Alexa Ehlert and Faculty Mentor: Evan Thacker, Department of Health Science
Introduction. Cognitive impairment is a concerning issue among the population of older adults in the United States. This problem is compounded by the fact that science is largely uncertain of how to prevent cognitive decline; however, if we knew that particular modifiable behaviors associated with cardiovascular health were also related to cognitive health, then efforts to prevent cognitive impairment could potentially be developed. There are three principal goals of this project: 1) determine how cardiovascular health evolves in older adults over several years; 2) determine the relationship between cardiovascular health and cognitive decline in older adults; 3) determine how changes in cardiovascular health affect cognitive decline in older adults. During my time as an ORCA grant recipient, I focused on addressing the first goal by developing variables for measuring cardiovascular health using data collected from >5,000 participants of the Cardiovascular Health Study. The specific work I did is described below, with a focus on the American Heart Association’s “Life’s Simple 7” score, the use of physical activity data in that score, and the use of dietary data in that score.
Life’s Simple 7. To measure cardiovascular health, we used Life’s Simple 7 (LS7), a cardiovascular health measurement tool. This paradigm is comprised of seven components, namely: smoking status, body mass index, physical activity, diet, blood cholesterol, blood pressure, and fasting glucose. While the process for determining the scores for smoking status and body mass index was fairly straight-forward, the other components required more attention. Measurements for blood cholesterol, blood pressure, and fasting glucose all required information on medications being taken to treat those ailments. For example, someone with a blood pressure of 108/74 mg/dl who was not taking blood pressure medications would receive a different score than someone who had the same systolic and diastolic values but was on medication to manage their blood pressure. Furthermore, fasting glucose scores also depended on whether or not a participant was actually fasting.
Physical activity. The LS7 score for physical activity was more complicated. We had information on activity type, frequency of participation in the activity, number of minutes of a typical bout for that activity, and months per year they participated in the activity. From this information, we were able to obtain the average number of minutes spent per week on a particular activity. Because LS7 differentiates between moderate physical activity and intense physical activity, it was also crucial to categorize each activity into an intensity level. This was done using previously published studies. The physical activity minutes were then combined to create a best estimate for each participant’s time spent on exercise per week.
Diet. Determining diet score was the biggest challenge. The data from the Cardiovascular Health Study contains two time points of dietary data. The first time point holds data on a group that we refer to as the “original cohort”. These individuals are primarily Caucasian Americans. The second time point contains data on the original cohort as well as on an additional group called the “new cohort”, which consists mainly of African American participants. One issue we experienced in creating the dietary variables was that the surveys administered at the two different time points differed drastically in what foods they asked about and how they measured intake. In order to make both time points comparable, I analyzed the data that we had at each time point. I then created a program that used weighted averages of each specific food intake to determine the most comparable value for food intake at each time point. This enabled us to compare scores between the two time points, in spite of the differences in the food questionnaire.
Next steps. Because we are interested in cardiovascular health changes over time, each participant needs dietary data from two different time points. To do this, we will utilize multiple imputation methods to create the best estimate for a person’s diet score for the year where they lack one. Preparing and examining the data for this process has been mostly completed, and the imputation process will begin soon. Following the utilization of multiple imputation for not only the dietary data but also other random missing data, our project will be ready for analysis. We will use regression models to determine the relationship between cardiovascular health (according to the LS7 paradigm) and cognitive decline.
My role going forward. I have now graduated from BYU, which means that I will no longer have access to the data for this project. Therefore, I will not continue to work directly on the data management or statistical analysis that I was doing as an ORCA grant recipient. I will, however, still contribute as an author on forthcoming papers and presentations that result from this project. The work that has already been completed has merit in the world of cardiovascular health research. We are the first researchers we know of to be modifying the data from the Cardiovascular Health Study so that it can been measured in accordance with LS7. Our subsequent findings will help reveal not only how cognitive decline relates to cardiovascular health, but also if the LS7 paradigm is an appropriate measure among the population of older adults in the United States. This fall, I will be attending the Harvard T.H. Chan School of Public Health, where I will be enrolled in a two-year Masters in Nutritional Epidemiology. The work that I have done as an ORCA grant recipient has helped prepare me for this next phase of my education.
Other work. During my time as an ORCA grant recipient, I also devoted substantial effort to another project with Dr. Thacker on cardiovascular health, specifically the clinical epidemiology of atrial fibrillation. I completed a project entitled “Administrative Billing Codes Accurately Identified Occurrence of Electrical Cardioversion and Ablation/Maze Procedures in a Prospective Cohort Study of Atrial Fibrillation Patients.” I presented a poster of this research at the American Heart Association Epidemiology and Prevention conference in Portland, Oregon, in March 2017. I am also first author on a full-length manuscript describing this project, which is currently under review at a peer-reviewed medical journal. Although not part of my original ORCA proposal, this additional work has enhanced my research and writing skills, allowed me to make a significant contribution to epidemiologic research, and strengthened my applications for graduate school.