Revamp data sources for measuring performance
Revamp data sources for measuring performance
Learn what’s available and what you don’t have
By Patrice Spath, ART
Consultant in Health Care Quality and Resource Management
Forest Grove, OR
For years, caregivers have said, "If it wasn’t written down, it wasn’t done!" With our increased emphasis on performance measurement, a slightly different principle applies: "If it wasn’t written down, it can’t be measured."
The data elements necessary for patient care activities usually were sufficient for quality management activities in the past. Data elements for many of today’s process and outcome measures are not routinely documented by caregivers, however. For example, to report the measure, "Average length of time from request for services to the first face-to-face meeting with a health care professional," two pieces of information are needed: A numerator "The total time between request for services and the first face-to-face contact with a healthcare professional for all people requesting services" and a denominator "The total number of requests for service." Data to create this performance measure may be unavailable in existing records or may not be uniformly documented.
Increasingly, the data needed to answer the following performance measurement questions are not readily available in the typical patient record or other data sources:
• Did the clinician do the right thing at the right time?
• Was effective care provided to each patient?
• Was care provided safely and in an appropriate time frame for each patient?
• Was the outcome as good as could be expected, given each patient’s condition and personal characteristics and the current state of medical science?
New information management strategies are necessary to support intensified performance measurement initiatives.
The collection of performance measurement data involves two steps. First, the data element must be documented (either in the patient record or in some other format). Second, the data element must be extracted from the source document and entered into a data file. Before embarking on a collection effort, it is important to answer the following questions:
1. What data elements are necessary to create the performance measure? (e.g., numerator and denominator)
2. Are the data elements currently documented somewhere? (e.g., in the patient record, in a computerized database, in a log or register, etc.)
3. If the data elements are not currently documented somewhere, what is the best strategy for obtaining the necessary documentation? (e.g., who will document them and what procedure will be used?)
Identify necessary data elements, sources
Start by identifying the data elements that will be required for the performance measures and the source(s) for these data elements. Map out a performance measurement profile for your organization. Using a format similar to the one shown in the table on p. 1 of the forms inserted in this issue, record the organizationwide performance measures in use and those you wish to start using. Complete a separate form for each of the Joint Commission’s major functions (or other categories that might be appropriate). Record all the measures relevant to the functional category, the data elements necessary to create the measure, and the data source. This exercise will help identify all the process and outcomes measures being used to provide a meaningful profile of the organization’s performance in that function.
This exercise also will help you identify data definition problems and data elements that are not readily available in existing data sources. The performance measurement profile on p. 1 of the inserted forms reveals that data for one of the measures is currently unavailable in any data source. Where data elements are not readily available in existing information systems (manual or computerized), supplementary data sources will need to be developed.
Don’t overlook existing information systems when searching for data sources. All too often, clinicians aren’t familiar with the data contained in the hospital’s clinical/financial information system, and they unnecessarily create new documentation tools. The table on p. 69 shows the common data elements contained in a health care organization’s information system. Many of these could be used to create performance measures without the need for additional data collection.
Develop new data sources
When the necessary data elements are already being documented by caregivers or are available in a computerized database, no new data sources need to be developed. It is likely, however, that the organization will discover that data elements necessary to create some performance measures are unavailable from any source.
In the example shown in the table on p. 1 of the inserted forms, hospital leaders felt it important to measure the amount of symptom improvement occurring for their patients. Unfortunately, the data needed for this performance measure were not reliably documented in the patient record or any other location. Therefore, a new data source had to be developed.
In this instance, the caregivers began using the PsychSentinel system, developed by researchers at the University of Connecticut and University of Washington medical schools. PsychSentinel measures patient outcome as a change in the number of symptoms that take place between initiation and the end of treatment. A multi-page form is completed by caregivers and submitted to researchers at the University of Connecticut, where the data are entered into an information system and comparative outcome reports produced.
In other instances, the data necessary for performance measures may be documented in a format that makes them difficult or labor-intensive to gather. For example, data for the performance measure, "Percentage of patients admitted with psychotic behavior who have B12, folate, & thyroid function studies completed within 24 hours of admission," are available in patient records. To gather this information, however, each patient record would need to be handled by a data abstractor. It might be more efficient to have caregivers document the needed data elements in another format, i.e., on a clinical path variance report. While the data still would need to be gathered off the variance report, the abstractor would not need to hunt through a voluminous patient record.
When caregivers move to using computerized patient record documentation systems, all the data elements necessary to the patient care experience will become automatically available for performance measurements. Until then, it may be necessary to use supplementary data sources.
Shown on p. 2 of the forms inserted in this month’s issue is an example of a form used by hospital caregivers to document circumstances surrounding a patient’s death. Prior to designing this form, the nurses had obtained necessary signatures for release of the patient’s body and autopsy authorization on a shorter version of the form. An expanded form was created to enable quality management staff to easily collect performance measurement data related to autopsies.
The nurse now indicates whether the patient’s death met criteria for autopsy consideration, if an autopsy was performed, or if the patient’s family refused the autopsy request. Because the caregiver already has personal knowledge of the details surrounding the patient’s death, these questions can be quickly answered. It would take quality management staff much longer to familiarize themselves with the case and capture the information necessary for the performance measure, "Percent of deaths meeting autopsy criteria for whom autopsy was performed or family refused autopsy authorization."
Be prudent in data-gathering efforts
Consider all available data sources before adding additional documentation burdens. Departments should not collect data in isolation from the rest of the organization. For instance, if the open heart surgery clinical path team wishes to know the number of patients who remain in the coronary care unit beyond one day post-op, they can ask the nurses to document longer stays on the path variance report form.
Perhaps it is more efficient to ask that the data be gathered for the path team by the health information management (HIM) department staff. These staff routinely code and abstract data from all records of discharged patients. Adding an additional data element to the HIM department’s abstract may be a less burdensome alternative.
Keep everyone apprised of the data elements contained in your computerized information systems. Otherwise, departments may design manual data-collection systems to capture data elements that are already available in the computer. Don’t gather data in a vacuum. Coordinate efforts to identify the most efficient method for gathering the information that is needed for today’s intensive performance measurement initiatives.
Reference
Davis DE, Fong MI. Measuring outcomes in psychiatry: An inpatient model. Jt Comm J Qual Improv 1996; 22(2):125-133.
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.