Objective
The purpose of this study is to evaluate the judgment of the Fire Department of NY Emergency Management System dispatcher’s statistical data on whether there was a categorical distinction in severity in critical and high risk cardiovascular and respiratory call prioritization during the Covid-19 lockdown period in the five boroughs of New York City.
Methods
Two models, an ordinal logit regression model with a proportional odds specification and a non-proportional odds ordinal model with log-scale effects are fitted and compared using Fire Department of New York EMS Incident Dispatch Data. Goodness-of-fit diagnostics—such as the test of scale effects and the convergence method—for ordinals models are also discussed.
Results
There were 730 more cardiovascular calls (or a 48.28% increase) in 2020 with a priority rank of P1, P2, or P3 when compared to 2019 call volumes. Similarly, there were 544 more respiratory calls in 2020 than 2019, a 74.22% increase. The vast majority of the increased cardiovascular and respiratory call types (1,241) in 2020 now included a fever-cough designation. For all call types, there was 24.28% increase in 2020.
When a log-scale ordinal model is estimated and adjusted for scale effects, our data showed that respiratory calls had the highest odds ratio of 43.23% when compared to cardiovascular calls of 31.24% which meant that higher priority levels were more than often applied to these higher call volumes calls when compared to all other call types, everything else remaining constant.
Among the boroughs, Queens showed the higher odds (25.26%) of being assigned a higher priority level than the other boroughs, given that Staten Island was omitted. The odds ratios for all categories of increased call volumes, including those fever-cough calls that appeared only in 2020, showed that higher priority levels were more often associated with lower call volumes. In the month of May, the odds ratio of 7.27% was the highest, while the month of June with an odds ratio of -8.77% saw that it was less likely calls were assigned when compare to the other months. In 2020, the odds ratio for that year was 11.72% when meant that higher priority call were assigned when compared to the previous 2019.
Average response time showed lower odds at higher priority levels. Surprisingly, all disposition levels showed that higher priority cardiovascular and respiratory calls meant that patients were less likely to be transported. The number of calls cardiovascular and respiratory calls with priority 1, 2 or 3 in which patients were transferred totaled only 566 or 16.06% when patients pronounced dead at the scene are excluded.
Conclusions
Our analysis showed that FDNY’s EMS dispatchers made a severity distinction in their assignment of priority levels (P1, P2 and P3) to potential high risk cardiovascular and respiratory cases during the Covid-19 lockdown, even though very few patients were transported. A log-scale effects ordinal model is preferred over a proportional odds specification, particularly when potential heterogeneity is unobserved.