Choice of Intensive Lifestyle Change and/or Metformin after Shared Decision Making for Diabetes Prevention: Results from the Prediabetes Informed Decisions and Education (PRIDE) Study

Introduction

The Diabetes Prevention Program (DPP) study demonstrated that intensive lifestyle intervention (ILI) reduced type 2 diabetes incidence by 58%, and metformin reduced type 2 diabetes incidence by 31%.¹ Since prediabetes can be a preference-sensitive decision,² adults with prediabetes can consider several factors in deciding whether to start ILI and/or metformin: e.g., relative efficacy, convenience, potential medication side effects, likelihood of associated benefits beyond diabetes prevention, etc. Understanding preferences for diabetes prevention is useful for system-level prevention efforts, for example in order match the availability of ILI classes with patients who prefer ILI versus metformin. However, there is little data how patients in real-world settings choose between available diabetes prevention options, and overall uptake of ILI and/or metformin remains quite low.^3–5 Thus, patient-centered approaches for diabetes prevention are critically needed for health systems, insurers, policymakers and program planners.

To our knowledge, the Prediabetes Informed Decisions and Education (PRIDE) study was the first study to test the effectiveness of shared decision making (SDM) for diabetes prevention.⁶ The SDM intervention informed participants at risk for type 2 diabetes about their evidence-based options for diabetes prevention to help them choose an option aligned with their preferences and needs. Since very little is known about how adults at risk for diabetes make decisions regarding available diabetes prevention strategies, our goal was to examine predictors of choosing ILI or metformin among PRIDE participants who received a novel SDM intervention for diabetes prevention.

Methods

Our study examined data collected from the intervention arm of the Prediabetes Informed Decisions and Education (PRIDE) study. Details of the PRIDE study, which was a cluster-randomized trial testing a novel SDM intervention for diabetes prevention within 20 primary care clinics within a regional health system from 2015–2018, have been previously published.⁶ SDM participants from 10 intervention clinics met the following electronic health record data (EHR) eligibility criteria; age 18 –74 years, body mass index (BMI) ≥24 m/kg² (≥22 m/kg² if Asian), HbA1c 5.7–6.4% [39–46 mmol/mol]. We excluded patients with diabetes (i.e., any HbA1c ≥6.5% (48 mmol/mol), ICD-9 codes of 250.xx, or antiglycemic medications), chronic kidney disease stage 4 or higher (i.e., estimated glomerular filtration index [eGFR] ≤29 ml/min [0.48 mL/s])), active eating disorder(s), and women who were pregnant or planning to get pregnant. Eligible participants received invitation letters followed by telephone calls from the study team to opt-in for an SDM session.

Participants who opted in for SDM visits met face-to-face with pharmacists trained in SDM who used a decision aid (DA) titled, “Prediabetes: Which Treatment Should I Use?,” produced by Healthwise, a provider of patient decision support tools for over 35 years.⁷ The DA provides comprehensive information about ILI and metformin as two evidence-based options for diabetes prevention, summarizing the evidence regarding relative risk reduction in pictographs along with logistical considerations and potential side effects. Patients were presented with relevant evidence regarding diabetes prevention options, given the opportunity to clarify their values and deliberate the decision with pharmacists. EHR progress note templates for SDM visits triggered the pharmacists to proceed in step-by-step fashion through 6 standardized DA steps (1-Get the facts, 2-Compare Options, 3-Your Feelings, 4-Your Decision, 5-Quiz yourself, and 6-Your Summary). The DA identifies patients under 60 years of age, with a history of gestational diabetes, and/or with BMI ≥35 kg/m2 as subgroups for whom metformin is most effective, consistent with American Diabetes Association guidelines.^7,8 Pharmacists used the DA to describe four prevention options – 1) ILI alone, 2) metformin alone, 3) ILI with metformin, or 4) continuing usual care - and helped the participants choose a diabetes prevention strategy. Excluding research activities such as informed consent, SDM visits lasted approximately 37 minutes. At the end of the SDM visit, participants received DPP contact information and/or metformin prescription based on their choices. EHR note templates documented that the SDM visit took place and were forwarded to primary providers for review and signature. Previously published PRIDE findings showed that adjusted weight loss among the first 351 PRIDE participants was greater than matched participants in control clinics at 12-months follow-up regardless of their chosen strategy (−5.3 lbs.[2.4kg] vs. −0.2 lbs. [0.09], p<0.001).⁶

We report here on findings of logistic regression models based on the generalized estimating equation (GEE) approach to account for clustering at the clinic level examining 1) predictors of choosing ILI with or without metformin, compared to metformin alone or usual care, and 2) predictors of choosing metformin with or without ILI, compared to ILI alone or usual care. These multivariate models included five self-reported variables: age (<50 years, 50–59 years, ≥60 years), gender, race/ethnicity (Non-Latino White, Latino, Asian, African American, other), education (high school education or less, some college, college degree), and income (<$45,000, $45,000–$84,999 and ≥$85,000). The models also included last recorded BMI (as a continuous variable) and HbA1c (5.7–5.9% vs. 6.0–6.4% [39–41 vs. 42–46 mmol/mol]) from the EHR. We performed multiple imputation to estimate missing values for education (n=7) and income (n=83).

Participants were invited to complete optional surveys immediately after SDM and at 4-months follow-up to assess general prediabetes knowledge, understanding of available options, what matters most and overall experience with the SDM intervention.

The study was approved by the University of California, Los Angeles Institutional Review Board, IRB protocol #15–000310.

Results

We identified 1,763 potentially eligible PRIDE SDM participants who were mailed invitation letters. We reached 1,071 by phone to confirm eligibility and 515 (48% of those reach by phone; 29% of those who received letters overall) opted in for SDM visits. The mean age of SDM participants was 56.3 years (SD 11.4), BMI of 30.3 kg/m2 (SD 5.2), and HbA1c of 6.0% (42 mmol/mol, Table 1). Among 515 PRIDE participants, 283 (55.0%) chose ILI alone, 44 (8.5%) chose metformin alone, 77 (15.0%) chose ILI with metformin, and 111 (21.6%) continued usual care. The DA included several standardized clarifying questions to help SDM participants identify “what matters most” to them in choosing ILI and/or metformin (Figure 1). Participants responses to these questions tended to align with their final choices. For example, when SDM participants were asked clarifying questions about their willingness to take medication or concerns regarding medication side effects, participants who chose metformin or ILI and/or metformin were more likely to give their highest importance rating to “If medicine can help me, I’ll take it” as compared those who chose ILI alone (metformin = 41%, ILI and/or metformin = 51% vs. ILI alone= 7%). Similarly, participants who chose metformin or ILI and/or metformin were also more likely to give their highest importance rating to “If I can avoid diabetes by using medicine, I’ll deal with any side effects” than those who chose ILI alone (metformin = 31%, ILI and/or metformin = 42% vs. ILI alone = 9%).

Table 1.

Descriptive Characteristics of PRIDE Study Participants (n=515)

Mean age	56.3 (SD 11.4)
Mean BMI	30.3 kg/m2 (SD 5.2)
Mean hemoglobin A1c	6.0%
Female gender	56.0%
Race/Ethnicity
Non-Latino white	40.4%
Asian	20.2%
Latino	18.1%
African American	16.3%
Other	5.1%
Choice of Diabetes Prevention Option/s
Intensive lifestyle change alone	283 (55.0%)
Metformin alone	44 (8.5%)
Both intensive lifestyle change and metformin	77 (15.0%)
Usual care	111 (21.6%)

Figure 1.

Screenshot of the “Prediabetes: Which Treatment Should I Use?” DA (included with permission from Healthwise)

Step 3 of the DA included several standardized clarifying questions to help SDM participants identify “what matters most” to them in choosing ILI and/or metformin.

In the adjusted multivariate analyses (Table 2), women had significantly higher odds than men of choosing ILI (OR=1.60, p=.023). Participants with higher BMI also had higher odds of choosing ILI (OR=1.07 per unit increase in BMI, p=.005) versus other options as well as choosing metformin (OR=1.06 per unit increase in BMI, p=.008) versus other options. Participants 60 years and older had lower odds of choosing metformin (OR 0.50, p=0.028) compared to patients less than 50 years of age. We did not find any association between race/ethnicity and choice of diabetes prevention options or any contradictory findings between predictors and the DA content.

Table 2.

Predictors of Choice of Intensive Lifestyle Change and/or Metformin after Shared Decision Making for Diabetes Prevention

	Intensive lifestyle change +/− metformin, N = 304 (versus metformin alone or usual care, N = 126)		Metformin +/− intensive lifestyle change, N = 101 (versus intensive lifestyle change alone or usual care, N=329)
	OR (95% CI)	p-value	OR (95% CI)	p-value
Age		.233		.071
<50 (reference)	---		---
50–59	0.77 (0.44–1.34)		0.87 (0.49–1.54)
≥60	1.25 (0.70–2.23)		0.52 (0.29–0.94)
Gender		.043		.938
Male (reference)	---		---
Female	1.58 (1.01–2.47)		0.98 (0.61–1.58)
Race/ethnicity		.465		.901
Non-Latino white (reference)	---		---
Latino	1.52 (0.76–3.02)		1.03 (0.52–2.02)
Asian	1.23 (0.66–2.32)		0.86 (0.44–1.69)
African American	1.04 (0.56–1.91)		0.76 (0.39–1.48)
Other	0.49 (0.16–1.53)		1.11 (0.34–3.66)
BMI (continuous)	1.09 (1.03–1.15)	<.001	1.08 (1.03–1.13)	.002
A1c		.207		.423
5.7–5.9% (reference)	---		---
6.0–6.4%	0.74 (0.46–1.19)		1.22 (0.74–2.02)
Income		.855		.205
$85,000 or more (reference)	---		---
<$45,000	1.20 (0.59–2.45)		1.01 (0.48–2.14)
$45,000-$84,999	1.12 (0.65–1.94)		1.64 (0.95–2.83)
Education (reference)		.857		.661
College Degree	---		---
HS Graduate or less	0.79 (0.34–1.85)		1.00 (0.41–2.48)
Some College	1.00 (0.61–1.62)		0.80 (0.48–1.35)

Significant values in bold font

Immediately after the SDM intervention, 99% of participants were aware that making either lifestyle changes or taking metformin could help prevent diabetes (irrespective of their final choice). All participants reported understanding of the options available to them and 97% felt clear about which benefits and side effects mattered most to them (Table 3 – Part A). Almost 97% of SDM participants also rated the content covered in the SDM intervention as very helpful or helpful. We later stopped asking these survey questions immediately after the SDM intervention given the limited variation in responses we received. All participants (N=515) were invited to complete a 4-month follow-up survey with a response rate of over 80% (Table 3 – Part B). In the 4-month follow-up survey, 97% of SDM participants reported having a better understanding of prediabetes after the SDM visit and 97% also felt the appointment with the pharmacists met all their expectations. Irrespective of their final choice, 84% of participants remained aware that lifestyle changes could help them lower their risk of getting diabetes and 82% still felt the DA helped them make a decision.

Table 3.

Survey Questions on Prediabetes Knowledge, Diabetes Prevention Options, What Matters Most and Overall Experience with SDM (A= items asked immediately after SDM and B= items asked 4-months after SDM)

A. Survey Questions After SDM Intervention	Yes	No	Not Sure
Does either making lifestyle changes or taking metformin work well to prevent type 2 diabetes? N=297	99.6% (293)	1.3% (4)	NA
Do you understand the options available to you? N=302	100% (302)	0% (0)	NA
Are you clear about which benefits and side effects matter most to you? N=302	99.6% (301)	0.3% (1)	NA
	Very Helpful or Helpful	Somewhat Helpful	Not Helpful
How helpful was this content? N=308	96.7% (298)	2.9% (9)	0.3% (1)
B. Survey Questions at 4-Months Follow-up	Yes	No	Not Sure
Do you feel that you have a better understanding of prediabetes as a result of meeting with the pharmacist? N=423	97% (409)	3% (14)	0% (0)
Did your appointment with the pharmacist meet your expectations? N=425	97% (412)	1% (4)	2% (9)
Did the decision aid help you make a decision? N=421	83% (354)	4% (18)	12% (49)
Do you think lifestyle changes would decrease your chance of getting diabetes? N=428	84% (361)	1% (2)	15% (65)

Discussion

We found that when presented with relevant evidence regarding diabetes prevention options, given the opportunity to clarify their values and deliberate the decision of a diabetes management strategy with pharmacists, 70% of patients with prediabetes chose ILI and 23% chose metformin. However, uptake of ILI and/or metformin among patients at risk for diabetes remains quite low in real-world setting. For example, prior studies of large group health plan data, less than 4% of patients with prediabetes were prescribed metformin.³ Very few studies to date have reported patient preferences for diabetes prevention and none have examined the impact an SDM intervention on diabetes prevention choices. Although prediabetes affects one in three US adults, uptake of evidence-based strategies for diabetes prevention remains quite low in real world settings.^4,5,3 Given the significant individual and societal burdens associated with the development of type 2 diabetes, it is important to help patient make fully informed decisions by putting “all the options on the table.” Informed patients are capable of considering the issues involved and working with their providers to make the best decision.³ We found that patient decisions were generally aligned with recommendations and expected benefits from the published literature; younger patients were more likely to choose metformin and patients with higher BMIs were more likely to choose either strategy as opposed to usual care.^1,3,7,9 Thus, our study highlights the importance of conducting SDM for diabetes prevention in order to increase uptake of evidence-based strategies that are aligned with both patients preferences and with established guidelines/recommendations.

In today’s multidisciplinary and team-based approach to primary care, there is an opportunity to spread the provision of preventive healthcare services among the team of allied clinical providers frequently integrated in primary care settings, including pharmacists. The pharmacists who delivered this novel SDM intervention were trained in SDM and diabetes prevention and used a DA to share the task of considering the issues, deliberating and making a consensual choice with SDM participants. Following the SDM intervention, pharmacists also provided transparent, bidirectional communication with the primary care provider. This novel approach ensured that patients at risk for diabetes had the opportunity to participate in an SDM intervention for diabetes prevention without increasing the burden on primary providers who face significant time constraints.

Despite these strengths, our study includes some limitations. First, this study was conducted in one health system with integrated pharmacists, so findings many not be generalizable to all settings. Second, participants had to opt in and provide consent for SDM visits, so we were not able to collect information about diabetes prevention decision making among those who opted out. However, this is one of the first studies highlighting the role of SDM and predictors of choice for diabetes prevention.

A critical goal of diabetes prevention should be helping patients with prediabetes adhere to an option that aligns with their preferences and needs. Successful weight loss in the National Diabetes Prevention Program (NDPP) was strongly associated with engagement, as participants lost 0.3% of body weight for every session attended.¹⁰ In the DPP study, 71% of participants randomized to the metformin arm who took at least 80% of their prescribed dose,¹¹ but long-term studies indicate that only 50% of patients with chronic conditions take their medications as prescribed.¹² A patient-centered approach that presents both ILI and metformin as evidence-based diabetes prevention options and follows up with support for adherence will be a significant advance in the effort to slow and ultimately reverse the epidemic of type 2 diabetes in the United States.