Conversations and Reactions Around Severe Hypoglycemia (CRASH) Study: Results From People With Diabetes and Caregivers in the United States

The CRASH study was a multinational, cross-sectional, online survey conducted with people diagnosed with type 1 or type 2 diabetes and caregivers of people with diabetes from Canada, China, France, Germany, Japan, Spain, United Kingdom, and the United States (N = 2,625). Survey responses were reported by people with diabetes and caregivers who were recruited separately and considered independent respondents with no relationship to each other (no dyads). Here, we report results from the United States.

The study was conducted using purposive sampling from online research panels; we enrolled people with diabetes who were ≥18 years of age, self-reported either type 1 or type 2 diabetes treated with insulin via injection or pump, and experienced one or more severe hypoglycemic event in the past 3 years. Severe hypoglycemia was defined as a low blood glucose event that the person with diabetes could not treat by him- or herself.

The study population also included caregivers, defined as adults who were relied on during a severe hypoglycemic event of a person >4 years of age who was diagnosed with type 1 or type 2 diabetes and met the above criteria for severe hypoglycemia and treatment. Caregivers may have been living in the same household as a person with diabetes (e.g., family members, roommates, domestic helpers, or relatives) or may have been other people in the life of a person with diabetes (e.g., coworkers or teachers). Eligible respondents provided electronic consent through a Web survey interface before the administration of any study procedures. Exclusion criteria for this study included a diagnosis of schizophrenia, bipolar disorder, or gestational diabetes. Professionally trained and licensed HCPs were also excluded from participation. People with diabetes who were treated with insulin could also be treated with oral antihyperglycemic medications except for sulfonylureas. Exemption of ethics approval for the U.S. survey was requested and received from the Chesapeake Institutional Review Board (Columbia, MD).

Respondents completed the online survey, which included questions about demographics, diabetes management, recent medical history, and hypoglycemia awareness. Details of respondents’ most recent severe hypoglycemic event were also recorded, including the setting, symptoms, actions taken, and emotional and life impacts (e.g., effects on physical activities, mood or emotional status, social or leisure activities, work or school, daily activities, relationships with friends and family, financial matters, or sleep). People with diabetes who were not conscious during the severe hypoglycemic event reported on what they were told about the event. As with the people with diabetes, caregivers’ survey responses reflected a report of their experiences during or what they were told occurred at the time of the severe hypoglycemic event. Most caregivers’ responses reported what had happened with the person with diabetes, but caregivers also reported their own experiences, conversations, feelings, and life impacts.

The survey included a standardized and validated self-reported measure of hypoglycemia awareness called the Gold score (9). Hypoglycemia awareness was reported by people with diabetes and caregivers based on the question, “Do you know when your hypoglycemia is commencing?” On a scale ranging from 1 (always aware) to 7 (never aware), impaired hypoglycemia awareness was defined as a score ≥4 (9). People with diabetes reported on whether they were aware of their own hypoglycemia commencing, whereas caregivers reported on awareness of hypoglycemia commencing for the person with diabetes for whom they cared.

Analyses were conducted for both people with diabetes and caregivers by type of diabetes. Continuous variables were summarized as mean ± SD. Categorical variables were summarized as numbers and percentages. Overall differences on measures among the four study groups (type 1 or type 2 diabetes among people with diabetes or caregivers) were assessed using the Kruskal-Wallis test, a nonparametric test for continuous data, χ² test for binary categorical variables, and a two-way Cochran-Mantel-Haenszel test for general association between two nominal variables. If the P value from the omnibus test was significant, differences between type 1 and type 2 diabetes were assessed separately for people with diabetes and caregivers. Statistical significance was set at P <0.05. Statistical analyses were performed using SAS, v. 9.4, statistical software (SAS Institute, Cary NC).

In total, 429 individuals based in the United States responded to the survey, including 219 people with diabetes (110 with type 1 diabetes and 109 with type 2 diabetes) and 210 caregivers (110 caring for someone with type 1 diabetes and 100 caring for someone with type 2 diabetes) (Table 1). The mean age of people with diabetes was 54.0 ± 13.1 years (quartile 1 [Q1] 45, median 56, Q3 63) for those with type 1 diabetes and 58.4 ± 10.5 years (Q1 54, median 59, Q3 64) for those with type 2 diabetes. Sixty percent of the people with type 1 diabetes were female, whereas 60.6% of those with type 2 diabetes were male. Overall, the majority (54.3%) of caregivers were reported as being a spouse or partner of a person with diabetes. A small percentage of caregivers were parents or guardians caring for a minor (>4 and ≤17 years of age with type 1 diabetes [20.9%] or type 2 diabetes [1%]). The majority of respondents (76.7%) reported having a college or university degree.

A1C levels for people with diabetes were self-reported by all respondents: 41.8% reported levels ≤7% (≤53 mmol/mol); 34.5% reported levels between 7.1 and 8.0% (54 and 64 mmol/mol); 15.0% reported levels between 8.1 and 9.0% (65 and 75 mmol/mol); and 8.7% reported levels >9.1% (>76 mmol/mol). The majority of people with type 2 diabetes (59.6%) had used insulin for >5 years. As expected, insulin pump use at the time of the last severe hypoglycemic event was reported more often with type 1 diabetes than with type 2 diabetes (among all type 1 diabetes 30.9%, among all type 2 diabetes 2.9%, P <0.0001). Impaired hypoglycemia awareness (Gold score ≥4) was reported significantly more often with type 1 diabetes than with type 2 diabetes (people with type 1 diabetes 36.4%, people with type 2 diabetes 20.2%, P = 0.0079).

The number of severe hypoglycemic events that occurred within the past 12 months (i.e., 0, 1, 2, ≥3 events) and within the past 3 years (i.e., 1, 2, ≥3 events) did not significantly differ between diabetes types, as reported by people with diabetes and caregivers (Table 1). Overall, 14.5% of respondents reported on their first severe hypoglycemic event.

The majority of respondents (62.5%) reported being with a spouse/partner or caregiver during the most recent severe hypoglycemic event. The severe hypoglycemic events reported by respondents occurred primarily at home (80.7%). Severe hypoglycemia occurred in the afternoon for 21.0% of respondents, evening and before midnight for 25.4%, and after midnight (before morning) for 25.4%. Almost one-third of all respondents (32.6%) reported being asleep when the severe hypoglycemic event occurred.

The most commonly reported cause of the most recent severe hypoglycemic event was that the person with diabetes ate less than planned or usual (39.9%); this occurred significantly more often in those with type 2 diabetes than in those with type 1 diabetes (46.8 vs. 27.3%, P = 0.0028). Other perceived causes included that the person with diabetes exercised more than planned or realized (18.2%) or took an incorrect dose of insulin (12.1%).

Glucagon injections were rarely used as rescue therapy (reported by 5.8% of respondents) (Table 2). A substantial minority of all respondents (40.4%) did not have a glucagon prescription, and, for people with diabetes, there was a statistically significant difference between those with type 1 diabetes and those with type 2 diabetes (38.0 vs. 65.4%, P <0.0001) (Supplementary Table S1). In addition, significantly more caregivers of people with type 2 diabetes were unaware of the option to administer glucagon than caregivers of people with type 1 diabetes (38.3 vs. 9.7%, P <0.0001). The proportion of respondents reporting any health care utilization during the most recent severe hypoglycemic event was low (20.3%). Receiving recommendations from HCPs before the most recent severe hypoglycemic event to use health care services (e.g., call ambulance, go to emergency department, or call HCP) in case of a severe hypoglycemic event was reported by a modest proportion of people with diabetes (type 1 diabetes 16.4% and type 2 diabetes 22.0%) and caregivers (of people with type 1 diabetes 29.1% and of people with type 2 diabetes 31.0%, P = 0.0494).

The emotional impacts (emotions that arose at the time of the event) for each respondent type are reported in Table 3. Overall, experiencing or witnessing the severe hypoglycemic event made the majority of respondents (70.9%) feel scared. Just less than half of respondents reported that the most recent severe hypoglycemic event made them feel unprepared (42.7%) or helpless (46.9%). First-person reporting was completed by people with diabetes and caregivers on eight life domains in response to the impact of the most recent severe hypoglycemic event, as shown in Table 3.

Several different actions (not mutually exclusive) were taken in response to the most recent severe hypoglycemic event (Table 4), including respondents starting to carry glucose-containing candy or sweet foods/drinks and/or adjusting their meal plan (60.8%), measuring blood glucose more often (43.8%), changing the insulin regimen or timing or dosing of insulin (35.0%), and wearing a continuous glucose monitoring (CGM) device (9.3%). A very small proportion of respondents obtained glucagon or kept glucagon close (3.7%) and/or kept glucagon kits in areas that are frequented or carried glucagon (2.8%). The actions of carrying candy or sweet foods/drinks and/or adjusting the meal plan (P = 0.0013) and wearing a CGM device were significantly different between groups (P <0.0001). All participants with type 2 diabetes reported carrying candy, foods/drinks, or adjusting their meal plan more than all participants with type 1 diabetes (68.9 vs. 53.2%). All participants with type 1 diabetes reported that they were more likely to start wearing a CGM device than those with type 2 diabetes (16.4 vs. 1.9%).

Respondents were surveyed on recommendations and discussions that occurred before their most recent severe hypoglycemic event. Approximately one-third of people with diabetes (type 1 diabetes 38.2%, type 2 diabetes 38.5%) reported having a discussion about severe hypoglycemia at every visit with their HCP (Table 5). Half of the respondents (50.6%) reported having conversations about severe hypoglycemia with their HCP at some visits. A small proportion of respondents (17.7%) reported never having a discussion with their HCP before their most recent severe hypoglycemic event.

The majority of people with diabetes (type 1 diabetes 81.8%, type 2 diabetes 84.4%) and caregivers (for someone with type 1 diabetes 72.7%, for someone with type 2 diabetes 65.0%) reported that consumption of carbohydrates and/or a form of glucose was advised to treat hypoglycemia during a discussion with an HCP before the patient’s most recent severe hypoglycemic event. Approximately one-fourth of respondents reported that their HCP recommended glucagon injection as a treatment option and, of these respondents, there was a significantly larger proportion with type 1 diabetes than with type 2 diabetes (44.1 vs. 7.2%, P <0.0001).

A substantial proportion of all respondents (40.3%) reported not having had a discussion with their HCP after the most recent severe hypoglycemic event occurred. The primary reasons reported as to why hypoglycemia was not discussed with an HCP included “knew cause of the severe hypoglycemia event” (62.4%), “no big deal” (24.2%), “did not think it would happen again” (9.6%), and “did not want to talk about it” (9.6%).

Actions for preventing or preparing for future severe hypoglycemic events that were recommended by HCPs after the most recent severe hypoglycemic event included (not mutually exclusive) changing the insulin regimen or timing or dosing of insulin (33.8%), more intensive glucose monitoring (28.2%), carrying sweets or some form of glucose (27.5%), adjusting meal plans (timing of meals, snack more, or amount of food intake) (18.6%), and increasing the availability of glucagon (obtain glucagon or keep it closer) for recovery from severe hypoglycemia (7.7%).

The CRASH study examined important aspects associated with severe hypoglycemia in people with type 1 or insulin-treated type 2 diabetes and caregivers of people with diabetes, including situational contexts, psychological experiences, and discussions with HCPs about prevention and treatment strategies. Communication and shared decision-making between people with diabetes and their HCPs are important for successfully managing diabetes and promoting quality of life. Results from the U.S. CRASH survey indicate that conversations are not taking place as often as recommended by ADA guidelines, which is consistent with a recent publication from Pilla et al. (10) who found that communication about hypoglycemia occurred in only 24% of health care visits. Of the respondents who did not have any conversation with their HCP regarding the most recent hypoglycemia event (40.3%), nearly two-thirds reported that the reason they did not discuss it was because they knew the cause of the event.

Clearly, the most recent severe hypoglycemic event had an acute impact on respondents’ emotions (i.e., feelings of being scared, helpless, and/or unprepared). Our survey did not assess whether discussions with HCPs dealt with these issues, but more attention to the psychological impact seems warranted. In addition, our survey did not examine the precise nature of enduring life impacts, although some impacts seem to involve lifestyle changes that attempt to prevent future hypoglycemia (e.g., impacts to physical, social, work/school, and daily activities), and some impacts are negative consequences of the events themselves (e.g., impacts to mood/ emotions, family/social relationships, sleep, and finances). Fulcher et al. (11) reported similar results from their survey, in which individuals with type 1 or type 2 diabetes reported a large financial and psychosocial impact resulting from experiencing a nocturnal or daytime nonsevere hypoglycemic event. Again, HCP assessments of such psychosocial impacts are essential to identify any need for additional attention, whether through counseling by diabetes care providers or mental health specialists. It is important to note that negative consequences were common among caregivers as well as people with diabetes.

Glucagon is used to treat severe hypoglycemia when a person with diabetes can no longer safely swallow oral carbohydrates or cannot tolerate carbohydrates because of nausea or vomiting. Specifically, the ADA recommends that glucagon should be prescribed for all individuals at increased risk of level 2 hypoglycemia, which is defined as blood glucose <54 mg/dL (<3.0 mmol/L), or level 3 (severe) hypoglycemia (7) to ensure that glucagon is available when needed. Results from this study demonstrate that a low percentage of people with diabetes and caregivers reported glucagon use for their most recent severe hypoglycemic event, and a low percentage of respondents obtained glucagon or made sure to keep it close even after the most recent event. Instead, eating or drinking a form of glucose was the most common first response at the time of the severe hypoglycemic event (6).

The assessment by a caregiver regarding the safety of giving food or drink by mouth is made at a panicked time, when emotions can be intense and decision-making may be impaired. In interviewing caregivers of people with diabetes, Stuckey-Peyrot et al. (12) found that caregivers felt a rising sense of panic and questioned their actions during severe hypoglycemic events. Some caregivers described needing to “encourage them along” or urging the person with diabetes to “please drink more,” while some said they had to take the lead on administering the carbohydrates, which one caregiver described as “dumping” soda down the person’s throat. If administering oral carbohydrates did not resolve severe hypoglycemia, then caregivers sometimes felt helpless and said that at a “certain point” it seemed that the “only option [was] to call for emergency help” (12).

Since the CRASH study was completed, new innovations in glucagon delivery for the treatment of severe hypoglycemia have been approved in the United States and elsewhere (13–15). New, ready-to-use drug/device combinations do not require a user to reconstitute the glucagon, thus allowing for more successful administration of a full dose of glucagon, and are easier to use than conventional injectable glucagon emergency kits (12,16–18). Certainly, after a severe hypoglycemic event, conversations among people with diabetes, caregivers, and HCPs should include assessment of unexpired glucagon ownership and review of when and how to use it.

Strengths of this study are the relatively large numbers of people in each group (people with type 1 or type 2 diabetes and caregivers of people with type 1 or type 2 diabetes) who reported on their experience of severe hypoglycemia. Reports by caregivers of a severe hypoglycemic event, which by definition requires help from another person for recovery, provide additional insight and remind us of the need to support caregivers’ burden. The study was also comprehensive, focusing on activities taking place before, during, and after the most recent hypoglycemic event.

Limitations of the CRASH study include the self-reported nature of the data, as potential biases may have affected responses. Information discussed with HCPs was also not available to the investigators. For example, HCPs may have discussed severe hypoglycemia or recommended seeking emergency health care, but respondents may not have recognized or remembered such discussions. In addition, because this study was an online survey requiring internet access, only members of a medical research panel were eligible to participate. Survey respondents were highly educated, with 76.7% reporting having a college/university degree. The sample population was therefore unlikely to be representative of all adults experiencing recent severe hypoglycemia. Although respondents were asked to respond about events that met the formal criterion for severe hypoglycemia, they were not required to demonstrate that these events met those criteria.

The CRASH study results can be used to improve the preparedness of people with diabetes and caregivers and increase their understanding of the medical importance of risk and avoidance of severe hypoglycemia. The actions that people with diabetes and caregivers take are influenced by conversations about severe hypoglycemia that occur with HCPs; therefore, it is important that HCPs consider these findings and apply them to their practice.

The authors thank the study respondents, investigators, and contributors from each of the study sites. Mark Peyrot (Loyola University, Baltimore, MD) provided consultation regarding the design of the statistical analysis and interpretation of results. The authors also thank Antonia Baldo (Syneos Health) for editorial assistance.

This study was funded by Eli Lilly and Company.

F.J.S. served as consultant and is on an advisory board for Eli Lilly and Company. E.S., C.J.C., S.B., and B.D.M. were employees and shareholders of Eli Lilly and Company when the CRASH survey was conducted and analyzed. B.A.N. is an employee at Syneos Health. Z.B., D.M.B., and R.P. are employees at Evidera. No other potential conflicts of interest relevant to this article were reported.

F.J.S. and B.D.M. contributed to the conceptualization. F.J.S., B.A.N., and B.D.M. wrote the original draft. E.S., B.A.N., Z.B., D.M.B., R.P., and B.D.M. were involved in data curation. E.S., Z.B., D.M.B., and B.D.M. devised the methodology. E.S., B.A.N., and B.D.M. supervised the crafting of the manuscript. E.S., Z.B., D.M.B., and R.P. were responsible for validation. B.A.N. led the project administration. C.J.N., D.M.B., and B.D.M. helped with the investigation of the manuscript. Z.B. and B.D.M. contributed resources to the development of the manuscript. All authors contributed to the formal analysis and reviewed and edited the manuscript. B.D.M. is the guarantor of this work and, as such, takes responsibility for the integrity of the data and the accuracy of the data analysis.