Huge Inconsistencies in COVID Death Totals for Colorado Counties

According to the Colorado Department of Public Health and Environment, 9,563 people have died because of COVID-19 and 9,337 have died with the disease. But that doesn't mean you can add the two numbers together to determine how many people have died while infected in Colorado. There's some overlap in the figures, and counties also use different systems to record who died with the disease and who died from COVID.

The stats from many Colorado counties show the number of fatalities among infected individuals and those for whom COVID was the direct cause of death to be almost even. But in plenty of small, conservative counties that have bristled at mask mandates and other safety protocols, the figures contend that many more people died with the disease than from it.

This disparity in fatality figures raises questions about the reliability of COVID death statistics in Colorado. While the CDPHE stands by its numbers, its explanation reveals that the process includes judgment calls from an array of parties — and those judgments can differ widely.

Death stats in many Colorado counties show a close correlation between deaths with and from COVID. For example, as of today, December 3, Douglas County lists a total 413 deaths from COVID, and another 325 deaths of people who were infected at the time they passed away. Denver, meanwhile, counts 989 deaths from the disease and 1,022 with the disease.

Contrast those figures with the numbers from Bent County, which counts 32 people dying with the virus, but just fifteen because of the virus, and Conejos County, which acknowledges 42 deaths among cases but only sixteen because of COVID. Kit Carson, Washington and Las Animas counties count 23, 24 and 27 deaths, respectively, with the disease but fewer than ten from it. The last number is imprecise because the federal Centers for Disease Control and Prevention suppresses county death counts in the single digits.

Here are the COVID death totals for every county currently listed by the CDPHE. When an asterisk appears, it represents deaths due to COVID, since the CDC doesn't release the figure for deaths from the disease if it's under ten.

Hinsdale — 0*
Jackson — 0*
San Juan — 0*
Lake — 1*
Mineral — 1*
Gilpin — 3*
Kiowa — 4*
Dolores — 5*
San Miguel — 5*
Sedgwick — 6*
Baca — 6*
Clear Creek — 6*
Ouray — 6*
Pitkin — 6*
Cheyenne — 7*
Archuleta — 8*
Lincoln — 8*
Saguache — 9*
Rio Blanco — 10*
Rio Grande — 10 from the disease, 32 with the disease
Crowley — 11 from the disease, 27 with the disease
Custer — 12*
Gunnison — 12*
Summit — 12*
Elbert — 13 from the disease, 26 with the disease
Park — 13*
Costilla — 13*
Montezuma — 15 from the disease, 48 with the disease
Bent — 15 from the disease, 32 with the disease
Phillips — 15*
Yuma — 15 from the disease, 18 with the disease
Conejos — 16 from the disease, 42 with the disease
Grand — 16*
Huerfano — 18 from the disease, 21 with the disease
Routt — 18 from the disease, 24 with the disease
Teller — 20 from the disease, 43 with the disease
Prowers — 21 from the disease, 38 with the disease
Kit Carson — 23*
Washington — 24*
Las Animas — 27*
Chaffee — 32 from the disease, 36 with the disease
Eagle — 32 from the disease, 17 with the disease
Moffat — 36 from the disease, 47 with the disease
Otero — 48 from the disease, 103 with the disease
Broomfield — 55 from the disease, 86 with the disease
Alamosa — 56 from the disease, 49 with the disease
Garfield — 66 from the disease, 72 with the disease
Logan — 70 from the disease, 92 with the disease
Fremont — 72 from the disease, 139 with the disease
Delta — 77 from the disease, 98 with the disease
Morgan — 79 from the disease, 115 with the disease
La Plata — 90 from the disease, 63 with the disease
Montrose — 113 from the disease, 113 with the disease
Douglas — 413 from the disease, 325 with the disease
Weld — 429 from the disease, 540 with the disease
Mesa — 464 from the disease, 410 with the disease
Boulder — 475 from the disease, 286 with the disease
Larimer — 507 from the disease, 378 with the disease
Pueblo — 591 from the disease, 593 with the disease
Denver — 989 from the disease, 1,022 with the disease
Jefferson — 1,002 from the disease, 1,001 with the disease
Arapahoe — 1,023 from the disease, 918 with the disease
Adams — 1,099 from the disease, 943 with the disease
El Paso — 1,441 from the disease, 1,267 with the disease

Why are there such enormous differences in the counts of deaths from and with the disease? To learn more, we reached out to CDPHE spokesperson Gabi Johnson.

"Epidemiologists track deaths among cases for a number of diseases, including COVID-19," Johnston notes. "This includes all deaths within thirty days of being identified as a case of COVID-19, regardless of cause. This methodology will include people who died of COVID-19, as well as some who died of causes other than COVID-19. It also will exclude deaths that occur more than thirty days after being identified as a COVID-19 case, even if the death was due to COVID-19. The strength of this methodology is that the data are available quickly, so that we can understand the disease in real time."

She adds: "Deaths due to COVID-19 are deaths where COVID-19 is an underlying cause of death or named as a significant contributing condition on the death certificate. This data rests on the judgment and assessment of the individual completing the death certificate (usually the provider if the patient is in the hospital, or a county coroner). The information is then sent to the CDC, which codes the death. That can take two to three weeks. This data lag makes 'deaths due to COVID-19' less useful for real-time assessment of the impacts of the pandemic."

At first blush, it would seem that people who died from COVID-19 would be a subset of those who died while infected. But the truth is considerably more complicated.

"A single individual who dies after contracting COVID-19 may appear in either or both data sets," Johnston points out. "If they die of COVID-19 shortly after diagnosis, they will be both a 'death among COVID-19 cases' and 'death due to COVID-19.' Alternatively, they may only appear in one or the other data set, depending on the time or cause of death. It is never appropriate to add these data sets together, since that will result in double-counting of deaths among people who appear in data generated via both methodologies."

As for the wide variations of death stats from county to county, Johnston offers this: "'Deaths among cases of COVID-19' reflects deaths among Colorado residents, or residents of a specific county, irrespective of where the death occurs. ‘Deaths due to COVID-19’ reflect deaths occurring in Colorado, or within a specific county, irrespective of the decedent’s state or county of residence. The latter is the definition used by the CDC/National Center for Health Statistics, from whom these data are initially obtained and reproduced on CDPHE’s COVID-19 data website. For example, assuming criteria are met for both ‘death among a case of COVID-19’ and a ‘death due to COVID-19’, a decedent who is a resident of Denver County but dies in Jefferson County will appear in Denver County’s ‘deaths among cases of COVID-19’ numbers, while appearing in Jefferson County’s ‘deaths due to COVID-19’ numbers."

Another factor cited by Johnston: "In counties that may not have hospitals, residents who die from COVID-19 often die in a hospital in a different county, so for those counties, the difference between 'deaths among cases' and 'deaths due to COVID' can be large. The differences are not due to variability in how cases or deaths are defined across counties." Still, she continues, "per Colorado law, the determination and recording of cause of death on death certificates may be done by a Colorado-licensed physician, or the elected county coroner or appointed medical examiner (the latter limited to Denver City/County)."

Vital records statistics, which include deaths due to COVID-19, are "based on CDC coding of death certificates, and the deaths could have occurred at any time after diagnosis," Johnston says. "Physicians, medical examiners or coroners rule COVID-19 as a cause of death on the death certificate, according to guidance from CDC/National Center for Health Statistics on how cause of death certification of a death involving COVID-19 should be completed."

The bottom line, Johnston concludes: "Deaths 'among' and deaths 'due to' represent two different ways of counting deaths, and each of the two methodologies we use has strengths and limitations."