Pragmatic Risk Score for Severe Hypoglycemic Events
Performer: Kaiser Foundation Research Institute
Principal Investigator: Andrew J. Karter
Project Duration: 9/22/14 - 3/31/17
Regulatory Science Challenge
More than 25 million Americans have type 2 diabetes. Although advances in diabetes clinical care and medical treatment have greatly reduced long-term complications and death among these patients, hypoglycemia--an unintended consequence of glucose-lowering treatment--has become a critical public health and drug safety concern.
Severe hypoglycemia (SH) is a low plasma glucose level for which a patient requires assistance. Fear of SH is a significant factor that prevents some patients from starting insulin, and SH is an important determinant of poorer quality of life and can be associated with cognitive decline and dementia. SH has been associated with serious falls and automobile accidents, cardiovascular complications, and a several-fold increased risk of death. SH is also costly--25% of all emergency hospitalizations for adverse drug events are due to SH, with rates even higher in older patients.
In this project, a risk assessment mechanism to assist clinicians in predicting a patient’s risk for SH was designed and tested. The Kaiser Foundation Research Institute developed a pragmatic risk-scoring tool that helps clinicians identify patients at highest risk of SH and guides shared decision-making. The scoring tool draws on an existing wealth of clinical data from electronic medical records from a large, integrated, health care delivery system.
The tool was designed for use in the outpatient setting. It simply and efficiently provides quantitative estimates of hypoglycemic risk, which together with the patient’s preferences and clinician knowledge, helps providers and patients weigh the risks of SH against reduction in long-term diabetes complications, reduces the incidence of SH events, and improves patient safety. The risk-scoring tool is publicly available.
Improve diabetes patient safety for the more than 25 million Americans with diabetes:
- Construct a research database, populate the database, and validate risk scores.
- Develop the risk-scoring tool for use by clinicians to identify patients at high risk for SH.
- Prepare a manuscript to describe the project and the tool for publication in a peer-reviewed journal.
The final report from the researchers to the FDA included the following results:
- A model was developed using six inputs to define a patient risk for SH: 1) total number of prior episodes of hypoglycemia-related ED or hospital utilization (0, 1-2, ≥3 times); 2) number of ED encounters for any reason in the prior 12 months (<2, ≥2 times); 3) insulin use (yes/no); 4) sulfonylurea use (yes/no); 5) presence of severe or end-stage kidney disease (yes/no); and 6) age younger than 77 years (yes/no).
- This model was translated into a checklist, to allow clinicians to quickly assess risk.
- The model successfully differentiated between patients at high-risk for SH (2% of patients, > 5% risk of SH) and low risk patients (87% of patients, <1% risk).
- The model was constructed based on ICD-9 codes. (A subsequent project re-validated the model using ICD-10; see reference below).
Karter AJ, Warton EM, Lipska KJ, Ralston JD, Moffet HH, Jackson GG, Huang ES, Miller DR. Development and Validation of a Tool to Identify Patients with Type 2 Diabetes at High Risk of Hypoglycemia-Related Emergency Department or Hospital Use. JAMA Intern Med. 2017 Oct 1;177(10):1461-1470. doi: 10.1001/jamainternmed.2017.3844. PMID: 28828479; PMCID: PMC5624849.
Karter AJ, Warton EM, Moffet HH, Ralston JD, Huang ES, Miller DR, Lipska KJ. Revalidation of the Hypoglycemia Risk Stratification Tool Using ICD-10 Codes. Diabetes Care. 2019 Apr;42(4):e58-e59. doi: 10.2337/dc18-2154. Epub 2019 Feb 14. PMID: 30765427; PMCID: PMC6429629.