Cost-Related Insulin Underuse Among Patients With Diabetes

Abstract

This survey study examines the association of higher insulin costs with nonadherence in patients with diabetes.

Insulin is lifesaving for people with diabetes and is included on the Model List of Essential Medicines formulated by the World Health Organization.¹ This means it should be available at all times at a price the individual and the community can afford.¹ However, over the past decade, insulin prices have tripled in the United States, while out-of-pocket costs per prescription doubled.^2,3 High costs of medications can contribute to nonadherence,⁴ but the prevalence of cost-related insulin underuse is unknown.

Methods

We administered a survey to patients with type 1 or type 2 diabetes for whom insulin was prescribed within the past 6 months and who had an outpatient visit at the Yale Diabetes Center (YDC) between June and August of 2017. The YDC serves a diverse patient population from New Haven, Connecticut and surrounding counties. The survey questions were based on previously validated surveys^5,6 and review of prior literature and refined based on cognitive interviews. The Yale University Human Investigations Committee approved the study. Written informed consent was obtained from participants.

The primary outcome was cost-related underuse in the past 12 months, defined by a positive response to any 1 of 6 questions: did you…(1) use less insulin than prescribed, (2) try to stretch out your insulin, (3) take smaller doses of insulin than prescribed, (4) stop using insulin, (5) not fill an insulin prescription, or (6) not start insulin…because of cost? We examined the association between sociodemographic, economic, and clinical factors and cost-related underuse using multivariable logistic regression.

We then examined the association between cost-related underuse and poor glycemic control (HbA1c ≥9% obtained at time of visit or within 3 months) adjusting for sex, body mass index (BMI, calculated as weight in kilograms divided by height in meters squared), diabetes duration, and income using a separate multivariable logistic regression model. We performed all analyses using R statistical software (version 3.1.1, R Foundation).

Results

Of 354 eligible patients (184 [52.0%] women, 191 [54.0%] white, 123 [34.8%] type 1 diabetes), 199 (56.2%) completed the survey (101 [50.8%] women, 121 [60.8%] white, 83 [41.7%] type 1 diabetes). Of these patients, 51 (25.5%) reported cost-related insulin underuse. The type of prescription drug coverage was not significantly associated with cost-related underuse (Table). Patients with cost-related underuse were more likely to report lower incomes; 31 [60.8%] of these patients discussed the cost of insulin with their clinician and 15 [29.4%] changed insulin type owing to cost. Patients who reported cost-related underuse (vs those who did not) were more likely to have poor glycemic control in the multivariable analysis (22 [43.1%] vs 41 [28.1%]; odds ratio = 2.96; 95% CI, 1.14-8.16; P = .03). Of the 199 patients, 2 had missing HbA1c levels.

Table. Characteristics of Survey Participants.

Characteristic	Participants, No.	With Cost-Related Insulin Underuse, No. (%)	Odds Ratio (95% CI)
All Participants	199	51 (25.5)
Age, y
18-44	63	20 (31.7)	1 [Reference]
44-64	84	20 (23.8)	1.14 (0.40-3.35)
≥64	52	11 (21.2)	0.90 (0.25-3.28)
Sex
Female	101	28 (27.7)	1 [Reference]
Male	98	23 (23.5)	1.14 (0.49-2.66)
Race/ethnicity
White	121	26 (21.5)	1 [Reference]
Latino/Hispanic/Latin American	14	4 (28.6)	1.00 (0.17-4.86)
Black/African American	49	16 (32.7)	1.71 (0.58-5.05)
Other/did not report	15	4 (33.3)	1.66 (0.38-6.83)
Diabetes
Type 1	83	22 (26.5)	1 [Reference]
Type 2	115	29 (25.2)	0.91 (0.33-2.55)
Did not report	1
Type of insulin used
Analog	181	44 (24.3)	1 [Reference]
Human	12	4 (33.3)	3.17 (0.68-13.36)
Both	6	3 (50.0)	2.79 (0.30-37.85)
Prescription drug coverage
Medicare Part D	40	7 (17.5)	1 [Reference]
Employer-sponsored	66	21 (31.8)	1.03 (0.29-3.95)
Medicaid with/without Medicare	85	19 (22.4)	3.05 (0.80-13.01)
None/other/unknown	8	4 (50)	2.19 (0.24-19.37)
Annual Income, $
100 000 and greater	24	1 (4.2)	1 [Reference]
50 000-99 999	26	10 (38.5)	12.51 (1.83-255.85)
25 000-49 999	37	14 (37.8)	11.50 (1.62-239.06)
10 000-24 999	53	13 (24.5)	9.79 (1.15-220.25)
<10 000	36	7 (19.4)	6.42 (0.65-154.07)
Did not reporta	23
Difficulty buying diabetes medical equipment
No	144	21 (14.6)	1 [Reference]
Yes	55	30 (56)	5.89 (2.52-14.50)

^aMissing Income values were excluded from multivariable analysis.

Discussion

One in 4 patients at an urban diabetes center reported cost-related insulin underuse and this was associated with poor glycemic control. These results highlight an urgent need to address affordability of insulin.

More than one-third of patients who experienced cost-related underuse did not discuss this with their clinician. These findings are consistent with a previous study, which found that 37% of patients did not speak to clinicians about cost issues.⁴ Patients with lower incomes were more likely to report cost-related underuse; nearly two-thirds of these patients also experienced difficulty affording diabetes equipment, indicating broader cost barriers to diabetes management.

This study has limitations. This single-center study may be limited in its broader generalizability. Given its cross-sectional design, a causal relationship between cost-related underuse and poor glycemic control cannot be established.

Insulin is a life-saving, essential medicine, and most patients cannot act as price-sensitive buyers. Regulators and the medical community need to intervene to ensure that insulin is affordable to patients who need it. At minimum, individual clinicians should screen all patients for cost issues to help them address these challenges.