Do blank rows of data in data sets released by HHP mean value unknown or value is zero? How do we know data inputter did not forget to put in values?

What is the average error rate for a data inputter-how often do they get data wrong ? How often do they input a wrong number or word string?

Do doctors input data or staff employed by them or both?

 

 

 

Empty cells could be caused by many things: unknown, typo, missed, supressed, etc. I certainly wouldn't assume it's zero.

I did a quick search through some papers I've collected regarding data quality in health care (I'm sure there are plenty more):

A systematic review of discharge coding accuracy (2001) by Campbell, Campbell, Grimshaw, Walker;

Accuracy of data in computer-based patient records (1997) by Hogan, Wagner;

Assessing accuracy of diagnosis-type indicators for flagging complications in administrative data (2004) by Quan, Parsons, Ghali;

Systematic review of scope and quality of electronic patient record data in primary care (2003) by Thiru, Hassey, Sullivan.

It's primarily billing data. You can assume a pretty high level of accuracy on the data points REQUIRED in order to get paid correctly/maximally.

Other data points will be of varying quality.

The easy assumption is that whatever the imperfection level of the data, it's more or less consistent throughout the covered period. On the other hand, maybe it's improving. Certainly there's been greatly increased attention in recent years to deriving quality and process measures from this kind of data, and thus some incentives to improve data beyond the base minimum necessities.