The proposed project will develop, implement, and test methodology for the design of an acceptable, effective intervention for diabetes prevention in a real-world setting. By determining the effectiveness of the intervention, the study will serve to guide the translation of research into routine public health prevention programs and policy. And, the hands-on experience will directly support the development of expertise, tools and training to advance translational science as a discipline. The study aims to test the effectiveness of an environmental and individual level intervention to prevent diabetes at Dhulikhel Hospital-Kathmandu University Hospital (DH-KUH), Nepal. Specific aims are to measure:

Effectiveness of a canteen intervention on a composite score based upon improvement in 3 cardio-metabolic risk factors (0-3) [HbA1c decrease ≥0.5%; a systolic blood pressure decrease (SBP) ≥5 mm Hg; and plasma triglycerides decrease ≥10 mg/dl] Effectiveness of a behavioral intervention on a composite score based upon improvement in 3 cardio-metabolic risk factors (0-3) [HbA1c decrease ≥0.5%; SBP decrease ≥5 mm Hg; and plasma triglycerides decrease ≥10 mg/dl] Effectiveness of a canteen intervention on change in HbA1C and healthy food intake after six months of the canteen only (CO) intervention compared to the change over six months during the control period.

Effectiveness of a behavioral intervention on change in Hba1c and healthy food intake after six months of behavioral intervention plus the canteen (CB) intervention compared to the change over six months during CO.

We will recruit 366 adult employees of DH-KUH. At baseline (T1), 6 months (T2), 12 months (T3) and 18 months (T4), we will administer a standard questionnaire to record relevant characteristics of the participants (age, sex, education, income, marital status, and family history of CVD, physical activity, smoking, alcohol intake and diet). We will abstract food consumption data from the administrative database in DH-KUH. Blood samples will be collected and analyzed for HbA1c, fasting glucose, and lipid profile (HDL, LDL, total cholesterol, triglycerides). We will measure height, weight, waist circumference, hip circumference, and blood pressure. After 6 months of control period, the participants will receive the canteen intervention: (a) form and train a canteen improvement team; (b) train canteen staff on healthy cooking; (c) add healthy food and remove unhealthy food; (d) information and communication of canteen changes to employees; (e) monitoring of the interventions. After six months of the canteen intervention, half of the participants will be randomized to receive the behavioral intervention. The behavior intervention will be a combination of intensive education sessions, group counselling, goal setting and monitoring based on the Diabetes Prevention Program (DPP). The primary analysis will use χ^2test use to compare (a) Proportion of individual with score ≥ 2 during CO intervention to the proportion of individual with score ≥ 2 during the control period; (b) Proportion of individual with score ≥ 2 during CB intervention to the proportion of individual with score ≥ 2 during CO intervention. We will conduct a logistic regression with the proportion of individual with score ≥ 2 as outcome and CO vs CB as exposure at T4. Secondary analysis will use paired t-test to compare (a) the change in healthy food intake and HbA1c during CO to the change in healthy food intake and HbA1C during the control period (b) the change in healthy food intake and HbA1C% during CB to the change in healthy food intake and HbA1C% during CO period. We will conduct linear regression with HbA1c% as the outcome and CO vs CB as the exposure variable at T4. Further analysis will adjust for confounding in time-varying variables and assess effect modification.

INTRODUCTION

The burden of non-communicable diseases (NCDs), such as cardiovascular disease (CVD), diabetes, cancer, and chronic obstructive pulmonary disease (COPD),are on the rise in low- and middle- income countries. CVD is the leading cause of morbidity, mortality, and disability in South Asia, where 20% of the world's population resides.

Sedentary lifestyle, poor diet, and excessive body weight are reported to have a large effect on the risk of developing NCD's. Lifestyle interventions addressing diet and exercise have reduced cardiovascular risk. Despite the evidence supporting the use of lifestyle interventions to prevent hypertension and diabetes and to improve glucose tolerance, their translation into real world settings has been challenging. Worksites provide unique opportunities for health promotion and disease prevention programs since people spend a majority of their time at work, and they allow access to large segments of the population. Worksites also provide an infrastructure and natural environment for social support. Worksite-based health programs have shown positive impacts on employee health, and they have shown significant improvements in cardiovascular risk factor profiles. Worksite interventions encompassing environmental changes (i.e. low-cost healthy food options), places for physical activity (i.e. fitness centers or gyms), , and group-based health education classes, have been highlighted as components of successful worksite interventions. As an important way to translate CVD prevention efforts, worksite interventions can promote healthy food choices, facilitate health education, and provide social support. Thus, we planned a pioneer study that will measure the effectiveness of a cafeteria-based intervention and a behavioral intervention on cardio-metabolic risk among employees of a hospital based in Nepal, by evaluating the change in number of individuals reaching two or more cardio-metabolic risk goals, specifically reductions in blood pressure, triglycerides, and glycosylated hemoglobin (HbA1c,).

METHODS AND MATERIALS

Study design:

The Nepal Pioneer Worksite Intervention Study is a two-step intervention study. In the first step, we will use a pre-post design to assess how a cafeteria-based intervention, that provides a healthier diet, effects cardio-metabolic risk. In the second step, we will conduct an open-masked, two-arm randomized trial by allocating half of the participants to a cafeteria and behavioral (CB) intervention on prevention of cardio-metabolic risk, while the other half of participants will receive the cafeteria-only (CO) intervention. The study protocol has been approved by the institutional review committee at Harvard T.H. Chan School of Public Health, Nepal Health Research Council, and Kathmandu University School of Medical Sciences.

Study Setting:

The study will be conducted at Dhulikhel Hospital -Kathmandu University Hospital (DH-KUH) in central Nepal. Dhulikhel Hospital is an independently owned, not-for-profit institution which was conceived and supported by the Dhulikhel community. The hospital has approximately 1040 employees and four functional cafeterias that are in operation 16 hours a day.

Recruitment:

We will conduct a 20-minute information session in all of the hospital departments. Employees will be invited to attend sessions through flyers posted around the hospital and through announcements in department meetings. During the information session, we will explain the purpose and the expectations of the study. We will explain the ethical considerations, emphasizing the importance of protecting participant's privacy during research-related interactions and outcomes, the volunteer nature of the study, the option to drop out of the study at any time and we will establish there should not be any coercion from supervisors to participate in the study. After the information session, research assistants (RAs) will set up an appointment for eligibility screening among interested participants.

The inclusion criteria are: (a) adults 18 years or older; (b) full time employees of DH-KUH; with (c) systolic blood pressure of >=120 mmHg or diastolic pressure >=80 mmHg; or HbA1c of 5.7% to 6.4%, or fasting blood sugar of >=100 mg/dL. The exclusion criteria of the study are: (a) pregnant women since dietary habits may change during pregnancy, (b) taking diabetes medication, or (c) taking hypertension medication. The details of the screening process are described below. The RA will obtain written informed consent in a private room for those individuals who are deemed eligible to participate in the study. A short oral consent process will be utilized for participants who cannot read the consent form. This will entail presenting all of the elements of the consent form verbally to the participant, in the presence of a witness. The witness will be required to sign a document stating the consent form has been verbally presented to the participant.

Data collection: The schedule of enrollment and assessment is presented in Table 1.

Screening: The RAs will conduct a 2 phased screening to identify the eligible participants. In the first screening, the RAs will screen participants by measuring blood pressure, and administering a questionnaire along with measures to calculate an Indian Diabetes Risk Score (IDRS).13 The IDRS takes age, abdominal obesity, self-reported physical activity, and family history of diabetes into account. The IDRS score ranges from 0 to 100, with a high number indicating higher risk. The IDRS has been considered a reliable instrument to screen the risk of diabetes in Asian-Indian populations.13 An IDRS of 30 or more has been shown to have 95% sensitivity and 45% specificity to detect prediabetes using fasting blood sugar criteria (100-120 mg/dL); and 87% sensitivity and 47% specificity using HbA1c criteria (HbA1c 5.7-6.5%) in an analysis of 560 free residents of Dhulikhel (Unpublished).

Blood pressure will be measured in the right arm of seated participants, after a five- minute rest period. Three measurements of systolic and diastolic blood pressure will be taken using a Microlife automatic blood pressure measuring device. The mean of three blood pressure measurements will be used. The participants with systolic blood pressure of >=120 mmHg or diastolic pressure >=80 mmHg will be invited to participate regardless of their IDRS score. In the second screening, the participants scoring 30 or more on IDRS will be asked to provide a blood sample to measure HbA1c and fasting blood glucose. The individuals with HbA1c between 5.7% to 6.4% or fasting blood sugar of >=100 mg/dL will be invited to participate. All eligible individuals that provided informed consent will be enrolled. No compensation or reimbursement will be provided to the participants.

Baseline assessment: At baseline, RAs will interview the participants using a standardized electronic questionnaire using Open Data Kit software.14 RAs will receive two weeks training on data collection and ethical issues.

The questionnaire will assess socioeconomic characteristics including age, sex, ethnicity, religion, marital status, annual income, education, and lifestyle factors including smoking, alcohol intake, and physical activity. We will use the Global Physical Activity Questionnaire,15 and calculate the metabolic equivalent of task (MET) minutes per week. A weekly MET equivalent of 600 would represent 30 minutes of brisk walking five times per week or 15 minutes of running five times per week.

Twenty four hour diet recall: To measure dietary intake, we will conduct two interviewer-administered 24-hour dietary recalls within a week. Each 24-hour dietary recall will take approximately 25 minutes to complete. First, activities of the previous day will be documented to refresh the participant's memory. Then, a dietitian will ask the participant to recall everything s/he consumed from the first meal to last meal. The time and place of each meal will be noted, followed by detailed information on each food including: specific brands, ingredients, and/or recipes. Participants will be asked to report their food portions using colorful examples of sizes or household measures such as spoon, bowl, etc. If a participant reports using their own recipe, then complete information on each individual ingredient will be inquired about. Energy and nutrient intakes will be calculated using the food composition table for Nepal.16

Anthropometry: Body weight will be measured with minimum clothing and without shoes using an Omron Model hbf-400 scale and recorded to the nearest 0.1 pounds. The weighing scale will calibrated to zero every day. Participants' heights will be measured, without shoes, while the participants stand against a wall. Height will be measured using a tape measure and recorded to the nearest 0.1cm.

Laboratory: Blood samples will be analyzed for HbA1c, fasting glucose, low density lipoprotein (LDL) cholesterol, high density lipoprotein (LDL) cholesterol, triglycerides, and total cholesterol. All of the laboratory procedures will be carried out in the biochemistry laboratory of DH-KUH. Blood samples will be collected using evacuated blood collection tubes. Participants will be asked to fast overnight (8-14 hours). The HbA1c will be measured using Boronate affinity chromatography (Axis-Shield, Norway); fasting blood glucose using hexokinase method (Dialab, Austria); LDL and HDL using the elimination method (Dialab, Austria); triglyceride using gpo-pap (Dialab, Austria); and total cholesterol using chod-pap (Dialab, Austria). For each type of assay, the laboratory has quality control (QC) materials (sing commercially available assayed and unassayed control material) from biorad Laboratories, USA. Each QC is run at least in duplicate. External QC is arranged by internationally recognized reference laboratories that distribute batches of samples of various concentrations for each assay. The laboratory performs External Quality Assurance Scheme from unknown assayed sample from the Department of Clinical Biochemistry, Christine Medical College, Vellore, India for 23 routine parameters, 5 immunological parameters and HbA1c. Additionally, 5% of the blood samples will be obtained in duplicates and sent for testing all parameters, blinded to the laboratory personnel.

Control period: We will have a control period of six months before implementing the intervention. There will not be any contact with the enrolled participants six months after the study enrollment. The control period will allow us to conduct the difference in difference analysis; hence minimizing potential biases in effect size, including increases or decreases in a health condition with time.

Interventions Step 1: Cafeteria Intervention: After six months of the control period, all of the participants will receive the cafeteria intervention (Table 2). The cafeteria intervention was developed based on the findings from four focus group discussions with cafeteria users and nine in-depth interviews with cafeteria operators and managers, about strategies to promote healthy foods in the worksite. The four cafeterias in the hospital will improve the quality of their meals by (1) increasing the availability of fresh fruit (not fruit juice) and vegetable options, (2) avoiding sales of sugar sweetened beverages, (3) replacing whole grains with refined grains in cooking; (4) using healthy vegetable oils such as soy and sunflower ; (5) minimizing the sale of fried foods; (6) trimming animal fats from meats before cooking; (7) using healthier protein sources such as chicken, beans, and nuts (8) making potable water free of cost; and (9) reducing salt in cooking. These guidelines are based on the recommendations for a healthy diet to improve cardiovascular health.22 To facilitate these changes, a cafeteria operation team will be formed and they will be trained on procedures to implement, supervise, and monitor the worksite's healthy changes. In addition, we will train the cafeteria staff on healthy eating, and how to modify recipes to incorporate healthy options.

Step 2: The behavioral intervention: After six months of the cafeteria intervention, half of the participants will be randomized using computer generated random numbers to receive a cafeteria and behavioral intervention (CB), the other half of the participants will continue to receive the cafeteria only (CO) intervention. The behavioral intervention will be comprised of intensive education sessions, group counseling, and goal setting and monitoring exercises based on the Diabetes Prevention Program (DPP)23 tailored to local needs. The curriculum includes 24 sessions of 16 cores weekly sessions during the first four months of the intervention followed by 8 weekly maintenance sessions. Each session will be 1 hour long, facilitated by a dietitian. Broadly, the curriculum covers the subject matters of importance of healthy weight, eating a healthy diet, increasing physical activity, stress management, and challenges of lifestyle changes. Similar to the cafeteria intervention, healthy eating messages include increasing fresh fruits and vegetable intake, avoiding added sugar, choosing whole grains, choosing healthier sources of protein, reducing sodium, and monitoring portion sizes. Participants will be encouraged to keep food and activity diaries throughout the course of the study. The maintenance period's focus will be on overcoming declines in motivation and maintaining long-term healthy behaviors. All of the sessions will be conducted at the worksite during the workday. There will be about 20 participants in each education class. Participants will set a minimum of two lifestyle-change goals (i.e. Half of their total grain intake will be whole grains, walking 30 minutes a day, or reducing 7% of their body weight).

Follow up:

We will assess outcomes at 6 months (at the end of the control period), 12 months (at the end of the cafeteria intervention), and 18 months (at the end of the behavioral intervention). During each follow up, fasting blood samples will be collected and analyzed for HbA1c, fasting glucose, and lipid profile (HDL, LDL, total cholesterol, triglycerides). We will re-administer the global physical activity questionnaire, and the two 24-hour diet recalls. We will re-measure height, weight, waist circumference, and blood pressure at each time point

Primary and secondary outcomes:

The primary outcome will be the proportion of individuals reaching two or more of their cardio-metabolic risk goals, namely, reductions in blood pressure, triglycerides, and HbA1c. Participants will be scored on the number of improved risk factors (0-3) as defined by decreases in (1) HbA1c ≥0.5%; (2) systolic blood pressure ≥5 mm Hg; or plasma triglycerides ≥10 mg/dl. These outcomes were selected because blood pressure, HbA1c, and triglycerides are commonly measured in clinical settings. This makes their use clinically-appropriate and translatable, as other CVD risk scores such as the Framingham Risk Score, do not perform well in South Asian populations.24 Moreover, the composite outcome allows for individuals to reduce different factors based on their variable risk profiles at baseline. For example, an individual with a baseline systolic blood pressure of 120 mmHg many not reduce this risk factor by as much as 5 mmHg, but may succeed in reducing HbA1c or triglycerides. The secondary outcomes are absolute changes in HbA1c, systolic blood pressure, diastolic blood pressure, and triglycerides.

Data Management:

We have taken four robust provisions to ensure data quality. First, the electronic questionnaire will be closed only if the data collection is complete to avoid partial or missing values. The data will be cleaned and checked every month. The answer fields for all integer variables will be constrained to ensure entry of only valid numbers. Second, RAs will receive intensive training on data collection and ethical considerations. Third, the site investigator will supervise the RAs on a day-to-day basis. Fourth, the principal investigators will hold weekly meetings with site investigators, and if required, with the RAs, to discuss the course of the intervention and to address any issues.

Data Analysis Plan: The primary analysis will be intention to treat. The quantitative data analysis will follow the Consolidated Standards of Reporting Trials (CONSORT) guidelines.25 Flowcharts will include the number of participants seen at each stage, including the number screened, eligible, randomized, and analyzed for the primary outcomes.

Effectiveness of the cafeteria intervention: At the 12-month follow-up, we will compare the proportion of participants who achieved two or more of their cardio-metabolic risk factor reduction goals during the cafeteria intervention period, to the proportion of participants in the control period, using the generalized estimating equations (GEE) approach with the binary distribution, logit link function and an exchangeable working correlation structure.26

Effectiveness of the behavioral intervention: At the 18-month follow-up, we will compare the proportion of participants who have achieved two or more of their cardio-metabolic risk factor reduction goals in CO arm to the proportion of participants in the CB arm using chi-square test. In the event that randomization does not control for differences between the treatment and control groups on baseline characteristics, we will statistically control for those differences using the GEE approach with the binary distribution, logit link function and an exchangeable working correlation structure.26

Our secondary analysis will compare the change in HbA1c, systolic blood pressure, and lipids during the cafeteria-only intervention to the change during the control period using a paired t-test. Similarly, we will compare the change in HbA1c, systolic blood pressure, and lipids between CO group and CB group using a paired t-test. In the event that randomization does not control for differences between the treatment and control groups on baseline characteristics, we will conduct the GEE approach with the Gaussian distribution, identity link function and an exchangeable working correlation matrix26 to adjust for potential confounding factors.

Sample size and power:

The pre-post design, with at least 366 eligible and enrolled participants27 and 5% loss to follow-up (LTF) after 12 months, will have over 90% of power to detect the primary effectiveness endpoint of this trial, of 31.5% or greater, compared to the 21% change in the control group in a South Asian population.28 For the analysis of two randomized groups (CO and CB arms), with 320 participants, we will be able to detect a relative risk of 1.5 given 0.315 risk probability in the control group, 5% LTF rate with 80% power and 5% level of significance.29

In the secondary analysis will be the comparison of changes in HbA1c, systolic and diastolic blood pressure, and lipids observed in the cafeteria intervention and the behavioral intervention. Given the sample sizes of 366 and 320 respectively, 5% LTF rate with 80% power and 5% level of significance, we will be able to detect the minimum differences presented in Table 1. For the before-after design, we considered three possible values of the correlation (ρ=0.5,0.6,0.7) between changes during the control period and during the intervention period.

Minimization of contamination:

The risk and level of contamination during the randomized study phase for the behavioral intervention will be monitored at the participant level. The participants will be instructed not to share information about the study and not to provide any support to people in their worksite, other than their class mates. The behavioral classes will be conducted in a separate building outside of their workspace to ensure privacy. Further, we will measure possible contamination by asking if the participants have received any information / advice regarding diet and lifestyle changes from any of their peers and if they did, collect the name and contact information of the peer. If the contamination is found to be significant, we will make adjustments while estimating the effect.

Intervention fidelity:

Intervention fidelity refers to the extent the intervention is delivered as it was intended.30 In the proposed study, we will warrant and quantify the fidelity of both the cafeteria and the behavioral intervention. First, we will provide training sessions on the study objectives, details of the interventions, and the importance of fidelity to the study staff, cafeteria improvement committee and cafeteria staff. . Second, we will monitor the implementation of changes in the cafeteria using a structured checklist on a weekly basis. The checklist is provided in the supplementary materials. The deviance in fidelity will be discussed during the monthly meeting of the cafeteria improvement team where necessary action will be taken. For the behavioral intervention, we will measure attendance from participants and we will test pre- and post- knowledge on diabetes prevention, covered in the health education sessions. Finally, a site investigator will supervise the implementation of intervention.

Interim analyses:

We do not have a plan for interim analysis and do not expect a situation that would lead us to stop the study. The participants with newly reported pregnancy during the course of the study will be excluded from the study.

Reportable Events and Indemnities:

Serious adverse events are not expected. Other events related to study participation (such as an infection at the site of a blood draw clinic, or a breach of confidentiality) will be documented. All unanticipated problems (non-medical occurrence) that involve risk to subjects or others will be reported to the institutional review board at Kathmandu School of Medical Sciences and Harvard T.H. Chan School of Public Health.

Inclusion Criteria:

• Full time employees of DH-KUH
• 18 years or above of age
• Score 30 or more on Indian Diabetes risk score, No confirmation of diabetes and, not on diabetes medication. OR Have HbA1c of 5.7% to 6.4%. OR Prediabetes group with Fastening Blood Sugar (FBS) of 100- 126 mg/dL
• Systolic blood pressure 120 mm Hg or more or Diastolic blood pressure 80 mm Hg or more and not on blood pressure medication

Exclusion Criteria:

• Less than 18 years
• On diabetes medication
• On hypertension medication