How to manage clinical path variance data
How to manage clinical path variance data
Identifying correctable system breakdowns
By Patrice Spath, RHIT
Brown-Spath & Associates
Forest Grove, OR
The purpose of collecting clinical path variance data is to identify correctable system breakdowns that cause patients to vary from the path, test the validity of the path recommendations, and correlate the processes of care with patient outcomes. However, hospitals involved in clinical path initiatives commonly find themselves with mountains of path variance data they don’t know what to do with. Case management departments can use simple data management techniques to overcome information overload.
Up-front planning
First, limit the number of variables that people are asked to collect manually. Go to the patient record and determine whether the data are currently available there. Don’t ask people to document information on a separate form when they are already writing the data in the patient record. Rather than asking caregivers to document the same information in two places, you can usually extract the information from the record more efficiently after the patient is discharged.
Thinking ahead of time about what information is really necessary will also help to ensure the variables being gathered will be sufficient to answer your study question. It is expensive to add variables to the data set that may have been overlooked in the original study design. The study question should be the principal determinant of variable selection. For example, if caregivers wish to know whether or not patient satisfaction has improved as a result of a process change, then a focused survey may be the best way to gather that information. Gathering "wouldn’t it be nice to know" data is labor-intensive, and, in many cases, a waste of time.
The variables needed to analyze clinical path variances may be contained in many data sources. Don’t ask people to manually gather data that are already available in another information system. For instance, the number and type of diagnostic tests performed for a patient are data that can usually be found in the financial database. Information that comes from more than one source of data can be merged into a single record for each observation.
Pick one variable as focus for data merge
The method for merging data from various sources will vary depending upon the analyses being performed. If you are conducting a study at the individual patient level, a unique individual identifier (for example, medical record number or Social Security number) is the best choice for a merging variable. If one is analyzing information about providers (for example, physicians or facilities), then the identifier for the provider may be the best merging variable.
If unique identifiers are not available, merging data sources can be time-consuming and expensive. For example, one might use the patient’s Social Security number to link records from the financial system with information stored in the case management system. Such links require an exact match, and any data input errors in either system will result in a failure to match. Unless great care is taken in linking files and merging data sets, biases of unknown magnitude or direction may be introduced into the conclusions.
After collecting the data for a month, organize the data elements into a data file (sometimes called a data set, or, if computerized, a database). A data file arranges the information like a spreadsheet. Each column represents a different variable, and each line represents a different case, or vice versa. An individual number in a data set represents the value of a particular variable for a particular case. Ideally, the data file is designed prior to variance data collection so the path variance report form can be arranged to expedite the efficiency and timeliness of the data entry process.
The data file is built by copying the information from various data sources onto a template you’ve created on a piece of paper (or into a computer database) using one line per case. Once all the spaces on the form are filled out, the data file is automatically created. A portion of a data file from a study of patients’ use of postoperative analgesics is shown in the table accompanying this article. (See table, above.) Caregivers wanted to determine if patients used less pain medication if they are taught simple relaxation techniques preoperatively.
The variables collected during the study are listed across the top according to the order in which they appeared on the data collection form. The results for each case are added, one line per case. You may choose to input the case data in no particular order if entering into a computerized database where sorting is easily accomplished. If building a manual data file, you may wish to arrange the information by case number or the answer to a specific variable (for example, "Y" for patient taught relaxation techniques).
Once the information is entered into the database, it is easier to perform quality checks. Check for missing data and identify why they are missing and if corrections are required. Check for measurement errors caused by poorly worded study questions or misinterpretations of data collection instructions. Try to identify possible sources of bias that may affect the study results. Now is the perfect time to clean up the data before summarizing the results. If the quality checks reveal problems with the data, that will have implications for the results.
Consider automating your data files
When the data are organized in a data file, whether on paper or in a computer file, it’s much easier to view the relationship between independent and dependent variables. The data file should facilitate ease of entry, storage, editing, and preparation of the data for analysis. The design should allow for querying and sorting, should interface with all potential users, and should be flexible enough to respond to data field modifications, patient types, measurement intervals, and new studies that evolve as a result of your data analysis. For this reason, consider automating the data file using statistical software or spreadsheet programs.
Path variance and patient outcome data can be a valuable source of information for clinical process improvement initiatives. With up-front planning and application of common data management techniques, collection and analysis of these data can become both more efficient and more worthwhile.
Editor’s note: Patrice Spath, author of this column, has written a new 135-page book titled How to Measure and Improve Case Management Performance, published by Brown-Spath and Associates. Cost: $35. To order, call (503) 357-9185, or order on the Internet at www.brownspath.com.
Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.