Piloting Diet and Exercise Interventions in Older Hispanics With Diabetes
Piloting Diet and Exercise Interventions in Older Hispanics With Diabetes
Type 2 diabetes affects 9.3% of the U.S. population and health disparities are evident. Compared to non-Hispanic Whites, twice as many Hispanics have type 2 diabetes. The estimated cost of diabetes was $327 billion in 2017. Older adults with type 2 diabetes often have increased risk of falls. Type 2 diabetes standard of care includes one annual self-management session, which rarely include assessments of mobility impairments and falls, and diabetes education programs are less likely to be offered and more likely to be discontinued in economically disadvantaged communities. Consequently, diabetes-related hospitalizations have doubled in the past 20 years. The objective of this study is to pilot test nutrition and exercise interventions to improve diabetes management (e.g. glycemic control and diet quality) and physical function, reduce falls and related healthcare use by disadvantaged older Hispanics with type 2 diabetes. The interventions will be tested separately and in combination to assess their individual and combined effects. We will recruit 60 older Hispanics with type 2 diabetes from 4 senior centers. The exercise intervention was adapted from the evidence-based Otago Exercise Program for falls prevention in older adults. The nutrition intervention aims to improve glycemic control and diet quality. We will cluster-randomize the participants into the following arms (n=15/arm) based on the congregate meal sites they attend: 1) Control (no intervention); 2) Exercise only; 3) Nutrition only, and 4) Exercise+Nutrition. The project will last 15 months; the duration of the interventions will be 6 months, and assessments will be completed at baseline, 3, 6, 9, and 12 months. The impact of the interventions will be assessed based on hemoglobin A1C levels (glycemic control), diet quality (Healthy Eating Index), blood pressure, physical function (balance, gait, strength, body composition, and fear of falls), and self-reported falls and healthcare use.
The pilot study will show the potential individual and combined effects of nutrition and exercise programs for disadvantaged older Hispanics with type 2 diabetes. The benefits to the participants include: 1) enhanced type 2 diabetes management; 2) improved diet quality and nutritional status; 3) improved physical function; 4) reduced rates of falls; and, 5) reduced rates of diabetes and fall-related emergency room visits and hospitalizations. The expected product is an effective type 2 diabetes management and falls prevention program for underserved Hispanic older adults with type 2 diabetes. If effective, the program may be expanded to additional congregate meal sites and communities serving older Hispanics.
The proposed study is a pilot cluster-randomized pilot trial for socioeconomically disadvantaged older Hispanics with type 2 diabetes. The participants will be cluster-randomized to one of the following four groups (n=10/group) based on the facilities they go for congregate meals: 1) Control (no intervention); 2) Exercise program only; 3) Nutrition program only; 4) Exercise and nutrition program. The interventions groups (2-4) will also receive basic diabetes education. The interventions will be randomized to sites to avoid cross-contamination. The interventions will last 6 months and assessments will be completed at baseline, 3, 6, 9, and 12 months.
The proposed pilot study is innovative because there are no exercise and nutrition programs for older Hispanics with type 2 diabetes; because it targets understudied socioeconomically disadvantaged older Hispanics with type 2 diabetes; it tests the separate and combined effects of nutrition and exercise interventions, and because we will assesses the outcomes as well as the mechanisms of falls reduction and improved type 2 diabetes management based on hemoglobin A1C levels, Healthy Eating Index scores, blood pressure, balance, gait, strength, and health care use.
We will assess the effects of the programs when delivered in isolation and in combination in relation to a control group. The client-centered programs will be delivered in the congregate meal sites and the participants will track their exercise, diet and medication adherence behaviors. We will evaluate the effects of the programs in physical function (gait, strength and balance), diet quality, nutritional status, falls rates and associated emergency room visits and hospitalizations. Glycemic control, physical function and falls in older people with type 2 diabetes can be addressed with exercise and adequate nutrition, but the uptake and implementation of these evidence-based interventions in the community is limited, especially among socioeconomically disadvantaged minority groups.
The interventions will be tailored culturally and to needs of older Hispanics with type 2 diabetes by respecting the participants' language choice, physical abilities and impairments, and food preferences. The nutritional intervention will consider the participants eating habits, preferences, food availability and costs, and the participants own goals. The exercise sessions will consider the participants limitations and health conditions.
Exercise and nutritional programs are rarely offered in congregate meal sites like the ones we will recruit our participants from. Each congregate meal site will host only one arm of the trial, but up to three groups of participants may be recruited from each site. The anticipated Mechanism of Action of the diet and exercise interventions involves improved diet quality, increasing protein, vitamin D and calcium intake, increased antioxidants from fruits and vegetables, reduced hemoglobin A1C and hospitalizations, increased energy for physical activity, increased muscle mass, and less falls. Increased exercise, will increase muscle mass, reduce A1C, improve physiological function, and reduce falls.
METHODS
Participants Older (≥ 65 years) Hispanics with type 2 diabetes will be recruited at congregate meal sites. We anticipate that approximately 65% of all participants will be women based on the sex distribution provided by our community partners.
Recruitment and intervention sites 60 eligible participants will be cluster-randomized based on the facilities they attend for one of the four trial arms (n=15/group to account for attrition). It is anticipated that recruitment will take up to 4 months. Eligible participants will be cluster-randomized by participating site into one of the four arms: 1) Control (no intervention); 2) Exercise program only; 3) Nutrition program only, or 4) Exercise and nutrition program.
Assessments will be carried out at baseline, 3, 6, 9 and 12 months at the sites during the morning in a private room. The participants will complete questionnaires regarding their medical health, falls history, fear of falls - Falls Efficacy Scale, socio-demographics, diabetes self-management, nutritional status - Mini Nutritional Assessment, food consumption (the Willett food-frequency questionnaire will be used to calculate the Healthy Eating Index), and health care use (emergency room visits and hospitalizations). In addition, the participants' gait, balance and strength will be measured. Gait will be assessed using an instrumented mat (GaitRite) to collect data on gait parameters during preferred walking speed and during fast walking conditions. A familiarization trial will be completed, followed by three recording trials under each condition in random order. Participants will be free to stop participating at any time. Balance will be assessed using force plates, and lower limb strength will be assessed based on the number of chair raises completed in 30 seconds. We will track any changes in food intake and/or exercise levels during the study period to control for potential confounders.
Interventions The interventions will be delivered at the congregate meal sites 2x/week in the afternoons, and attendance will be recorded. Participants in the 3 intervention groups (2-4) will participate in an 1 hour long diabetes education session as part of orientation at the beginning of the program. Primarily Hispanic undergraduate and graduate students will be trained and help deliver the exercise (physical therapy students) and nutrition (dietetics and nutrition students) interventions under supervision. Participants in all 4 groups (including those in the control group) will receive a physical activity log to record the days they exercise in group sessions and individually, and the days and duration of walks. Participants will track exercise using logs and we will assess medication history, medication changes, and medication adherence for all trial arms. The control arm will not receive any intervention other than possibly the standard of care provided by their physician. Information regarding the participation in diabetes self-management and other activities will be collected during the assessments. Adjustments will be made for duration of diabetes. Participants in all arms will track their exercise and medication adherence will be assessed.
Trial Arms: Duration, Frequency and Details.
Arms Interventions
The Exercise Intervention will be tailored for needs and capabilities of the participants and will involve short conversations (~5 min) about physical function, falls risks and prevention, exercise effects on diabetes and other exercise-related topics, followed by 30-minute exercise sessions. The intervention was developed based on the Otago Exercise Program, which is effective in reducing falls in community-dwelling older adults. It was piloted and well accepted by older Caribbeans, but it has not been tested in older Hispanics, nor in people with type 2 diabetes. The exercises involve lower and upper limb strengthening, balance training and a walking plan. Based on the baseline assessment, the participants will receive individualized exercise parameters and instructions on the weights to use, walking time (up to 30 minutes 2x/week) and safety instructions. The participants will receive illustrations of the exercises to do. Each participant will exercise for 30 minutes, 2x/week during supervised sessions in the facilities and be encouraged to do the exercises by themselves on days the group sessions are not offered. At the end of the sessions, we will have discussions of safe places to walk, proper footwear, adequate lighting (avoid down, dusk and evening walks), and safe environments (sidewalks in good conditions or walk in maintained trails in a park). We will also teach self-monitoring for signs of low-blood glucose to ensure safety. Bilingual research team members will be present to ensure that participants that only speak Spanish can understand the instructions. Progression will be done every two weeks during the first 2 months, once at month 3, and once at month 6 by increasing intensity and/or difficulty and the walks by increasing continuous walking time and walking speed without compromising safety. Assessments will be completed at baseline, 3, 6, 9, and 12 months.
The Nutrition Intervention will be tailored for the participants and based on Dietary Guidelines for Older Americans and Diabetes.org's food and Nutrition Recommendations, and will consider the participants' food preferences and concerns for managing glycemic control and maintaining muscle-mass. Each topic will be presented interactively and the participants will be encouraged to attend 2 times/week (attendance will be logged). The topics will be presented as a combination of evidenced-based materials for older adults with type 2 diabetes. We anticipate 3 sessions/topic during the 6-month intervention and will modify the topics and time to fit the needs of our audience. The lessons and materials will be offered in English and Spanish. The food quality change will be measured using a Food Frequency Questionnaire validated in Hispanic populations. The goals are to increase the participants' self-efficacy (belief and self-confidence) that they have the tools necessary to manage diabetes and reduce mobility impairments.
DATA ANALYSIS All primary analyses will be done on an "intent to treat" basis, and adjustments will be made for duration of diabetes. We will perform descriptive analysis on the demographic variables to describe the sample. Additional analysis will be conducted to assess differences between completers and dropouts. We will examine the potential cluster effect among the sites using the intraclass correlation (ICC), and will adjust the cluster effect in the analyses below if the ICC is significant. The effects of age and sex will be assessed during the analysis. The regression model for Aim 1 and 2 will include sociodemographic and biometric factors (age, sex, medication adherence, years with diabetes, insurance coverage, income, living alone, smoking, depressive symptoms, diabetes education, waist circumference or percent body fat as covariates for adjustment. All statistical analyses will be performed using SPSS v. 21 (SPSS Inc., USA).
Specific Aim 1: We will use linear mixed model to examine the change of hemoglobin A1C and HEI-2015 score from baseline to 3, 6, 9, and 12 months respectively by treating time as categorical variable. We are expecting there will be significant improvement from baseline to 3 and 6 months, a subsequent slight decline from 6 to 9 month, and then a sustained plateau at an improved level at the 12-month follow-up compared to baseline. We will compare the change of A1C and HEI-2015 score for the nutrition + exercise group with the other three groups by testing the interaction effect between time and treatment arms in the linear mixed model with Bonferroni correction as the multiple testing adjustment.
Specific Aim 2: We will compare the percentage of falls within each treatment arm between baseline and 3, 6, 9, and 12 months respectively using a McNemar test. And we will employ a repeated measures logistic regression to compare the change of risk of falls for nutrition + exercise group with the other three groups by the testing the interaction effect between time and treatment arms with Bonferroni correction as the multiple testing adjustment.
Ancillary Measures: We will use the same statistical models in Aim 1 and 2 for the secondary outcomes based on the type of the outcome variables. We will also perform a mediation analysis to test whether the gains in physical function, including gait characteristics, lower limb and grip strength, muscle mass, balance, and reduce fatigability and fear of falls, mediate the effect of the intervention on the percent of falls.
IMPLEMENTATION
Assessments will be carried out at baseline, 3, 6, 9 and 12 months at the sites during the morning in a private room. The participants will complete questionnaires regarding their medical health, falls history, fear of falls, socio-demographics, diabetes self-management, nutritional status, food consumption, and health care use (emergency room visits and hospitalizations). In addition, the participants gait, balance and strength will be tested. Gait will be assessed using an instrumented mat (GaitRite) to collect data on gait parameters during preferred walking speed and during fast walking conditions. A familiarization trial will be completed, followed by three recording trials under each condition in random order. Participants will be free to stop participating at any time. Balance will be assessed using force plates, and lower limb strength will be assessed based on the number of chair raises completed in 30 seconds. We will track any changes in food intake and/or exercise levels during the study period to control for potential confounders.
Interventions The interventions will be delivered at the congregate meal sites 2x/week in the afternoons, and attendance will be recorded. Primarily Hispanic undergraduate and graduate students will be trained and help deliver information on diabetes self-management (nursing students; i.e. what is diabetes, importance of lifestyle management and handouts), the exercise (physical therapy students) and nutrition (dietetics and nutrition students) interventions under supervision. Participants in all 4 groups (including those in the control group) will receive a physical activity log to record the days they exercise in group sessions and individually, and the days and duration of walks. Participants will track exercise using logs and we will assess medication history, medication changes, and medication adherence for all trial arms. The control arm will not receive any intervention other than possibly the standard of care provided by their physician. Information regarding the participation in diabetes self-management and other activities will be collected during the assessments.
Attrition Management The following will be done to minimize attrition: 1) we will have sessions on the health benefits of completing the study; 2) we will tailor the interventions to the needs and preferences of the participants; 3) the participants will be compensated for their time ($50: $10 per assessment), and 4) the participants will receive reports after each assessment including information on their physical performance and the lab results for A1C, lipid panel, albumin and hemoglobin, and body composition (Bioelectric Impedance Analysis), and the results will be explained to them.
Data analysis: Potential problems include non-linear data. The first approach would be to transform the data to achieve linearity and normality. If this is not achieved, non-parametric tests such as non-linear regression and a suitable alternative to path analysis will be conducted.
Missing Data: To reduce sample bias, we will use "intent to treat' analysis and will carry-over values for participants who dropped out and who have missing data points. For the data missing at random, we will use maximum likelihood imputation algorithms (SAS PROC MI), which are applicable to both missing outcomes and missing covariates. While, for data not at random missing, we will use complex modeling for missing values estimation.
Self-reported questionnaires: Since there may be bias, we will include objective measures of diabetes management and physical function. The Willett's food frequency questionnaire is a validated tool and calculation of HEI-2015 is a validated method; however, social desirability bias may impact the results. We will compare changes in A1C with diet quality.
General Measures:
Socio-demographic data will be collected by trained bilingual interviewers (English/Spanish). Subjects will be asked to complete a socio-demographic questionnaire constructed by the PI and used in previous studies, which will include questions related to gender, age, education, income, employment status, health insurance, smoking, and medications. The effects of the interventions will be assessed by age group (decade of life) and sex.
Anthropometric Measures: Height and weight will be measured using a SECA balance scale with stadiometer (Seca Corp, Columbia, MD). Body Mass Index will be calculated as weight in kg/height in m2. To attain the waist-to-hip ratio, the subject will stand erect with the weight evenly distributed. Waist circumference will be measured at the narrowest part between the lowest rib and iliac crest; the hip circumference will be taken around the widest portion of the buttocks. The ratio will be calculated by dividing waist circumference by hip circumference. Each measurement will be taken twice and averaged. Estimation of body fat mass and percent muscle will be made using Bioelectric Impedance Analysis with the InBody (Biospace, Inc. Los Angeles, CA, U.S.) model 529.
All participants will complete a questionnaire including information on health status, hospitalizations, doctors' and emergency room visits, falls history and fear of falls. The 1-1:30-hour long assessments will be conducted at baseline, 3, 6, 9 and 12 months and will also include the evaluation of physical activity level, grip, and lower limb functional strength, gait and balance.
Falls history: will be assessed based on self-reporting using the following questions: How many times have you fallen in the last 3, 6, and 12 months? Approximately when did you fall each of these times reported (e.g. complete date or just the month and year)? Did you get injured in any fall (please state the injury type)? Did it result in ER visits or hospitalization?
Fear of falls: will be based on the Falls Efficacy Scale. Falls self-efficacy (an indication of fear of falls) is associated with increased risk of subsequent falls. Falls can trigger a cycle of fear of falls, reduced physical activity, deconditioning, functional decline, social isolation, reduced quality of life, depression, and increased risk of subsequent falls.
Grip Strength: will be assessed using a Jamar dynamometer allowing adaptable grip lengths. The isometric hand grip strength measure using a Jamar dynamometer was chosen not only because it is strongly related to lower extremity muscle power, knee extension torque and calf cross-sectional muscle area, but also because decreased grip strength is a clinical marker of poor mobility. The participants will complete one five-second familiarization trial followed by two testing trials. The grip strength measures will be used to classify the frailty status and will also be classified as normal, intermediate or week based on the following criteria for men (>32, 26-32, <26 kg) and women (>20, 16-20, <16 kg).
Lower Limb Functional Strength: will be assessed based on the number of chair-stands completed in 30 seconds without hand assistance from a strait back chair without armrest, with seat height of 17 inches. Completing less than 8 repetitions in 30 seconds is a frailty functional marker.
Gait: will be assessed during preferred speed and fast walking conditions using a GAITRite® system (SN: Q209, CIR Systems Inc). In general, the reliability of the GAITRite® system measures of temporo-spacial parameters of gait has been found to be excellent for both young and older adults. The participants will complete one familiarization walking trial followed by two testing trials. The following gait parameters will be assessed:
Balance: will be assessed based on postural sway frequency, velocity and area during bipedal and single-leg stance for 30 seconds with eyes open and closed on a force plate. The participants will complete one familiarization trial followed by two testing trials.
Emergency Department Visits and Hospital Admissions: Self-reported measures of recent health care utilization agrees with actual records in more than 90% of the times and therefore are adequate for use in clinical and epidemiological studies.
Inclusion Criteria:
Exclusion Criteria: