While the Centers for Disease Control and Prevention (CDC) attributed only three percent of recent deaths to dementia, the condition contributed to 13 percent of deaths among the population the researchers studied.

"We were actually surprised at what we found," Dr. Mary Tinetti, who led the study at Yale University, told Reuters Health. "Dementia was the second most common contributor to death, which is a much higher proportion than when only a single cause of death is considered."

When doctors in the U.S. fill out death certificates, they follow World Health Organization guidelines to decide how a person died, and then record just one underlying reason for death.

That may make sense for young people who die, said Tinetti, but attributing death to just a single disease isn't accurate for older adults.

"The vast majority of older adults have two, three, four or even five chronic diseases," said Tinetti, whose team published their report in the Journal of the American Geriatrics Society.

For the new study, Tinetti and her colleagues looked at treatment claims submitted for 22,890 Medicare beneficiaries and found 97 chronic diseases and acute problems.

During the 41-month study, 2,245 people died; two-thirds of those were over 80 years old.

When researchers looked at Medicare claims for those people who died, they found that fifteen diseases had contributed to 70 percent of deaths. Heart failure was involved in 20 percent of deaths, making it the most common contributor. Dementia contributed to 13 percent of deaths.

Lung diseases, such as bronchitis, asthma and emphysema, were found in 12.4 percent of deaths and pneumonia in 5.3 percent of deaths. Cancer contributed to just 5.6 percent of deaths.

Mortality data from death certificates are published by the CDC and widely used to guide health policy and healthcare research.

Tinetti and her colleagues found that the main contributors to death in older adults are similar to those listed by the CDC, but there were important exceptions: the CDC attributed only three percent of deaths to dementia in 2004, while cancer accounted for 22 percent of deaths.

Diabetes, the sixth most important cause of death on the CDC list, was absent from the Yale team's list. In their report, the researchers say that's most likely due to the close association between diabetes and other conditions like heart and kidney diseases.

Researchers couldn't get access to death certificates for the people who died during the study, so the official causes of death couldn't be compared to the claims-based findings.

A MORE ACCURATE APPROACH?

In their report, the researchers point to a need for a more accurate method of measuring how different diseases contribute to death to establish public health, clinical and research priorities for an aging population.

But Tinetti thinks that change is unlikely to come any time soon.

"It would probably take a lot of time and lot of people to decide that the current approach doesn't work and that we need to change it. But at least this starts people thinking that we need to do this differently," she said.

Dr. James Burke, who has studied multiple causes of death at the University of Michigan, agreed.

"It would require a huge amount of resources to use that data more meaningfully, and it would depend on the investment society wants to make," Burke told Reuters Health.

SOURCE: bit.ly/OpqZ0u Journal of the American Geriatrics Society, online June 26, 2012.

• A Permanent Economic Emergency