(PDF) Exploring Additional Intention Predictors within Dietary Behavior among Type 2 Diabetes

World Academy of Science, Engineering and Technology International Journal of Psychological and Behavioral Sciences Vol:7, No:3, 2013 Exploring Additional Intention Predictors within Dietary Behavior among Type 2 Diabetes D. O. Omondi, M. K. Walingo, G. M. Mbagaya reduce both the incidence and progression of Type 2 diabetes Abstract—Objective: This study explored the possibility of [1]-[4]. However, a major challenge is identifying the best integrating Health Belief Concepts as additional predictors of approaches to effectively promote a healthy diet among people intention to adopt a recommended diet-category within the Theory of with Type 2 diabetes. Planned Behavior (TPB). Methods: The study adopted a Sequential The current communication methodologies used to motivate Exploratory Mixed Methods approach. Qualitative data were generated on attitude, subjective norm, perceived behavioral control Type 2 diabetics towards healthy eating in most clinics in and perceptions on predetermined diet-categories including perceived Kenya neglect patient’s cognitive related factors [5]. Most International Science Index, Psychological and Behavioral Sciences Vol:7, No:3, 2013 waset.org/Publication/9996982 susceptibility, perceived benefits, perceived severity and cues to approaches focus on patients as passive players in the decision action. Synthesis of qualitative data was done using constant making process with regard to healthy food choice. Cognitive comparative approach during phase 1. A survey tool developed from related theories such as the Theory of Planned Behavior [6] qualitative results was used to collect information on the same and the Health Belief Model [7] have great potential to concepts across 237 legible Type 2 diabetics. Data analysis included use of Structural Equation Modeling in Analysis of Moment identify key intrinsic modifiable patient-related factors that Structures to explore the possibility of including perceived influence healthy dietary practices among Type 2 diabetics. susceptibility, perceived benefits, perceived severity and cues to The Theory of Planned Behavior (TPB) puts emphasis on action as additional intention predictors in a single nested model. attitudes, subjective norms and perceived behavioral control as Results: Two models-one nested based on the traditional TPB model immediate predictors of behavioral intentions. Behavioral {χ2=223.3, df = 77, p = .02, χ2/df = 2.9; TLI = .93; CFI =.91; intention in turn predicts behavior. The Health Belief Model, RMSEA (90CI) = .090(.039, .146)} and the newly proposed Planned Behavior Health Belief Model (PBHB) {χ2 = 743.47, df = 301, p = on the other hand, identifies perceived susceptibility, .019; TLI = .90; CFI=.91; RMSEA (90CI) = .079(.031, .14)} passed perceived severity, perceived benefits, perceived barriers, cues the goodness of fit tests based on common fit indicators used. to action, and self efficacy as key predictors of behavior. Conclusion: The newly developed PBHB Model ranked higher than Previous empirical evidence found out that attitude, the traditional TPB model with reference made to chi-square ratios subjective norm and perceived behavioral control only (PBHB: χ2/df = 2.47; p=0.19 against TPB: χ2/df = 2.9, p=0.02). The accounted for 58% of the variance on intention [8]. This lack integrated model can be used to motivate Type 2 diabetics towards healthy eating. of full accountability of intention by key variables within the TBP is shared in work of Blue et al. [9]. It therefore implies Keywords—Theory, intention, predictors, mixed methods design. that some additional proxy factors not specified in the model may be important drivers of behavioral intentions. The Health I. INTRODUCTION Belief Model outlines five key factors as direct and immediate predictors of behavior and excludes the mediating role of T YPE 2 diabetes remains one of the leading health problems highly prevalent in many nations both in the developing and developed world. This condition has negative intention within the framework. It relates largely to the cognitive factors predisposing a person to health behavior, concluding with a belief in one's self-efficacy for the behavior. impact on individual health status and if not detected and We noted that the model leaves a gap to be explained by prevented early could lead to serious health complications factors enabling and reinforcing one's behavior, in this case such as blurred vision, amputation, fatigue and death. The intention. These factors become increasingly important when burden of disease stemming from Type 2 diabetes now calls the model is used to explain and predict more complex for multidisciplinary interventional approaches for effective lifestyle behaviors that need to be maintained over a lifetime. management. Successful management of this condition can be A systematic, quantitative review of studies that had applied enhanced through lifestyle change interventions. Healthy the Health Belief Model among adults into the late 1980s dietary behavior is one intervention that has had promising found it lacking in consistent predictive power for many kinds results. Five randomized trials have shown a healthy diet can of health behavior, probably because its scope is limited to predisposing factors [10]. One study that specifically D. O. Omondi is Lecturer at the School of Public Health & Community Development, Maseno University, P. O. Box 333, Maseno, Kenya (e-mail: compared its predictive power with other models found that it

). accounted for a smaller proportion of the variance in diet, M. K. Walingo is Professor (PhD) in the School of Public Health & exercise, and smoking behaviors than did the Theory of Community Development, Maseno University, P. O. Box 333, Maseno, Planned Behavior [11]. Based on these two parallel Kenya (e-mail:

). G. M. Mbagaya is Associate Professor (PhD) in the Department of Family weaknesses concerning performance of the two theories & Consumer Sciences, Moi University, P. O. Box 1125, Eldoret, Kenya (e- applicable to individual cases, we test the utility of the mail:

). International Scholarly and Scientific Research & Innovation 7(3) 2013 807 scholar.waset.org/1307-6892/9996982 World Academy of Science, Engineering and Technology International Journal of Psychological and Behavioral Sciences Vol:7, No:3, 2013 Planned Behavior Health Belief Theory which builds on two population of 400 patients forming the sampling frame for the theories with an aim of closing the pre-intention gap within entire dietary behavior cohort. Simple random sampling the TPB. The proposed theory is a hybrid generation of the technique was used to select individual participants out of 400 Theory of Planned Behavior with traditional health belief legible participants using a table of random numbers. We concepts integrated as competing predictors of intention. ensured that this group of patients was the one involved during Emphasis was put on quantitative assessment of a model built the past two month of qualitative study to ensure homogeneity on the newly formulated theory based on a hypothesis that: “a and consistency. nested model structured from planned behavior health belief D. Data Collection Tools and Measurement theory fits the data on dietary behavior acceptably among Type 2 diabetics”. The survey instruments detail focused on measuring the key concepts intended to be nested into models as key latent II. METHODS variables. The latent variables from the traditional concepts in the TPB model includes attitude towards dietary behavior, A. Setting of the Study subjective norm towards dietary behavior, perceived The study was conducted at Kisii Level-V Hospital between behavioral control towards dietary behavior, intention towards June and November, 2009. This is a provincial referral dietary behavior and dietary behavior itself. Latent variables International Science Index, Psychological and Behavioral Sciences Vol:7, No:3, 2013 waset.org/Publication/9996982 hospital and the largest in Kisii County in Kenya. Until the from the Health Belief Model intended to be included in the period of data collection the clinic was operated by one newly developed model includes perceived susceptibility, consultant doctor who was not a residence of the Hospital, five perceived benefits, perceived severity and cues to action. residence doctors, six clinical officers, four nurses and one Measurements of these variables were done using a 7-point nutritionist. Diabetic management clinic are held every likert scale and more details are in Table I. Tuesdays and Fridays. During each clinic patients arrive at 8.30 am. Health education sessions are conducted by a health professional as patients wait for clinical processes to begin. The patients then undergo various medical tests including blood to determine sugar levels, followed by receipt of medical prescriptions. After receiving their prescriptions, the patients undergo counseling conducted by the nurse in charge then proceed to collect drug and other supplies from the pharmacy. This means that our interviews were ethically scheduled one hour before 9.00 am and after the all the clinical processes following initial identification of the patient to be interviewed. B. Research Design This study used a three-phase sequential exploratory mixed methods approach. We first gathered qualitative data during Phase 1 using Focus Group Discussions (FGDs). Transcripts of the FGDs were analyzed using constant comparative approach in grounded theory analysis. The approach involved collecting data using FGD guide and analyzing the transcripts collected, then collecting more data and analyzing. In each case attempts were made to identify main categories measuring key concepts in the original and proposed theories until saturation. The techniques used included open, axial and selective coding process. The findings from the qualitative component were then used to develop a quantitative survey tool during Phase 2. The survey tool developed was then used to collect relevant data for an exploratory study during Phase 3 [12], [13]. In this article we only focus on the results of the quantitative survey. C. Study Population and Sampling The desired sample size was computed using the Creative Research Systems [14] formula, which has been used in several studies [15], [16]. A sample of 217 participants was computed as the minimum sample size given a finite International Scholarly and Scientific Research & Innovation 7(3) 2013 808 scholar.waset.org/1307-6892/9996982 World Academy of Science, Engineering and Technology International Journal of Psychological and Behavioral Sciences Vol:7, No:3, 2013 TABLE I MEASUREMENT OF KEY CONCEPTS Concepts Measurement criteria Dietary Measured on the by the number of times in a week a patient consumes foods in high fat diet (Beef, chicken with skin, egg yolk, fried potato behavior chips, roast meat, fatty meats, chapatti, and cream) , high sugar diet (Sweet potato, Irish potato, white rice, white rice, white sugar, soda and sweet soft drinks, cakes, ice cream, chocolate, sugared beverage, jam, glucose, honey, arrow roots and boiled maize) and recommended diet (Whole grain rice, green vegetables, low fat milk, chicken without skin, fish, beans, green grams, carrots, minnow fish (omena), sweet banana, pineapple and mangoes) categories as identified during FGDs. A score of 8 was given to patients who consumed food from recommended diet category, no foods from high fat diet and high sugar diet. (e.g. How frequently do you consume foods from high fat diet (Diet class-1) category) Attitude Computed by summing up the product of five salient belief strengths and corresponding evaluation weights for attitude towards “high fat diet” (attitude-1), “high sugar diet” (attitude-1) and “recommended diet” (attitude-3). (e.g. Consuming class 1 foods make you go into a comma-1=Strongly disagree, 7=strongly agree) Subjective Computed by summing up the product of five normative belief strengths and corresponding motivation to comply weights for subjective norm norm towards “high fat diet” (subjective norm-1), “high sugar diet” (subjective norm-2) and “recommended diet” (subjective norm-3). (e.g. My doctor/nurse/nutritionist think that ________________ consume of fruits and vegetables, fish, poultry without skin, whole wheat flour, maize flour and unpolished rice grain when diabetic- 1=I should, 7=I should not) Perceived Computed by finding the product between control belief strength (barriers) and control power weight, for perceived behavioral control behavior towards “high sugar diet” (Perceived behavioral control-1), “high sugar diet” (perceived behavioral control-2) and “recommended diet” control (perceived behavioral control-3). In this study only barriers emerged as factors which hinder following appropriate dietary recommendation. International Science Index, Psychological and Behavioral Sciences Vol:7, No:3, 2013 waset.org/Publication/9996982 (e.g. How often do you encounter factors such as social influence, unavailability of foods, conveniences among others that prevent you from reducing consumption of food items rich in fat such as red meat, fried potatoes among others?-1=very rarely, 2=very frequently) Intention Measured by the degree of willingness to reduce fat and sugar intake or increase consumption of recommended diet (vegetables and fruits) by half. (e.g Intend to reduce the intake of foods including red meat, fried potatoes among others by half.-1=not at all, 7=very much) Perceived Measured the perceived level of risk the participants attached to negative outcome of their conditions in relation to the three diet categories. susceptibility Risk indicators included elevated blood sugar, blurred vision and loss of strength. (e.g. Failure to reduce intake of diet rich in fats increases the chances of experiencing elevated blood sugar levels (hyperglycemia), blurred vision and loss of strength-1=totally disagree, 7=totally agree) Perceived Measured the participants’ perception of how severe their conditions could be if they failed to follow dietary recommendations. Severity severity indicators included amputation, going into a comma and skin irritation. (e.g. Adhering to the recommended diet consistently maintains blood sugar level within normal range-1=totally disagree, 7=totally agree) Perceived Measured the participants’ perception of the benefits they could get if they followed appropriate or recommended diet. Benefit indicators benefits included improved strength and work productivity, risk avoiding risks and severe levels of disease and maintenance of blood sugar level. (e.g. If I don’t reduce intake of diet rich in fats I risk being amputated, going into a comma and suffering from skin irritation.-1=totally disagree, 7=totally disagree) Cues to Focused on whether the participants were aware of materials and processes that promote appropriate dietary practice. Cues to action action indicators included reading materials (booklets, magazines among others), visual materials (posters, television among others) and diabetic education day. (e.g. There are enough reading materials (booklets, magazine among others) explaining the relationship between diet and Type 2 diabetes in this clinic.-1=totally disagree, 7=totally agree) E. Data Collection Tools and Measurement III. RESULTS Preliminary analysis was done to statistically test for the Reliability analyses: It is a common practice that survey reliability of the survey tool using Cronbach’s alpha as a items measuring same concept need to be reliable before measure for internal consistency. Descriptive statistics were nesting a model for SEM. We assessed the internal also reported to show the range of measurements and test for consistency of the questionnaire using Cronbach’s alpha normality of data since we intended to use Structural Equation reliability coefficient. Cronbach’s alpha reliability coefficient Modelling (SEM). SEM is based on the assumption that well normally ranges between 0 and 1. The closer Cronbach’s fitting data should be normally distributed. Finally we alpha coefficient is to 1.0 the greater the internal consistency performed SEM in AMOS 7.0 using Maximum Likelihood of the items in the scale. George and Mallery [20] rules of (ML) estimation during final analysis to test for the godness of thumb was used to classify the Cronbach’s alpha coefficients fit of the models nested. During SEM analysis we included generated. These rules of thumb provide the following: “> .9 – Confirmatoty Fit Index (CFI), Turkey Lewis Index (TLI), Excellent, > .8 – Good, > .7 – Acceptable, > .6 – Root Mean Square Error of Approximation (RMSEA), Chi- Questionable, > .5 – Poor, and < .5 – Unacceptable” (p. 231). square P-Value, Relative Chi-square and Hoelter’s critical N All the Cronbach’s alpha exceeded the 0.5 threshold criteria as fit parameters. CFI and TLI values greater than 0.90 were we set except for dietary behavior. Lower Cronbach’s alpha considered satisfactory [17]. RMSEA less than 0.08 was also level for dietary behavior suggested the possibility of varied considered satisfactory [18]. Relative chi-square was dietary practice being displayed and these categories were considered fit under 3:1 range [19] and more accuate when mutually exclusive. Table II shows internal consistency closer but not less than 1. Hoelter’s critical N was considered reliability coefficients for all the grouped factors measuring low below 75 cases and bootsrap samples were set at 200 each concept within the questionnaire. (Garson, 2009). ALL presentations were made in tables and nested model. International Scholarly and Scientific Research & Innovation 7(3) 2013 809 scholar.waset.org/1307-6892/9996982 World Academy of Science, Engineering and Technology International Journal of Psychological and Behavioral Sciences Vol:7, No:3, 2013 TABLE II perceived behavioral control measures, attitude-1 and RELIABILITY TEST FOR DIETARY QUESTIONNAIRE perceived susceptibility-3. Number of Cronbach’s Concepts measurement alpha (main TABLE III items survey, (n=237) ENDOGENOUS AND EXOGENOUS VARIABLES Exogenous Variables Dietary behavior measures 3 0.387 Endogenous Variables (Unobserved) Indirect attitude 3 0.570 Observed Perceived Benefits Indirect subjective norm 3 0.940 Attitude-1 (A1) Cues to Action Indirect Perceived Behavioral 3 0.590 Attitude-2 (A2) Perceived Susceptibility Control Attitude-3 (A3) Perceived Severity Dietary intention 3 0.587 Perceived susceptibility-1 (PS1) PBC Pre-intention moderators Perceived susceptibility-2 (PS2) Attitude Perceived susceptibility 3 0.514 Perceived susceptibility-3 (PS3) e3 Perceived severity 3 0.688 Perceived severity-1 (SE1) e2 Perceived benefits 3 0.844 Perceived severity-2 (SE2) e1 Cues to action 3 0.713 Perceived severity-3(SE3) c6 International Science Index, Psychological and Behavioral Sciences Vol:7, No:3, 2013 waset.org/Publication/9996982 Subjective norm-1 (SN1) c5 Exploring models using SEM: In order to address the core Subjective norm-2 (SN2) c4 objective of this study, we intended to dwell more on the Subjective norm-3 (SN3) c3 newly generated model and do a comparison with the PBC-1 (PC1) c2 traditional TPB model outcome. To perform this structural PBC-2 (PC2) c1 analysis, we tested the hypothesis that including perceived PBC-3 (PC3) Subjective Norm Intention-1 (IN1) e6 susceptibility, perceived severity, perceived benefits and cues Intention -2 (IN2) e5 to action as additional intention predictors within the TPB Intention -3 (IN3) e4 model applied to dietary behavior (the hybrid planned Cues to action-1 (CA1) e13 behavior health belief theory), would enhance the variance Cues to action-2 (CA2) e14 accounted for in predicting dietary behavioral intentions. All Cues to action-3 (CA3) e15 items in the model were accepted following the initial Perceived benefits-1 (PB1) c12 Confirmatory Factor Analysis (CFA). CFA is a measurement Perceived benefits-2 (PB2) c11 model whose purpose is to obtain factor loading estimates of Perceived benefits-3 (PB3) c10 the parameters of the model, the variances and covariance of Diet class-1 (D1) c9 the factors and the residual error variances of the observed Diet class-2 (D2) c8 variables. Usually it helps in exclusion of observed variables Diet class-3 (D3) c7 that load poorly into a model. Both item measurements Unobserved e9 analysis and measurement model analysis were performed Intention e8 using observed/unobserved endogenous and unobserved Dietary Behavior e7 exogenous variables (Table III). other1 All variable cases were subjected to both univariate and other2 e10 multivariate screening to test for the normality of the data for e11 each variable observed before fitting the model. The means e12 and standard deviations for all the measures within model are e/c= error; other=other factors 1=High fat diet 2=High sugar diet displayed (Table IV). All the measures were subjected to 3=Recommended diet skewness tests and based on the recommended ±2 range for normal distribution measures of dietary behavior were negatively skewed except for diet class-1 which appeared to be normally distributed. Measures of intention were all negatively skewed. All measures of cues to action and perceived behavioral control were normally distributed, while subjective norm measures, perceived benefits, perceived severity appeared to be negatively skewed. Perceived susceptibility measures were negatively skewed except for perceived susceptibility-1 which was normally distributed. Attitude measures were all normally distributed. On the overall data violated normality assumption based on skewness. Kurtosis also indicated that all measures were outside the ±2 range for normal distribution except for diet class-1, and International Scholarly and Scientific Research & Innovation 7(3) 2013 810 scholar.waset.org/1307-6892/9996982 World Academy of Science, Engineering and Technology International Journal of Psychological and Behavioral Sciences Vol:7, No:3, 2013 TABLE IV {TLI = .90; CFI =.91; RMSEA (90CI) = .079(.031, .14)} also MEASUREMENT LEVEL DESCRIPTIVE STATISTICS, UNIVARIATE AND demonstrated a good model fit. Hoelter's critical N values MULTIVARIATE NORMALITY FOR THE MODEL (N= 237) suggest that the model would have been accepted at the .05 Variable skew c.r. kurtosis c.r. IN3 -3.071 -19.298 11.485 36.091 significance level with 161 cases and the upper limit of N for IN2 -4.636 -29.136 28.659 90.058 the .01 significance level is 197. Because the data violated the IN1 -3.097 -19.467 10.696 33.613 normality assumption, bootstrapped chi-square values were PC1 .279 1.754 -1.617 -5.082 also calculated and the model fits better in 200 bootstrapped PC2 .045 .285 -1.777 -5.583 samples. The Bollen-Stine p = 0.02 provided further PC3 1.070 6.722 -.489 -1.537 reassurance about the model fit. Based on the goodness of fit PB1 -5.248 -32.984 27.635 86.840 statistics an attempt was made to advance the planned PB2 -4.549 -28.591 21.067 66.203 behavior health belief theory using structural model. PB3 -3.558 -22.362 11.422 35.895 Standardized regression weights indicated that attitude was CA1 .362 2.276 -1.710 -5.373 a better predictor of knowledge (β=0.56, p<0.0), followed CA2 .629 3.953 -1.420 -4.463 subjective norm (β=0.38, p<0.05). Perceived behavioral CA3 -1.367 -8.589 .255 .800 control showed very minimal negative change on (β=-0.01, D3 -3.242 -20.378 9.942 31.242 p>0.05) intention and minimal positive change on dietary International Science Index, Psychological and Behavioral Sciences Vol:7, No:3, 2013 waset.org/Publication/9996982 D2 -2.799 -17.594 10.447 32.829 behavior (β=0.01, p>0.05). Perceived susceptibility (β=0.03, D1 -.970 -6.093 .815 2.562 p>0.05), perceived severity (β=0.02, p>0.05), perceived SN1 -1.728 -10.859 2.637 8.286 benefits (β=0.07, p>0.05) and cues to action (β=0.06, p>0.05) SN2 -2.079 -13.064 4.348 13.663 SN3 -2.098 -13.184 4.978 15.642 poorly predicted intention while intention still had a strong PSE1 -2.810 -17.661 6.981 21.938 prediction for dietary behavior (β=1.00, p<0.001). PSE2 -3.915 -24.607 15.352 48.244 PSE3 -2.205 -13.858 3.416 10.733 IV. DISCUSSION PS1 -3.049 -19.166 8.468 26.610 This study put to test the four concepts drawn from the PS2 -5.202 -32.694 28.998 91.124 Health Belief Model [7]. These concepts were included in the PS3 -1.867 -11.732 1.858 5.837 original Theory of Planned Behavior [6] in order to build a A1 -.847 -5.324 .365 1.147 new behavior model for the Type 2 diabetics. The concepts A2 -1.837 -11.548 5.800 18.225 included perceived susceptibility, perceived severity, A3 -1.688 -10.612 8.288 26.045 perceived benefits and cues to action. These concepts were Multivariate 425.543 82.774 incorporated into the TPB model to advance a new theory labeled planned behavior health belief theory (PBHB). Item level measurements were then performed for the Goodness of fit comparisons (based on relative chi-squares model due to the difference in the measurement scales. The and p-values) between the planned behavior health belief model was recursive with a df=301. Standardized regression theory and the traditional Theory of Planned Behavior ranked weights for the endogenous variables were determined and the new theory higher within dietary behavior (PBHB: χ2/df = screened. Items defining attitudes, subjective norms, perceived 2.47; p=0.19 against TPB: χ2/df = 2.9, p=0.02 under similar behavioral control, perceived susceptibility, perceived conditions). Chi-square ratio is usually the most commonly severity, perceived benefits, cues to action, intention, and used index for model comparison and the closer value is to 2 dietary behavior had very high regression weights close to the better the performance. There were also considerable 1.00. The squared multiple correlation indicated that reduction (-0.24) in the prediction power of attitude [21] predictors of subscales accounted for >90 percent except for implying these variables were possible moderators. This perceived behavioral control (PBC-3) for the recommended goodness of fit comparison leads us to the next phase of diet and cues to action-3 where the predictors accounted for 44 discussion where we now take a critical analysis of the percent and 77.8 percent of the variances respectively. contribution each of the additional intention proposed Correlations between variables in the model were strong predictors made in the model. (p<0.001) and positive except PBC3 which registered lower Health belief concepts performed rather well as predictors but significant positive correlation coefficient (p<0.01). of intention within dietary behavior domain. However, the Modification indices suggested specifying relationships variance of intention accounted for by each of the four among items within and between the scales, which suggest concepts was not significant but gain larger than zero. The multicolinearity. relationship between perceived susceptibility and health The goodness of fit statistics were statistically non- related behavior is well researched [22] but puts more significant at the .01 level but the model should be rejected at emphasis on the direct link with health behavior. However, in the .05 level (χ2 = 743.47, df = 301, p = .019, χ2/df = 2.47). this study we examined perceived susceptibility as an indirect However, the relative chi-square was under the recommended determinant of dietary behavior. Perceived susceptibility 3:1 range indicating acceptable fit after significant focused on how Type 2 diabetic patients’ view the risks modification indices were uncorrelated. Modification Index related to dietary practices and explained up to 3 percent of the was set at the customary cutoff value of 4.00. Other fit indices variance in dietary intention. This shows that Type 2 diabetic International Scholarly and Scientific Research & Innovation 7(3) 2013 811 scholar.waset.org/1307-6892/9996982 World Academy of Science, Engineering and Technology International Journal of Psychological and Behavioral Sciences Vol:7, No:3, 2013 patients always had intention whenever they perceived negatively accounted for 6 percent of the variance in dietary themselves to be at high risk. Perceived susceptibility is one of intention. The average mean score (µ=4.08±0.16) indicates the motivator for people to adopt healthier behaviors. When that most patients were undecided on whether enough perceived risk is high, individuals tend adopt healthier materials exist to explain relationship between diet and Type 2 behaviors to decrease the risk [23], [24]. In close ties with our diabetes. To some extent some patients disagreed if TVs and results, Type 2 diabetics would become more likely to follow posters were relevant to their conditions. Watching and recommended diet whenever they feel more are at risk to hearing TV or radio news stories about food borne illness and worse outcomes of their condition. On the on the contrary, reading the safe handling instructions on packages of new when the patients believe that they are not at risk at all or meat and poultry are cues to action associated with safer food- minor risk, they tend to resort to unhealthy dietary practices. handling behaviors [28]. Similarly having posters and showing Among people whose parents had or have the Type 2 diabetes, patients TV pictures relevant to Type 2 diabetes are cues to the perception of risk of developing the condition was action associated with prevention of severe conditions of the predictive of more heath-enhancing behaviors and were more disease. likely than others to engage in behaviors to control their Conclusion and international health relevance of the study: weight [25], since weight is a known risk factor to Type 2 This study has revealed that health belief concepts such as perceived susceptibility, perceived benefits, perceived severity International Science Index, Psychological and Behavioral Sciences Vol:7, No:3, 2013 waset.org/Publication/9996982 diabetes. The construct of perceived severity also referred to as and cues to action when combined with attitude, subjective perceived seriousness in some studies also showed elements of norm and perceived behavioral control as intention predictors accountability for Type 2 diabetics’ dietary intentions. It within the original theory of planned behavior improves the appeared that perceived severity positively accounted for 2 model’s competitive power. This was arrived at after percent of the variance in dietary intention. This study considering improved performance of a model nested based on revealed that while the perception of severity of a disease is the concepts within planned behavior health belief theory often based on medical information or knowledge, it may also compared to a model nested based on the traditional concepts come from beliefs a patient has about the difficulties a disease of the Theory of Planned Behavior. Even though the would create or the effects it would have on his or her life in contributions of attitude and subjective norm were still on the general [26]. For instance, some Type 2 diabetics view their lead as predictors of intention, their prediction power were condition as relatively minor ailment during initial stages. much reduced when concepts such as perceived susceptibility, When they are diagnosed with the condition, they simply walk perceived severity, perceived benefit and cues to action were to the clinic get medication and get better. However, when loaded into the model. This suggests that researchers need to their conditions worsen, they realize the seriousness of the re-focus on the pre-intention phase of the TPB model, disease and seek serious medical help. otherwise users of the TBP would assume that it is only The construct of perceived benefits focused on the patients’ attitude, subjective norm and perceived behavioral control that opinions on the value or usefulness of a new behavior in are key intention predictors yet this may not be the case. The decreasing the risk of developing severe conditions of Type 2 newly integrated model in addition to being patients’ centered diabetes. Perceived benefits performed better than other and marries key concepts from two theory and can be used concepts in the planned behavior health belief model by educate Type 2 diabetics at individual level. We believe that explaining up to 7 percent of the variance in dietary intention. implementation of this model may have a remarkable impact This implies that Type 2 diabetics tend to develop high on positive dietary behavior change and opens up a window intentions to follow recommended diet when they realize the for debate in the scientific world. benefits of healthy eating. Perceived benefit plays a greater role in the adoption of secondary prevention behaviors. For ACKNOWLEDGMENT example, Graham [27] discovered that the earlier breast cancer This research was funded by African Doctoral Dissertation is found, the greater the chance of survival and when a breast Research Fund (ADDRF) offered by African Population and self exam (BSE) is done regularly early detection early is Health Research Centre (APHRC) in partnership with guaranteed. However, not all women do BSE regularly and International Development Research Centre (IDRC). somehow they have to believe there is a benefit in adopting this behavior. Similarly in this study Type 2 diabetics must REFERENCES know there are benefits before they make decision to adopt a [1] A. Ramachandran, C. Snehalatha, S. Mary, B. Mukesh, A.D. Bhaskar & prescribed dietary prescription. V. Vijay. The Indian Diabetes Prevention Programme shows that Cues to actions are events, people, or even things that move lifestyle modification and metformin prevent type 2 diabetes in Asian people to change their behavior. For example, illness of a Indian subjects with impaired glucose tolerance (IDPP-1). Diabetologia, Vol 49, issue 2, pp. 289-297, 2006. family member, media reports [27], mass media campaigns, [2] J. Lindström, A. Louheranta, M. Mannelin, Rastas, V. Salminen, J. and reminder postcards from health care provider or warning Eriksson et al. The Finnish Diabetes Prevention Study (DPS): Lifestyle labels on a food product. This study focused on three intervention and 3-year results on diet and physical activity. Diabetes Care, vol. 26, issue 12, pp. 3230-3236, 2003. categories of cues to action, including posters and materials, [3] X.R. Pan, G.W. Li, Y.H. Hu, J.X. Wang, Wy Yang, Z.X. An et al. television or radios and weekly education program as elements Effects of diet and exercise in preventing NIDDM in people with of cues to action. The study has shown that cues to action International Scholarly and Scientific Research & Innovation 7(3) 2013 812 scholar.waset.org/1307-6892/9996982 World Academy of Science, Engineering and Technology International Journal of Psychological and Behavioral Sciences Vol:7, No:3, 2013 impaired glucose tolerance. The Da Qing IGT and Diabetes Study. dependent diabetes mellitus. Behavioral medicine, vol. 23, issue 3, pp. Diabetes Care, vol. 20, issue4, pp.537-544, 1997. 112-120, 2007. [4] K.F. Eriksson & F. Lindgarde. Prevention of type 2 (non-insulin- [26] K. McCormic-Brown (1999). Health Belief Model. (Retrieved 2009-12- dependent) diabetes mellitus by diet and physical exercise. The 6-year 3).http//hsc.usf.edu/~kmbrown/Health_Belief_Model_Overview.htm. Malmo feasibility study. Diabetologia, vol. 34, issue 12, pp.891-898, [27] Graham ME (2002). Health beliefs and self breast examination in black 1991. women. Journal of Cultural Diversity, vol. 9 issue 2, pp. 49-54, 1999. [5] National Diabetic Control Programmeme. National Clinical Guidelines [28] R.E. Schumacker & G.L. Richard. A beginner’s guide to structural for Management of Diabetes Mellitus. Ministry of Public Health, equation modeling (Second Edition). Mahwah, N.J: Lawrence Erlbaum Nairobi, Kenya, 2010. Associates, 2004. [6] I. Ajzen. The Theory of Planned Behavior. Organizational Behavior and Human Decision Processes, Vol. 50 pp. 179-211, 1991. [7] J.B.F. de Wit, R. Vet, M. Schutten, & J. van Streenbergen. Social cognitive determinants of vaccination behavior against hepatitis B: An assessment among men who have sex with men. Preventive medicine, Vol.40, issue 6, pp. 795-802, 2005. [8] N.A. Åstrøm & I. Okullo. Temporal Stability of the Theory of Planned Behavior: A Prospective Analysis of Sugar Consumption among Ugandan Adolescents. Community Dentistry and Oral Epidemiology, Vol. 32, issue 6, pp 426-434, 2004. [9] L.C. Blue. Does the Theory of Planned Behavior Identify Diabetes- International Science Index, Psychological and Behavioral Sciences Vol:7, No:3, 2013 waset.org/Publication/9996982 related Cognitions for Intention to be Physically Active and Eat a Healthy Diet? Public Health Nursing; vol. 24, issue 2, pp. 141-150, 2007. [10] J.A. Harrison, P.D. Mullen, & L.W. Green. A meta-analysis of studies of the health belief model with adults. Health Education Research, vol 7, pp. 107-116, 1992. [11] P.D. Mullen, J. Hersey & D.C. Iverson. "Health Behavior Models Compared." Social Science and Medicine, vol. 24, pp. 973–981, 1987. [12] J.W. Creswell & V.l. Plano Clark (2007). Designing and conducting mixed methods research. Thousand Oaks, CA: Sage. [13] C.L. Lysack & L. Krefting. Qualitative methods in field research: An Indonesian experience in community based practice. The Occupational Therapy Journal of Research, vol.14 issue 20, pp. 93-110, 1994. [14] Creative Research Systems. The survey system: Sample size calculator, 2003 [Online]. Available: {http://www.surveysystem.com/sscalc.htm} [Accessed on November 15, 2004]. [15] L. Ibironke. Scholarly Communication: The Use and Non-Use of E-Print Archives for the Dissemination of Scientific Information. Science and Technology Librarianship; Fall 2002. [Online]. Available: {http://www.istl.org/02-fall/article3.html} [Accessed on November 15, 2004].. [16] R. Mugnaini, A. L. Packer & R Meneghini. Comparison of scientists of the Brazilian Academy of Sciences and of the National Academy of Sciences of the USA on the basis of the h-index. Brazilian Journal of Medical and Biological Research, vol. 41, issue 4, pp. 258-262, 2008 [17] G.D. Garson. Structural equation modeling,2009 (Retrieved on 2009-12- 16): http://faculty.chass.ncass.ncsu.edu/garson/Pa 765/structur.htm [18] R.E. Schumacker & G.L. Richard. A beginner’s guide to structural equation modeling (Second Edition). Mahwah, N.J: Lawrence Erlbaum Associates, 2004. [19] R.B. Kline. Principles and practice of structural equation modeling. New York. Guilford Press, 1998. [20] D. George, & P. Mallery. SPSS for Windows step by step: A simple guide and reference. 11.0 update (4th ed.). Boston: Allyn & Bacon, 2003. [21] D.O. Omondi, M.K. Walingo, G.M. Mbagaya & A.O.L., Othoun. Understanding Physical Activity Behavior of Type 2 diabetes Using the Theory of Planned Behavior and Structural Equation Modeling. International Journal of Human and Social Science (Journal of social science), vol 5, issue 3, pp. 160-167, 2010. [22] S.T. Kershaw, M.L. Niccolai, A.K. Esther, B.J. Lewis & R.J. Ickovics. Perceived susceptibility to pregnancy and sexual transmitted disease among pregnant and nonpregnant adolescents. Journal of Community Psychology, vol.31, issue 4, pp. 419-434, 2003 [23] J.B.F. de Wit, R. Vet, M. Schutten, & J. van Streenbergen. Social cognitive determinants of vaccination behavior against hepatitis B: An assessment among men who have sex with men. Preventive medicine, vol. 40, issue 6, issue, pp. 795-802, 2005. [24] L. Belcher, M.R. Sternmberg, R.J. Wolotski, P. Halkitis, & C. Hoff. Condom use and perceived risk of HIV transmission among sexually active HIV positive men who have sex with men. AIDS Education and Prevention, vol. 17, issue 1, pp. 79-89, 2005 [25] L.H. Forsyth & V.L. Goetsch (1997). Perceived threat of illness and health protective behaviors in offspring of adults with non-insulin International Scholarly and Scientific Research & Innovation 7(3) 2013 813 scholar.waset.org/1307-6892/9996982