Sloan, Caroline ECampagna, AdaTu, KarissaDoerstling, StevenDavis, J KellyUbel, Peter A2024-02-032024-02-032023-070003-48191539-3704https://hdl.handle.net/10161/30134Background: More than 40% of patients with diabetes in the United States have trouble paying their medical bills. Among patients with health-related financial hardship, 56% have delayed or foregone care (1). By one estimate, an insulin-dependent patient with insurance spends about $4800 annually on medications, physicians’ visits, supplies, hospitalizations, and lost wages, equivalent to 15% of the median U.S. per capita income (2). Crowdfunding is increasingly being used by patients struggling with medical costs (3). Crowdfunding campaigns can shed light on the types of expenses that patients with diabetes may struggle to afford (3). Objective: To leverage the rich real-world testimonies available in crowdfunding campaigns in order to characterize the financial challenges of patients with diabetes who seek fundraising support. Methods and Findings: We scraped a random sample of 89 645 active U.S. medical crowdfunding campaigns posted on GoFundMe from 2010 to August 2020, using a previously published natural-language algorithm (4). Campaigns were included if they were in English and requested money for a single person whose primary condition was diabetes. We randomly selected 350 of 807 campaigns that met inclusion criteria, with oversampling of type 1 diabetes campaigns to obtain roughly equal proportions of type 1 and non–type 1 diabetes campaigns and ensure a wide breadth of experiences. Campaigns described patients’ medical situations, expenses, and fundraising goals. We conducted a directed content analysis to evaluate for the presence of predetermined fundraising request categories (Table 1). Two authors double-coded the first 100 campaigns and 10% of the remaining 250 campaigns. Intercoder reliability using the Cohen κ coefficient was 0.67 or higher for all coding categories and direct medical expense subcategories. After exclusions, our final sample included 313 campaigns; 89% were posted during 2015 to 2020. The median fundraising goal was $10 000 (IQR, $4500 to $15 000), the median fundraising amount was $2600 (IQR, $1670 to $4760), and 14% of campaigns reached their goal. Table 2 highlights campaign characteristics overall and by stated diabetes type. One quarter of fundraisers reported having insurance; of these, 49% said their insurance covered their medical expenses but out-of-pocket costs were still too high. Almost half of direct medical expenses were not directly related to glucose control (99 of 206 [48%]). Only 6% requested money specifically for insulin. One fifth of campaigns (21%) requested money for diabetic alert dogs; almost all were campaigns for type 1 diabetes. Indirect medical expenses included lost wages, healthy food, moving to a new city to be closer to state-of-the-art medical care, car repairs to enable transportation to physicians’ appointments, personal trainers, home modifications to support new physical disabilities, and funeral expenses. Campaigns that were not specifically for type 1 diabetes (type 2 or unspecified) mentioned indirect medical expenses more often than campaigns for type 1 diabetes (63% vs. 34%). Discussion: Crowdfunding campaigns provide a window into the wide range of financial struggles that patients with diabetes may face. First, many aspects of diabetes care beyond insulin can be cost-prohibitive, including indirect expenses that clinicians may not be equipped to address. Although the Inflation Reduction Act’s cap on out-of-pocket costs for insulin is an important step in making care more affordable, policymakers should address other diabetes-related costs as well. Second, even people who have insurance use crowdfunding due to lack of coverage for certain expenses or unaffordable copayments. Third, 35% of patients with type 1 diabetes started fundraising campaigns for diabetic alert dogs, which cost about $15 000 and are not covered by insurance because of high variability in effectiveness (5). Clinicians who learn of a patient’s intent to purchase a dog could redirect them toward proven management strategies, such as continuous glucose monitors. Our conclusions are not generalizable to financially distressed patients who may not seek crowdfunding because of older age (people who use crowdfunding are generally younger [3]), poor internet access, or other reasons. Our use of content analysis limited our ability to consider textual context. We could not evaluate for fraud or consider patients’ reported versus actual financial needs. The expenses for which patients use crowdfunding include life-sustaining expenses, such as food and hospitalizations; unproven therapies, such as diabetic alert dogs; and less essential indirect costs, such as personal trainers. Future research should evaluate whether and how these expenses contribute to financial distress in the larger population of patients with diabetes, including those who do not use crowdfunding.HumansDiabetes MellitusHealthcare FinancingJournal article