Surveillance of Hypoglycemia—Limitations of Emergency Department and Hospital Utilization Data

Hypoglycemia is a common and life-threatening adverse drug event associated with glucose-lowering medications. Hypoglycemia is classified as severe when assistance from another person is required to actively administer carbohydrates, glucagon, or other resuscitative actions.¹ Surveillance in the United States has relied on data from electronic medical records (EMR) or administrative claims from emergency department (ED) visits and hospital admissions for hypoglycemia, which are presumably the most severe and costly events.^2,3 However, this surveillance fails to account for events treated outside of the health care system (eg, by family members). We estimated the proportion of all severe hypoglycemic events that are captured by surveillance based on health care utilization alone among pharmacologically treated patients with diabetes.

We estimated the frequency of self-reported severe hypoglycemic events based on responses to the question, “In the past year, how many times have you had a severe low blood sugar reaction such as passing out or needing help to treat the reaction?” included in the Diabetes Study of Northern California (DISTANCE) survey of Kaiser Permanente members.⁴ Survey responses were linked with EMR data during the matching 12-month time frame covered by the survey. Hypoglycemia-related utilization was based on the primary diagnosis of hypoglycemia in the ED or principal discharge diagnosis in the hospital using a validated, International Classification of Diseases, Ninth Revision (ICD-9)-based algorithm.⁵ This research was approved by the Kaiser Foundation Research Institute institutional review board.

In total, 20 188 (62% of eligible patients with diabetes) responded to the DISTANCE survey.⁴ We excluded 3018 patients who were not treated with glucose-lowering medications and 3811 who did not provide a valid response to hypoglycemia questions, yielding a study sample of 13 359. The mean age was 58.9 years, 6597 of 13 359 (49.4%) were women, 3396 of 13 214 (25.7%) were non-Hispanic white, 9818 of 13 214 (74.3%) were ethnic minorities, 13 037 (97.7%) of 13 349 had type 2 diabetes, 3527 (26.4%) of 13 359 were insulin-treated, and 8446 (63.2%) of 13 359 were sulfonylurea-treated. Of 13 359 patients included in the study, 1566 (11.7%) reported having 1 or more severe hypoglycemic events requiring third-party assistance in the previous 12 months, whereas only 102 (0.8%) had a documented hypoglycemia-related ED or hospital utilization during the same time period (Table). Only 82 (5.2%) of those self-reporting severe hypoglycemia also had an ED or hospital admission for hypoglycemia during the same 12-month time period.

Surveillance of severe hypoglycemia in the United States is currently based on hypoglycemic events resulting in ED or hospital utilization recorded in the EMR. Although hypoglycemia-related ED visits or hospitalizations are rare (approximately 0.8% annually), self-reported severe hypoglycemia events are relatively common (approximately 12% annually). Based on our comparison of these 2 sources of data, we estimate that only approximately 5% of self-reported events among pharmacologically treated patients with diabetes are captured by health care utilization-based surveillance. Thus, current surveillance grossly underestimates the overall burden of severe hypoglycemia.

Because most severe hypoglycemic events occur outside of the health care system, a more complete understanding of the burden of severe hypoglycemia will require collecting self-reported events. Currently, there are no validated instruments to collect these patient-reported outcome data. A brief assessment tool that reliably captures severe hypoglycemic event history should be developed and tested in clinical settings. Importantly, self-reporting in our study did not capture all hypoglycemia-related utilization, suggesting that a combination of modes of data collection may be required to ascertain the total burden of severe hypoglycemia. Although the proportion of total severe hypoglycemic events that result in ED or inpatient utilization likely differs across health care delivery settings, our qualitative observation that most severe hypoglycemic events are not captured in ED and hospital utilization data may be widely generalizable. Determining the total burden of severe hypoglycemia, not just the small subset of events leading to health care utilization, will better inform clinical decision making, public health policy, and quality measure development to enhance patient safety.

Corresponding Author: Andrew J. Karter, PhD, Division of Research, Kaiser Permanente, 2000 Broadway, Oakland, CA 94612 (andy.j.karter@kp.org).

Accepted for Publication: February 16, 2018.

Published Online: April 16, 2018. doi:10.1001/jamainternmed.2018.1014

Author Contributions: Dr Karter had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Karter, Moffet, Lipska.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Karter.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Karter, Liu.

Obtained funding: Karter.

Administrative, technical, or material support: Moffet.

Supervision: Karter.

Conflict of Interest Disclosures: Dr Karter and Mr Moffet received unrelated funding from AstraZeneca. Mr Moffet and Ms Liu also received unrelated funding from Regeneron. Dr Lipska receives support from the National Institute on Aging and the American Federation of Aging Research through the Paul Beeson Career Development Award (K23 AG048359) and the Yale Claude D. Pepper Older Americans Independence Center (P30 AG021342). Dr Karter is also supported by the NIDDK Centers for Diabetes Translational Research (P30 DK092924).

Funding/Support: This project was supported by the National Institute of Health (R01 DK103721, R01 DK065664).

Role of the Funder/Sponsor: The National Institute of Health had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.