Politicians Use Statistics: The Devil Can Cite Scripture For His Purpose

There is a movement in the administration to challenge the accuracy of the mortality statistics that are commonly used, in an attempt to dilute the political impact being felt from the growing total of COVID-19 deaths. The theory is that if the number is disputed, the criticisms being leveled against the President will be similarly suspect. We’ve covered the misuse and misrepresentation of statistics in several previous blogs and articles, but there are elements to this particular situation that are worth analyzing.

Combination of Differing Standards

To be clear: the mortality number and rate are inherently wrong. A combination of differing standards for associating deaths with the virus on a state by state basis (sometimes from location to location), along with delays in reporting and significant absences in nursing home documentation tells us that whatever the real number is, it won’t be what we read every day in the official statistics. That said, whatever has been reported has been consistent, and changing the methodology at this point in time is likely to fail on a number of levels.

First, if the intent is to reduce the accepted number of fatalities, it is relevant to understand that the actual number of deaths is almost assuredly higher. While there may be some fatalities that have occurred from multiple causes, or from causes other than the virus but mistakenly included in the counts, there are several factors that would suggest that we have been undercounting, perhaps significantly.

Use a Statistical Measure

We have a fairly good idea of how many people have died over this period, regardless of what the cause of death might be. Health professionals use a statistical measure here that is relevant, called excess mortality; this is the difference between recorded deaths and the historical norms for a given period. In both national and more defined areas, the excess mortality is greater than the number of deaths attributed to the Coronavirus, suggesting that at least some portion of the excess can be attributed to the virus. There is virtually no statistical evidence that there has been an over-accounting; if so, then the statistical anomaly would be even further out of whack.

We can know from past experience and present reporting by municipalities and health care facilities that there is a substantial delay in the reporting of causes of death, as much as several weeks in some cases. In times when the activity level is beyond normal (an understatement here) the delays can be substantially greater. In prior epidemics, the actual mortality figures were only understood in hindsight, months after the illness had subsided and records from hospitals, morgues, and other sources were able to catch up. There is no historical evidence of counts being revised lower following a subsequent analysis.

We also know from a large sampling that nationally, nursing homes and related facilities form a significant portion of the fatalities, well above their percentage of the general population. Analysis has shown that, of the over 79,000 deaths attributed to the virus overall, about 27,600 (35%) have been confirmed as either patients or staff at facilities of that nature. The mortality rate of seniors, particularly with other factors, is by far the highest of any group; given the estimates of about 150,000 infected patients in those locations, the rate of death is above 18%, and individual facility numbers show even higher rates. During much of the first month where the virus was impactful, these facilities lacked testing capabilities and the capacity to quantify COVID deaths specifically, designating the casualties as respiratory-related fatalities. Since these facilities commonly cremate their deceased, there is no ready opportunity to reclassify those deaths.

Additionally, our present understanding of the symptoms and impacts of the coronavirus continues to evolve. Once thought of as a respiratory ailment, there have been confirmations of it having severe neurological effects, causing kidney failure, blood clotting, and coronary damage. Recently, the CDC had to reissue its list of possible symptoms to used to identify cases; doctors have been expanding their understandings of how the disease infects and kills, to include the broader range of connections and afflictions. It is highly likely that — if we haven’t already — we will soon be expanding the number of patients whose cause of death was Covid-19 well beyond what we thought the virus was responsible for based on further understandings.

As we’ve discussed in previous posts, when confronted with a debate over statistical evidence or the context of its use, it is often worth looking at the related motivations and benefits of one or the other interpretation. In this matter, the motivations are particularly telling.

Receive Additional Funding

To date, the primary suggestion of a reason for over-reporting is a recognition that hospitals receive additional funding for patients that either is infected with or have died of, Covid-19. There is some reason to consider this possibility: under the CARES act and other legislation, hospitals will receive additional funds for uninsured patients where they are diagnosed with the disease. That presents a clear financial motivation for hospitals to count patients suspected, but not confirmed, of having the virus as infected. The issue here is a series of reductions that make the ultimate impact likely to be fairly small.

First of all, in the areas most afflicted such as New York, New Jersey, and other states, hospitals have routinely tested suspected patients since the levels of incoming elevated. Second, the population of deaths include a substantial portion that ended their lives on ventilators, carried as a separate category in the compensation. Third, the primary parties in providing a diagnosis or death certificate receive no financial benefit from designating a patient in one form or another; by the time a hospital administrator is involved, the record has usually been completed. This suggests that while there is a theoretical motivation, it is likely only a factor in a very small minority of cases.

Considerable Motivations

In cases where states are involved in the provision of causes of death to the federal government, there are considerable motivations for less diligent reporting. Governors of either party are under intense scrutiny for their handling of the mitigation process (lockdowns, distancing, masks, etc.), and a common number used in the evaluation is deaths; inflating those numbers are hurtful to their public approval. A good number of the Governors would like to open their states up for business, a primary source of revenues for the state; elevating the counts for cases and particularly deaths would extend the period before they are able to do that.

By that same token, the motivation for the Administration to argue for smaller numbers are obvious and significant. The Administration is — similar to the Governors — likely to be judged on their response and success in containing the virus. Deaths are the most dramatic and impactful of the numbers that will be measured; the greater the mortalities, the less likely that their efforts will be judged favorably. In addition, the President has frequently offered his opinions on what he would consider a successful outcome; as the number of reported deaths reaches and exceeds those totals, his prior statements become available as self-incrimination.

Relying on the CDC for Information

Finally, efforts of the Administration to recast the statistics will likely fail to move the needle of public perception, regardless of their pronouncements. Media sources have become used to acquiring the data that they report from private sources, relying on the CDC for information less frequently than ever before. Any decision by the Administration to change the mechanism will likely be ignored by most of the media, who will continue to report the now well-known figures. The only media likely to alter their reporting would be the most partisan outposts, where the approval of the President’s work is already overwhelmingly positive; neutral or alternatively partisan reporting will not change opinions, merely reinforce the existing perception that the Administration is playing politics with the virus and its effects.

Science, statistical analysis, logic, and historical norms all suggest that whatever numbers we see today will be prove to have been less, not more, than the actual counts. The motivation of the Administration is transparent — it is seeking to reduce the public perception of a critical piece of evidence as to their success in responding to the pandemic, particularly relevant as the election continues to approach.

It seems wise to go with the science, the logic, and the history over the Administration’s wishes to change the conversation. As it often does, an understanding of the underlying particulars and an analysis of motivations reveals a simple equation for knowing what to believe.