Over the past year, Multnomah County has used one key metric to defend the quality of its ambulance system: the survival rate of people who are revived by paramedics after their heart stops pumping, a condition known as cardiac arrest.
Cardiac arrest survival is considered a key metric for first responders nationwide. They collaborate with hospitals to track it.
In Multnomah County in 2022, 15.7% of patients who paramedics treated for cardiac arrest survived long enough to make it home from the hospital. That number may sound dismal, but it’s high compared to the national average, just 9.3% that year.
Then in April, Willamette Week reported data that showed that in Multnomah County, that number was going in the wrong direction: In 2023, it had fallen to 12.8%.
That gave ammunition to the county’s critics, who said its ambulance service is in a crisis because of standards that require two paramedics on each ambulance, rather than one paramedic and one less-trained emergency medical technician. Those critics, including the county’s ambulance provider, American Medical Response, say the staffing requirement has made it impossible to field as many ambulances as the Portland area needs, contributing to slow response times. During a cardiac arrest, rapid treatment is essential for survival.
County EMS Medical Director Dr. Jonathan Jui rebutted that criticism, and explained that his data showed overdoses were responsible for the falling survival rates, not slower ambulance response times.
It’s an explanation that made intuitive sense. But it was wrong.
Last month, OPB identified inconsistencies in the data that had been originally prepared by Jui as part of training for paramedics and then shared with Willamette Week. Two national experts on out-of-hospital cardiac arrest also said overdoses were an unlikely explanation for a decline in the survival rate.
At OPB’s request, Multnomah County’s public health team conducted a new review of the past 10 years of cardiac arrest survival rates in the county.
That analysis revealed numerous inaccuracies in the cardiac arrest data that the county had provided to Willamette Week and OPB in April. The updated data makes it clear overdoses are not likely to explain what happened between 2022 and 2023.
The county has since retracted Jui’s statement claiming overdoses were to blame for the decline in survival rate.
“Survival and return of circulation percentages vary from year to year, but are generally stable in Multnomah County,’’ county communications director Julie Sullivan-Springhetti wrote in an email confirming the retraction.
Data that didn’t add up
Every year, emergency dispatch centers, first responders and hospitals submit data on cardiac arrests to a national database known as CARES, or the Cardiac Arrest Registry to Enhance Survival. Oregon participates at the state level.
Participants can query the database to analyze trends in their own communities. But it’s a private, confidential registry: Reporters and the general public don’t have access to it.
So when Willamette Week got hold of a data table that included the 2023 cardiac arrest survival rate for Multnomah County, it was a rare look at a metric that can be hard to get.
The table showed 10 years of cardiac arrest survival rates and included a detailed breakdown of what heart rhythms the patients had when first responders hooked them to a defibrillator.
Those rhythms are a critical piece of information because patients tend to have very different odds of survival based on their heart rhythm.
Jui, the county EMS medical director, told the alt-weekly that survival rates had only fallen for a single group of patients: people whose hearts were flat-lining. The medical term for that is asystole. Those patients always have the lowest odds of surviving. Overdose patients are most often found with that flat-line heart rhythm.
For other patients, Jui said, survival rates in 2023 had actually gone up.
That math, though, seemed implausible. The data table the county provided to Willamette Week showed the survival rate for patients in asystole falling by less than half of a percent between 2022 and 2023. The slight dip in asystole survival rates seemed unlikely to be enough to offset larger gains in survival for other types of cardiac arrest.
The county’s response
On April 14, OPB asked the county to clarify whether a change in survival for flat-lining patients alone could explain a 3% drop in patient survival from 2022 to 2023.
The county provided OPB the same data table it had given Willamette Week, and Jui repeated his explanation that drug overdoses had caused the decline in survival.
“The devastating effect of fentanyl on our community is an external factor EMS cannot control,” he wrote in an April 17 email. “Regarding quality of care, we are actually at 2022 levels or higher when you look at the presenting rhythms data and survival outcomes.”
On May 10, after repeated requests for clarification, the county agreed to look into the apparent contradictions in the data.
A data analyst ran the numbers again using the most recent data available in the CARES registry. That fresh analysis showed that in 2023, survival rates fell for a broader set of patients than Jui’s previous explanation covered.
The county saw a small improvement for just one group of patients, those with a rhythm known as PEA, or pulseless electrical activity. The change amounted to about two more survivors in that group in 2023, according to the county.
Survival dropped for asystole patients — amounting to about five more deaths in 2023.
But the biggest decline in survival was for patients whose hearts were in a weak, fluttering rhythm known as ventricular fibrillation. About 15 more patients in ventricular fibrillation died in 2023, compared to the year before.
Ventricular fibrillation is also called a “shockable” rhythm because a patient can be kept alive with chest compressions and then treated with a defibrillator.
It is widely understood to be the most survivable type of cardiac arrest, according to two independent experts who spoke to OPB.
“You’re talking about fivefold plus better outcomes,” said Dr. Cameron Dezfulian, a senior faculty member at Baylor College of Medicine in Texas and an expert in cardiac care.
It’s also a rhythm that is less frequently associated with overdoses, according to Dezfulian.
Sullivan-Springhetti, the county spokesperson, said the original data table hadn’t been vetted before it was publicly released. The document was only intended to be used in CPR training for paramedics.
“Certainly we regret the error,” she said, adding that any mistakes were inadvertent.
“Dr. Jui updates his slides annually for training in the winter months but the CARES database is not finalized for each year until April or May because final outcomes (death, survival to discharge, missing data) are not complete until then,” Sullivan-Springhetti wrote. “He does not go back and update the previous years because the intention is training not formal reporting.”
The new analysis showed in total, about a third of the numbers in the data table Jui had been using in his training presentation were off by a percentage point or more.
The county EMS team said its new analysis is accurate, as of the date it was conducted — May 13, 2024. Oregon’s CARES coordinator said that the state’s 2023 data was finalized in late April and that no cases submitted to the registry were lost to follow-up.
Survival paradox for overdose patients
The experts OPB spoke with said even if the county’s data hadn’t proven inaccurate, there’s a second reason to doubt that overdoses caused a drop in cardiac arrest survival in the Portland area last year.
EMS systems are treating increasing numbers of overdose patients who’ve gone into cardiac arrest due to fentanyl. And those patients do often have a flat-line, asystole heart rhythm, they said. But survival rates for overdose patients are better than one might expect, based on that rhythm alone.
“The weird paradox is that drug-related out-of-hospital cardiac arrests tend to do better than non-drug-related out-of-hospital cardiac arrests. They have a higher rate of survival,” said Dr. Ralph Wang, professor of clinical emergency medicine at the University of California, San Francisco who recently published a paper on overdose and cardiac arrest.
Dezfulian, who serves on a committee of the American Heart Association, said the same surprising trend appears in national data. It may be, he said, that the combination of Naloxone and CPR is leading to better outcomes for overdose patients.
“The explanation of opioids being the cause doesn’t really make sense,” he said.
At the same time, Dezfulian cautioned against drawing any hasty conclusions from a single-year drop in survival.
Dezfulian said it’s normal for EMS systems to have an “off” year from time to time, and the data can’t always clearly explain why. Multiple factors impact cardiac arrest survival rates.
Without question, he said, the speed of first responders can be a major factor.
But two other key ingredients driving survival can fluctuate year to year: Cardiac arrest patients are much more likely to survive if they receive chest compressions from a bystander right away, and if there’s an AED defibrillator nearby.
Nationwide, survival rates improved last year: 10.2% of patients lived to make it home from the hospital.
If Multnomah County wants to get to the bottom of what caused its decline in survival, it could look at whether bystander intervention and AED use rates changed.
That’s information that first responders collect about each cardiac arrest patient and submit to the CARES database.
“If your survival is down and your bystander rates are down or your AED use is down, that could be the explanation in of itself,” Dezfulian said.
On the other hand, if CPR rates or AED use remained steady or improved in 2023, as survival fell, it would be a red flag that response times could be the culprit.
“At that point,” Dezfulian said, “You start wondering about the response.”
