Healthcare

Data-Driven Decisions - Lies & Statistics

by Joseph Cahill

One of the perennial buzzwords in governance is “data-driven.” This concept implies that work performed by emergency medical services (EMS) and other government agencies can be judged by statistics and that decisions can be made based on an aggregation of previous outcomes. EMS systems are frequently judged on statistics – for example, the “save rate,” which is the percentage of cases when a patient has no heart beat upon arrival of emergency medical technicians (EMTs) but he or she survives the episode. However, the ways in which agencies use these data measurements can make a significant difference.

Lives Saved & Data Collected Save rate is the type of statistic that plays well in the media, but may not be realistic. As with all statistics, the save rate depends on what it is compared to and how this type of statistic is used. In one example, a major city in the Pacific Northwest received accolades because it had a save rate around 25 percent, while most cities were hovering in the single digits. A review of the rules of that city revealed that if the patient was flatline on a cardiac monitor, also known as asystole, they were pronounced dead and no workup or medical care was rendered. Meanwhile, in other cities with low save rates, asystolic patients were routinely treated. Since such patients have the lowest chance of survival, their save rates significantly lowered their cities’ overall rates.

Based on these statistics, EMS managers could act upon such data by conforming their policies in two ways. First, they could mandate that patients with asystole not be treated, thus pushing the agency save rate up. Although the save rate for flatline is low, it is not zero. This move would condemn some to die, who might otherwise have been saved. Second, they may consider the cost-benefit ratio. Workup – or code of cardiac arrest – is the most costly work EMS performs so, by removing a large number of low success cases, an agency could fund units they would not otherwise be able to afford, thus lowering response times and saving lives.

Setting the Bar Another statistical analysis that EMS agencies may perform would determine the useful lifespan of an ambulance. “Useful lifespan” could be defined as the service life after which the costs of needed repairs and maintenance exceed a particular benchmark, or the acceptable amount of out-of-service time within a year. Once this benchmark is established, vehicle replacement would be planned within a particular timeframe.

Both of these statistical examples – save rate and useful lifespan – seem reasonable but, in both cases, a decision must be made about the level of adherence and the severity of the stakes. In the case of save rates, some people may die if the bar is set too high. In addition, managers must have the latitude to make common-sense decisions. In the case of vehicle replacement schedules, fleet managers should be able to flag a vehicle that has outlived its lifespan prior to the expectation. Without these flexibilities, a statistically driven government is as bad as a statistically blind government.

 

Joseph Cahill is the director of medicolegal investigations for the Massachusetts Office of the Chief Medical Examiner. He previously served as exercise and training coordinator for the Massachusetts Department of Public Health and as emergency planner in the Westchester County (N.Y.) Office of Emergency Management. He also served for five years as citywide advanced life support (ALS) coordinator for the FDNY – Bureau of EMS. Before that, he was the department’s Division 6 ALS coordinator, covering the South Bronx and Harlem. He also served on the faculty of the Westchester County Community College’s paramedic program and has been a frequent guest lecturer for the U.S. Secret Service, the FDNY EMS Academy, and Montefiore Hospital.