Exploring demographic, cognitive, and behavioral dif-ferences. Diabetes Educ 2010;36:801e15.

12. Plotnikoff RC, Taylor LM, Wilson PM, et al. Factors associated with physicalactivity in Canadian adults with diabetes. Med Sci Sports Exerc 2006;38:1526e34.

13. Plotnikoff RC, Johnson ST, Loucaides CA, et al. Population-based estimates ofphysical activity for adults with type 2 diabetes: A cautionary tale of potentialconfounding by weight status. J Obes 2011;2011:561432.

14. Plotnikoff RC. Physical activity in the management of diabetes: Population-based perspectives and strategies. Can J Diabetes 2006;30:52e62.

15. McManus R. Time for action: A Canadian proposal for primary prevention oftype 2 diabetes mellitus. Can J Diabetes 2012;36:44e9.

16. Umpierre D, Ribeiro PA, Kramer CK, et al. Physical activity advice only orstructured exercise training and association with HbA1c levels in type 2 dia-betes: A systematic review and meta-analysis. JAMA 2011;305:1790e9.

17. Vuori IM, Lavie CJ, Blair SN. Physical activity promotion in the health caresystem. Mayo Clin Proc 2013;88:1446e61.

18. Kirk A, Mutrie N, MacIntyre P, Fisher M. Effects of a 12-month physical activitycounselling intervention on glycaemic control and on the status of cardiovas-cular risk factors in people with type 2 diabetes. Diabetologia 2004;47:821e32.

19. Garrett S, Elley CR, Rose SB, et al. Are physical activity interventions in primarycare and the community cost effective? A systematic review of the evidence. BrJ Gen Pract 2011;61:125e33.

20. Elley CR, Kerse N, Arroll B, Robinson E. Effectiveness of counselling patients onphysical activity in general practice: Cluster randomised controlled trial. BMJ2003;326:793.

21. Fortier MS, HoggW, O’Sullivan TL, et al. Impact of integrating a physical activitycounsellor into the primary health care team: Physical activity and healthoutcomes of the Physical Activity Counselling randomized control trial. ApplPhysiol Nutr Met 2011;36:503e14.

22. Dillman CJ, Shields CA, Fowles JR, et al. Including physical activity exercise indiabetes management: Diabetes educators’ perceptions of their own abilitiesthe abilities of their patients. Can J Diabetes 2010;34:218e26.

23. Gornall A, Lévesque L, Sigal R. A pilot study of physical activity educationdelivery in Diabetes Education Centres in Ontario. Can J Diabetes 2008;32:123e30.

24. Fowles JR, Murphy RJL, Shields CA, Fryia SA. Diabetes Physical Activity andExercise Toolkit. Halifax: Diabetes Care Program of Nova Scotia, 2008.

25. Shields CA, Fowles JR, Dunbar P, et al. Increasing diabetes educators’ confidencein exercise counselling: The effectiveness of the “physical activity and exercisetoolkit” training intervention. Can J Diabetes 2013;37:381e7.

26. Fowles JR, Shields CA, Murphy RJL, Durant M. Building competency in diabeteseducation. Vol 3. Physical activity and exercise. Toronto: Canadian DiabetesAssociation, 2012.

27. Reason P, Bradbury H. Handbook of action research. 2nd ed. London, UK: Sage;2007.

28. Estabrooks PA, Gyurscsik NC. Evaluating the impact of behavioural in-terventions that target physical activity: Issues of generalizability and publichealth. Psych Sport Exerc 2003;4:41e55.

29. Craig CL, Marshall AL, Sjostrom M, et al. International Physical Activity Ques-tionnaire: 12-country reliability and validity. Med Sci Sports Exerc 2003;35:1381e95.

30. Wen CP, Wai JP, Tsai MK, et al. Minimum of physical activity for reducedmortality and extended life expectancy: A prospective cohort study. Lancet2011;378:1244e53.

31. Blanchard CM, Fortier M, Sweet S, et al. Explaining physical activity levels froma self-efficacy perspective: The Physical Activity Counseling Trial. Ann BehavMed 2007;34:323e8.

32. Rzewnicki R, Auweele YV, Bourdeaudhuij ID. Addressing overreporting on theInternational Physical Activity Questionnaire (IPAQ) telephone survey with apopulation sample. Publ Health Nutr 2003;6:299e305.

33. Plotnikoff RC, Trinh L, Courneya KS, et al. Predictors of physical activity inadults with type 2 diabetes. Am J Health Behav 2011;35:359e70.

34. Gray E, Shields CS, Fowles JR, Oh P. An uphill climb: Building competency inphysical activity and exercise in diabetes education. Insight from a nationalinitiative. Can J Diabetes 2013;37(Suppl 4):19.

35. Sweet SN, Fortier MS, Guerin E, et al. Understanding physical activity in adultswith type 2 diabetes after completing an exercise interventional trial:A mediation model of self-efficacy and autonomous motivation.2009;14:419e429.