This article is more than 5 years old. Information may no longer be current.

Novel approach determines US mortality rates at county level

A new analysis has identified variations in mortality rates and causes of death at the county level over the past 35 years, according to data from an analysis published in JAMA.

Researchers reported they used algorithms from the Global Burden of Disease study to assign garbage codes to deaths that occurred between Jan. 1, 1980 and Dec. 31, 2014. The distinct garbage code, or related groups of these codes, were given what researchers called “biological plausible target causes”.

Then, according to researchers, the deaths that got the garbage codes were redistributed to the target causes in one of several ways: according to proportions first seen among the targets; those gathered by expert opinion and/or previous existing publications; regression models that connected changes in the proportion of deaths assigned to a given target code or given garbage code; and for those deaths with codes known to be associated with HIV/AIDS, the rate of mortality in each 5-year window was compared with that in 1980, with those deaths beyond a 5% increase getting an HIV/AIDS designation and the others receiving a separate, biologically possible target.

“Previous efforts to generate country-wide county-level estimates of cause-specific mortality have generally focused on only a single cause or group of closely related causes,” Laura Dwyer-Lindgren, MPH, Institute for Health Metrics and Evaluation, Seattle, Wash., and colleagues wrote. “The approach to county-level analyses with small area models used in this study has the potential to provide novel insights into [United States] disease-specific mortality time trends and their differences across geographic regions.”

A study researcher explained three ways that physicians could use the data found to help their patients.

Some of the findings included:

• mortality rates from chronic respiratory disease were highest in Western West Virginia and eastern Kentucky;
• mortality rates from neurological disorders saw large increases in southern counties from Alabama to Oklahoma to Eastern Texas;
• mortality rates from interpersonal violence and self-harm in Southwestern counties increased;
• mortality rates were exceptionally high from mental and substance use disorders in a section of counties in both North and South Dakota, Alaska, Eastern Kentucky and Southwestern West Virginia, as well as Southwestern states with Native American reservations;
• mortality rates from diarrhea, lower respiratory and other common infectious diseases saw large increases in counties in Southern Illinois, Eastern Kentucky, Alabama, Mississippi, Arkansas and Louisiana;
• mortality rates were highest from cardiovascular disease along the Southern half of the Mississippi River; and
• mortality rates from cirrhosis and other chronic liver diseases saw large increases in Northwestern Texas and Southwestern Oregon.

Ali Mokdad, PhD, of the Institute for Health Metrics and Evaluation, University of Washington, Seattle, outlined the ways physicians can use the findings.

“One, increase awareness of the main risk factors among their patients and communities; smoking, diet, obesity, alcohol, physical activity, prenatal checkup – pregnant women or considering to [become pregnant] – etc. Two, work on early detection of diseases through screening and regular checkups, and three, follow with the patient in their care to ensure that their conditions are controlled and refer to other doctors/health facilities when needed,” Mokdad told Healio Family Medicine.

In a related editorial, Cheryl R. Clark, MD, ScD, assistant professor of medicine, Harvard Medical School and David R. Williams, PhD, MPH, of the department of social and behavioral sciences, Harvard T.H. Chan School of Public Health, wrote that though the data in the new study is “powerful”, other forms of data should be used in conjunction with Dwyer-Lindgren’s work.

“The careful analyst and policy maker should observe the limitations of these data, use participatory science to interpret these patterns, and find additional data to understand geographic trends relevant to diverse groups for the purpose of assigning priorities for further etiologic investigations and interventions,” Clark and Williams wrote. – by Janel Miller