Getty Images

AI EHR Integration in ED Reduces Med Rec Errors, Boosts Patient Safety

An AI EHR integration helped a Michigan hospital boost patient safety by cutting down on medication reconciliation errors by 90 percent in the emergency department.

As the digital health transformation continues, artificial intelligence (AI) EHR integrations are helping streamline medication reconciliation for improved patient safety.

Medication reconciliation is the process of compiling the most accurate list of a patient’s medications. Inaccurate medication lists can lead to adverse drug events (ADE) that threaten patient safety.

Medication reconciliation has always been a chore, especially in the emergency department, said Aaron Smith, MD, chief medical informatics officer at Covenant Health in Saginaw, Michigan.

Traditionally, Covenant’s ED nurses have been in charge of medication reconciliation, who are also responsible for triage, administering meds and IVs, among many other tasks.

“In the ED, there are lots of competing interests,” Smith, who is also a practicing emergency physician, told EHRIntelligence. “It can be a little bit chaotic at times, so getting an accurate medication list can be challenging.”

Sometimes, patients come in with a list of their current medications. However, that's the exception as opposed to the norm, Smith noted.

“If the patient's been within our health system, then their physicians more than likely prescribe all the medications through the EHR, but that's not foolproof,” Smith explained. “We're not all things to all people. Patients see specialists that are outside of our health system, so we can't account for those meds either. It's fraught with errors to find an accurate list of what the patient's actually on.”

To address this issue, ED clinical staff requested that the hospital executive team hire staff dedicated to the medication reconciliation process.

“We proposed a trial of a program to bring pharmacists within our emergency department to address this problem,” Smith said.

“We tracked how many med errors were occurring in a patient's chart for a period of time compared to a complete medication reconciliation once they were admitted, then used literature to show what each potential med error could cause in terms of adverse drug events and the cost per adverse event.”

Then, ED staff estimated the cost of having a pharmacy program within the ED staffed by pharmacists and pharmacy technicians whose responsibility is to gather an accurate medication list upfront.

“We were able to get approval from the executive team to have a trial, which was overwhelmingly successful,” Smith said.

The pharmacy technicians helped identify patients that were likely to be admitted to the hospital.  

“We see on average anywhere from 250 to 300 patients a day, so we can't complete med recs on every one of those patients,” Smith explained. “We built some scoring mechanisms and added an order into the EHR for a physician to say if they were going to admit a patient to alert pharmacy technicians so they could start the med reconciliation process.”

Initially, the pharmacy technicians would go in and talk with the patients for the medication reconciliation. If they couldn't talk to the patients, they would call family members and local pharmacies.

While the addition of the pharmacy program helped alleviate clinician burden, compiling an accurate medication list was still a very manual process, Smith explained.

The hospital decided to leverage an EHR integration that brings in prescription fill data from pharmacies to mitigate the administrative burden of medication reconciliation.

Smith emphasized that the best indication of what a patient is currently taking is her prescription fill history.

The first EHR integration the hospital tapped showed significant gaps in coverage, Smith explained. For instance, the vendor did not contract with small, independent pharmacies in the area or Walmart.

“They only had about a 60-something percent hit rate of meds the patient may actually be taking,” Smith noted. “We were finding our techs were still doing a lot of calls to small pharmacies and family members.”

“Maybe that vendor would be better in a different market where there's a different payer mix and a different stronghold of national pharmacies,” he added.

The hospital then leveraged an EHR integration from health IT vendor DrFirst.

Smith explained that the vendor worked with some of the independent pharmacies to bring them on board, boosting the medication hit rate to 93 percent.

“That's huge time savings there,” Smith said. “The pharm techs have access to that, but the physicians also have access to that in their workflows. I can go in my workflow and I can look at our medications prescribed by our providers listed on the EHR, but I can also look at outside feeds, so it's been really valuable in that way.”

Smith explained that the integration helps take the human error element out of medicine reconciliation. Instead of focusing on manual investigation, the integration allows pharmacy technicians to validate medication lists to ensure lists are accurate.

The pharmacy technician presence and the AI EHR integration led to an almost 90 percent reduction in errors of patient medication lists, Smith said.

While literature offers a range of what potential ADE can cost healthcare organizations, the hospital used $8,750 as the figure for how much each an ADE costs the hospital. The hospital estimated that one in 10 adverse errors cause patient harm, leading to $6 million in savings.

Smith noted that as COVID-19 continues to spread across the country and overwhelm healthcare organizations, health IT that streamlines workflows can help support clinicians.

“Healthcare is under a tremendous amount of stress right now,” Smith said. “I'm optimistic that we can survive this, but there's going to be a cost to the system. We're really burning out providers. Any kind of technology we can use to offload some of the cognitive work and to streamline their processes is more valuable now than ever.”

Next Steps

Dig Deeper on Interoperability in healthcare

Cloud Computing
Mobile Computing