|
|
Click here to visit our web site |
INTRODUCTION BACKGROUND History has shown us that databases can be very valuable, or utterly useless. For many years, the AO documentation center in Bern collected data on fracture patients. When the AO tried recently to analyze this data, they discovered that they were unable to answer questions with the data they had collected, because they had not correctly predicted at the outset what questions they wanted to answer. Thus, they did not collect useful data. Conversely, BWHs Total Joint Replacement Registry has been a very successful research tool. In less than 10 years, this registry has accommodated more than 1500 requests for clinical research projects. These requests have produced 110 publications and 140 presentations at annual meetings of the American Association of Orthopaedic Surgeons. The database has made it possible for BWHs Department of Orthopaedic Surgery to become a leader in the implementation of new patient education programs, clinical pathways, and cost-efficiency programs. (1) A notable problem with existing administrative databases is that the quality of their data is often sub-optimal. Hospitals collect data on all patients, by medical record review, after discharge. These administrative databases include patient demographics and data on diagnoses, diagnostic testing, and procedures using ICD-9 codes. This data is used primarily for billing purposes and the data are usually abstracted and compiled by medical records or billing personnel. More recently, these large databases have been used for purposes other than billing, including evaluating therapies and quality assurance. (2) There is much in the literature to suggest these administrative databases lack the type of data integrity needed for sound research purposes. (3-7) Dr. Vrahas experiences in New Orleans with an orthopaedic fracture registry yielded a few lessons about data quality. The most important lesson was that those who are asked to enter the data must believe that they are benefiting from the exercise in some way. If they do not realize any benefit, they are less likely to enter the data. Secondly, the data entry process must be quick and efficient. In the clinical setting where time and resources are often limited, these are key considerations. Our goal was to create a database that would address these issues. We first decided that we should collect a very focused set of data. From a research point of view, this would allow us to sort and identify groups of patients more easily. Second, we wanted a limited and straightforward data entry requirement for residents. Attending physicians would be expected to check the quality of this data, and would also be given a focused data set to enter. Finally, we designed elements into the system that provide value to the person entering the data - thus encouraging him/her to enter the data. Ultimately, we believe we have created a very workable application. OVERALL STRUCTURE The application uses the patients medical record number as the key identifier, and links to PCIS and BICS, the hospitalbased computerized clinical information systems, allowing for transfer of basic demographic information. Ortho DUDEs main data points are injury classification, initial management, initial follow-up, definitive management, rehabilitation, and complications. Data is selected from drop-down boxes, radio buttons, pop-up selections, and list boxes. Injuries are classified using the AO fracture and soft tissue classification systems. While most of the DUDEs data collection relates to the early stages of a patients injury, there is also a utility that records complications. This allows us to track the success rates of various techniques. FUNCTIONAL FLOW
The resident is then presented with the initial management screen where s/he clicks on whether or not a reduction was performed, the type of immobilization used, weight bearing status, and whether or not the patient was admitted (Figure 3). If the patient is admitted, the residents data entry ends. If the patient is being discharged to home, the resident enters follow-up information such as when to follow up, with which service. This final exercise will produce a letter for the patient that contains information about the injury, its care, symptoms to look for, and clinic follow-up information. The next business day, administrative staff will generate a list of patients seen the previous day in the ED for whom they must schedule appointments.
At the time of a patients definitive management, the attending surgeon records the type of injury management technique and its corresponding CPT code (Figure 4). He/she may also further classify the fracture, and enter additional information specific to the surgery about the approach and type of fixation. Finally, the surgeon selects the appropriate rehabilitation instructions (weight bearing and range of motion limits). The surgeons secretary will prepare billing documentation for surgeries performed based upon the coding entered at the time of definitive management. If a patient has a complication, the surgeon can record the type of complication, its treatment, and its CPT code. This complication will be linked to the patients initial presentation history. Ortho DUDE also records information about patients who initially present with a complication.
VALUE ADDED VALUE TO ATTENDINGS VALUE TO RESIDENTS VALUE TO PATIENTS THE FUTURE ACKNOWLEDGEMENT Suzanne M. Morrison, MPH is Program Coordinator, Partners Orthopaedic Trauma Services, Brigham and Womens Hospital, Massachusetts General Hospital. Mark Vrahas, MD is Partners Chief of Orthopaedic Trauma Services, Brigham and Womens Hospital, Massachusetts General Hospital, Assistant Professor, Harvard Medical School. Address correspondence to: |
Print Manuscript View References Download PDF version Close window |