RELYING ON HANDWRITTEN MEDICAL RECORDS RESULTS IN INSULIN OVERDOSE

RELYING ON HANDWRITTEN MEDICAL RECORDS RESULTS IN INSULIN OVERDOSE

By Amy Rosenberg, Pharm.D., and Eric Rosenberg, M.D.

Admitting Physician:

“The patient didn’t know his meds very well, so I copied his home medications from his discharge instructions that he got from his last hospitalization.”

In fact, the patient was receiving 10 units of insulin glargine (Lantus) during the prior hospitalization and at home. The “g” above “Lantus 10 units” was misinterpreted as an additional “0” — Lantus 100 units.

Pharmacist:

“100 units seemed like a large dose of Lantus. So I double-checked. I also thought the online home medication list said the patient was taking 100 units of Lantus at home so I approved and dispensed the dose.”

The next morning, the patient’s blood glucose was critically low. Although the patient recovered with no long-term adverse outcomes, dextrose infusions were required that day to maintain the patient’s blood glucose.

Discussion:

Unusually high doses of insulin should prompt clinicians to investigate further and verify dosages. In this case the pharmacy department’s computerized medication order entry systems generated an alert because insulin glargine (Lantus) dose was greater than 80 units (this is the level at which the system is set to alert the pharmacist). However, because some patients require high doses of insulin (100 or more units per day), the alert is not accurate for all patients. In this case, the pharmacist received the alert and investigated further, but saw the same handwritten record the physician saw and overrode the alert.

Errors due to interpretation of handwriting are common and widely reported across the country. The most reliable strategy to eliminate errors due to handwriting or legibility is to eliminate handwriting as much as possible from the medical record.

Although there are many other potential benefits to a fully electronic medical record, elimination of legibility errors is one major benefi t of these systems.

Actions taken to prevent this error from occurring in the future:

Implementation of the Epic fully integrated electronic medical record will eliminate legibility errors at Shands at the University of Florida.