Patient Safety: Getting the Name Wrong May be Fatal
Guest commentary by Dr. Bruce Lambert, University of Illinois at Chicago professor, Department of Pharmacy Administration, and president, Pharm I. R., Inc.;
and Dr. Leonard A Shaefer, IBM chief scientist, Global Name Recognition
After taking Flomax, used to treat the symptoms of an enlarged prostate, instead of Volmax, used to relieve bronchospasm, a 50-year-old woman was hospitalized.
After he was given clozapine instead of olanzapine, two drugs used to treat schizophrenia, 19-year-old man showed signs of potentially fatal complications.
After receiving methadone instead of methylphenidate, a drug used to treat attention deficit disorders, an 8-year-old died, according to MedicineNet.com.
Getting patient and drug names right is crucial to the success of health care IT.
More than 4.8 million wrong drug errors occur per year in the United States, according to the Journal of the American Pharmacists Association. And as more drugs get approved, this problem can only get worse, making us even more vulnerable.
Physicians, pharmacists, nurses and patients are more likely to confuse drug names that look or sound alike (e.g., ephedrine and epinephrine, Taxol and Paxil, vinblastine and vincristine, Actos and Actonel).
We see errors happen at every stage of the medication process: prescribing, transcribing, dispensing, administration and monitoring. They can occur with handwritten and typewritten prescriptions and they can occur when prescriptions are communicated by phone, fax and computer.
The risks posed by similar drug names are well known, but until recently it has been difficult to do much about it. Historically, the problem has been approached by compiling lists of confusing names, publishing those lists, and warning practitioners (and, to a lesser extent, patients) to be extra vigilant when dealing with confusing names.
Computer systems for ordering and dispensing drugs sometimes come with built-in warnings. Sometimes pharmacists post warning signs on the shelf next to particularly confusing names, or they store such drugs in separate areas of the pharmacy.
Having objective, numerical similarity scores provided by analytics makes it possible to predict and prevent drug-name confusions in new ways such as using similarity measures as the basis for a drug-name search engine.
For example, the manufacturer of a new drug can submit the new drug name into the search engine and retrieve a list of existing drug names, with the new names on top of the resulting list and the least similar names at the bottom.
Now, new drug names can be carefully screened prior to approval, and highly similar, potentially confusing names can be kept off the market. Name similarity software also can be used to look at existing drug names -- for example, all drugs used by a given hospital -- and the most similar pairs can be identified and targeted for error prevention.
And drug name confusion is just one of the problems that threaten patient safety.
A 67-year-old woman received invasive cardiac electrophysiology. The problem was, she didn't need it. Seventeen discrete errors resulted in a woman receiving a cardiac electrophysiology procedure intended for another patient with a similar last name, according to the Annals of Internal Medicine.
The use of fully digitized medical records has introduced greater consistency in health information systems. But the possibility for rapid propagation of errors can start during data entry of a patient's name. Effective and accurate merging of patient medical records hinges on whether we capture patient information accurately.
As new stimulus investments drive the effort to digitize medical records, we are seeing the need for better analytics to ensure consistent, accurate capture of patient name data at point of intake by flagging potential inconsistencies and miscategorized name-parts (e.g., last name in first name position) while the patient is still present to clarify or correct.
And while collecting patient name data accurately is critical for treatment, accurate data also allows analytic insights into patient demographics. We can glean a higher level of intelligence so hospitals can provide doctors and nurses up-to-date information on specific illnesses and diseases; tracking diagnostic and treatment successes enables more effective treatment of patients.
Improved name-processing techniques for health care also can provide important analytical insights into ethnicity, national origin, native language, and even unanticipated relationships or connections in the entire patient demographic, such as locating an organ donor or rare blood type.
For example, if a hospital sees from patient records that 30 percent of its demographic is Korean, then it can ensure it has staff who can speak the language, a critical factor when working in an emergency room where seconds can mean the difference between life and death.
For larger health care providers, effective strategic use of name analytics can yield valuable insights to detect demographic trends in the surrounding community and to predict where there will be a need for improved customer-care and communications. Outreach, education and prevention programs can be kept fresh and effective by staying in touch with the changing blend of cultures and languages that are being served.
Many medical errors associated with name confusion can be prevented. For instance, alerts can be sent to caregivers when a name has been incorrectly captured into the master patient index (MPI) system, or when there is significant potential for errors due to names shared by two or more patients currently under care within that system.
We need to do a better job of providing the right training to caregivers, and medical intake and record-keeping personnel, with special attention paid to capturing data of patients from unfamiliar cultural/ethnic backgrounds whose names may fit poorly or not at all into the typical First-Middle-Last pattern used in many Western countries.
In addition, extra quality-control must be placed on patient names as they are converted from spoken/recorded to handwritten forms, then converted from handwritten to electronic form. These multiple steps are fertile ground for potentially harmful transcription errors.
At the same time, key health care information technology (HIT) assets such as master patient index (MPI) systems need to provide state-of-the-art support for searching and analysis of names across all segments in patient demographics.
Sustainable progress in U.S. health policies and medical practices should include a renewed emphasis on getting names right. Driving out unnecessary medical risks and treatment costs associated with drug-name and patient-name errors can have a pervasively life-saving impact; using patient-name analytics can be a key preventive step for both patients and health care organizations.