The Doctor’s Advocate | Fourth Quarter 2015
An Ounce of Prevention
Heed Those EHR Alerts
Patient harm caused at least in part by the use of electronic health records (EHRs)—or e-iatrogenesis—emerged as a factor in a closed claims study conducted by The Doctors Company. The study of 71 claims closed by The Doctors Company between 2007 and 2013 revealed that 65 percent involved EHR-related user issues, and 42 percent identified system technology design risk factors. Some claims included both user issues and system technology design risk factors.
The ECRI Institute (formerly Emergency Care Research Institute) recognized alarm hazards as the number one IT-related problem in 2015. Alarm hazards occur not only with physiologic monitoring systems,1 but also with alarm-generating devices, such as EHRs. Unfortunately, human factors may prevent healthcare providers from responding appropriately or using the alarms that are readily available to them.
EHRs have multiple benefits—from improved patient outcomes and improved care coordination to practice efficiency and cost savings. However, inappropriate use of or ignoring EHR alarms/alerts has been connected to patient harm. On the issue of over-alerting (systems generating too many alerts), Dr. Tejal K. Gandhi, president and CEO of the National Patient Safety Foundation, noted “most studies have found that only 20 percent of alerts are actually accepted,” so an important alert could be missed. She added that studies have shown that reducing the number of alerts “by streamlining the ones that you decide to alert on, by tiering, to only interrupt for things that have a certain amount of significance, you can actually reduce the number of interruptions significantly.”2
E-health data come from external sources, such as websites or through health information exchanges (hospital charts, consultant reports, and laboratory and radiology reports). Doctors also have access to data through e-prescribing community medication histories—which can expose them to liability for potential interactions with drugs prescribed by other clinicians. For example: Dr. A renews a medication, and his e-prescribing program sends an alert advising him that the medication could interact with another medication the patient is taking. He has not prescribed that drug, so his office staff will have to contact the patient to identify who has prescribed it, and then Dr. A will have to contact Dr. X to “negotiate” which drug will be discontinued or changed. If failure to take action results in patient injury from a drug-drug interaction, Dr. A may be liable.3
Drug-drug interaction lists are often so comprehensive and generate alerts with such frequency that they can become disruptive and annoying. Doctors may develop “alert fatigue” and ignore, override, or disable them. However, if it can be shown that following an alert would have prevented an adverse patient event, the physician may be found liable for failing to respond.4 Optimized, clinically meaningful drug-drug interaction lists that focus on a smaller set of interactions most frequently associated with harm or expert consensus lists may address this problem. However, EHR vendors may resist eliminating the low-risk warnings, fearing that doing so could increase their liability.
The following two claims illustrate some of the issues surrounding alerts and human factors.
An elderly female saw an otolaryngologist for ear/nose complaints. The physician intended to order Flonase nasal spray. The patient filled the prescription and took it as directed. Ten days later, she went to the ER for dizziness. Two weeks after that, the pharmacy sent a refill to the physician at his request. It was for Flomax—a medication prescribed for enlarged prostate—which has a side effect of hypotension. When ordering the prescription, the physician had typed “FLO” in the medication order screen. The EHR automatched Flomax, and the physician selected it without realizing the mistake. Flomax is not FDA-approved for females. There was no EHR Drug Alert available for gender. To prevent this type of mistake, the provider should have reviewed the prescription with the patient and read what was ordered. By writing the indication for the medication on the prescription, the pharmacist would have been alerted that the medication was not appropriate to the condition being treated.
A dialysis patient was transferred to a skilled nursing facility. There was an active hospital transfer order for Lovenox. A physician evaluated the patient on admission but made no comment about the Lovenox order. During the first dialysis treatment, there was active bleeding at the fistula site. The anticoagulant heparin had not been given. The nursing staff did not inform the physician of the bleeding. During the second dialysis treatment, there was uncontrolled bleeding from the fistula, and the patient exsanguinated and expired. Experts were critical that there was no EHR high-risk medication alert. Medication reconciliation might have prevented this error.
Strategies for Reducing Alert Hazards
- Understand alarm fatigue. When caregivers become overwhelmed, distracted, or desensitized to an alarm or an alert, determine the most important alarms, and work with your vendor to ensure that unnecessary alarms or alerts are not built into your system.5
- Determine if alerts are appropriately configured so that alert conditions are not missed or ignored.6
- Assess your EHR for frequent drug-drug interaction alerts, which have been shown to lead to alert fatigue that can cause the alerts to be disregarded, ignored, or disabled. Work with your EHR vendor to use key data elements to design EHR alerts for high-risk drug-to-drug interactions. The result will be more meaningful alerts that are less likely to be ignored or disabled, thus avoiding a possible error.
- Be aware that clicking through drug-to-drug therapeutic duplicates or drug/allergy alerts with little review can be interpreted to mean that the physician ignored the safety alerts.
- Read the alerts. EHRs record how much time is spent reviewing information. If the time is very brief and there is a negative patient outcome, the physician could be perceived as sloppy or hurried.
- Don’t turn off alerts. If a hospital-employed physician and hospital turn off alerts that could have avoided a patient problem, the hospital and physician may both be found liable.
- Always document why a clinical decision support (CDS) prompt was overridden. CDS may conflict with a medical specialty’s clinical standards of care or practice guidelines or with the information in FDA-approved drug labels.7
Alerts are a necessary safety mechanism when used, heeded, and configured appropriately for your practice. Check your alerts to make sure they provide adequate information and are not overly burdensome to your practice.