Recently the Wall Street Journal alerted its readers to a report by the ECRI Institute in the US on the common problem of patient misidentification. [1,2] The study found 7613 “wrong patient” incidents reported across 181 organizations in the period January 2013 to July 2015. Selecting the wrong electronic record was one example of this type of incident.

The consequences of record mis-selection can be serious and include patients receiving the wrong medications, incorrect results, diagnosis, and history of illness. The incidents always involve an error of commission (e.g. giving treatment that should not have been given) and one of omissions (e.g. not giving the treatment as intended), and two patients: the patient whose wrong record has been selected, and the patient who was the intended recipient.

The consequences for patients may go beyond their healthcare, for example it can affect their health insurance policy, credit rating, or other financial matters. [3] Arguably it is also ethically wrong to enter wrong data into people’s records. [4] Technology does not always help in this type of incidents. In fact, in the ECRI study in 15% of the cases (1,148) technology contributed to the errors. [2]

These are some of the specific issues with identification errors in digital patient record systems:

• Once the record is closed, it remains invisible. The issue of having used the wrong patient record may only be found long after it happened, when the record is opened again, for example at the next patient admission to hospital.
• Given integrated data interchanges, errors in data entry easily cascade across interconnected digital systems, with potential repercussions across a number of settings. By the time the error is found in one system and one setting, it may already have had repercussions for the patient in another healthcare setting. The error may be fixed in one system but not in other systems that have made copies of the data.
• The software may allow opening multiple records at the same time—an error prone system feature for busy, constantly interrupted healthcare work.
• Electronic records fix the legibility problems of handwriting but make patients (and their records) “all look the same.”
• Patient banners may not make the identifiers clearly visible—indeed some older systems may have screens missing the patient identifiers all together.
• Digital screens may not display “same name alerts,” for patients with same or similar names in the same place at the same time
• Electronic records can be accessed remotely, giving opportunities for extra safety and convenience; but remote access also means away from the patient, who may be able to provide the extra clue for the recognition that the record is wrong.
• The use of barcoded patient wristbands and “scanning for safety” initiatives can only benefit the safety of activities involving patients directly—e.g. at bedside, or at the time of the procedure. However a large part of clinical work is done away from the patient.
• Errors in data entry cannot be easily fixed: for medical-legal reasons the wrong data may be left in the record and only a note added to say it was entered by mistake. Depending on system design, this note may remain disconnected from the data and the information still be used for a next encounter.

Selection errors are likely to be under-reported, and may remain unknown. For example, in an effort to capture un-reported mis-selections of records, the Montefiore Medical Centre implemented a Retract-And-Reorder tool to log orders sent and cancelled within minutes as a clue of selection errors beyond incidents reported. The researchers identified 6885 retract-and-reorder events (in a three months period) they classified as near misses, “self-caught by the provider before causing patient harm.” [5]

This type of error seems also very frequent in the NHS. Research in progress on the analysis of NRLS incidents (National Reporting and Learning System) involving the use of technology for supply and use of medicines has found a number of cases of wrong record selection. Often they involve patients with the same or similar names, which is a well known challenge. [6] Similarly, incidental findings during an ethnographic study of the use of electronic systems in hospitals (research in progress) include observation of a patient records mix-up: lab tests requested and recorded in the wrong record by junior doctors, who were then told they “have to be more careful.” It was good that, although perhaps embarrassed and afraid, the doctors in training spoke of the mistake.

Standardised wristbands and scanning barcodes links patients with their records more safely. When medicines are administered, barcodes can be used to track and match medicines with the patient and their records. However a large part of clinical work involving patient records is done away from the patient. Furthermore, mistakes in record selection and their consequences emerge from complex sociotechnical systems.

The ECRI report suggests a number of prevention strategies.[2]  Others may be identified from safety and reliability literature. Among the most urgent and most challenging to implement and evaluate may be three patient safety approaches:

1. Evaluation of technology design prior to, or during, implementation in view of how easy it may be to select the wrong record and not recognise it is the wrong record. Evaluation needs to be theoretically informed and founded on evidence. [7]
2. Give patients (and their advocates, friends or relatives when applicable) real time access to their records in all settings, including hospitals. [8]
3. Organise for collective mindfulness. [9] This involves, for example, accepting the constant questioning of one’s own assumptions, preoccupation with failure and training in interpersonal skills as part of reliability-enhancing work practices (REWP). [10]

Valentina Lichtner Valentina Lichtner is a lecturer in Information Management at the University of Leeds.

Acknowledgements: The research in progress mentioned in this article is done in collaboration with NHS Improvement and the London School of Economics. The NRLS study is a Knowledge Transfer Secondment funded by the University of Leeds EPSRC Impact Acceleration Account. The Delivering Digital Drugs project (http://digital-drugs.org) is funded by the EPSRC/Research Council UK under the Digital Economy Programme.  The funders were not involved in the writing and submission of this blog.

Competing interests: The author declares no competing interests.

References:

1] Beck M. Medical Record Mix-Ups a Common Problem, Study Finds. Wall Street Journal 2016.
ECRI Institute. PSO Deep Dive: Patient Identification, 2016.

2] Lee TH. Coming Back from the Dead. NEJM 2016;375(6):507-09

3] Wynia M, Dunn K. Dreams and Nightmares: Practical and Ethical Issues for Patients and Physicians Using Personal Health Records. The Journal of Law, Medicine & Ethics 2010;38(1):64-73

4] AHRQ Grant Awarded to Study the Impact of Health IT on Patient Safety at Montefiore and Einstein and Brigham and Women’s Hospital [News Release]. 2014. http://www.montefiore.org/body.cfm?id=1738&action=detail&ref=1185.

5] Adelman, J. S., et al. (2013). “Understanding and preventing wrong-patient electronic orders: a randomized controlled trial.” Journal of the American Medical Informatics Association : JAMIA 20(2): 305-310. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3638184/

6] Walker J, Darer JD, Elmore  JG, Delbanco  T. The Road toward Fully Transparent Medical Records. New England Journal of Medicine 2014;370(1):6-8

7] Weick KE, Sutcliffe KM, Obstfeld D. Organizing for High Reliability: Processes of Collective Mindfulness. Research in Organizational Behavior 1999(21):81-123

8] Vogus TJ, Iacobucci D. Creating Highly Reliable Health Care: How Reliability-Enhancing Work Practices Affect Patient Safety in Hospitals. ILR Review 2016;69(4):911-38