Abstract
Background:
Older patients (≥65 years) with mild cognitive impairment (MCI) are undertreated for cardiovascular disease (CVD). One reason for this disparity could be that patients with MCI might underestimate the chances of CVD and overestimate dementia.
Objective:
To compare conceptions of health risk between older patients with MCI and normal cognition (NC) and their care partners.
Methods:
We conducted a multi-center mixed-methods study of patient-care partner dyads completing written quantitative surveys (73% response rate; 127 dyads: 66 MCI and 61 NC) or semi-structured interviews (20 dyads: 11 MCI, and 9 NC). Surveys assessed two-year patient risks of dementia, heart attack, stroke, and fall. Interviews assessed similar health risks and reasons for risk perceptions.
Results:
On surveys, a similarly low proportion of MCI and NC patients felt they were at risk of stroke (5% versus 2%; p = 0.62) and heart attack (2% versus 0%; p = 0.99). More MCI than NC patients perceived dementia risk (26% versus 2%; p < 0.001). Care partners’ survey findings were similar. Interviews generally confirmed these patterns and also identified reasons for future health concerns. For both MCI and NC dyads, personal experience with cognitive decline or CVD (personal or family history) increased concerns about each disease. Additionally, perceptions of irreversibility and lack of treatment for cognitive decline increased concern about dementia.
Conclusion:
Less use of CVD treatments in MCI seems unlikely to be driven by differential perceptions of CVD risk. Future work to improve awareness of CVD risks in older patients and dementia risk in patients with MCI are warranted.
INTRODUCTION
Up to 20% of older adults (65+) have mild cognitive impairment (MCI) today [1]. The prevalence of MCI is projected to triple by 2050 [2]. MCI is characterized as measurable cognitive decline that does not severely impair activities of daily living [3]. Some people with MCI progress to dementia (conversion rates range from 3% to 15% per year) [4, 5], while others remain stable or revert to normal cognitive status (14% -55%) [6]. As a result, many people with MCI live years with good quality of life and face competing health risks of aging, particularly cardiovascular disease (CVD) [7–9]. In a community-based study of older adults with MCI, CVD was a leading cause of death with dementia causing relatively fewer deaths [7]. Yet evidence suggests that those with MCI receive fewer guideline-concordant treatments for CVD and other conditions than people with normal cognition (NC) [10–12].
One possible reason for MCI patients’ lower use of guideline-concordant treatments for CVD could be misperceptions of either absolute or relative dementia and CVD risks. In particular, people with MCI and their care partners might overestimate risk for dementia and underestimate risk for CVD because people tend to overestimate rarer events and underestimate more common events [13]. People with poor self-rated memory and lower cognitive function perceive a greater threat of Alzheimer’s disease and dementia [14]. The extent to which risk perceptions differ between those with MCI and NC is unknown, but it is critical to understand how to target efforts to improve the quality of care in the large and growing population of older adults with MCI.
We conducted a multi-center, convergent, mixed-methods study [15] using surveys and interviews of older adults with MCI and NC and their care partners to determine how patient MCI diagnosis influences patients’ and care partners’ understanding of health risks, including dementia and the two most common CVD events, acute myocardial infarction (AMI) and stroke. We hypothesized that people with MCI and their care partners would underestimate CVD risk compared to people with NC and their care partners. In addition, we sought to understand reasons why participants were concerned about dementia or CVD and compare these reasons across MCI and NC groups. Finally, among people with MCI and their care partners, we also explored their understanding of the chances of memory problems getting worse.
MATERIALS AND METHODS
Study population
This study was approved by the Institutional Review Boards of the University of Michigan (U-M) and Duke University. Participants were recruited by mail and telephone from two study sites (U-M and Duke University) to complete either a written survey or a semi-structured interview. People with MCI and their care partners were identified from cognitive/memory disorders clinics, geriatrics clinics, medical records, active cognition studies, and Alzheimer’s Disease Centers at each site (the Michigan Alzheimer’s Disease Center (MADC) and the Joseph and Kathleen Bryan Alzheimer’s Disease Research Center at Duke University Medical Center). Participants with NC and their care partners were identified from established registries and active studies at U-M and Duke. At Duke, patients were prioritized for contact if they lived within a reasonable distance from the study site and had a family member or friend who was involved with their medical care. While the interview and survey samples were recruited from the same pool, enrollment in each portion was separate (e.g., the same individual did not do both survey and interview), as the interviews were conducted first to inform other separately reported portions of the survey design. In the survey portion of this study, we oversampled by race to have 50% Black and 50% White people with MCI, as other portions of the survey not reported in this manuscript were specifically focused on racial differences [16]. Survey participants recruited through sampling [17] of existing registries of people with and without cognitive impairment from each site as previously described [16]. At Duke, participants completed the survey in-person at a research office, therefore patients who lived in close proximity to Duke were prioritized in survey recruitment. To clarify terminology, we will use the term “patients” to refer to the designated individual with known MCI or NC when necessary to distinguish from care partners. The term “participant” will be used to refer to people without distinguishing between patients and care partners.
At both sites, the classification of patient MCI and NC were based on neuropsychological testing, clinical examinations, and expert clinician diagnosis following the 2011 National Institute on Aging–Alz-heimer’s Association (NIA-AA) diagnostic guidelines for MCI [18]. Eligible patients were required to be ≥65 year of age, have a diagnosis of MCI or NC within the last 12 months, read and speak English, self-identify as either Black or White race, and have a care partner ≥18 years of age who was able to read and speak English. There were no specific inclusion or exclusion criteria based on other medical or psychiatric co-morbidities. Eligible care partners were defined as a friend or family member identified by the patient as involved in his or her medical care; no formal cognitive screening was done for care partners. The survey sample was designed to comprise approximately equal numbers of people with MCI and NC.
Recruitment for survey and interview
Eligible patients for either the survey and the interviews were sent a letter introducing the study and subsequently called by study staff. Once a patient agreed to participate, study staff confirmed the contact information for their care partner and contacted them in the same manner. Survey participants were provided written information on key elements of informed consent, with return of the survey indicative of consent under an IRB-approved waiver of documentation of consent. Written informed consent was obtained for all interview participants. All survey and interview participants received a $50 incentive. At U-M, the survey was mailed once with an unconditional cash incentive included in the envelope; non-responders received reminders by telephone or email and were mailed a second survey without cash incentive up to 6 weeks later. At Duke, the survey was completed in-person at Duke University Medical Center (incentive provided after completion). The survey recruitment target was 50 dyads per site (total of 200 participants), based on sample size estimates for other analyses reported elsewhere. [16]
Survey data collection
The primary survey question asked participants about their self-perceived chance of developing specific health conditions (dementia, heart attack, stroke, and fall) during the next 2 years. Each of the four health conditions were considered separately in questions of the form, “I feel that I’m going to have a heart attack in the next 2 years.” Response options were on a 5-point Likert scale (disagree strongly, disagree, neither agree nor disagree, agree, agree strongly). Care partner questions were framed asking about the patient’s future risk. Prior work on self-reported risk perception has shown that similar questions about feeling at risk have better predicted behavior and outperformed risk magnitude judgements [19]. The survey also included questions on socio-demographics, clinical history, and treatment preferences that will be reported separately. Surveys were refined based on 2 rounds of cognitive interviews with 28 participants (including 8 with MCI) to ensure understanding in the target population. The wording of the outcome questions in the survey are available in Supplementary Material 1. The survey data was maintained in REDCap [20, 21] (Research Electronic Data Capture) tools hosted at U-M using double data entry.
Survey analysis
Descriptive statistics were summarized with frequency and percent or median and interquartile range (IQR). Responses to the questions about the 2-year risk perceptions for dementia, heart attack, stroke, and fall were summarized by patient cognitive status (MCI versus NC) and participant type (patient or care partner). Responses were then collapsed into a binary variable (agree strongly/agree versus all other responses) and analyzed by cognitive status with Fisher’s exact test. Logistic regression was initially planned but not done due to low cell counts in the strongly agree/agree group. Analysis was done with Stata SE (version 16) [22].
Interview data collection
Interview methods adhered to the consolidated criteria for reporting qualitative research [23]. Trained qualitative interviewers (co-authors CDK and JW, females) conducted a single in-person one-on-one interview with each individual separately. Participants did not know the interviewers before the study and only knew their occupation and that they were not medical professionals (CDK: qualitative research consultant and JW: PhD student and qualitative interviewer). The interviews took place at MADC and the Joseph and Kathleen Bryan Alzheimer’s Disease Research Center at Duke University Medical Center. Prior to the interview, participants completed questionnaires to collect age, gender, race, ethnicity, marital status, education level, and history of AMI and stroke. Care partners’ questionnaires also included details about the relationships (type, duration, frequency of contact), and their partner’s functional abilities using the Dementia Severity Rating Scale (DSRS) [24].
The interviewers used a standardized guide [25, 26]. The interview covered several topics not included in this manuscript. The current report focuses on responses to three questions on concern about future health problems in the next 2 years (biggest health concerns in general; whether they were more concerned about dementia, heart attack, or stroke; and whether they felt that dementia, heart attack, or stroke were most likely). MCI participants were also asked what they were told about the possibility that their memory problems may get worse and how they felt about this. Details of question wording is shown in Supplementary Material 2. All interviews were audio recorded and the interviewers made field notes during or after the interview. The interview duration was approximately 60 min. All interviews were professionally transcribed, with de-identified transcripts uploaded to the Dedoose web application (http://www.dedoose.com/) for analysis [27]. Data collection was continued until thematic saturation was achieved based on the inductive approach [28].
Interview analysis
We used descriptive qualitative methodology gro-unded in a naturalist philosophy, wherein the goal is to be “data-near,” reporting findings in their everyday terms, rather than more highly theorized [29]. The underlying epistemology is subjectivism [30]. In using inductive qualitative content analysis as our approach to coding [26, 31], we accepted data as representing our participants’ subjective perceptions, and our role as researchers in using our skills to interpret the phenomenon based on these perceptions as described by the participants [30]. This approach supports our goal of producing concrete findings for real-world practice.
The research team identified unifying and recurrent themes using qualitative content analysis [31]. The coding team consisted of a vascular neurologist (DBZ), internal medicine physician (DAL), qualitative researchers (JF, CDK), and study staff. The coding team initially read through the first several transcripts to identify participant concerns about future dementia and CVD as well as reasons for these concerns. Codes were not pre-specified but emerged from the data and were iteratively discussed by the coding team to produce stable code definitions, which were then used to code the transcripts. Once the initial coding definitions were established, weekly meetings were held with any questions about coding resolved by consensus. In addition, one-third of interview transcripts were coded by two coders, with any differences adjudicated by the entire coding team. Themes were developed by reviewing the individual codes and comparing findings across groups (MCI versus NC)
Mixed methods analysis
For the mixed methods analysis, we compared the survey findings on the proportion of MCI and NC participants who felt they were at risk of each condition (dementia and CVD) to the interview responses asking which conditions they were most concerned about and which they felt were most likely. In the discussion, we present our interpretation of how the interview findings on factors underlying future health concerns help explain the survey findings.
RESULTS
Survey population
A total of 190 dyads were screened for survey participation with 173 found to be eligible. Of those eligible, 136 dyads agreed to participate with complete responses obtained in 127 dyads (66 MCI and 61 NC). The response rate was 73.4% (127/173) and the survey completion rate was 93.4% (127/136). Details of the recruitment are shown in a STROBE [32] diagram in Supplementary Material 3. Participant characteristics for both the interview and survey samples are shown in Table 1, with dyad characteristics shown in Table 2. The majority of care partners were spouses, and the median duration of the relationship was 52 (40, 60) years for MCI dyads and 43 (28, 53) years for NC dyads.
Participant characteristics from interviews and surveys
MCI, mild cognitive impairment; NC, normal cognition; IQR, interquartile range; MoCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination; DSRS, Dementia Severity Rating Scale (higher = greater impairment); ADLs, Activities of Daily Living. * Indicates statistical significance at p < 0.05. Obtained using Mann-Whitney U and χ2 tests. **MoCA scores available for n = 7 interview participants. †MMSE scores available for n = 15 interview participants. ‡Symptoms of depression were defined based on a score of 2 or greater on the Patient Health Questionnaire-2.
Dyad characteristics from surveys and interview
MCI, mild cognitive impairment; NC, normal cognition; UM, University of Michigan; IQR, interquartile range. * Indicates statistical significance at p < 0.05. Obtained using Mann-Whitney U and χ2 tests.
Survey: Perceptions of future health risks by MCI status
Participant perceptions of future health risks in the next two years from the survey are summarized in Table 3. Few patients with MCI and NC felt that they were at risk of developing stroke (5% versus 2%; p = 0.62) or heart attack (2% versus 0%; p = 0.99) in the next two years. Patients with MCI were more likely than those with NC to feel that they would develop dementia in the next two years (26% versus 2%; p < 0.001). There was no difference in the perceived risk of fall between patients with MCI and those with NC (18% versus 13%; p = 0.47). Care partner responses were comparable to those of patients, with few care partners concerned about heart attack or stroke in both MCI and NC groups (Table 3). Care partners of patients with MCI were more likely to report feeling that the patient was going to develop dementia than care partners of patients with NC (46% versus 10%; p < 0.001).
Proportion of survey respondents concerned about 2-year risk of dementia, heart attack, stroke, or fall
Each cell reports the percentage of respondents who answered “Agree” or “Strongly Agree” to questions of the form “I feel that I’m going to have a X in the next 2 years” with response options on a 5-point Likert scale (disagree strongly, disagree, neither agree nor disagree, agree, agree strongly. *Care partners responded in their assessment of the patient’s risk. MCI, mild cognitive impairment; NC, normal cognition. *p-value results from Fisher’s exact tests. A test could not be performed on care partner’s assessment of heart attack risk due to low cell counts.
Interview population
A total of 20 dyads completed interviews (11 MCI, 9 NC). Details of the recruitment are shown in a STROBE [32] diagram in Supplementary Material 3. The majority of care partners were spouses and the median duration of the relationship between patient and care partners was over 50 years.
Interview findings
Comparison of qualitative findings with survey results
When participants were asked to compare whether they felt that dementia, heart attack, or stroke were most likely in the next two years, the responses generally confirmed findings from the survey. Among MCI dyads, dementia was mentioned more commonly than heart attack or stroke, especially among MCI care partners where over 70% felt that dementia was most likely. Among NC dyads, one-third indicated that dementia was most likely, while slightly less than half indicated that stroke or AMI was most likely.
Reasons underlying future health concerns
We identified several overarching themes regarding reasons for participants’ future health concerns. Regardless of patient MCI status: 1) Personal experience with cognition or CVD (i.e., observations of cognitive decline versus stability, personal history of CVD or CVD risk factors, family history) increased future health concerns about each disease; 2) The perceived irreversibility of and lack of available treatment for dementia led to greater concern about dementia; and 3) Some participants’ general attitude of trying to stay healthy and not worry about the future decreased future health concerns. Below we further describe these themes and how they compared across MCI and NC groups, with additional exemplar quotes summarized in Table 4.
Interview themes and exemplar quotes of individuals with MCI, normal cognition, and care partners
MCI, mild cognitive impairment; NC, normal cognition; CVD, cardiovascular disease.
Personal experience: observations of cognitive decline increased concern about dementia, while cognitive stability decreased concern
Personal observations regarding the trajectory of memory changes seemed to underlie either increased or decreased concerns about future dementia depending on whether cognition was observed to be worsening or stable. This MCI patient with worsening memory was typical:
My memory, number one. I’m scared I won’t be able to function, and it’s getting worse every day. That would be the number one problem. (Dyad DU003, MCI patient)
Not surprisingly, worsening memory trajectory leading to concern about dementia was more commonly seen in MCI participants than NC participants, though concerns about worsening cognition were also noted among those classified as NC:
I would say my number one concern would be declining cognitive function. We have noticed a little bit of a difference in the past several years. (Dyad DU001, NC care partner)
Along the same lines, participants who perceived cognition as stable were relatively less concerned about dementia:
I think a heart attack or stroke would be a more major concern than dementia. Because I think she is still pretty sharp and I think that she does have health issues though, yeah. (Dyad UM0018, NC care partner)
Personal experience: Personal history of CVD or CVD risk factors increased concerns about CVD
Similar to concerns about personal observations of cognitive trajectory, personal history of stroke, CVD, or risk factors for these conditions were important drivers of CVD risk perceptions. Personal history of CVD or risk factors informed risk in both MCI and NC groups.
And the only thing I’m just scared of, really scared of my heart. You know, because that’s the main –you know, the main organ in your body . . . that’s where I had my problems at. (Dyad UM013, MCI patient)
I’m more likely to have a heart attack or stroke.
Interviewer: And you base that on?
My diet. My food. My lack of exercise. (Dyad UM004, NC patient)
Personal experience: Family history increased concerns about CVD and dementia
In addition to personal history, family history of either CVD or dementia influenced CVD and dementia risk perceptions for many participants. This finding was seen across MCI and NC groups, as each group mentioned family history of dementia or CVD as an important factor contributing to health risks of either condition.
Pop had Alzheimer’s. His father did . . . it runs in the family, which scares me. (Dyad DU004, MCI patient)
Probably the dementia because I saw what it did to my mother. (Dyad DU008, MCI care partner)
Perceived irreversibility of and lack of available treatment for dementia led to greater concern about dementia
The lack of available prevention or treatment strategies led participants to be more concerned about dementia relative to CVD. Several participants also seemed to be concerned about the irreversibility of dementia. These concerns occurred at a similar frequency in the MCI and NC groups.
I guess, my attitude is with stroke or heart attack, I can do something to try to be proactive with that. I don’t feel that same way with dementia, so I guess my biggest concern would be dementia. (Dyad UM009, NC patient)
Probably dementia as opposed to heart attack or stroke, because heart attack and stroke, they seem to be able to bring people back from that more so than dementia. So, I would be concerned more about dementia than anything else because they don’t seem to be able to reverse that or make it any better. (Dyad UM10, MCI care partner)
General attitude of trying to stay healthy and not worrying about the future decreased future health concerns
Other participants seemed to have a general attitude of not worrying about the future or about things that were out of their control. These sentiments also occurred to a similar degree among MCI and NC participants.
You know, we don’t wait for bad things to happen. We deal with it if it happens. But, you know, can’t predict the future, so I’m not worried about it. (Dyad DU002, NC care partner)
I don’t think about those things. Because it if weighs on your mind, it’ll just make you sick. So I always bury them. You think about them . . . you can drive yourself crazy. You know, drive yourself into a condition where you just be in terrible health. (Dyad DU008, MCI patient)
MCI participant understanding of chances of memory worsening
Patients with MCI and their care partners were asked what they had been told about the “chances of [your/his/her] memory problems getting worse.” The most common response, seen in about half of participants, was that they had been told “nothing” regarding their chances of memory problems worsening. Only 6 of 22 participants indicated that they had been explicitly told about the possibility of memory problems worsening, with the remainder indicating that they had only been told something vague.
Patients’ and care partners’ reactions to not knowing the prognosis for memory difficulties were generally split among feeling scared, upset or frustrated, and some level of acceptance. Many who were accepting of not knowing the future commented on their inability to change the future. This is exemplified by the participant who said, “Well, you’re concerned, but it’s part of life. It’s like somebody saying, you know, ‘Next week you might get in a serious car accident and be crippled.’ You know, gee, that’d be horrible, but . . . what can I do about it today? I have to face things when they happen. So, and I don’t waste a lot of time today worrying about what I’m going to forget about tomorrow.” (Dyad UM003, MCI patient)
Other MCI patients and care partners were accepting of the uncertain future as they had transitioned to making decisions together or made other arrangements in case the patients were unable to make decisions themselves due to cognition dysfunction.
Interviewer: So how do you feel about making medical decisions [ . . . ] not knowing sort of how it’s going to progress?
I think by making the decisions together which we do. We make everything together; it’s based on the here and now. That’s the way we do it. What will happen down the road, we will deal with as effectively as possible [ . . . ] and the children will be part of the decision as well. I’m very aware of some of his wishes. But until I truly believe that he’s incapable in participating in those decisions then he will be a part of the discussion always. (Dyad UM0016, MCI care partner)
Oh yeah, we already have all that in place, and goodness, our power of attorney and his do not resuscitate. And he has his living will and I’m his power of attorney there. But yeah, I know it’s going to come down to I’m just going to be the one to make all of the decisions. And all I can do is do it in as informed manner as I can. (Dyad DU004, MCI care partner)
DISCUSSION
In this convergent mixed methods study, we investigated whether differences in self-perceived health risks might explain observed differences in CVD treatment between people with MCI and those with NC. In the quantitative survey, we found no evidence of differences in the proportion of participants who felt they were at risk of stroke or MI between those with MCI and those with NC. In fact, the low level of concern for CVD in both groups limited our ability to analyze any differences. Further, MCI patients and care partners were more likely than those with NC to feel they were going to develop dementia. These findings are generally consistent with other literature indicating a tendency for people to underestimate risks of CVD [33] and overestimate the risk of dementia [34–36]. The low level of concern for CVD in both MCI and NC is remarkable given that study patients were largely in the eighth decade of life, where CVD risks are high and CVD is the leading cause of death and disability. While further investigation may be warranted in larger samples or with other methods of assessing self-perceived risks, our current data do not support the hypothesis that differences in self-perceived risks of CVD are a major driver of the MCI patients’ lower receipt of guideline-concordant treatments for CVD compared with NC patients [10, 11].
Qualitative interview findings generally confirmed the survey results, with MCI dyads indicating greater concerns about dementia than NC dyads, with generally similar concerns for CVD across MCI and NC groups. While the greater concern about dementia among MCI participants may seem intuitive, our qualitative interview data provide some insight into what factors inform patient and care partner concerns about future health risks between MCI and NC groups. These underlying reasons will be important for clinicians and researcher to consider when counseling individual patients on health risks or developing materials to improve their accuracy of risk perception.
The most common factor that we noted to influence future health concerns was personal experience with disease, which was seen in both the MCI and NC groups and for both dementia and CVD. Thus, personal observations of cognitive decline (or stability) led to increased (or decreased) concern about dementia risk regardless of whether the patient was classified as MCI or NC. As people diagnosed with MCI tended to more commonly report observations of cognitive decline, this factor likely contributed to the greater concerns about dementia in this group seen in both qualitative and quantitative data. Similarly, people with personal history of CVD or risk factors cited these issues as reasons for concern about future risk of stroke or heart attack. Family history was important for both CVD and dementia risk, consistent with prior literature demonstrating the important of family history in informing self-perceptions of risk [37–39]. Taken together, these findings are consistent with individuals’ experience with disease in themselves or their family being an important driver of future health risk perceptions [40].
Another factor contributing to concern about future dementia risk was the perceived lack of effective prevention or treatment strategies for dementia, in contrast to myocardial infarction and stroke. Prior research has shown that patients with MCI are frustrated at the lack of ‘treatments’ available for MCI and are often anxious to halt or slow cognitive decline [41]. Only a small number of participants discussed preventative treatments and actions to minimize the progression from MCI to dementia, though strategies do exist [3]. Pharmacologic treatments have not been definitively shown to slow the progression of MCI and dementia [42], though it may be worth investigating whether attitudes about dementia risk change as new treatments are developed. However, stroke and cardiovascular risk factors increase the risk for dementia [43–45]. Interventions aimed at preventing stroke, and reducing cardiovascular risk factors, particularly hypertension, are promising strategies to reduce risk for dementia [3, 46]. Our findings are consistent with other research showing that many people do not know that CVD and CVD risk factors increase risk for dementia [47].
Relatively few of the MCI dyads in our group reported having a good understanding of the chances of cognitive worsening in MCI. Prior work has shown that patients may lack clarity surrounding the terminology, prognosis, and treatments of MCI [41, 48]. Our finding that many of those with MCI indicated that they had been told “nothing” of their future dementia risks supports this prior work. While these data are dependent on self-report, and we cannot verify whether prognostic information was requested by participants or provided by clinicians, these findings are still important as they represent the current perceptions of MCI patients and their caregivers. It is unclear from our data if the problem is that clinicians do not discuss prognosis, or that patients and families do not desire, understand, or retain the information that is being provided. As the MCI dyads in our study were recruited from specialized Alzheimer’s clinics and had an established MCI diagnosis, we would expect them to be among the most informed about MCI. It is likely that this lack of understanding of MCI prognosis and treatment is even worse in others not managed at specialized centers. Overall, our findings support continued and expanded efforts[49] to optimize communication and understanding of MCI prognosis to try to limit distress and support future planning and decision making.
Strengths and limitations
Our study has several strengths. We met recruitment goals and achieved a high response rate for the survey. We used clinician’s diagnosis of MCI and NC using established methodology. We pre-tested the survey instrument with cognitive interviewing to ensure patients, particularly those with MCI, and care partners could understand the survey questions. By conducting in-depth interviews with MCI patients and their care partners as well as NC patients and their care partners, we were able to obtain a rich and informative dataset from a variety of perspectives.
We acknowledge there are limitations in the present study. Results may not be generalizable to the broader population of people with MCI or NC due to our recruitment from dedicated Alzheimer’s clinics and registries; In particular, we did not collect detailed data on psychiatric comorbidities, diabetes, medication use, or sleep patterns which limited our ability to compare to other populations or consider these factors in our analysis. Survey and interview data came from different participants though they were recruited from the same populations. While it does not appear that differential risk perceptions of CVD contribute to less use of CVD treatments among those with MCI, it is still possible that MCI patients’ higher sense of dementia risk may affect their decision making and preferences in other ways for other treatments including CVD. Our quantitative assessment of future health risks did not include specific numeric risks, though this was intentional due to concerns about comprehension and numeracy in this population. The low overall perceptions of CVD risk limited our ability to conduct multivariable analyses, such as investigating whether importance of personal history of heart disease or stroke seen in the qualitative interview group was also seen in the survey sample.
Conclusions
Less use of guideline-concordant CVD treatments among those with MCI seem unlikely to be driven by differential perceptions of CVD risk. MCI and NC dyads similarly perceived a low CVD risk which may not be accurate given the high CVD risk in older adults. Future work to develop interventions to improve awareness of future health risk and dementia prevention strategies are warranted for both people with MCI and those with NC.
Footnotes
ACKNOWLEDGMENTS
This work was supported by NIH/NIA grant R01 AG051827 (Levine DA, PI). Ms. Kollman is employed by Kollman Research Services, Ann Arbor, MI. The funder provided support in the form of salaries for authors but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. REDCap at the University of Michigan is supported by UL1TR002240.
