40% drop in BC deaths, 1989-2015, a question

Momine

http://time.com/4967148/breast-cancer-death-rates/

I have now read several articles about this study (but not the study itself). I am not understanding how they calculated this to reach the conclusion that the death rate has dropped 40%. Anyone?

SummerAngel

The actual study is here: http://onlinelibrary.wiley.com/doi/10.3322/caac.21...

From the study: "The estimated number of female breast cancer deaths averted because of the reduction in breast cancer death rates was calculated by first estimating the number of cancer deaths that would have occurred if the death rate had remained at its 1989 level. The expected number of deaths was estimated by applying the 5-year age-specific cancer death rates in 1989 to the corresponding age-specific female populations from 1990 through 2015. The total number of breast cancer deaths averted was the sum of the difference between the expected number and recorded cancer deaths in each age group and calendar year."

Momine

Summer Angel, thanks, so am I right to read that as "death rate" meaning "dead within the first 5 years after DX"?

SummerAngel

No, the death rate was from the SEER database. The death rate statistics in the SEER database are taken from the US Mortality Files from the National Center for Health Statistics, CDC, not adjusted by year of diagnosis.

https://seer.cancer.gov/csr/1975_2014/browse_csr.p...

So, in 1989, the age-adjusted death rate per 100,000 individuals was 33.23. In 2014, it was 20.55. This is very good news!

Momine

Then why does it say: "The expected number of deaths was estimated by applying the 5-year age-specific cancer death rates in 1989"?

It had been my understanding that the SEER data only tracks 5-yr survival, not beyond.

SummerAngel

The "5-year" refers to 5-year age grouping, not tracking specific individuals for 5 years.

From the study: "Mortality data were obtained from the SEER program's SEER*Stat database, as provided by the National Center for Health Statistics."

National Center for Health Statistics mortality data: https://www.cdc.gov/nchs/nvss/deaths.htm

The mortality reports used in the study tracked 5-year groups, which is why the data was "age-adjusted", meaning that the data was adjusted as if the population had the same age distribution as the "standard" population. They are summary measures adjusted for differences in the 5-year age distribution.

The mortality data listed on the site above is collected by the states following the regulations stated here (page 7, which describes how the death is documented and by whom, along with the cause of death), just fyi: https://www.cdc.gov/nchs/data/misc/mvsact92aacc.pd...

Momine

Thanks very much, and sorry for being so dense.

SummerAngel

No problem, the wording in studies can be very confusing! I'm just used to reading them, I've been interested in medical studies for many years.

TectonicShift

.

ErenTo

Thanks for posting this, interesting.

What was the U.S. population decades ago? If the mortality rate stayed the same and population went up then the percentage has dropped (assuming incidence didn't go down). It's still 40,000 too many, but improvement in treatment outcomes should be acknowledged as well. Yes, it's incremental and way too slow.