A Daily Physical Activity and Diet Intervention for Individuals with Type 2 Diabetes Mellitus :

communities because walking in rural areas is used as a means of transport (Mbanya 2009). Type 2 Diabetes Mellitus commonly occurs in persons who are obese and insulin­resistant in combination with impaired beta cell function (Haslett et al 2002). Little time is spent on patient care and education because of inadequate staff­ ing in South Africa (Whiting et al 2003). Clinical observations showed that the risk factors of the majority of patients attending Diabetes Outpatient Clinic at the Steve Biko Hospital were not well­controlled. The question asked was how to improve self­management and the clinical outcomes, taking into account the complexity of diabetes­related com­ Introduction South Africa is one of the five countries in Sub­Saharan Africa with the highest number of people affected by diabetes (Mbanya 2009). Urbanisation, a seden­ tary lifestyle and increasing obesity due to westernization and changed eating habits have been identified as independ­ ent risk factors for diabetes in the South African population (Levitt et al 1999). The South African Demographic and Health Survey (2003) established that 63% of women and 48% of men were inactive and that the prevalence of inac­ tivity was higher in the urban than in the non­urban communities (Department of Health 2003). Physical activity, mainly consisting of walking, is significantly less in urban populations than in rural Correspondence to: Prof AJ van Rooijen Department of Physiotherapy P O Box 667 Pretoria, Gauteng South Africa 0002 Email: Tania.VanRooijen@up.ac.za AbSTrAcT: Urbanisation, a sedentary lifestyle and increasing obesity due to westernization and changed eating habits have been identified as independent risk factors for diabetes in the South African population. To establish the effectiveness of a daily walk and diet education intervention program. A randomized controlled trial was performed. The study population consisted of men and women of all races, ages 40 to 65 with Type 2 Diabetes Mellitus (DM) of duration at least one year attending the Steve Biko Diabetes Outpatient clinic. Patients of all weights were considered. Patients who had an HbA1c > 8 – 9.5% were included in the sample group. Four weekly group classes consisted of education considered essential for ongoing nutrition self-management and physical activity. Yamax pedometers and walk prescriptions based on the average number of steps walked in three days were used. The participants had a follow-up assessment at 16 weeks and one year. The intervention and control groups were compared with respect to changes from baseline, using analysis of covariance (ANCOVA) with baseline values as covariates. The difference between the intervention and control groups in the change in HbA1c from the baseline was significant at the 16-week follow-up assessment (p=0.041) and in the total cholesterol and LDL-cholesterol at the one-year follow-up assessment (p =0.047; p =0.014). These results suggest that HbA1c can be improved over a period of four months. More frequent contact with the patients is necessary.


Introduction
South Africa is one of the five countries in SubSaharan Africa with the highest number of people affected by diabetes (Mbanya 2009).Urbanisation, a seden tary lifestyle and increasing obesity due to westernization and changed eating habits have been identified as independ ent risk factors for diabetes in the South African population (Levitt et al 1999).
The South African Demographic and Health Survey (2003) established that 63% of women and 48% of men were inactive and that the prevalence of inac tivity was higher in the urban than in the nonurban communities (Department of Health 2003).Physical activity, mainly consisting of walking, is significantly less in urban populations than in rural AbSTrAcT: Urbanisation, a sedentary lifestyle and increasing obesity due to westernization and changed eating habits have been identified as independent risk factors for diabetes in the South African population.To establish the effectiveness of a daily walk and diet education intervention program.A randomized controlled trial was performed.The study population consisted of men and women of all races, ages 40 to 65 with Type 2 Diabetes Mellitus (DM) of duration at least one year attending the Steve Biko Diabetes Outpatient clinic.Patients of all weights were considered.Patients who had an HbA 1c > 8 -9.5% were included in the sample group.Four weekly group classes consisted of education considered essential for ongoing nutrition self-management and physical activity.Yamax pedometers and walk prescriptions based on the average number of steps walked in three days were used.The participants had a follow-up assessment at 16 weeks and one year.The intervention and control groups were compared with respect to changes from baseline, using analysis of covariance (ANCOVA) with baseline values as covariates.The difference between the intervention and control groups in the change in HbA 1c from the baseline was significant at the 16-week follow-up assessment (p=0.041) and in the total cholesterol and LDL-cholesterol at the one-year follow-up assessment (p =0.047; p =0.014).These results suggest that HbA 1c can be improved over a period of four months.More frequent contact with the patients is necessary.

Van Rooijen Agatha J, PhD 1
Viviers Christa M, MSc 1 Becker Piet J, PhD Boule et al ( 2001) did a metaanalysis of controlled clinical trials on the effects of exercise on glycaemic control and weight in T2DM patients.A significant reduction of 0.66% in HbA 1c after aerobic exercise of 8 or more weeks was reported.
Sigal et al ( 2004) suggested that exer cise should be performed at least 3 days per week and that there should not be more than 2 days between exercise bouts.This recommendation was included in the American Diabetes Association's recommendations for physical activity.
Di Loreto et al ( 2005) have shown that physical activity is an effective costsaving tool in the care of T2DM in their study on the longterm impact of different amounts of physical activity.Brisk walking was the most common form of leisure time physical activity performed by their participants.The use of pedometers is a relative inexpensive and userfriendly way to provide goals for patients in steps per day.Pedometers have been shown to be valid for assessing physical activity in obese individuals by Swartz et al ( 2003).They have shown that pedometers correlate (r =0.800.93) with more expensive accelerometers in controlled and field conditions.
More recently Yates et al ( 2009) did a review on the role of physical acti vity in the management of T2DM.They concluded that cost effective ways of increasing physical activity in persons with T2DM are investigated such as structured education to improve self management by these individuals.
Taylor JD et al (2009) investigated the impact of physiotherapydirected exer cise counseling in persons with T2DM.They reported that improvement in exer cise outcomes comparable to those of supervised exercise could be obtained.
The aim of our study was to establish the effectiveness of a costeffective daily walk and diet education inter vention program in T2DM patients, attending the Steve Biko Diabetes Outpatient clinic.1998); measurement of the blood pres sure (Bailey et al 1993)

Randomization
Participants were randomized by means of block randomization (http:www.randomisation.com)Allocation con cealment was ensured by means of sequentially numbered, sealed, opaque envelopes.The principal researcher assigned the participants to their groups.The research assistants and laboratory personnel measuring the primary and secondary outcomes were blinded to group assignment.The flow of the par ticipants is illustrated in Figure 1.

Intervention group
The participants in the intervention group received usual care plus the study intervention.Participants attended four weekly group sessions.A staged approach to education based on the SkilledHelper Model was used to empower the par ticipants with knowledge and skills necessary for medical nutrition therapy with the ultimate goal to help participants to become better at helping themselves in their everyday lives (Frost 2003).A diagram of the model is demonstrated in Figure 2.
Topics included: planning, purchasing and preparing food and meals; sources of carbohydrate, protein and fat; reading of food nutrition labels; grocery shopping guidelines; modifying fat intake; use of sugarcontaining foods; diabetic foods and sweeteners, as well as the Glycaemic

The setting and participants
The study was conducted at the Diabetes Outpatient Clinic of the Steve Biko Academic Hospital in Gauteng Province, South Africa.This clinic is managed at a tertiary level and most patients have diabetesrelated complications and are from lowincome communities.
The study population consisted of men and women of all races, ages 40 to 65 years with T2DM for at least one year.Patients of all weights and who had an HbA 1c 1 > 8 -9.5% were included in the sample group.Patients on diet therapy only, on glucoselowering agents and/or insulin, were included.
Participants were screened for chest pain on effort, previous myocardial infarction, intermittent claudication, cerebrovascular incidents, severe arthri tis, retinopathy and general health by means of the London School of Hygiene Cardiovascular Questionnaire (Rose et al 1982).Cases where there were uncer tainties regarding the aforementioned conditions, were referred to the attend ing specialist physician for clinical evaluation and advice about inclusion in the sample group.Participants, who were illiterate, were excluded, because completion of the Diabetes Knowledge questionnaire was required at all assess ment points in the study.Each subject in this group received a Yamax SW200 pedometer©.Participants had to wear the sealed pedometers from the time they woke up until they went to bed at night for two week days and one weekend day to establish the baseline average number of steps for each patient.Personal goal setting was calculated on the average of the threeday step record (TudorLocke et al 2005).The aim was to motivate participants to walk at least 10 000 steps per day five days per week.Participants received a logbook to document daily steps.

Ethics
Our study was unique in the sense that all participants were examined by a podiatrist.The examination included vascular, neurological, dermatological and orthopedic assessments.Therapeutic insoles based on the shape and plantar pressure management and the type of shoe the subject wore were supplied to the participants if necessary (Owings et al 2008).It was hoped that this inter vention would contribute to improved compliance to the walk program and comfort of the participants.
After the four group sessions, partici pants continued at home from week five of the program.They received motivating 3c -Plan (What next and when?) q q q q text messages fortnightly.Participants were also encouraged to phone or to visit the researchers at the clinic whenever they had problems or questions.
The participants had a followup assessment at 16 weeks, during which time they had to return the pedometers, but continued with their respective walking and eating programs (Wolf et al 2004).Final assessment was conducted at one year.

Control group
The control group received the usual care and was waitlisted for an intervention similar to the intervention group after the study was completed.Participants in the control group were also followed up at 16 weeks and one year.

Outcomes
The primary outcome of this study was to establish the effectiveness of a daily walk and diet intervention program to decrease HbA 1c over a period of one year in T2DM individuals aged 40 to 65 years with an HbA 1c > 89.5%.
Secondary outcomes were to establish the effect of the intervention on the blood lipid values, body mass index, and the diabetes knowledge of the participants.

Statistical methods
The intervention and control groups were compared with respect to changes from baseline, using analysis of covariance (ANCOVA) with baseline values as covariates using Stat Release 10 statistical software.Testing was done at 0.05 level of significance.Summary statistics were reported for the observed data, while confidence intervals were based on adjusted means following the ANCOVA.The 95% (CI) was based on adjusted analysis.No other analysis was performed.Parameters were analyzed individually, with the 12month analysis being of primary importance.
Each item of the Diabetes Knowledge scale was assigned a score of one for a correct response and zero for an incor rect response.The total score for the scale was calculated by summing the scores of each of the 15 items to give a potential score range of 0 to 15, with higher scores indicating better diabetes knowledge (Beeney et al 1994).
et al (2008) reported: "patients felt well cared for, better supported and more successful and confident" when partici pating in group classes.
The participants in our study did not reach a stage of selfefficacy in the self management phase of the study (Bandura 1997).Within the framework of the SkilledHelper Model selfdetermination and selfcontrol are essential for action.Participants who were unable to exert decisional selfcontrol complemented by protracted selfcontrol might have had an inability of manage their lives better (Frost 2003).Our sample may have been in the action stage of change but were not able to maintain the behavior after four months.This finding supports that of a study by Davis et al. (2009), who compared a oneyear dietary interven tion of a low carbohydrate to a lowfat diet on weight and glycaemic control and reported the greatest reduction in HbA 1c within the first three months of their study.
The twoweekly text messages were not intensive enough for the interven tion group to sustain their motivation to maintain the change in physical acti vity and diet.Schillinger et al. ( 2009) conducted a oneyear study on self management and reported that although glycaemic control improved and BMI changed, these changes were not sig nificant enough in comparison with the changes seen in usualcare patients.
The difference from baseline in the total cholesterol and LDLcholesterol was significant at the one year followup assessment (p =0.047; p =0.014).This is in agreement with Miller et al (2002) who reported that patients who followed an intensive diabetes education program had optimal total cholesterol values at post test.Small weight losses result in improved LDL and HDLcholesterol levels even when the ideal BMI is not achieved (Krummel 2008).Moderateintensity physical activity of 6090 minutes per day plays a major role in the maintenance of weight loss (Sigal et al 2004).
The difference between the baseline and both the 16week and one year followup assessments in terms of the diabetes knowledge of the two groups were significant (p <0.001; p<0.002).

Baseline data and numbers analyzed
The baseline and final analysis bio graphic characteristics of the two groups were comparable.Two parti cipants moved to other cities in South Africa and two participants who were previously unemployed for extended periods of time started new fulltime jobs.In the control group one subject passed away.Two other participants dropped out because of socioeconomic circumstances.The baseline biographic characteristics of all participants in the trial, those lost to followup, and those remaining in the trial to the end are demonstrated in Table1.
The difference in means between the groups was significant for HbA 1c at 16 weeks but not after one year.The decrease in the means of Total cholesterol and LDLcholesterol was significantly lower between the groups after one year.The intervention group's scores on diabetes knowledge improved signifi cantly at 16weeks and 1year.
The change in the means from baseline in the primary and secondary outcomes is demonstrated in Table 2.

Adverse events
No adverse events or sideeffects were reported.

Discussion
The aim of this study was to establish the effectiveness of a costeffective daily walk and diet education intervention program on HbA 1c in T2DM patients.
The difference in means between the intervention and control groups in the change in HbA 1c from the baseline was significant at the 16week follow up assessment (p=0.041).The differ ence between the groups could not be maintained at the one year followup assessment.
The decrease in the HbA 1c during the first four months could firstly be con tributed to the group visits during this period.The interaction between parti cipants contributed to sharedproblem solving and encouraged the exchange of advice.Furthermore, participants could report problems with the walk program or feet, which were then addressed by the researchers and the podiatrist.Davis 90( 18) 80( 16)

Conclusions
The effectiveness of a daily walk and diet education on HbA 1c, blood lipid levels, BMI and diabetes knowledge in Type 2 T2DM patients was investigated in this study.The intervention consisted of four weeks of weekly contact with the research team and a selfmanagement phase with textmessage contact fort nightly up to one year.The results sug gests that HbA 1c can be improved over a period of four months in a sample of patients with diabetesrelated compli cations and who are from a poor socio economic background.More frequent contact with the patients and individu alized nutrition prescriptions during the selfmanagement phase may be more effective.et al 2001;Thomas et al 2006).This finding once again emphasizes the role of the physiotherapist in the manage ment of T2DM patients.South Africa has few dedicated walk ing paths that may encourage people to walk more.The safety and security situa tion in the various neighborhoods where our sample resides may have contributed to limit walking from the intervention group (Department of Health 2007).Costeffective ways of physical activity in this patient population is needed.One such possibility is resistance training in combination with aerobic exercise.A recent metaanalysis showed signifi cant improvements in HbA1c and blood pressure with the aforementioned exer cise (Snowling et al 2006).

Limitations
The attrition rate was high due to circum stances that we could not control.The results should be viewed as a pilot study, because it cannot be generalized to popu lations other than the Steve Biko patients.

Recommendations for clinical practice
The relapse of patients in the self management phase suggests the need for more frequent facetoface contact with the participants when prescribing physical activity.Furthermore 6090 The protocol was approved by the Ethics Committee of the Faculty of Health Sciences of the University of Pretoria (Number 17/2005), as well as the Chief Executive Officer of the Steve Biko Hospital and was registered as a clini cal trial with the National Department of Health (Number DOH2701071390).Informed consent was obtained.Baseline evaluation Baseline evaluation consisted of: fasting Glycated Haemoglobin (HbA 1c ) and lipid blood samples (SYNCHRON LX 1 Glycated haemoglobin provides an accurate and objective measure of glycaemic control over a period of weeks to months (Haslett et al 2002) index and glycaemic load.No individu alized dietary programs were given.

Table 1 :
The baseline biographic characteristics of all participants in the trial, those lost to follow-up, and those remaining in the trial to the end.Values are percentages (numbers) unless stated otherwise.

523(-0.25;1.23)
This study was supported by the Medical Research Council of South Africa and in part by the South African Sugar Association.No conflict of interest exists.This increase in knowledge in the intervention group could however not be translated to successful sustain able weight loss and physical activity.`ThisfindingconcurswithRaynoretal(2008)who reported that a diet with an increased consumption of fruit and vegetables without a prescription to decrease energy intake resulted in a small weight loss that could not be maintained.People with Type 2 T2DM might have more difficulty in losing weight due to different genetic susceptibilities(Kaput  et al 2007; Harris et al 2003)].A genetic ßcell defect together with a reduction in insulin release could lead potentially to an increase in body weight(Ioannides  2008).The obesity phenotype of parti cipants might furthermore have played a role in their response to regular exercise(Sasai et al 2009).Several metaanalyses have shown that physical activity inter ventions improve glycaemic control independently of weight loss.(Boule