This article is more than 5 years old. Information may no longer be current.
US electronic diabetes registers lag behind UK
Click Here to Manage Email Alerts
Researchers have identified a substantial number of patients in the United States with uncoded, and potentially undiagnosed, diabetes by applying algorithms to primary care electronic records, according to findings published in the Canadian Medical Association Journal.
“Undiagnosed diabetes is a major problem and very costly both in human and economic terms,” Tim A. Holt, PhD, of the University of Oxford, United Kingdom, told Endocrine Today. “Early diagnosis followed by proactive, structured care is important to minimize risk of long-term complications.”
Tim A. Holt
Quality of care appeared to be higher for patients with records containing a diabetes code, which includes them in a register, Holt said. An underuse of electronic registers observed in US primary care could open possibilities for systematic approaches as seen in England.
“Wider use of electronic records to support diabetes management is an opportunity in the United States,” Holt said.
Holt and colleagues from other institutions conducted a cross-sectional study and retrospective observational cohort analysis of primary care electronic medical records from the nationally representative US database GE Centricity; 11,540,454 records from more than 9,000 clinics across the country were included.
The researchers applied simple algorithms to the electronic data to assess the feasibility of identifying patients with undiagnosed diabetes. Using a set of 16 indicators, the quality of care provided to US patients with diabetes (coded and uncoded) for at least 15 months was compared with that provided in England.
There were 1,110,398 records that indicated diagnosed diabetes, but only 61.9% contained a diagnostic code. There were 10,430,056 records for patients without diabetes, with 0.4% (n=40,359) demonstrating at least two abnormal fasting or random blood glucose values; of the remaining records, 0.2% (n=23,261) had at least one documented HbA1c value ≥6.5%.
Quality-of-care indicators were available for 622,260 patients; those with a coded diabetes diagnosis had a higher quality-of-care level than those with uncoded diabetes (P<.01), but the level was generally lower than that indicated in England, the researchers wrote.
The total number of patients with diabetes, including potentially undiagnosed, was 1,174,018; of these, 63,620 (5.4%) had undiagnosed diabetes. Parts of Arizona, North Dakota, Minnesota, South Carolina and Indiana showed higher proportions.
“It is possible to identify readily, and at low cost, people with undiagnosed diabetes in the United States using information already present in electronic health records,” Holt said. – by Allegra Tiver
For more information:
Tim A. Holt, PhD, can be reached at University of Oxford, Department of Primary Care Health Sciences, Radcliffe Observatory Quarter, Woodstock Road, Oxford, OX2 6GG; email: tim.holt@phc.ox.ac.uk.
Disclosure: GE Healthcare provided sole funding for this study.
Perspective
Back to Top
Gerald Bernstein, MD, FACP
Dr. Holt’s study is 100% on-point. In the United States, we do not screen often enough, and many patients make it through the health care system undiagnosed until they have another significant illness. Including prediabetes in the pool of subjects would bring our numbers up to 42.3% of the US population over the age of 20 years who have diabetes and prediabetes — about 140 million people.
The message is that there needs to be universal screening, especially as we form mega-institutions who would have the tools to do it. People who are seen for a sore throat should be screened for abnormal glucose, just as we did syphilis screening 60 years ago. This will allow all the undiagnosed to be found and preventative or therapeutic interventions started. If we don't do this, we will be a very sick country in another 2 decades or so.
The study is a statistical analysis with projections, all of which are probably true. If we are to really know, we need a cradle-to-grave study like the Framingham. It would take a long time, but the trends would show up early.
Click Here to Manage Email Alerts