Message:

<p>[QUOTE=&quot;beef1020, post: 2805532, member: 24544&quot;]Getting a bit off topic, but a couple of you may be interested, and it happens to be an area I am familiar with.&nbsp; When talking about and using ‘Big Data’ it’s important to keep in mind that it has very large potential benefits while also being easily abused and misused.&nbsp; I don’t think most people appreciate this conceptually, but where it’s most valuable is in generating more accurate estimates as opposed to finding subtle new causations/correlations.</p><p><br /></p><p>For example, if we know that some behavior increases your risk of cancer but the confidence bounds of our estimate place that increased risk between 120% and 600% then we may not be able to act on that information.&nbsp; Along comes big data, and now instead of a study with 10,000 people we have one with 2.6 million people, we can use it to lower our confidence bounds and get an estimate of 530% to 560% increase risk.&nbsp; Programs that improve health at an increased risk above 400% can now be implemented whereas before we couldn’t be sure of their value.</p><p><br /></p><p>Again, the premise here is that big data increases our fidelity, which is often valuable, but not revolutionary or game changing.&nbsp; The size of the signal you can find is related to the amount of noise in your data, lots of noise means you can only find big signals, less noise allows you to see small signals.&nbsp; As we move along to lower and lower noise levels, big data is giving us insights into smaller and smaller signals, however these smaller signals are also less valuable.&nbsp; In the field of medical research I work, these small signals we can now find are just not that important.&nbsp; We are looking for what increases risk by a factor of 2 to 3, or higher, i.e. smoking for basically all cancers, BRCA for breast cancer, HPV status for cervical cancer.&nbsp; Now we can see things like slight decrease in risk for women who breast feed, but the risk reduction is 4.3%.&nbsp; How do you develop a public health program that incorporates that risk reduction.&nbsp; I believe this may be the tie in for [USER=71723]@V. Kurt Bellman[/USER], Big data in and of itself is not bad, unfortunately people have shown a history of misusing whatever tools they have at their disposal.[/QUOTE]</p><p><br /></p>

Your name or email address:

Do you already have an account?

• No, create an account now.
• Yes, my password is:
• Forgot your password?

Stay logged in