Finally, some worthwhile data.
In our current culture —especially in the medical sphere, acquiring data for data’s sake has become its own illness whose insidious contagion serves further to fracture and fragment our health care delivery.
Though I don’t routinely find good news in the topic of death, being the skeptical optimist that I am enables me to see the potential in a new study published in JAMA detailing the mortality rates for major causes of death from 1980-2014.
Why the cheer? Because the report is documenting United States county-level trends. Recognize it is a tempered one, but cheer nonetheless. Until we start to recognize that policy decisions and implementation need to be adjusted and adaptable to the population at hand, we will never spur momentum away from the status quo or make a necessary dent in cure rates. Uniform health policies for the entire nation will rarely effectively improve conditions. In-depth understanding of regional disparity is crucial to success.
So, kudos to this effort for attempting to expand knowledge that might actually provide a path to reducing geographic disparities in disease. How it is inevitably used and employed is a whole other matter. At the very least, it is informative.
The researchers’ objective was “to use a novel method for county-level estimation and to estimate annual mortality rates by US county for 21 mutually exclusive causes of death from 1980 through 2014.” They analyzed the death registration data from the National Vital Statistics System of the 80,412,524 deaths recorded in the United States during that time frame from 3110 counties or groups of counties—19.4 million were redistributed as garbage codes, aka “implausible or insufficiently specific cause of death codes.” They concluded “this analysis found significant variation in mortality rates and changes in mortality rates among counties for all causes of death.”
In my article Population Well-Being Tied to Longevity (Shocker!), I address the life expectancy inequalities by region that is consistently researched. In this latest study, cause-specific mortality per area is further explored and extended to more top causes of death.
Cardiovascular disease-related deaths were elevated along the southern half of the Mississippi River. Self-harm and interpersonal violence highest in southwestern counties, while eastern Kentucky and western West Virginia claimed the title for chronic respiratory disease. The county trends of cause-specific mortality rates also fluctuated throughout the thirty-four year span of the study for most causes.
For example, neoplasms (abnormal growths or tumors, often cancerous) caused 24.3% of all deaths during the study’s temporal course. The areas of greatest occurrence were along the southern half of the Mississippi River, eastern Kentucky, western West Virginia and western Alaska. Many counties stretching from Idaho and Wyoming in the north to western Texas in the south enjoyed rates well below average. The largest decreases in rates included central Colorado, southern Florida, Alaska, parts of New England, and California coastal communities.
Perusing this work will likely interest many, if not merely for the pretty, color-identified maps of the United States featuring the counties corresponding to top causes of death that range from diabetes, blood, liver, neurological and urogenital disorders to transport injuries, mental and substance use disorders.
Now, to my Debbie Downer portion of the story. This data, like most, is no panacea. Death certificates are not the most detailed account of the tale. That’s why so many “garbage codes” existed. They are a snapshot and the way the form itself is written supports the most concise completion by the health care provider. The researchers seem aware of this limitation hence their calculated endeavor to “redistribute” these based on public accounts and so forth which enables further estimation or speculation, but not precision. Perfect is not possible. Autopsy is the most reliable methodology.
Additionally, in their attempt to be more inclusive of multiple causes of death, they did still limit the number of diseases addressed by focusing on 21. Hence, my earlier mention of a tempered enthusiasm. Because availability of data was not uniform and, in part, based on projections throughout the long time span, some was taken from the US Census Bureau (1980-89) and the rest from the National Center for Health Statistics (1990-2014). As a result, disparities in collection can present and the information be subject to error.
This JAMA study used the codes according to the International Classification of Disease, Ninth and Tenth Revisions (ICD-9, ICD-10). ICD-9 for deaths prior to 1999 and ICD-10 for those in 1999 and after. You must be aware that the former contained roughly 14,000 codes (4,000 for procedures) and the latter swelled to 68,000 (procedure version an added 87,000). So, great variability possible in how things were coded, aka deemed cause of death, which opens the data up to extensive interpretation. There can also be unintended bias in coding given the climate of awareness of certain diseases at the time of the recording (e.g. Alzheimer's vs dementia).
Turning physicians into data entry specialists, for one, is a du jour narrative that threatens the entire health care system. To learn more about the insanity of the government coding system and the unrealistic, overregulation of medical practice and how it is impacting care, review my article Government ‘Torture’ of Physicians: Where’s the Outrage? where I extensively address the subject matter. In so doing, you will better understand the inherent flaws in coding and how it impacts the utility of this study.
Nice study on the effort front, imperfect in its execution. Don’t take the data at face value, appreciating the way it is obtained and interpreted is essential to utilizing a discerning eye in its application to the real world. Regardless, the goal and objective is on the right path to approaching eradication of disease and modifiable risks and causes of death.
Casting a wide net and painting with a big brush are relics of ineffective public health policy. Personalization and population-specific measures are where health innovation begins. Lifestyle and treatment plans by region, for instance, could be tweaked to shift the pendulum toward greater health and well-being. A step in a beneficial direction.