Diabetes Medication Changes in Older Adults With Type 2 Diabetes: Insights Into Physician Factors and Questions Ahead

Older adults with diabetes have a higher risk of hypoglycemia than younger adults (1). To combat this risk, clinical practice guidelines recommend deintensification of hypoglycemia-inducing medications, namely, insulins and sulfonylureas, if this can be achieved within individualized glycemic targets (1).

To guide providers in deciding when deintensification is appropriate for older adults with type 2 diabetes, the American Diabetes Association (ADA) has published a framework that is based on patient health status (1). This framework includes three health status groupings for older adults: 1) healthy, with few coexisting chronic illnesses and intact cognition and functional status; 2) complex health, with multiple coexisting chronic illnesses or moderate functional and/or mild to moderate cognitive impairment; and 3) poor health, with residence in long-term care facilities or having severe functional impairment and/or moderate to severe cognitive impairment. The ADA framework also recommends relaxation of HbA_1c targets when health status declines, which in turn provides opportunities for medication deintensification in older adults with complex and poor health (1).

Despite evidence showing that the risks of continuing hypoglycemia-inducing medications and targeting tight glycemic targets outweigh benefits in many cases (2–5), many older adults remain on sulfonylureas and complex insulin regimens, and they have HbA_1c levels that fall below recommended targets (3,5–9). A major challenge for deintensification research has been moving beyond descriptive analyses toward a deeper understanding of the complex and interdependent factors that underlie observed trends.

In this issue of Diabetes Care, Pilla et al. (10) present findings from a national survey of U.S. physicians that underscore the gap between deintensification guidelines and real-world practice while also yielding valuable new insights into why this gap exists. The authors surveyed 445 physicians, including 133 general medicine practitioners, 73 geriatricians, and 239 endocrinologists (response rate 37.5%). The survey provided three case scenarios describing older adults on insulin and/or sulfonylurea therapy who had healthy, complex, or poor health status in accordance with the ADA framework. Analyses revealed that most physicians would not deintensify or switch hypoglycemia-inducing medications and that most would select lower HbA_1c targets than recommended by guidelines. Consistent with findings from prior studies (4,11), the authors report that physicians were, on average, more likely to deintensify therapy in healthy older adults than in those with complex health, a trend that is notably misaligned with guidelines (1). Pilla et al. (10) found occurrence of hypoglycemia to have the strongest influence on deintensification decisions, with >99% of surveyed physicians opting to modify therapy in response to even mild hypoglycemia.

This article provides additional glimpses into physician behaviors that drive diabetes medication decisions. For example, the authors note that even when physicians selected guideline-concordant HbA_1c targets, this did not necessarily correlate with a decision to deintensify or switch medications. Despite recognizing the need to relax HbA_1c targets, only 9% and 27% of physicians opted to deintensify therapy for individuals with complex and poor health, respectively. Why physician medication management did not align with selection of treatment targets is unclear, but this finding points to the complexities underlying physician decision-making and may have important implications for the future development of physician-facing interventions to promote guideline-concordant care.

Another key strength of the current study is that it offered respondents the option to either deintensify (i.e., stop or reduce) or switch medications. Medication switching may represent an important step toward simplification, a complementary but distinct concept from deintensification that likewise reflects guideline-concordant care for older adults (1,12). While deintensification implies lowering a regimen’s potency, simplification emphasizes making a regimen easier to follow (12). Both approaches can lessen hypoglycemia and polypharmacy, but simplification may have the added benefit of promoting medication adherence and improving quality of life (12). Although Pilla et al. (10) report that endocrinologists were the most likely to replace insulin or sulfonylurea with another diabetes medication, deeper exploration of preferred medication-switching strategies might help address current evidence gaps related to simplification. In addition, further research is needed to investigate how newer diabetes therapies (i.e., glucagon-like peptide 1 receptor agonists and sodium–glucose cotransporter 2 inhibitors) might facilitate simplification and deintensification for older adults (13), particularly in the setting of relevant comorbidities (e.g., cardiorenal disease).

As with any survey study, the findings of Pilla et al. (10) may be affected by selection bias. Furthermore, the survey responses provide limited opportunities to explore the reasoning or other nuances that may drive respondent medication decisions and selection of HbA_1c targets. In addition, survey questions cannot reflect the entirety of the real-world clinical decision-making process, and each of the case scenarios reflects only a subset of the complex patient, provider, and contextual factors that may hinder guideline-concordant deintensification. In particular, patient perspectives and preferences, which could not be explored in the current study design, are likely to have a sizeable influence on medication decisions in practice. Many older adults who achieve tight control without apparent hypoglycemia may not see any value in discontinuing medications or relaxing HbA_1c targets they have tolerated for years. Alternatively, patients may not understand the reasoning behind medication changes and interpret them as negative transitions rather than promoting quality of life. While the perspectives of older adults with diabetes have been elicited in small qualitative studies (14,15), further work is needed to understand patients’ beliefs and fears about deintensification as well as the resources that would be helpful to support changes in care.

Despite these limitations, the current study highlights opportunities to better support guideline-concordant care in older adults. For example, it may be difficult for physicians to elicit hypoglycemia history in routine practice due to competing clinical demands, limited time, and underreporting of hypoglycemia by patients. To this end, continuous glucose monitoring may provide a way for clinicians to objectively assess for problematic patterns of hypoglycemia among older adults. Additionally, more time may need to be allotted for clinic encounters focused on education and counseling on patient understanding, preferences, and emotions regarding medication deintensification and simplification.

The current study underscores the need for additional research to understand when and how medication decisions are made. Figure 1 illustrates potentially important factors related to medication decision-making in older adults with diabetes that may merit rigorous study. No singular research methodology will, by itself, elucidate how medication decisions are made or provide the solutions. Rather, the complexity underlying diabetes medication decisions invites more studies that integrate patient perspectives as well as quantitative and qualitative methods, the development of new resources for patients and providers, inclusion of multiple members of the care team, and ongoing evaluation of both guidelines and practice to ultimately close the gap between them.

In summary, Pilla et al. (10) should be commended for capturing national deintensification practices and providing valuable insights into physician decision-making. This study paves the way toward addressing the gaps and fostering appropriate deintensification and simplification, and it may spur momentum for future studies designed to delve into provider as well as patient and system-level factors influencing medication decisions. Only by understanding the drivers of observed practice can we develop effective, evidence-based tools to support clinicians and patients in shared decision-making, be it for deintensification, simplification, or any other medication change aimed at balancing risks and improving the lives of older adults with diabetes.