In yesterday’s article about challenges with obtaining Elmhurst Hospital emergency department visit data, I included a footnote (7) about a discrepancy involving monthly ED numbers colleagues & I used in Does New York City Make Any Sense? (Verduyn et al, 2023).
Here I show the discrepancy in the context of other data, which I think could be a sign that data has been misrepresented or manipulated.
Source 1: SPARCS Audit Reports accessed October 2023
Figure 10 in Verduyn, et al (copied below) shows the monthly ED visits at Elmhurst alongside inpatient admissions and outpatient visits from January 2018-December 2020 as reported in audit reports from the New York Statewide Planning and Research Cooperative System (SPARCS).
The 2020 numbers were accessed from the then-recent version of the relevant SPARCS page in October 2023; data from older reports were accessed via Wayback Machine following an unsuccessful attempt to get them directly from the agency.
An ED-visits only version of the data with the raw numbers displayed shows volumes were highest in January 2020 than at any other point that year. The massive decline in April 2020 and low levels through December 2020 mirror the general pattern shown across other sources.
Source 2: HHC Monthly Visits
How do those numbers compare to the data NYC Health + Hospitals gave me recently in response to an FOI request (shown in figure 6 here)?
Not very well in early 2019, late 2019, and early 2020.
The biggest gaps between the two sources are during months associated with seasonal or “COVID pandemic” respiratory illness. The SPARCS audit reports accessed a year ago show a busier Elmhurst emergency department in January 2019, and in November 2019 through January 2020, than do the HHC data. The greatest difference percentage-wise is in April 2020, with SPARCS showing 37% fewer visits than HHC.
Source 3: SPARCS Audit Reports accessed November 2024
When I use data from 2020 SPARCS audit reports currently available on the state health department website, I see numbers both comparable to and discrepant from reports accessed in October 2023 and the data provided by HHC.
Limiting the view to 2020 (the only year for all three sources) shows close agreement between the SPARCS audit reports for every month but February. Because the audit report data is updated periodically, minor differences are expected; a gap of more than 1,400 visits is harder to reconcile, especially when the more “recent” version of the report shows a higher number, as though a slew of visits were suddenly found, moved, or reclassified.
Combination
Although I don’t have monthly data for the complete January 2018-December 2023 timeline from all three sources, seeing what I do have on the same graph nevertheless raises questions about what the agency that controls Elmhurst (HHC) provided to me versus what they reported to SPARCS in the past and more recently. Why there is relative agreement for much of 2020, yet significant discrepancies in late 2019/early 2020 – and again after April 2021 – I’m not sure, but taxpayers deserve better than ‘because pandemic.’
Quarterly Comps
The first set of Elmhurst ED visit data HHC gave me (in January 2023) were quarterly numbers. As shown in the previous post, those data conflict with the monthly data the agency provided a few weeks ago.
When I turn all monthly data into quarterly, it’s still the internal conflict between HHC records that is most problematic. All sources converge in the second quarter of 2020 and are comparable through the end of the year before diverging again.
Explanations?
Given the dynamic nature of the SPARCS audit reports, it could be argued the data reported therein is less valid and reliable as an “official” source. I agree with that view when it comes to a current calendar year. Four-plus years later, the numbers should be relatively final and not deviate significantly from HHC numbers, assuming that HHC is reporting directly to SPARCS.
The fluidity is probably why the audit report page says the reports “should not be used for public reporting of counts related to understanding hospital inpatient, outpatient, emergency department, and ambulatory surgery admissions/visits” and to request via email counts by facility for “public reporting needs.” When I tried to do so, I didn’t get anywhere. (I emailed again today.)
Other than definitional differences, incompetence, or data entry errors, I have no charitable reason for why the SPARCS and HHC data sources are discrepant in the ways they are, or at the points shown, but welcome explanations I may not have considered.
I lean toward viewing the Elmhurst ED data as potentially manipulated mainly (but not only) because
- agencies and researchers have been cagey about releasing records (e.g., ED visit data, pre-2020 death data),
- like the citywide hospital spike, Elmhurst spring 2020 death toll is speciously high from a body management perspective alone,
- bed occupancy data sources are discrepant and/or incomplete, and
- the entire NYC mass casualty event is rife with serious problems, has not been substantiated with basic proof (e.g., names, public death records), or by federal aid for disaster-related funeral assistance.
Leave a Reply