Should US doctors embrace electronic health records?
BMJ 2017; 356 doi: https://doi.org/10.1136/bmj.j242 (Published 24 January 2017) Cite this as: BMJ 2017;356:j242
- George A Gellert, associate system chief medical information officer1,
- S Luke Webster, vice president1,
- John A Gillean, executive vice president1,
- Edward R Melnick, assistant professor2,
- Hemal K Kanzaria, assistant professor3
- 1CHRISTUS Health, San Antonio, TX, USA
- 2Department of Emergency Medicine, Yale School of Medicine, New Haven, CT, USA
- 3Department of Emergency Medicine, University of California, San Francisco, CA, USA
- Correspondence to: G A Gellert ggellert33{at}gmail.com, E Melnick edward.melnick{at}yale.edu
Yes—George A Gellert, S Luke Webster, John A Gillean
Preventable healthcare related errors are responsible for 210 000-400 000 deaths a year in the US, making them a leading cause of death.12 Substantial evidence shows that electronic health records improve patient safety, quality of care, and outcomes.345678910 Computerized physician order entry (CPOE) accelerates the delivery of care, improves efficiency, and reduces the number of professionals in the clinical workflow, decreasing delays, adverse events, and errors arising from illegible handwriting and miscommunication. Electronic health records, CPOE with clinical decision support, and digital documentation reduce unintended acts of omission or commission, errors of execution or planning, and deviations from standard care; they also distribute and ensure the practice of evidence based medicine.
Reduced medical errors
A systematic review of 50 studies found CPOE in hospitals reduced medical errors and adverse drug events, especially when bundled with clinical decision support that alerts providers to laboratory or medical errors.3 Meta-analysis of 16 studies found that compared with paper order entry, CPOE reduced preventable adverse drug events and drug errors by 50%.4 Analysis of 3364 hospitals found 269 facilities implementing CPOE outperformed comparison hospitals on five of 11 measures related to ordering drugs.5 Review of 22 studies found use of CPOE and clinical decision support in emergency departments was associated with significantly fewer prescribing errors (reduced by 17-201errors per 100 orders), potential adverse drug events (reduced by 0.9 per 100 orders), and prescribing of excessive doses (reduced by 31%).6 Eight studies on medical imaging services showed reduced unnecessary or duplicative testing associated with use of CPOE and decision support and better adherence to evidence based guidelines on ordering tests; three showed statistically significant reductions in test turnaround times.7
Lyons found CPOE was a significant predictor of shorter hospital stay (by 0.9 days) and lower mortality (reduced by one to three deaths per 1000), and Schreiber reported a 20.2% shorter length of stay correlated with rising CPOE adoption (improving clinical, cost, and efficiency outcomes).89 Amarasingham evaluated 167 233 patients older than 50 in 41 Texas hospitals and found that patients cared for in hospitals with automated notes and records, CPOE, and decision support had fewer complications, lower mortality, and lower costs.10 Reported effects on mortality are conflicting, warranting further research.
Are current electronic health records ideally suited for implementation? No, they are not. But the evidence we have cited shows that they are well suited, and has any technology ever been ideal at inception? The evidence of benefits thus far is compelling, and electronic health records will evolve (or perish), as with all new technology. Competition between the more than 20 companies currently producing electronic records should be enough to drive innovation as it is similar to the numbers in other industries that have seen rapid product evolution, such as computer, mobile phone, and automobile industries.
No alternative
The promise of electronic health records resides not only in advancing personalized medicine, genomics and proteomics, population health, and predictive analytics tomorrow, but in meeting today’s paramount imperative: reducing the epidemic of preventable iatrogenic morbidity and mortality. Winston Churchill said, “Democracy is the worst form of government, except for all the others.” What alternative to electronic records—and the engagement required to drive their evolution—will help accomplish these essential objectives?
Clinicians have expressed concerns about the effect of electronic records on the quality and sanctity of their relationship with patients. However, the greatest risk to the patient-clinician relationship today is unsafe practice causing iatrogenic patient harm.
Healthcare is the last major American industry to move to an electronic information infrastructure. This relatively new technology is immature, needing substantial improvements in usability, functionality, and interoperability,1112 much as other information technologies that have transformed our lives did in their early years, such as the personal computer. But delaying introduction of this essential technology—and with it the journey of continuous learning and refinement inherent to all technology evolution—may result in countless patients being killed or injured because of preventable errors in their care. The imperative to reduce this harm argues against letting the pursuit of a perfect electronic health record delay deployment of a good one.
Physician dissatisfaction with electronic records is substantial, however.131415 For many they are the most dislocating changes in clinical practice and workflow in a generation. But who would have thought two decades ago that those 1.5 kg mobile phones—which barely delivered their single function effectively—would evolve into the highly intuitive multifunctional device and “internet in our pockets” that is the contemporary smartphone? Similarly, electronic record technology is not stagnant, and remains a critical, foundational contributor to advancing evidence based medicine, ensuring patient safety, and reducing epidemic healthcare related morbidity and mortality.
No—Edward R Melnick, Hemal K Kanzaria
First, do no harm. Though paper charts were intuitive and simple, they were criticized for being disorganized and illegible, leading to medical errors. Electronic health records promised to improve patient safety and outcomes by reducing errors. A rigorous, large scale observational study to determine the effect of implementing electronic health records found no negative association with mortality or adverse events across 17 US hospitals.16 But although electronic records are unlikely to be dangerous, their hasty implementation has done harm in many other ways.
The 2009 US Health Information Technology for Economic and Clinical Health Act provided $28bn (£23bn; €26bn) to encourage doctors and hospitals to adopt electronic health records. By this year, 90% of physician offices are predicted to have adopted an electronic record system.17 Unfortunately, in the rush to get systems up and running, clinicians’ and patients’ needs have fallen by the wayside. A small number of vendors have monopolized the market, stifling further innovation. Their products have evolved from billing systems, with documentation and ordering systems tacked on as an afterthought.18
Damaging the clinician-patient relationship
The emerging and rapidly growing market of 10 000 medical scribes over the past six years is evidence that the current electronic health record is not usable as intended, integrated into clinician workflow, or designed to support clinicians’ or patients’ needs.15 Furthermore, contracts are typically negotiated and agreed by administrators, not the system’s end users—practising clinicians and patients.18 As a result, electronic records have entered the exam room in a way that jeopardizes the clinician-patient relationship.
Anyone who has recently been a patient in the US knows that the desktop interface physically obstructs and separates the clinician from the patient. Well conducted observational studies suggest that current electronic records compromise the physician’s already minimal time for each patient—two hours with the electronic record for every hour spent in direct face time with patients and an additional 1-2 hours (outside of office hours) of computer time each night—further limiting communication.1920
Citing literature spanning 55 years, Czernik and Lin argue that time at the bedside has always been limited by “indirect patient care,” such as charting, sifting through redundant documentation, reviewing results, and writing orders.21 They contend that electronic records have not substantially altered the amount of time that clinicians spend directly with patients, though they acknowledge that there is decreased eye contact with patients. They recommend using computers to shift “indirect patient care” to the bedside to improve communication and engagement. We agree that technology can and should support time at the bedside,22 but presently it does not, and their argument misses the forest through the trees: the current system distracts and obstructs, threatening the quality of this time.
Failing on design criteria
Lidwell and colleagues have proposed a hierarchy of needs for design success in general.23 To thrive, a system must meet the following needs in order: functionality, reliability, usability, proficiency, and creativity. The first three needs are likely familiar. Proficiency allows people to do things better than they could previously, and with creativity “people begin interacting with the design in innovative ways.”
Though clunky, the current electronic record is functional and reliable. But it is not yet usable as intended, and we have a way to go before it can routinely and effectively promote bedside interaction between patients and clinicians. Technologic innovation must focus on optimizing both the quality and quantity of time at the bedside. Only through such an approach can we foster an environment in which clinicians can listen to, care for, and heal their patients.
To do so, the healthcare industry—across its continuum—must invest appropriate resources and follow established design principles that put the needs of users first.24 Only then will we have an electronic health record that allows patients and clinicians to do things better than they could on paper and to interact in innovative ways. In the meantime, the sanctity of the clinician-patient relationship remains at risk.
Footnotes
Competing interests: All authors have read and understood BMJ policy on declaration of interests and declare the following: ERM is supported in part by grant number K08HS021271 from the Agency for Healthcare Research and Quality. HKK has been supported by the US Veterans Affairs / Robert Wood Johnson Clinical Scholars program and has been a consultant for RAND Health and Castlight Health. The content does not necessarily represent the official views of any funding agency.
Provenance and peer review: Commissioned; not externally peer reviewed.