Catherine H. Ivory, BSN, RNC Finding Buried Treasure in Unit Log Books Data Mining Can unit log books help nurses use evidence in their practice? In a 2001 article, Youngblut and Brooten stated, Evidence-based practice provides for opportunities for nursing care to be more individualized, more effective, streamlined and dynamic, and to maximize effects of clinical judgment. As nurses, we are always looking for ways to provide better care for our patients. We d like better birth outcomes, more satisfying birth experiences, fewer operative interventions and fewer days in the NICU and better strategies for effective patient education. However, management at our institutions is pressured and demands that the nurse at the bedside do more with less, be more efficient and be a good steward of the health care dollar. Staff nurses voice dissatisfaction with the shrinking amount of time they have available to spend at the bedside and the increasing length of time they are required to spend documenting. 62 AWHONN Lifelines Volume 9 Issue 1
for Gold Additionally, the women we serve are increasingly well-read. They are technologically savvy, and most know where to go online to look for the latest trends and research related to their own health care needs or those of a family member. As the nurses caring for them, it s in our best interest, indeed in the best interest of the profession, to do our part to promote and contribute to evidence-based practice. Nurses contribute to this process daily by recording data about our patients. It s the ease or difficulty in accessing and utilizing these data that can challenge the quest to find the buried treasure needed to advance evidence-based practice. Historically, most data have been collected with the use of a hand-written log book kept on a nursing unit. For example, on a labor and delivery unit, admission and delivery times, method of delivery, anesthesia use and complications are common entries. Nurses in mother-baby units make data entries such as feeding method, circumcision, hearing screening, metabolic testing and others. February March 2005 AWHONN Lifelines 63
The process of extracting, sorting and analyzing these data is called data mining. Rob and Coronel describe data mining as a methodology designed to perform knowledge discovery expeditions over the... data (2004, p. 598). When done by hand, this process is cumbersome and time-consuming. Today, with the increased use of technology, many facilities have added electronic data collection methods. Perinatal software products and computerized documentation systems have built-in database capabilities, making it much easier to extract needed statistics and answer research questions. Table 1 lists perinatal systems companies that exhibited at the 2004 AWHONN conference. Data collection products are costly; many facilities have been slow to implement them. Therefore, hand-written data collection is still a very real part of everyday practice. This method presents significant challenges and requires that nurses spend huge amounts of time accurately extracting and making good use of the compiled data. When the time comes to actually use it, the nurse can truly think she is mining for gold. As graduate students in health care informatics at Georgia College and State University, a colleague and I were assigned the project of analyzing the current system of data collection and suggesting a computer-based solution. Immediately, I thought of the log book kept on the unit at my practice setting in suburban Atlanta, GA. The unit s a 12-bed High Risk Pregnancy Unit (HRPU) at a large not-for-profit tertiary care center. HRPU admits pregnant and postpartum patients with complications. The unit opened in April 2001 as part of the newly constructed women s pavilion expansion. In an effort to keep track of the physicians and group practices admitting patients to HRPU, the unit manager designed a simple log book. (Something similar may be used at your institution.) A sample of the log book page is included in Table 2. This was important for staff nurses as well. Since the unit was brand new, it was important to know that physicians felt comfortable with the unit and its nurses and to identify any who did not. The log book also tracked important diagnoses, length of stay and patient acuity, since staffing was acuity based. Finally, the log book included an entry for patient disposition because we wanted to know how long our patients were staying compared to other similar units. Unfortunately, in the 3 1/2 years the HRPU had been open, recording these data was all that the unit had ever done with this information. Using Microsoft Access, a database application, we created a table that contained all the categories present in the log book. Database applications refer to these categories as fields. Since my project partner was not affiliated with our health system, and to protect patient privacy, we eliminated patient names and used only medical record numbers as identifiers. In all, we entered more than 4,000 records into this table. During this process we began to see both the problems with the handwritten log and the buried treasure inside. Our biggest challenge turned out to be the way diagnoses were entered into the log. It became clear that the same types of diagnoses were listed many different ways. For example, a patient with high blood pressure might be admitted with diagnoses of increased BP, PIH, pre-eclampsia, severe pre-eclampsia or R/O any of these. A diagnosis related to bleeding might be categorized as bleeding, partial abruption, placenta previa or hemorrhage. Pre-term labor was diagnosed as PTL, short cervix, incompetent cervix or contractions. And the list went on. Ultimately, it became clear that if we ever hoped to actually use this information for research or even accurate statistics, we must somehow standardize the entries. Some patients had several diagnoses, and we learned that we must be able to list them in order of primary, secondary and subsequent. Sometimes entries were made more than once, such as when a patient became unstable, was transferred and then was returned to HRPU. In other cases, entries were left out all together. Other problems included illegible handwriting, Table 1. Examples of Perinatal Documentation Systems Clinical Computer Systems, Inc.: www.obix.net E&C Medical Intelligence, Inc.: www.e-and-c.com GE Healthcare: www.gehealthcare.com Lifecare Technologies, Inc.: www.lifecaretech.com Philips Medical Systems: www.medical.philips.com Spacelabs Medical, Inc.: www.spacelabs.com Catherine H. Ivory, BSN, RNC, is a graduate student in health care informatics at Georgia College & State University in Milledgeville, GA. DOI: 10.1177/1091592305275182 64 AWHONN Lifelines Volume 9 Issue 1
Table 2. Sample Log Book ADMIT ROOM # MR# ACCT # PATIENT NAME MD DIAGNOSIS ACUITY MgSO4 D/C DATE & TIME STATUS PHONE NUMBER SCORE Y/N DATE A / P A / P P / HR / OBV P / HR / OBV incomplete entries and missing physicians who had either left the system or were now with another group practice. We even found some information we were recording, such as whether the patient was on magnesium sulfate therapy, was information we didn t need at all. As a staff nurse working on this unit, I had been making entries into the log book myself (without a unit secretary, each staff member makes her own entries) and resented the extra time this data collection took. If it didn t get done, it was not high on my priority list. After all, it was never used. Now, as an investigator, I began to see that the data we recorded were indeed important. Effective use of the data we collect can provide the basis to learn more about our patients and to determine whether we nurses provide the best, evidence-based care. Implications for Practice When the original log book was designed, the unit manager identified several ways the data might be used. We discovered others in the subsequent three years of unit operation: 1. Keeping a unit-specific record of daily census, daily admissions and discharges. Doing so helps identify trends, staffing requirements and budgeting. Nurses want to work on a unit that is adequately staffed. Presenting hard data to those making hiring decisions may provide evidence of the need to hire additional staff nurses or change the number scheduled for any given shift 2. Tracking patient admission status. At this facility, billing is different for observation patients than it is for inpatients. Yet accurately tracking this information has been difficult because patients may be admitted or discharged from another area in the hospital with a change in status somewhere along the way. If accurate and unit-specific records were kept as to when an observation patient is converted to an inpatient, billing could be more accurate. Billing accuracy potentially leads to increased revenue, which could then lead to more nursing dollars available to the unit. In some institutions, nursing salary increases and bonuses may be linked to revenue. Ideally, a computerbased data collection system would interface with the hospital s admission, discharge and transfer (ADT) system 3.Optimal staff scheduling. HRPU staff is scheduled for each shift based on patient acuity. Analyzing the data collected could identify trends related to acuity. For example, the unit s charge nurse might choose to schedule another staff nurse for a particular day, even though unit census seems to indicate one less nurse would be adequate. Hard data about the actual patient acuity for that shift gives the charge nurse the tools needed to justify this decision 4. Tracking data collection related to diagnoses electronically. If accurate and consistent data are recorded regarding HRPU patient diagnoses, unit staff and the service-line CNS can more easily conduct research studies that contribute to evidence-based nursing knowledge, particularly regarding pre-term labor, pregnant patients with underlying diseases and cardiac complications. All of these diagnoses are common on HRPU 5. Tracking patient length of stay on HRPU. It s not unusual for one particular patient to be admitted and discharged from HRPU more than once, such as a patient who becomes unstable, is transferred to labor and delivery and then returns. Benchmark data, which compares HRPU with units of similar size and demographics, indicates that the average length of stay on HRPU is longer than in other facilities. Complicating the situation further is the fact that by the time some patients are discharged and billed, they have delivered babies. Therefore, from an accounting standpoint, the patient s primary diagnosis relates specifically to the delivery instead of the diagnosis for which she was admitted to HRPU. The data collected directly on the unit could certainly assist in the identification of trends and make it easier to compare this unit with others. When the staff nurse assists with this data collection and analysis, the nurse helps justify the unit s February March 2005 AWHONN Lifelines 65
value to the facility in which it resides. She also helps to justify its value to the larger community, potentially increasing the census. Keeping unit census up helps to ensure that the staff nurse will work as scheduled and decreases the likelihood that the nurse will be floated to another, busier unit 6. Collecting more viable data related to nursing functions. In her 2004 article in JOGNN, McCartney noted, Historically, nursing care has been invisible in the paper patient record. While our particular log book does not directly address nursing care activities, nursing can identify tasks related to those diagnoses and nursing time related to patient acuity by accurately tracking diagnoses and patient acuity. Tracking the tasks and nursing time, and later analyzing the information, validates the nurse s role and value to the patient and the institution In a 2004 report, Future Nursing Scan: Next Generation Patient Care Priorities, the Nursing Executive Center found that health care information technology spending in 2004 is expected to be more than $25 billion (Nursing Executive Center, 2004). In his 2004 State of the Union address, President Bush mentioned increased use of technology as a key to resolving the country s health care crisis. Yet, as nurses caring for women and newborns, we can sometimes think of technology as a nuisance rather than a valuable practice tool. However, I am convinced that the use of technology for data collection and analysis can help validate the job we do everyday and help us do it better. All that data can truly be a gold mine for nursing. Come along on the treasure hunt. References McCartney, P. R. (2004). Leadership in nursing informatics. Journal of Obstetric, Gynecologic, and Neonatal Nursing, 33(3), 371-380. Nursing Executive Center. (2004). Future nursing scan: Next generation patient care priorities. Washington, DC: The Advisory Board Company. Rob, P., & Coronel, C. (2004). The data warehouse. Database systems: Design, implementation & management (p. 598). Boston: Thomson Technology. Youngblut, J. M., & Brooten, D. (2001). Evidence based nursing: Why is it important? AACN Clinical Issues: Advanced Practice in Acute and Critical Care, 12(4), 468-476. 66 AWHONN Lifelines Volume 9 Issue 1