Abstract
Background:
To examine willingness of older patients to take less diabetes medication (de-intensify) and to identify characteristics associated with willingness to de-intensify treatment.
Methods:
Survey conducted in 2019 in an age-stratified, random sample of older (65-100 yrs) adults with diabetes on glucose-lowering medications in the Kaiser Permanente Northern California Diabetes Registry. We classified survey responses to the question: “I would be willing to take less medication for my diabetes” as willing, neutral, or unwilling to de-intensify. Willingness to de-intensify treatment was examined by several clinical characteristics, including American Diabetes Association (ADA) health status categories used for individualizing glycemic targets. Analyses were weighted to account for over-sampling of older individuals.
Results:
A total of 1,337 older adults on glucose-lowering medication(s) were included (age 74.2 ± 6.0 years, 44% female, 54.4% non-Hispanic white). The proportions of participants willing, neutral, or unwilling to take less medication were 51.2%, 27.3%, and 21.5%, respectively. Proportions of willing to take less medication varied by age (65-74 years: 54.2% vs. 85+ years: 38.5%) and duration of diabetes (0-4 years: 61.0% vs. 15+ years: 44.2%), both p<0.001. Patients on 1-2 medications were more willing to take less medication(s) compared to patients on 10+ medications (62.1% vs. 46.6%, p= 0.03). Similar proportions of willingness to take less medications were seen across ADA health status and HbA1c. Willingness to take less medication(s) was similar across survey responses to questions about patient-clinician relationships.
Conclusions:
Clinical guidelines suggest considering treatment de-intensification in older patients with longer duration of diabetes, yet patients with these characteristics are less likely to be willing to take less medication(s).
Background
Glucose lowering therapies are known to delay long-term complications of diabetes.1,2 However, clinical trials have had limited representation of the oldest patients (>75 years of age) and have excluded those with multi-comorbidities or at high risk of hypoglycemia.3 Post-hoc analysis of results among patients with long-standing diabetes in the ACCORD and ADVANCE trials have found an increased risk of complications such as hypoglycemia, hospitalizations, and death among patients who achieve an HbA1c <7%.4,5 Other epidemiologic6,7 and decision analysis studies have8 suggested that patients with multiple comorbidities and limited life expectancy are less likely to benefit from intensive glucose control. These studies have influenced current guidelines (e.g., ADA and Endocrine Society) emphasizing relaxed glycemic goals and less intensive therapies for those who are older, have long duration of diabetes, have poor health or limited life expectancy.9,11 Thus, regular reevaluation of treatment intensity is needed as patients age and health status changes.
Treatment de-intensification has been proposed as an approach to improve safety and limit burden of treatment in older, medically complex patients with diabetes and limited life expectancy. Yet, studies have shown a low rate of de-intensification among patients who are considered high risk for treatment-related complications, namely hypoglycemia.7,10 De-intensification requires both clinician and patient to engage in a shared decision-making process.11 A recent national survey reported physicians frequently would not de-prescribe hypoglycemia-causing medications and preferred lower Hemoglobin A1c (HbA1c) targets than recommended by guidelines for complex patients with diabetes and poor health.12 Furthermore, clinicians who propose de-intensification may be asking patients to stop medications which had been effective for many years. In one study, patients reported that they felt more aggressive treatment should be carried out specifically in patients with longer duration of diabetes, more established complications, and a greater number of other health conditions.13 These perceptions held by both clinicians and patients may contribute to potential difficulties in successfully implementing current guideline recommendations for treatment de-intensification. Several qualitative and mixed-methods studies assessed variables associated with patient willingness to de-intensify. Often patients valued provider’s opinion on deintensification, patient’s perception of achieved glycemic control, and medication benefits as factors influencing their decision for treatment de-intensfication.14,15,16,17 Accordingly, it is not clear how clinicians should approach de-intensification with their patients and what barriers may exist when communicating with patients about potential de-intensification.
Current studies examining patient preferences include middle age or younger patients with a small sample size and/or low survey response rate.12,13 An important gap in the current literature is that few studies have directly assessed the degree to which older patients are willing to take less medications and what characteristics are associated with willingness to de-intensify. In the current study, we examined the prevalence and correlates of willingness to de-intensify medications and how these align with guideline recommendations. The goal of this research is to help clinicians better understand patient goals, potential barriers to de-intensification and to achieve a more collaborative, patient-centered treatment plan.
Methods
The Diabetes & Aging Study is an ongoing, multidisciplinary research project which began in 2009 to study diabetes in older adults who are largely excluded from clinical trials and for whom the lack of clinical evidence has led to conflicting and ambiguous recommendations for care. In 2019, the Diabetes & Aging Study launched a survey (the Diabetes Preferences and Self-Care Survey) to characterize treatment preferences, self-management barriers, patient-reported outcome, and patient-provider relationships, data which are not typically available in an electronic medical record (EHR). The survey development began with finalizing our research questions that would inform the survey content, identification of appropriate validated survey items, development of de novo questions (e.g., willingness to de-intensify) followed by pilot testing and item modification when indicated. This survey was conducted among patients from the Kaiser Permanente Northern California (KPNC) Diabetes Registry, a large, diverse, real-world population which has been followed longitudinally since its inception in 1994.18 KPNC is an integrated healthcare delivery system providing Medicare Advantage to all patients over the age of 65 years. The KPNC EHR comprehensively captures clinical and laboratory data. It is updated annually based on diabetes indicators obtained from HER data including inpatient and outpatient diagnoses, lab results, and medication dispensing. Subjects were eligible for the survey if they were included within the diabetes registry, age ≥ 65 years, and their preferred written language was English, Spanish, or Chinese. Subjects were excluded if they had a diagnosis of cognitive impairment or were not community dwelling. An age-stratified (65-74, 75-84, 85-100 years of age as of 01/01/2019), random sample (N=6,000, 50.5% female) was invited to participate in the survey. A written survey was mailed to each subject in their preferred language (English [n=5,505], Spanish [n=314] and Chinese [n=181]) with an option to complete the survey by telephone interview or online in their preferred language. Older participants were oversampled with an equal number of respondents in each age group; hence analyses were weighted to address the complex sampling design.
Cross-sectional analysis of survey responses and EHR data was conducted for patients who were dispensed a glucose-lowering medication within 6 months prior to the survey date. The outcome of interest was “willingness to de-intensify” based on survey response to the following question, “I would be willing to take less medication for my diabetes,” and categorized into 3 groups: “strongly agree” or “agree” were categorized as “willing”; “neither agree nor disagree” was categorized as “neutral”; and “strongly disagree” or “disagree” were categorized as “unwilling.” The prevalence of willingness to de-intensify, after survey weighting, was calculated along with associations of willingness to de-intensify with sociodemographic and clinical factors. Age, sex, race ethnicity, and comorbidities were captured from electronic medical record at baseline. Most recent HbA1c present in the last 24 months prior to survey completion was used for analyses. Maintained medications filled 6 months prior to survey completion were included within medication count. Contextual socioeconomic status was based on a validated neighborhood deprivation index (NDI) calculated using geocoded census tract data.19,23 Diabetes duration, marital status, and living situation were based on self-report. The number of missing data were very low for baseline characteristics [diabetes duration (n=174), HbA1c (n=9), marital status (n=16), living situation (n=32), hypoglycemia in the past 4 weeks (n=78), and education (n=37)].
We also assessed self-reported patient-provider relationship based on responses to the following survey questions: “In the past 12 months, how often did your primary care provider (PCP) explain things in a way you could understand?” (validated CAHPS item for patient understanding), “… and how often did your PCP involve you in making decisions about your care as much as you wanted?(collaborative decision making), and “…how often have you felt confidence and trust in your PCP?( validated trust in physician item). Weighted generalized linear models (binomial distribution with identity link) with robust error variance were used to assess statistical significance of differences in willingness to de-intensify (“willing” versus “neutral” or “unwilling”) by patient characteristics. We also assessed health status (good, intermediate, or poor health) using a modified version of the American Diabetes Association health status categories, which are used to stratify patients and recommendations for care based on comorbidities, functional status, and cognitive impairment.11,22 Additional information about patient perceptions of the benefits and risks of taking less medication was gathered from these survey questions: “What would be the benefits to you of taking less medication for diabetes (check all that apply)? (Simplify my treatment routine, reduce my risk of side effects, reduce my medication costs, I do not think there would be any benefits, or other)” and “What would be the risks to you of taking less medication for diabetes (check all that apply)? (Increase my blood sugar, increase my risk of complications, I do not think there would be any risks, or other)”. The study received institutional review board approval from the Kaiser Foundation Research Institute (Oakland, CA).
Results
A total of 2,200 active Kaiser Permanente patients with type 2 diabetes responded to the survey and 469 patients were excluded for not filling at least one diabetes medication. An additional 391 patients were excluded for not answering or answering “don’t know” to the question about willingness to take less medications. A total sample of 1,337 was used for our analysis. Mean age of participants was 74.2 ± 6.0 years; 44% (n=593) were female, and 54.4% (n=744) non-Hispanic white. Mean HbA1c was 7.3% (SD ± 1.1%) and was similar among different age groups (Table 1). Most patients (72.1%) filled more than 5 medications on an ongoing basis, one of which was a diabetes medication; 34.1% (n=454) were on insulin and 69.9% (n=934) on metformin. Average duration of diabetes was 17.5 years (SD ± 12.0 years) with a higher proportion of individuals 85+ years having diabetes for 15+ years compared to those aged 65-74 (69.1% vs 41.9%, p <0.001). Compared to individuals 65-74 years of age, a higher proportion of individuals aged 85+ years had microvascular disease as well as several significant comorbidities including CKD, heart failure, history of cancer and of falls (Table 1).
Table 1.
Baseline Characteristics of Survey Respondents
| Characteristics | Age Category: n (%) | p-value† | |||
|---|---|---|---|---|---|
| All* | 65-74 years | 75-84 years | 85+ years | ||
| Age in years, mean ± SD | 74.2 ± 6.0 | 70.1 ± 2.7 | 79 ± 2.6 | 88.4 ± 2.9 | < 0.001 |
| Sex | 0.69 | ||||
| Male | 56% | 268 (55.6%) | 300 (56.8%) | 176 (53.8%) | |
| Female | 44% | 214 (44.4%) | 228 (43.2%) | 151 (46.2%) | |
| Race | 0.09 | ||||
| White | 54.4% | 254 (52.7%) | 298 (56.4%) | 192 (58.7%) | |
| Black | 7.9% | 37 (7.7%) | 50 (9.5%) | 12 (3.7%) | |
| Hispanic | 11.9% | 58 (12%) | 63 (11.9%) | 39 (11.9%) | |
| Asian | 19.5% | 100 (20.7%) | 89 (16.9%) | 64 (19.6%) | |
| Other | 6.4% | 33 (6.8%) | 28 (5.3%) | 20 (6.1%) | |
| Diabetes duration | <0.001 | ||||
| 0-14 years | 53.1% | 241 (58.1%) | 224 (48.1%) | 87 (30.9%) | |
| 15+ years | 46.9% | 174 (41.9%) | 242 (51.9%) | 195 (69.1%) | |
| HbA1c | |||||
| Mean ± SD | 7.3% ± 5.5% | 7.2% ± 1.1% | 7.3% ± 1.1% | 7.3% ± 1.1% | 0.46 |
| <7% | 41.9% | 206 (43%) | 212 (40.2%) | 131 (40.8%) | |
| 7-9% | 51.5% | 241 (50.3%) | 283 (53.6%) | 162 (50.5%) | |
| >9% | 6.6% | 32 (6.7%) | 33 (6.3%) | 28 (8.7%) | |
| Insulin use | 34.1% | 160 (33.2%) | 192 (36.4%) | 102 (31.2%) | 0.27 |
| Non-insulin hyperglycemic agent | |||||
| Metformin | 69.9% | 389 (80.7%) | 362 (68.6%) | 183 (56%) | <0.001 |
| Sulfonylurea | 46.5% | 220 (45.6%) | 249 (47.2%) | 174 (53.2%) | 0.09 |
| GLP-1 RA | 0.8% | 5 (1%) | 3 (0.6%) | 0 (0%) | 0.17 |
| SGLT2i | 0.3% | 1 (0.2%) | 2 (0.4%) | 0 (0%) | 0.52 |
| DPP4i | 0.7% | 3 (0.6%) | 4 (0.8%) | 4 (1.2%) | 0.64 |
| Medication count | 0.06 | ||||
| 1-2 medications | 7% | 41 (8.5%) | 23 (4.4%) | 21 (6.4%) | |
| 3-4 medications | 20.9% | 105 (21.8%) | 106 (20.1%) | 62 (19%) | |
| 5+ medications | 72.1% | 336 (69.7%) | 399 (75.6%) | 244 (74.6%) | |
| Preferred language | 0.24 | ||||
| English | 93.7% | 452 (93.8%) | 492 (93.2%) | 308 (94.2%) | |
| Spanish | 3.4% | 16 (3.3%) | 21 (4%) | 5 (1.5%) | |
| Other | 2.9% | 14 (2.9%) | 15 (2.8%) | 14 (4.3%) | |
| Deprivation index | 0.05 | ||||
| 1st quartile | 28.9% | 125 (25.9%) | 171 (32.4%) | 121 (37%) | |
| 2nd quartile | 38.6% | 192 (39.8%) | 192 (36.4%) | 117 (35.8%) | |
| 3rd quartile | 17.3% | 88 (18.3%) | 87 (16.5%) | 44 (13.5%) | |
| 4th quartile | 15.3% | 77 (16%) | 78 (14.8%) | 45 (13.8%) | |
| Marital status | <0.001 | ||||
| Married/living with someone | 65.2% | 319 (67%) | 341 (65.1%) | 145 (45.2%) | |
| Divorced/widowed/separated/single | 34.8% | 157 (33%) | 183 (34.9%) | 176 (54.8%) | |
| Living situation | <0.001 | ||||
| Lives with others | 69.8% | 341 (71.9%) | 353 (68.3%) | 173 (54.9%) | |
| Lives alone | 20.3% | 92 (19.4%) | 107 (20.7%) | 89 (28.3%) | |
| Senior/retirement facility | 1.4% | 3 (0.6%) | 8 (1.5%) | 27 (8.6%) | |
| Other | 8.6% | 37 (7.8%) | 49 (9.5%) | 26 (8.6%) | |
| Education | 0.44 | ||||
| No degree | 10.6% | 48 (10.2%) | 58 (11.3%) | 41 (13.1%) | |
| High school/GED | 42.5% | 199 (42.2%) | 220 (42.7%) | 147 (47%) | |
| College | 45.7% | 220 (46.6%) | 229 (44.5%) | 121 (38.7%) | |
| Microvascular disease present | 56.2% | 244 (50.6%) | 336 (63.6%) | 229 (70%) | <0.001 |
| Comorbidities | |||||
| Hypertension | 90.1% | 420 (87.1%) | 502 (95.1%) | 299 (91.4%) | <0.001 |
| Prior MI | 7.3% | 26 (5.4%) | 52 (9.8%) | 34 (10.4%) | 0.01 |
| History of CABG | 14% | 54 (11.2%) | 94 (17.8%) | 72 (22%) | 0.001 |
| Prior stroke | 3.1% | 12 (2.5%) | 22 (4.2%) | 11 (3.4%) | 0.34 |
| CKD, stage 3-5 | 28.8% | 100 (20.7%) | 207 (39.2%) | 169 (51.7%) | <0.001 |
| COPD | 23.1% | 90 (18.7%) | 159 (30.1%) | 90 (27.5%) | <0.001 |
| Heart failure | 12.6% | 44 (9.1%) | 88 (16.7%) | 80 (24.5%) | <0.001 |
| History of cancer | 14.2% | 58 (12%) | 92 (17.4%) | 60 (18.3%) | 0.02 |
| Depression | 21.5% | 112 (23.2%) | 102 (19.3%) | 46 (14.1%) | 0.01 |
| History of falls | 20.7% | 77 (16%) | 136 (25.8%) | 138 (42.2%) | <0.001 |
Data present within all columns represents weighted data.
p-value represents statistical differences between age categories.
In the weighted sample, the proportions willing, neutral, and unwilling to take less medication were 51.2%, 27.3%, and 21.5%, respectively. Individuals who were willing to take less medication compared to those neutral or unwilling were more likely to endorse anticipated benefits of taking less medication like simplification of treatment regimen, reduction in side effects, or reduction in cost (willing: 77% vs. neutral/unwilling: 42%), p<0.0001, see Figure 1). Though significant statistically (p<0.0001), the differences in reported risks (increased complications or increase in blood glucose) to taking less medication were relatively similar among willing (84%) compared to neutral or unwilling (92%).
Figure 1.
Patients perceived benefits(left) and risks(right) of de-intensification among individuals who were willing, neutral, or unwilling to take less medications for diabetes. SE: side effects. BG: blood sugar.
Willingness to take less medication varied significantly (p<0.001) by age group with respondents aged 65-74 more likely to be willing to take less medication compared to adults age 85+ (Table 2). Willingness to take less medication was inversely associated with duration of diabetes and number of medications.
Table 2.
Percent willing to take less medication (de-intensify) by patient characteristics.
| % Willing to de-intensify | p-value* | |
|---|---|---|
| Age (years) | p<0.0001 | |
| 65-69 | 64.2% | ref |
| 70-74 | 44.4% | <0.001 |
| 75-79 | 50.3% | <0.001 |
| 80-84 | 41.4% | <0.001 |
| 85+ | 38.6% | <0.001 |
| Diabetes Duration (years) | p=0.12 | |
| 0-4 years | 60.3% | ref |
| 5-9 years | 54.3% | 0.38 |
| 10-14 years | 52.1% | 0.22 |
| 15+ years | 47.2% | 0.03 |
| HbA1c | p=0.57 | |
| <7% | 50.0% | ref |
| 7-9% | 50.8% | 0.81 |
| >9% | 57.0% | 0.28 |
| ADA Health Status | p=0.09 | |
| Good health | 54.2% | ref |
| Intermediate health | 47.5% | 0.05 |
| Poor health | 55.1% | 0.87 |
| Charlson Comorbidity Index | p=0.18 | |
| 1 | 57.8% | ref |
| 2-3 | 49.9% | 0.11 |
| 4+ | 49.5% | 0.07 |
| Microvascular Disease | p=0.01 | |
| present | 47.5% | 0.01 |
| absent | 55.4% | ref |
| Medication Count | p=0.046 | |
| 1-2 meds | 62.1% | ref |
| 3-4 meds | 56.3% | 0.39 |
| 5-9 meds | 48.7% | 0.03 |
| 10+ meds | 46.6% | 0.03 |
| Insulin Use | p=0.31 | |
| yes | 48.7% | 0.31 |
| no | 52.3% | ref |
| Hypoglycemia in past 4 weeks | p=0.68 | |
| present | 52.0% | 0.68 |
| absent | 50.6% | ref |
| Education | p=0.09 | |
| no degree | 44.3% | ref |
| high school | 49.2% | 0.37 |
| college or more | 55.5% | 0.04 |
From binomial regression with identity link and robust standard error. Models patient willingness as “willing” vs. “neutral or unwilling” with the patient characteristic as the single independent variable and weighted by the inverse of the sampling fractions used in the complex survey design.
Health Status and Co-morbidities:
Based on ADA health status definitions,11 39.3% of respondents were classified as having good health, 50.3% intermediate health, and 10.4% poor health. Willingness to take less medications did not vary significantly by number of comorbidities as defined by ADA health status or the Charlson Comorbidity Index. Patients with a history of microvascular disease were less willing to take less medications compared to those with no microvascular disease (47.5% vs 55.4%, p=0.01).
Hemoglobin A1c:
Willingness to take less medications did not vary by HbA1c with half of patients in each category willing to take less medication for their diabetes (Table 2). In exploratory analyses, the proportion of those willing to take less medication was the highest among those with poor health and HbA1c >9% (Supplemental Table 1). Those with good or intermediate health and better glycemic control were markedly less likely to be willing to take less medication.
Hyperglycemia and Hypoglycemia:
A total of 36.4% respondents reported experiencing an episode of hypoglycemia within the 4 weeks prior to completing the survey. A higher proportion of respondents had hypoglycemia if their HbA1c was >9% compared to <7% (44.1% vs 28.9%, supplemental Table 2). Presence of hypoglycemia within the last 4 weeks was not significantly associated with a higher percentage of willingness to take less medications compared to those without hypoglycemia (52.0% vs 50.6%, respectively, p=0.68).
Polypharmacy and Insulin:
Willingness to take less medication decreased as the number of diabetes and non-diabetes medications a patient filled increased (1-2 meds: 62% vs. 10+ meds: 47%, p=0.03, Table 2). Lastly, willingness to take less did not differ by insulin use (Table 2).
Patient-Clinician Relationship:
Most patients reported they always had trust in their clinician (70.9%) and their clinician always explained things in a way they could understand (62.8%). However, only 57.3% of patients reported clinicians always involved them in making decisions about their care. Patients reported a similar percentage of willingness to take less medications across all levels of patient-clinician relationships (Table 3).
Table 3.
Percent willing to take less medication and patient-clinician relationship.
| % Willing to take less medications | Missing data: n (%) | p-value* | |
|---|---|---|---|
| Trust in clinician Q: In the past 12 months, how often have you felt confidence and trust in your PCP? |
198/1337 (14.8%) |
p=0.74 | |
| Always | 49.2% | ref | |
| Usually | 53.5% | 0.30 | |
| Sometimes | 53.3% | 0.56 | |
| Never | 50.0% | 0.95 | |
| Collaborative decision making Q: In the past 12 months, how often did your PCP involve you in making decisions about your care as much as you wanted? |
225/1337 (16.8%) |
p=0.76 | |
| Always | 51.8% | ref | |
| Usually | 48.8% | 0.46 | |
| Sometimes | 48.2% | 0.51 | |
| Never | 46.1% | 0.44 | |
| Patient understanding Q: In the past 12 months, how often did your PCP explain things in a way you could understand? |
204/1337 (15.3%) |
p=0.49 | |
| Always | 50.2% | ref | |
| Usually | 52.9% | 0.51 | |
| Sometimes | 52.3% | 0.72 | |
| Never | 39.5% | 0.19 | |
From binomial regression with identity link and robust standard error. Models patient willingness as “willing” vs. “neutral or unwilling” with the patient characteristic as the single independent variable and weighted by the inverse of the sampling fractions used in the complex survey design.
Conclusion
In this large survey of older patients with type 2 diabetes treated with glucose-lowering medications, 51.2% were willing to take less diabetes medication. Clinical guidelines recommend treatment de-intensification for older patients with longer duration of diabetes, poor health status or limited life expectancy, and HbA1c level <7%. Our study suggests only half of patients meeting these criteria were willing to take less medication. Particularly in older patients with longer duration with diabetes, a smaller percentage reported willingness to take less medication. This suggests applying algorithms to select patients for de-intensification is likely to be met with substantial resistance and highlights the importance of clinician-patient conversations about treatment goals in diabetes care.
In our study, duration of diabetes, age, and polypharmacy were associated with patient willingness to de-intensify in the opposite way from what is typically recommended by clinical guidelines. The discrepancy between a patient’s willingness to de-intensify and current clinical guidelines may be due to several factors. Patients with longstanding and stable treatment regimens may exhibit a status quo bias and be reluctant to accept the uncertainty associated with changing regimens – even if changes in their health may imply that their current regimen poses harm. This is particularly highlighted by older patients in our cohort reporting lower rates of willingness to de-intensify treatment compared to younger patients within the sample. We suspect older individuals may perceive de-intensification as “less care” from their provider and may explain the reluctance to de-intensify. In one survey, most patients believed that patients with long disease duration and more complications should receive more intensive therapy.17 Yet, further research is needed to understand barriers to deintensification in older patients with diabetes. It was surprising that recent experience with hypoglycemia was not associated with substantively higher rates of willingness to de-intensify. This may be related to several factors. Our survey reported a low rate of severe hypoglycemia and in turn patients may not see mild hypoglycemia as a sufficient reason to de-intensify. We and others have shown that individuals with poor glycemic control have elevated rates of hypoglycemia.10,20 Patients may accept mild hypoglycemia as a necessary tradeoff to improve their HbA1c. Furthermore, the presence of higher hypoglycemia in patients with poorer control may partially explain why poorer control was associated with greater willingness to de-intensify, possibly due to patient perceived failure of therapy, greater perceived treatment burden, or reduced patient capacity to follow current treatment plan. As per our study, reduction in treatment burden (e.g., reduction in side effects, treatment simplification) was an important factor associated with patient’s willingness to de-intensify. In contrast, patients with a lower HbA1c may accept the risk of hypoglycemia in hopes to avoid hyperglycemia16 and this was highlighted in our study as more than 70% of respondents in our study cited hyperglycemia as a risk to deintensification. Particularly with de-intensification, older patients who are accustomed to certain treatment goals may perceive de-intensification or a higher HbA1c target as their clinician providing less care or giving up on them.17
Our study also found patients were more willing to de-intensify if they identified a benefit to de-intensification. Additionally, reported harms to de-intensification were relatively similar between those who were willing to de-intensify compared to those who were not willing. This finding has been confirmed in prior studies.14,16,17Some patients can identify benefit(s) to taking less medication, however more often patients express their reluctance to de-intensify citing several fears to taking less medications including worsening glycemic control and complications related to hyperglycemia.16Clinician-patient relationship has been cited in the literature to be an important factor in patient’s willingness to intensify or de-intensify treatment.16 A recent survey which found clinicians preferred intensive treatment even when guidelines recommended de-intensification.12 Patients reluctance to de-intensify may be related to their clinician’s recommendations, which ran contrary to guidelines. However, in our survey patients who reported always trusting their clinicians were slightly more willing to take less medication, though not significantly.
Recommendations endorsed by guidelines or clinicians may not align with patient preferences. An approach prioritizing patient-specific concerns may provide clinicians with more insight into their patient’s experience and capture those who are likely to respond to and benefit from de-intensification efforts. Ultimately, ongoing shared decision making over the disease course and risks and time to benefit of glucose lowering medications is needed to create a treatment plan that evolves with the patient’s changing health and remains both effective and safe.
Limitations:
Survey responses were obtained from a single healthcare system of fully insured patients and may not be generalizable to other populations or age groups. Because the survey responders were all insured under Medicare Advantage, their out-of-pocket medical costs were low; thus, willingness to de-intensify may be lower than in other healthcare settings where patients experience greater medication expenditures. Although consistent or better than many other studies survey response rates, less than half of the sample responded to the survey.24 We used the survey question “I would be willing to take less medications for my diabetes” as a surrogate to assess willingness to de-intensify though survey responders may not have interpreted the question in that way. Patients may perceive de-intensification differently than clinicians and which may alter the interpretation of survey responses. Furthermore, most respondents were identified as having intermediate health and were somewhat sicker than the general population.21
In summary, approximately half of the older patients with type 2 diabetes within our study were willing to de-intensify treatment. Those willing to take less medication were more likely to identify a benefit to de-intensification than those who were unwilling to de-intensify treatment. However, there was no substantive difference in reported risks to de-intensification among those willing versus unwilling to take less medications. Patients with attributes that clinical guidelines identify as important criteria for considering de-intensification (e.g., advanced age, long duration of diabetes) were less likely to report willingness to de-intensify, which may pose a challenge for clinicians when discussing de-intensification. These results underscore the importance of the quality of patient-provider communication and shared decision making about the risks and benefits associated with continuing existing treatment regimens at their current intensity.
Supplementary Material
Key Points.
Older patients with longer diabetes duration, poor health status and limited life expectancy are less likely to benefit from intensive glycemic control. Clinical guidelines recommend treatment de-intensification when A1c levels are <7% in these patients.
Among older patients with type 2 diabetes on glucose-lowering medications, younger age and shorter duration of diabetes were each associated with willingness to take less medications.
Characteristics of patients willing to de-intensify diabetes treatment do not align with current clinical de-intensification guidelines.
Why does this paper matter?
This misalignment between guidelines and patient preferences highlights difficulties clinicians may face when suggesting deintensification and emphasizes the importance of involving patients in shared decision-making.
Sponsor’s role:
This study was supported by the NIDDK and NIA-funded Diabetes & Aging Study (R01-DK081796, R01-AG063391, and R56-AG051683), NIDDK Centers for Diabetes Translational Research (P30-DK092949 and P30-DK092924), and by a National Institutes of Health T32 grant (DK007058). The funding bodies did not have a role in the study design and did not take part in data collection, analysis, or interpretation for the study. The funding bodies did not participate in writing this manuscript.
Footnotes
Conflicts of Interest:
Authors have no conflicts of interest.
References
- 1.Turner R, et al. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes. The Lancet. 1998. 352(9131):837–53. [PubMed] [Google Scholar]
- 2.Patel A, et al. Intensive Blood Glucose Control and Vascular Outcomes in Patients with Type 2 Diabetes. The New England Journal of Medicine. 2008. 358(24):2560–2572. [DOI] [PubMed] [Google Scholar]
- 3.Cruz-Jentoft AJ, Carpena-Ruiz M, Montero-Errasquín B, Sánchez-Castellano C, Sánchez-García E. Exclusion of older adults from ongoing clinical trials about type 2 diabetes mellitus. J Am Geriatr Soc. 2013. May;61(5):734–8. [DOI] [PubMed] [Google Scholar]
- 4.Miller ME, Bonds DE, Gerstein HC, Seaquist ER, Bergenstal RM, Calles-Escandon J, Childress RD, Craven TE, Cuddihy RM, Dailey G, Feinglos MN, Ismail-Beigi F, Largay JF, O’Connor PJ, Paul T, Savage PJ, Schubart UK, Sood A, Genuth S. ACCORD investigators: The effects of baseline characteristics, glycaemia treatment approach, and glycated haemoglobin concentration on the risk of severe hypoglycaemia: post hoc epidemiological analysis of the ACCORD study. BMJ. 2010; 340:b5444. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Duckworth WC, Abraira C, Moritz TE, Davis SN, Emanuele N, Goldman S, Hayward R, Huang GD, Marks JB, Reaven PD, Reda DJ, Warren SR, Zieve FJ. The duration of diabetes affects the response to intensive glucose control in type 2 subjects: the VA Diabetes Trial. J Diabetes Complications. 2011; 25:355–361. [DOI] [PubMed] [Google Scholar]
- 6.Greenfield S, Billimek J, Pellegrini F, Franciosi M, De Berardis G, Nicolucci A, Kaplan SH. Comorbidity affects the relationship between glycemic control and cardiovascular outcomes in diabetes: a cohort study. Ann Intern Med. 2009. Dec 15;151(12):854–60. [DOI] [PubMed] [Google Scholar]
- 7.McCoy RG, Lipska KJ, Yao X, Ross JS, Montori VM, Shah ND. Intensive Treatment and Severe Hypoglycemia Among Adults With Type 2 Diabetes. JAMA Intern Med. 2016;176(7):969–978. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Huang ES, Zhang Q, Gandra N, Chin MH, Meltzer DO. The effect of comorbid illness and functional status on the expected benefits of intensive glucose control in older patients with type 2 diabetes: a decision analysis. Ann Intern Med. 2008. Jul 1;149(1):11–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.LeRoith D, Biessels GJ, Braithwaite SB, Casanueva FF, Draznin B, Halter JB, Hirsch IB, McDonnell ME, Molitch ME, Murad MH, Sinclair AJ. Treatment of Diabetes in Older Adults: An Endocrine Society Clinical Practice Guideline. JCEM. 2019; 104(5):1520–1574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lipska KJ, Ross JS, Miao Y, Shah ND, Lee SJ, Steinman MA, (2015). Potential overtreatment of diabetes mellitus in older adults with tight glycemic control. JAMA internal medicine, 175(3), 356–362. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.American Diabetes Association Professional Practice Committee; 13. Older Adults: Standards of Medical Care in Diabetes—2022. Diabetes Care. 2022; 45 (S1): S195–S207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pilla SJ, Jalalzai R, Tang O, Schoenborn NL, Boyd CM, Golden SH, Mathioudakis NN, Maruthur NM. A National Physician Survey of Deintensifying Diabetes Medications for Older Adults With Type 2 Diabetes. Diabetes Care. 2023; dc222146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Grant RW, Pabon-Nau L, Ross KM, Youatt EJ, Pandiscio JC, Park ER. Diabetes oral medication initiation and intensification: patient views compared with current treatment guidelines. Diabetes Educ. 2011. Jan-Feb;37(1):78–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hart HE, Ditzel K, Rutten GE, de Groot E, Seidu S, Khunti K, Vos RC. De-Intensification Of Blood Glucose Lowering Medication In People Identified As Being Over-Treated: A Mixed Methods Study. Patient Prefer Adherence. 2019. Oct 18;13:1775–1783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Green AR, Aschmann H, Boyd CM, Schoenborn N. Association between willingness to deprescribe and health outcome priorities among U.S. older adults: Results of a national survey. J Am Geriatr Soc. 2022. Oct;70(10):2895–2904. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Pilla SJ, Meza KA, Schoenborn NL, Boyd CM, Maruthur NM, Chander G. A Qualitative Study of Perspectives of Older Adults on Deintensifying Diabetes Medications. J Gen Intern Med. 2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Schoenborn NL, Crossnohere NL, Bridges JFP, Pollack CE, Pilla SJ, Boyd CM. Patient Perceptions of Diabetes Guideline Frameworks for Individualizing Glycemic Targets. JAMA Intern Med. 2019;179(12):1642–1649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Karter AJ, Ferrara A, Liu JY, Moffet HH, Ackerson LM, Selby JV. Ethnic disparities in diabetic complications in an insured population: The Northern California Kaiser Permanente Diabetes Registry. JAMA. 2002;287(19):2519–2527. [DOI] [PubMed] [Google Scholar]
- 19.Messer LC, Laraia BA, Kaufman JS, Eyster J, Holzman C, Culhane J, Elo I, Burke JG, O’Campo P. The development of a standardized neighborhood deprivation index. J Urban Health. 2006. Nov;83(6):1041–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Lipska KJ, Warton EM, Huang ES, Moffet HH, Inzucchi SE, Krumholz HM, Karter AJ. HbA1c and Risk of Severe Hypoglycemia in Type 2 Diabetes: The Diabetes and Aging Study. Diabetes Care. 2013;36:3535–3542. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.O’Hara B and Caswell K. Health Status, Health Insurance, and Medical Services Utilization: 2010. Household Economic Studies. U.S. Department of Commerce. 2013. [Google Scholar]
- 22.Weiner JZ, Gopalan A, Mishra P, et al. Use and Discontinuation of Insulin Treatment Among Adults Aged 75 to 79 Years With Type 2 Diabetes. JAMA Intern Med. 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Laraia BA, Karter AJ, Warton EM, Schillinger D, Moffet HH, Adler N. Place matters: neighborhood deprivation and cardiometabolic risk factors in the Diabetes Study of Northern California (DISTANCE). Soc Sci Med. 2012. Apr;74(7):1082–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.CDC. Behavioral Risk Factor Surveillance System. 2021 Summary Data Quality Report. 2022. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.