Computerized Cognitive Training for Diabetic Elderly Veterans (CCT_DEV)

Computerized cognitive training (CCT) is an intervention has improved cognitive functioning in the elderly with and without cognitive impairment. The investigators will study the effect of a CCT program over an active control, "classic" computerized games. The outcomes will be memory and executive functions/attention, diabetes elf-management and adherence to medications, and glycemic and blood pressure control. Non-demented elderly Veterans with diabetes mellitus, who are at high risk for cognitive impairment, will be from the James J. Peters, Bronx, NY and Ann Arbor, MI VAMCs. This novel potential service fits the portfolio of the Quality Enhancement Research Initiative for Diabetes Mellitus (QUERI-DM) for which the Ann Arbor VAMC is a primary center. If successful, the VA National Center for Prevention and MyHeatheVet will collaborate in disseminating results to encourage implementation throughout the VA

Project Background: Diabetes mellitus (DM) has consistently been associated with increased risk for cognitive decline, mild cognitive impairment, and dementia in the elderly. Even minor cognitive impairments in nondemented individuals dramatically affect disease self-management. This, in turn, is associated with poor glycemic and blood pressure control in diabetes, which by themselves increase the risk of dementia, provoking a reinforcing cycle of disease. Thus, it is imperative to find interventions to delay or prevent cognitive compromise in diabetic patients, that can be relatively easily and rapidly implemented, and that are not cost prohibitive. This is especially true in the VA, in view of the high incidence of both diabetes and dementia in the growing population of elderly Veterans. Epidemiologic evidence suggests modifiable life-style factors, including cognitive activity, may prevent or delay the onset of cognitive decline. Computerized cognitive training (CCT) is an intervention that has shown promising results in the improvement of cognitive functioning, more consistently in non-demented elderly, with additional benefits from booster training sessions. To date, studies of CCT have typically only examined cognitive outcomes, and only shortly after the intervention. The proposed CCT program, Personal Coach from Cognifit, is designed to improve cognition of elderly persons by targeting their weak cognitive functions, using a personally tailored training plan. The proposed study will provide the first evaluation of the effects of CCT on DM self-management behavior and clinical outcomes, in addition to cognition. Project Objectives: Aim 1A: To determine whether the CCT, relative to the active control games program, improves cognition (memory and executive functions/attention), DM-related behavior (DM self-management and medication adherence), and clinical outcomes (glycemic and blood pressure control), 6 and 12 months after the intervention. Aim 1B: To demonstrate efficacy by improvement in behavioral outcomes (DM self-management and medication adherence) 6 months after the intervention. Aim 2: To document the effects of CCT on the successive changes in memory and executive functions/attention, DM self-management and medication adherence, and glycemic and blood pressure control. Aim 3: To explore the impact of demographic (age, education, ethnicity, site) and health (ADL/IADL, health literacy, depression, dementia family history, lifestyle factors) characteristics, on the intervention effects. Project Methods: Non-demented DM elderly from the James J. Peters (Bronx, NY) and Ann Arbor (MI) VAMCs will be randomized to CCT or games intervention and perform the respective program 3 days per week (every other day), for 20 minutes, for 24 total sessions. Four months after the intervention, subjects will receive a 1-week booster training. Subjects will be assessed at baseline; and immediately, 6 months, and 12 months after the intervention. At each time point, assessments will be cognitive function, DM self-management, and blood pressure; blood will be drawn for HbA1c measurement. VA records will be used to monitor medication adherence. Longitudinal mixed model analyses will assess the effects of the intervention on change in outcomes over time. Path analyses will evaluate the inter-relationships among changes in cognition, DM self-management, and clinical outcomes for each intervention at 6 and 12 months.

Randomly selected subjects perform a variety of computer games tailored to address their personal cognitive deficits.

Randomly selected subjects perform a variety of computer games that are engaging but not designed to enhance cognitive skills.

The data below represent values of the Diabetes Self-Management Questionnaire (DSMQ) scores at 6 months post-intervention, the primary outcome. DSMQ scores range from 0 to 20, with higher scores indicating better self-management, the same as baseline.

Medication adherence was assessed by the Continuous Multiple interval measure of Gaps in therapy (CMG), using diabetes medication refill data from the CPRS for the full year post-intervention. To calculate CMG percentage, the total number of days without medication refills when due - refill gap-days - was divided by 365, the days in a year. Multiplying this fraction by 100 provides the CMG percentage, where a higher percentage indicates worse medication adherence. The data below represent the CMG percentage at 12 months post-intervention.

As with baseline memory score, the outcome score was the mean of Word List Memory and Logical Memory (Story A) tests, each providing three scores: immediate recall, delayed recall, and recognition tasks after conversion to a common unit of measurement. However, to standardize these outcome tasks, the statistics from which the baseline z-scores were calculated (mean_baseline and SD_baseline from the full sample before randomization) were used, rather than the means of the given outcome period: standardized score = (outcome task score - mean_baseline) / SD_baseline. Thus, while baseline task scores were conventional z-scores, the six-month post-intervention standardized scores were not, in order to best inform the outcome measure and better facilitate comparison with baseline. The data below represent outcome values at 6 months. Like the baseline measure, this measure does not have a theoretical minimum or maximum, and higher scores indicate better performance.

As with baseline attention/executive function score, the outcome score was the mean of the Target Cancellation (TMX, Diamond), Trail Making (A, B; reversed), Digit Symbol Substitution, and, Digit Span (Forward, Backward) tasks after conversion to a common unit of measurement. However, to standardize these outcome tasks, the statistics from which the baseline z-scores were calculated (mean_baseline and SD_baseline from the full sample before randomization) were used, rather than the means of the given outcome period: standardized score = (outcome task score - mean_baseline) / SD_baseline. Thus, while baseline task scores were conventional z-scores, the six-month post-intervention standardized scores were not, in order to best inform the outcome measure and better facilitate comparison with baseline. The data below represent outcome values at 6 months. Like the baseline measure, this measure does not have a theoretical minimum or maximum, and higher scores indicate better performance.

A measure of glycemic control derived from blood sample. The data below represent outcome values at 6 months. Higher values represent poorer control of glycation.

A measure of the extent of pressure that blood is exerting against an individual's artery walls when the heart beats, assessed using a blood pressure cuff. Higher values indicate worse outcome. The data below represent outcome values at 6 months.

A measure of the extent of pressure of the blood in the arteries when the heart is filling, assessed using a blood pressure cuff. Higher values indicate worse outcome. The data below represent outcome values at 6 months.

The "Appointments - Delayed Recall" subtest of the Rivermead Behavioral Memory Test - Third Edition (RBMT - 3) is a test of prospective memory - remembering to do something in the future. The examinee was instructed to ask two questions when an alarm rings 25 minutes later. For each question, a score of 2 was given when a participant responded without requiring a prompt, 1 point when a prompt was needed and elicited a correct response, and 0 when there was no appropriate response despite prompting. Scores were summed and the total score ranges from 0 to 4. Higher scores indicate better performance. The data below represent outcome values at 6 months.

Inclusion Criteria: 55 years old or above A diagnosis of type 2 diabetes Access to computer and internet Has an informant Self-management score 18 or below Exclusion Criteria: Dementia or prescribed AD medications Major medical, psychiatric, or neurological conditions that affect cognitive performance Severe impairment of vision, hearing or fine motor control necessary for computer operation

Karran M, Guerrero-Berroa E, Schmeidler J, Lee PG, Alexander N, Nabozny M, West RK, Beeri MS, Sano M, Silverman JM. Recruitment of Older Veterans with Diabetes Risk for Alzheimer's Disease for a Randomized Clinical Trial of Computerized Cognitive Training. J Alzheimers Dis. 2019;69(2):401-411. doi: 10.3233/JAD-180952.

Silverman JM, Schmeidler J, Lee PG, Alexander NB, Beeri MS, Guerrero-Berroa E, West RK, Sano M, Nabozny M, Rodriguez Alvarez C. Associations of hemoglobin A1c with cognition reduced for long diabetes duration. Alzheimers Dement (N Y). 2019 Dec 11;5:926-932. doi: 10.1016/j.trci.2019.11.009. eCollection 2019.