Ongoing CCNR Projects
The CCNR is Galvanic
Nurse leaders in health systems, operations, and informatics are working within the CU Patient Initiated Data system.
The objective is to extract various data from different EHR systems and then translate those data into a uniform, CU Patient Initiated Data nomenclature.
This will allow the CCNR to isolate/identify patient outcomes from EHR fields, EHR drop-down menus, and EHR narrative entries and then return the data to facilities for real-time decision-making.
CCNR members agreed from day 1 that no new data would be collected.
This, therefore, is our charge: conduct research and evidence-based projects using data already captured in the EHR.
Three ongoing CCNR proof-of-concept projects address, respectively, each data-structure type found in the EHR—structured, semi-structured, and unstructured.
First Proof of Concept (Structured Data): Cost & Quality Project
How much does nursing care cost? Actually, we know very little about patient-level nursing costs in hospitals or across the spectrum of healthcare. The goal of this proof of concept is to measure direct nursing-care costs per patient, beginning in the acute-care setting. The study is framed by the emerging abilities (a) to identify and link nurses to each patient and (b) derive within the EHR data actual direct-care time and associated costs for each patient. Data are captured within the assignment of nurses to patients from the electronic nurse scheduling and staffing modules. When these data are combined with other quality, operational, finance, and outcomes data, they can provide a robust and near real-time information environment for nurses, managers, and nursing leaders to improve decision-making (Caspers & Pickard, 2013; Pickard & Warner, 2007). This approach overcomes weaknesses in traditional cost-accounting practices (e.g., nursing hours and costs per patient day) that average nursing time and costs across many nurses and many patients.
Second Proof of Concept (Semi-structured Data):30-day Readmission Project
The CCNR is looking at readmission rates to hospitals within 30 days of discharge. Recent data show an average cost of $7,400 per readmission (Friedman & Basu, 2004), much of which, for Medicare patients, will not be reimbursed to hospitals. But while nurse staffing has been linked to readmission, there are few data on whether nursing quality in hospital is a major contributing factor. For example, are nursing interventions such as patient education, discharge planning, and pain assessment and management associated with hospital 30-day readmission rates?
While patient demographics are found in structured fields within the EHR, data points such as “education on discharge,” “pain management,” and “discharge planning” are in semi-structured fields. This study negotiates both data formats in a cross-sectional, retrospective observational study across several hospital systems. The CCNR 30-day Readmission Study will help inform nursing practice and identify potential additional predictors of hospital readmission. In addition, the study aligns with CCNR efforts to build collegiality among participating hospitals and healthcare systems.
Third Proof of Concept (Unstructured Data): Symptom Status/Trajectory Project
Existing research establishes the need to improve processes for symptom assessment documentation (Fan, Filipczak, & Chow, 2007). Furthermore, Fan et al. (2007) suggest that documenting assessments of symptom clusters may reveal more reliable information about an individual’s health status than documenting assessments of single-symptoms (such as pain). The CCNR Symptom Status proof of concept is currently assessing the type and level of symptom data documented in the EHR. Clinical colleagues in the CCNR record much of their symptom data in nurse narrative notes rather than specified fields (M. B. Makic, personal communication, May 20, 2014). Because so many of the symptom-related data are entered as unstructured data in EHR narratives, the CCNR is collaborating with natural language processing (NLP) consultants in developing a data-extraction algorithm to seek and find words/constructs within the EHR.