0
$\begingroup$

I work for a health insurance company and am planning on conducting a predictive modeling analysis on our claims data.

I'm thinking about publishing any interesting findings. However, since the data contains personal health information from our patients, it's unlikely the data could be made available to reviewers or interested readers.

Will this make my article submission a non-starter for journals?

$\endgroup$
  • 1
    $\begingroup$ Unfortunately, no, it's fine with most journals not to share the data. $\endgroup$ – Cliff AB Jun 19 '18 at 16:36
  • 2
    $\begingroup$ Is there a similar, public dataset that you could demonstrate your methods on? Alternatively you could consider anonymization of the data (or part of it for demonstration purposes) or at least publish the code so anyone can try on their own. While keeping the data private will probably not be an issue, doing any of the mentioned things will positively affect the real impact of your publication. $\endgroup$ – Jan Kukacka Jun 19 '18 at 16:44
  • $\begingroup$ @CliffAB I understand for the purposes of replicating results, not providing data isn't ideal. Although I'm not doing scientific research where replication is critical - rather I'd be demonstrating a useful technique in my field for predicting high cost members. $\endgroup$ – RobertF Jun 19 '18 at 16:50
  • $\begingroup$ @JanKukacka Anonymization of the data may be a possibility - remove all identifiers, recode other data (like State of residence). Certainly publishing my code would be feasible. $\endgroup$ – RobertF Jun 19 '18 at 16:53
  • $\begingroup$ @RobertF: to be clear, my earlier comment wasn't a poke at you but at journal standards. It's totally reasonable that an important lesson might be learned on proprietary data and is thus reasonable to share the lessons learned and whatever else can be shared with research community. As such, anonymization of the data is great: the researchers shouldn't care about who the individuals are, just how the methods work on the whole dataset. $\endgroup$ – Cliff AB Jun 19 '18 at 18:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.