Tuesday, May 08, 2018

All this statistics gymnastics, do they really give the true picture ?

Diabetes-related amputations: How we crunched the data
by Brandon Quester | September 20, 2017
TwitterFacebookEmailPrintPocketMore
To identify the increase in diabetes-related amputations in California, inewsource analyzed data from the California Office of Statewide Health Planning and Development, the agency that collects information about the care provided to patients discharged from California hospitals.
The data, compiled at the request of inewsource, includes more than 100 separate billing codes, by body part and procedure, detailing patients treated within hospitals across the state. The data represents the number of amputation procedures, not the number of patients treated.
http://data.inewsource.org/interactives/diabetes-related-amputations-2010-2016/
Explore the rate of amputations in California.
Because the number of diabetes patients is unknown at the state or county level, inewsource calculated the rate of occurrence for diabetes-related amputations as the number of amputations per 100,000 residents, by county and year. That calculation used population estimates from the California Department of Finance to account for population growth over time.
By calculating amputations as a rate of occurrence within the state’s overall population, by county and year, inewsource was able to compare and rank counties across California. Some counties had fewer than 11 diabetes-related amputations and were redacted for privacy concerns, so inewsource removed those from the statewide analysis.
In addition, inewsource removed counties that reported fewer than 20 diabetes-related amputations in either 2010 or 2016 because numbers that small are considered unreliable.

 if you remove the counties which showed low amputation numbers, then the rest of the counties will most likely show a higher number of amputations .n but does this mean these numbers of incresed amputations reliable? I am not an expert in statistics but somehow common sense tells me  there is something wrong with this number-crunching .

No comments: