Evaluating the Effectiveness of Problem Based Charting in Utilizing the Problem List to Structure Clinical Data in the Intensive Care Unit: A Retrospective Time Series Study

Background:

Accurate problem lists linked to electronic medical record (EMR) clinical data can be useful for chart review by clinicians for patient care, clinical decision support design, and data extraction for clinical research.  The EMR at Stanford Hospital allows the problem list to be modified in several ways: adding and deleting problems, and designating problems as active or resolved.  Nevertheless, modifying the EMR problem list is disruptive to clinician workflow under traditional charting methods, resulting in incomplete, inaccurate, and outdated, problem lists that do not reliably structure clinical data.  In 11/2013, Stanford Hospital’s intensive care unit (ICU) implemented problem based charting (PBC) a system of clinical documentation that uses the problem list as an anchor for physician notes.  PBC requires clinicians to document and update assessment and plans, instead of in a single note, under individual problems in the EMR problem list, therefore incorporating problem list management into the clinician’s workflow.  We report preliminary findings from our study that aims to evaluate the effectiveness of PBC in improving problem list utilization.  

Methods:

We conducted a retrospective time series using data from a six consecutive month period prior to PBC implementation in the ICU (pre-PBC) from 5/1/2013 to 11/22/2013 and after PBC implementation (post-PBC) from 11/23/2013 to 5/1/2014.  Patient data were queried and extracted from the EMR using SQL.  Frequencies of total, active, resolved, and deleted problems on the problem list were ascertained per patient encounter.  Statistical comparisons were performed between pre-PBC and post-PBC using chi2, rank sum, and t-tests with Stata software.  

Results:

A total of 735 and 758 patient encounters were identified in the pre-PBC and post-PBC periods, respectively.  There were no significant differences in the age, gender, and lengths of stay between patient encounters in the pre-PBC and post-PBC periods.  Compared to pre-PBC, post-PBC encounters had a higher total number of problems on the EMR problem list [7 (IQR: 7-12) vs 4 (IQR: 2-8), p<0.001).  Post PBC encounter problem lists had a lower mean proportion of problems labeled as “active” (0.86±0.23 vs 0.91±0.23, p<0.001), but had a slightly higher mean proportion of deleted problems (0.03± 0.1vs 0.02± 0.1, p=0.006) and resolved problems (0.12±0.2 vs 0.07±0.2, p<0.001).   

Conclusions:

This preliminary analysis of a retrospective time series suggests that problem lists were more often used and updated after the implementation of PBC in the ICU at Stanford Hospital.  Additional characterizations of problem list utilization are ongoing.  We are also studying whether PBC is effective in generating problem lists that can structure clinical data to more accurately phenotype common ICU hospital conditions, such as sepsis, and ICU clinical outcomes such as severity of illness and risk of mortality.