Design of a worksite intervention to lower cardiometabollc risk In South Africa

Abstract

Background: Non-communicable diseases (NCDs) continue to increase globally, with an unduly larger impact in low to middle income countries (LMICs). NCDs are the main cause of death worldwide. There is strong evidence that lifestyle changes, such as weight loss, increased physical activity, and improved diet quality can help avert or slow down type 2 diabetes and reduce cardiometabolic risk factors, for example, high blood glucose, plasma lipids, and blood pressure. Regardless of the verification of research data, supporting the use of lifestyle interventions to prevent diabetes (improve glucose tolerance and lower high blood pressure), and implementing interventions in real-life settings has been proven to be difficult. Aim: The purpose of the study was to guide the development of an acceptable, appropriate, and feasible worksite intervention targeting the food environment and behavioural intervention to reduce cardiometabolic risk at a worksite in South Africa (SA). Methodology: In this cross-sectional study both qualitative and quantitative methods were used. The capacity of the built environment was explored through structured observations of the food and physical environment to offer healthy food and promote physical activity. Semi-structured in-depth interviews (IDIs) were conducted with worksite managers and canteen managers to assess the appropriateness, acceptability and feasibility of changes at worksites and explore the perceptions, provisions, facilitators and barriers to healthy eating at the worksite environment. Purposive snowballing sampling was used to recruit worksite managers and canteen managers for the IDIs. Focus Group Discussions (FGDs) were conducted amongst employees to explore the perceptions, provisions, drivers of and barriers to healthy eating at the worksite. An Organisational Readiness to Implement Change (ORIC) questionnaire was administered face to face at the worksite canteens and online through emails to employees aged between 18 to 65 years to determine the worksite readiness to implement changes at the worksite. Canteen staff were approached to rate possible intervention components for the purpose of tailoring the interventions and identifying the best way to deliver the intervention. The interventions were rated by canteen managers or canteen staff on a scale from one to five with regards to the feasibility of implementing different components of the intervention: with one (1) being impossible to implement and five (5) being easy to implement. A scoping review was carried out to gather data from empirical findings on the categorisation of healthy foods through Front of Pack (FOP) labelling schemes and was reported using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) extension for Scoping Reviews (PRISMA-ScR) guidelines. Key findings from the formative work were used to select suitable food and physical environment interventions for the worksite. Worksite canteen intervention training was conducted to train canteen managers for the implementation of the food environmental intervention using the training manual that was developed by the researcher. Results: The study was conducted at six Unilever SA worksites, three in KwaZulu-Natal (KZN) and three in Gauteng province. The Unilever worksites comprised of five factories and one Head Office, namely, Unilever Maydon Wharf (Personal Care and Home Care), Unilever Indonsa (Savoury dry food plant - Knorr, Robertson’s, Rajah), Unilever Khanyisa (Home Care), Boksburg (Home care) and Lordsview (Ice cream -Ola) and La Lucia Head Office (Head Office). The study was also conducted at Retailability Head Office, a fashion apparel brand, which borders on the Unilever Head Office, making a total of seven worksites. The findings of the structured observations for the physical environment indicated that walking tracks were demarcated within the worksites; however, there were no signage prompts to encourage walking. Stairwells were clean, well lit, easily visible and accessible at each building; however, the health benefits of using the stairs were not displayed. The food environment was observed at six canteens including La Lucia Head Office and Retailability, Unilever Maydon Wharf, Unilever Indonsa, Unilever Khanyisa and Unilever Boksburg. There were two canteens at La Lucia Head Office: the main canteen and a coffee shop. There was no canteen at Unilever Lordsview, but there was a dining area with packaged food sold at the vending machine. Regarding the observation of the food environment, the worksite canteens offered five subsidised contract dishes among other unsubsidised dishes of which the healthy meal was the most expensive in a four-week menu cycle. Pre-made pre-packaged food options were sold at the canteens and displayed near to the point of sale. Ten semi-structured IDIs were conducted with worksite managers, to assess the appropriateness, acceptability and feasibility of changes at worksites and to explore the perceptions, provisions and facilitators of and the barriers to healthy eating at the worksite environment. Several themes emerged from the IDIs with worksite managers with regards to participation in a lifestyle intervention namely: availability, worksite resources and barriers to participation. Four semi-structured IDIs were conducted with canteen managers to assess the appropriateness, acceptability and feasibility of changes at worksites and barriers to healthy eating at the worksite environment. Numerous themes emerged from the IDIs with canteen managers including the enablement of a healthy food environment, information dissemination, employee preference and canteen enablers. Five FGDs were conducted, each group comprising of 4-6 employees, until data saturation was reached to understand the appropriateness, acceptability and feasibility of a range of possible changes at the worksite. Many themes emerged from the FGDs such as nutrient-dense foods, nutrient-poor foods and energy-dense beverages as well as the food environment and healthy longevity. Worksite readiness to implement change was determined through the ORIC questionnaire. The ORIC statements that Unilever employees agreed with most scored 4.51 on the scale of “I believe this change will benefit our worksite”, followed by 4.49 on the scale of “I am committed to implementing this change” and 4.41 on the scale of “I want to implement this change”. The lowest scores (disagreement) were 3.35 for “I have the equipment we need to implement this change”, 3.37 for “I have the resources we need to implement this change” and 3.46 for “I have the expertise to implement this change”. All statements were significantly different from ‘3’, on average. All mean values were >3 which indicates there was a significant agreement with all the statements. Factor analysis with promax rotation was applied to the 31 statements. A Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO) of 0.958 and a significant Bartlett’s test indicated that the data was adequate for successful and reliable extraction. The alpha value for factor 1 and 2 was >.7, therefore indicating reliability. Composite variables were formed by calculating the average of the agreement scores for all items included in a variable. It was observed that there was a significant agreement for commitment (COM) and implementation (IMP) (with agreement that they desired the change and were committed to it being significantly higher than their agreement that they could implement the change). Results from a paired ttest showed agreement that they desired the change and were committed to it and were significantly higher than agreement that they could implement the change, t (299) = 15.229, p<.001. Further analysis to determine differences across worksites was conducted. Analysis from ANOVA showed that there was a significant difference in COM (commitment) across worksites, F (3, 296) = 5.375, p=.001. Post hoc analysis using Tukey’s test indicated that agreement to commitment was significantly higher at Maydon Wharf than at La Lucia Head Office (p=.043) and at Khanyisa (p=.021); and at Indonsa than at La Lucia Head Office (p=.031) and at Khanyisa (p=.021). Furthermore, there was a significant difference in IMP (implementation) across worksites, F (3, 296) = 9.960, p<.001. The ORIC statements that Retailability employees agreed with the most scored 4.58 on the scale: “I am committed to implementing this change”, followed by 4.53 on the scale of “I believe this change will make things better”, 4.50 on the scale of “I feel that implementing this change is a good idea” and 4.50 on the scale of “I feel this change is compatible with our values”. The lowest scores (indicating disagreement) were 3.11 for “I have the equipment we need to implement this change”, 3.24 for “I have the resources we need to implement this change” and 3.46 for “I have the expertise to implement this change.” All statements were significantly different from ‘3’, on average. All mean values were >3 which indicated there was significant agreement to all the statements. Factor analysis with promax rotation was applied to these 31 items. Two factors were extracted which accounted for 72.19% of the variance in the data. A Kaiser-MeyerOlkin Measure of Sampling Adequacy (KMO) of .784 and a significant Bartlett’s test indicated that the data was adequate for successful and reliable extraction. The alpha value for factor 1 and 2 was >.7, therefore indicating reliability. It was observed that there was a significant agreement for commitment (COM) and implementation (IMP) (with agreement that they desired the change and were committed to it being significantly higher than their agreement that they could implement the change). Results from a paired ttest showed that agreement that they desired the change and were committed to it was significantly higher than agreement that they could implement the change, t (37) = 6.727, p<.001. The intervention rating scale was completed by 12 canteen managers and canteen staff using a scale from one to five for the feasibility of implementing different components of the intervention. Interventions that were easy to implement were reported by 100% (n=12) of participants for increasing fruit and vegetable choices, 92% (n=11) for the provision of free water, followed by 75% (n=9) for the addition of a salad bar and ready-to- eat healthy meals. The one-sample t-test was done to determine whether the average ‘implementability’ score differed significantly from the central score of ‘3’. The easiest interventions to implement were reported by 83% (n=10) of participants being the strategic positioning of healthier alternatives to make healthy items more accessible with 75% (n=9) recommending traffic light labelling and healthy option stations. The most difficult to implement was reported by 25% (n=3) of participants for the display of kilojoules of a product translated into the number of minutes to perform a certain physical activity. A scoping review was conducted on FOP labelling schemes to inform the selection of a FOP labelling scheme best suited for canteen foods at worksites in SA. Several articles (n = 2513) were identified and screened after excluding duplicates (n = 2474). Overall, 1347 articles were excluded from the study because their abstracts and titles did not match the qualifying criteria. A total of six articles was used in the qualitative analysis after a full-text review of the remaining articles. It was concluded that grading foods into categories of healthfulness through evidence of key nutritional dimensions is a practical tool to inform food environmental interventions that may assist in public health promotion by influencing consumer choice in workplace canteens and beyond. The results from the observations, IDIs, FGDs, ORIC questionnaire, canteen rating intervention scale and scoping review were used to guide the development of the physical and canteen intervention and canteen staff were capacitated with training to implement the food environment intervention. A list of interventions detailing the tools to be used and responsibilities for executing the six weeks’ intervention with two weeks’ maintenance classes for phase three of the study, titled the South African Pioneer Worksite Multicomponent Lifestyle Intervention Study (WMLIS) was developed by the research team. Conclusion: Interventions are a useful tool that can be used in worksite settings to improve employees' overall occupational well-being. Interventions targeting the food environment and behavioural intervention can promote healthy eating behaviours and reduce cardiometabolic risk. It is important to take into account organisational complexity and the built environment when designing an intervention. In this study, formative research methods were used to engage worksite stakeholders to develop an intervention plan that is both theoretically and practically grounded to foster institutionalisation of the intervention. The findings of this study was used to contextualise and guide the development of acceptable, appropriate and feasible worksite food environment and behavioural intervention to reduce cardiometabolic risk among South African employees. The methods used allows for drawing of general conclusions for the implementation of lifestyle and food environment changes at worksites in SA. It is recommended that the study should be scaled up to other worksites to determine the comparativeness of this study to the response of other worksites regarding the acceptability, appropriateness and feasibility of worksite interventions.