Begin with the end in mind: getting great data out of your system

Begin with the end in mind: getting great data out of your system – part 2

Lisa Anne Bove

In the first part of this article, Lisa Bove described a process to determine “who needed the ‘data’, what data was needed, and how that data should be organized.” Determining whether the system is supporting patient care is depended on the usefulness of the data collected, not just the efficiency and accuracy of the data itself. Begin with identifying the users of the data, the regulatory requirements for the data, the reimbursement, staffing and supply issues based on the data terms. Use a committee to decide how best to group and name terms, as well as define policies and procedures for the use of the data. These procedures will allow the DATA to be the focus and move you faster toward your reasons for implementing the system. Remember the most important thing when working with the users–don’t just replace the paper chart with a computer–make the computer do the work of data collection and aggregation, so the user can focus on patient care.–Ed.

After you focus on the users of the data and the users of the system, focus next on the other systems that are impacted by the data. These systems can be ancillary systems like Laboratory or Radiology. They can be general financial reporting systems used to report overall hospital statistics, or inventory or staffing systems. For example, many institutions interface Order Entry Systems to Laboratory systems. Important questions to ask during analysis include which system does the billing? Will users enter or review data from more than one system? How does hospital reporting (case management, finance, and overall quality assurance) get their data? Which system do the physicians use? Does the patient see data (other than the bill) from any of these systems?

All of these questions will help to determine how to make the data similar, if not the same on both systems. From a database viewpoint, it’s better to use one name for one term. The key here is standardization of the DATA. From a user perspective, that can be difficult. Often the person writing the order does not call an item the same as what will show up on the patient bill, or in the inventory systems. How will you reconcile this? Interfaces make this happen (and can make it both harder and easier). Many systems allow for alternate names or short and long descriptions in their files. Use these when you can, but use them sparingly. Again, focus on the data–does it matter if it’s called CBC or complete blood count? If the physician doing the order entry wants to order a CBC (a complete blood count), for example let the order entry system show a CBC for the ‘order name’ to simplify. Have the billing system print complete blood count (for the same order) on the patient bill.

Last, focus on the system requirements for the data. Often, this is the first thing that analysts do–determine how the data is collected by a particular system. Doing this last–after you have determined the user’s data needs–helps to give your users more useful data. For example, does the system have flowhseet charting with multiple or single choices? Can users default previous choices or do they need to re-key information for each assessment? Does the system work more like longhand documentation or flowsheet documentation? By answering these questions last, the data remains the focus. Users will have identified what they need and then you can determine the easiest way to have them do the data entry based on requirements. This way you can avoid problems due to software constraints like length of field, GUI vs character cell, and/or differing types of data.

Figuring out how to get what you want out of the data so you can build it right is the biggest challenge for clinical analysts. By beginning with the end in mind; that is, knowing what you want to get out of the data before deciding what you are going to collect, etc. you can create a system that gives you great data. Always think about the user’s needs, but use the database as your first qualifier and the system will be well received.

Lisa Anne Bove, MSN, RN,C

COPYRIGHT 2001 Capital Area Roundtable on Informatics in Nursing

COPYRIGHT 2008 Gale, Cengage Learning