Your guide to NYC's public proceedings.
Q&A
Update and explanation of heat-related mortality data error
0:28:08
ยท
3 min
Council Member Keith Powers inquires about the significant increase in reported heat-related deaths from 350 to 580. Carolyn Olson from the Department of Health explains the reason for this update and the process of discovering and correcting the error.
- The increase was due to a coding error, not an actual increase in deaths
- The error was discovered during routine updates and addition of a new analyst
- The corrected data shows approximately 580 heat-related deaths per year over a five-year period
Keith Powers
0:28:08
I got it.
0:28:11
I believe the Department of Health had updated their report from last year about individuals that had died related to heat.
0:28:22
Originally was three fifty.
0:28:24
I believe the number is now up to five eighty.
0:28:26
Can you talk about why that number got updated?
0:28:29
What was the cause for that number to go from three fifty to five eighty and the discrepancy?
0:28:34
And just talk us through that number change.
Carolyn Olson
0:28:38
Definitely, thanks for that.
0:28:40
So we discovered an error, quite literally a typo, in our estimation of heat exacerbated So this is an analysis that requires application of a risk estimate, which we didn't have any errors in, to the number of deaths that occurred in the city, and we are then able to estimate the number of people who have had a condition that exacerbated by the heat.
0:29:07
And so when we identified this error, we immediately corrected it and what we discovered was that previously our estimate was three fifty, approximately three fifty deaths per year and that number is closer to it's more than five hundred.
0:29:22
So in our most recent five year period, because we estimate for five years at a time, we see that it's about five eighty, as you said.
Keith Powers
0:29:34
For last year?
Carolyn Olson
0:29:35
No.
0:29:36
So I want to be clear that this doesn't represent an increase from last year.
0:29:41
This just was literally an error.
0:29:43
And so the last two years, which both took we do rolling five averages, so we add a new year each year.
0:29:52
And both of those were around five eighty, more than 500.
Christina Farrell
0:29:57
I
Keith Powers
0:29:57
see.
0:29:57
And where did the typo occur?
Carolyn Olson
0:29:59
It was literally in the code that we have.
0:30:02
So you know, we discovered that we had a character that was wrong and it was pulling in the incorrect data sets for in order to calculate that specific number.
Keith Powers
0:30:13
And have you guys gone back to look at previous years to see if the same error?
Carolyn Olson
0:30:17
Yeah.
0:30:18
So we so those those reports include trend analyses, which we actually started producing last year.
0:30:26
So this was the second year, the February so we corrected and updated the 2023 and 2024 reports.
0:30:34
And in both of those reports for the first time, we did trend analyses and we have made certain that all of those analyses are providing the correct numbers.
Keith Powers
0:30:42
And how did that get caught?
0:30:44
Like how how what was the what was the what was the cause of going back to look at
Carolyn Olson
0:30:49
Oh, we're constantly updating those analyses and this is an annual report that is required by city council.
0:30:57
We've been doing for for a number of years.
0:30:59
So every year, we work to expand our analyses and improve upon them.
0:31:04
And we always have at least two analysts who are working with the code.
0:31:08
And this year, we added an additional analyst to the team.
0:31:12
And in the process of doing that, we had a new set of eyes on the code, and it was discovered.