1

I'm involved in a project where medical data is intended to be captured via an online form. There is a bunch of essentially numeric variables like BP / Haemoglobin etc. that are to be captured.

I'd have captured them as numerical values but some others on the team have opined that they want to avoid numeric data entry and instead present ranges to (a) decrease the risk of bad data entry and (b) increase the speed of data entry

e.g.

  • <50,000
  • 50,000 to 100,000
  • 100,000 to 150,000
  • greater than 150,000

etc.

Is there merit in either of these fears? Is selecting much faster than entering? Isn't there a cognitive load added in reading a number and binning it mentally?

Some of their data entry fears can be alleviated by suitable validation of numeric boxes I feel. e.g. If a test value cannot exceed 100,000 then flag it etc.

The downstream use will be analytics and fraud detection. The data will be entered by data entry operators with some training (though not huge amounts) who will be accessing paper records to feed in the right data.

Any opinions?

1 Answer 1

1

1) Entering vs Choosing - As it seems to be for good large volume of data, choosing seems better. In putting effort in typing the value vs a bit of cognitive load, later wins.

2) Now, effort can be made towards making that choosing bit easy - In the style you mentioned, like ranges - a timeline approach could be done - roughly like attached. This makes it easier to process and fit the data. System response also helps in getting speed, and moving on.

enter image description here

1
  • That would have been my take. One positive reinforcement which could lower the error rate would be to indicate the clinical norms for that patient (including age, gender, etc variables) e.g. 75%, 95% percentile
    – Jason A.
    Nov 27, 2015 at 10:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.