Input Problems

Users may Develop a Dummy Value System: To get around pesky mandatory field requirements, users may develop techniques for entering junk. If users are forced to assign an employer for some patients, for example, the users may agree to use "Acme Construction" whenever the users do not have the employer's actual name. At some point, someone will look at the reports without understanding the data history and let loose a witch hunt against Acme Construction for unsafe working conditions. This seems too stupid to occur in real life, but there are rumors of such idiocy.

Basic Typos: A lot of information is simply entered incorrectly because paper forms are misread and values are misheard. Typing mistakes are a problem as well. Numeric lab results may be astronomical and Medicare ID values junk. The people doing data input are often overloaded and working in noisy environments. We should expect input mistakes.

Note that a few extreme outliers in the numeric data, due to typos, can cause large errors in statistics. If you review data provided by a client, you will inevitably find values which are completely unrealistic.

Users do not Enter and Maintain Data: Already healthcare providers spend much of their time feeding data into an EHR. The value of even basic data is doubtful, and providers are often rushed. It should not be a surprise that organization administrators overestimate the amount and quality of the data that users feed into the EHR. This is especially true if data is not needed for basic care, only special projects.

Unclear Error Reporting: Many users, especially managers, will not give useful descriptions of suspected bugs. To confirm a problem, technical people typically need an MRN, the location in the software, and the expected output/behavior.

• Helpful: The January 8, 2022, weight in pounds for the patient with MRN 5512390 appears as 82.6 in the "Last Visit" roster, but it should be 182.
• Useless: THE VITALS ARE BROKEN!

Perseverance is required to get suspected problem details from some people.

Visibility of Input Errors: Often we just assume that the input values are correct, and we are unaware of errors. A task such as merging patient data from two sources will show errors, though. We will find, for example, that a patient from hospital A is the same as patient from hospital B, even though data does not match exactly. A birth date, name, or Medicare ID is incorrect. Name suffixes, such as "Jr and Sr," as well as name changes for marriage and nicknames, all contribute to differences.