Background: Hospitals are nationally ranked on the basis of comparisons of quality of care, mortality, readmissions and health care associated infections.1 Center for Medicare and Medicaid Services (CMS) looks at these top 6 illnesses; stroke, pneumonia, heart failure, chronic obstructive pulmonary disease, acute myocardial infarction and coronary artery bypass graft (CABG) to compare hospital’s specific risk standardized mortality rates (RSMRs) to the US National Rate.2 Thirty day mortality is calculated by dividing observed mortality by expected mortality.3 Expected mortality is predicted by the documentation of patient’s severity of illness (SOI) which also translates into risk of morality (ROM).4 A high mortality score resulted in a decrease ranking at our hospital on this quality measure. Case analysis suggested that deficiencies in documentation of illness severity were a major factor behind our mortality score thus a quality improvement project was initiated to better capture illness severity of critically ill patients.
Methods: We reviewed provider documentation of all 51 cases with a 30-day mortality at our institution from 3/2018-9/2018. We excluded 29 cases with SOI and ROM score of 4 because 4 is the highest score that can be achieved and further improvement in those documentation wouldn’t impact expected mortality. We then developed a problem list scoring sheet to note which diagnoses were being charted and which ones were missed. Each chart was reviewed by 2 residents, any discrepancy in score was dealt by reviewing the chart together and coming to an agreed conclusion. Three factors were identified: hyponatremia, encephalopathy and malnutrition with a frequency missing that adversely impacted the SOI & ROM thereby under representing illness severity. Pre-Intervention data on documentation in the intensive care unit (ICU) was collected for one week (12/2/18 – 12/7/18). Our interventions, included an educational series on documentation of illness of severity for residents and hospitalist as well as changing the nutritionist’s template in our electronic medical record system (EMR) to make their assessment of malnutrition more visible in their notes. Additionally, posters reinforcing the key learning points regarding documentation of illness severity to include hyponatremia, malnutrition, and encephalopathy were displayed in ICU and resident’s charting area.
Results: Using our problem list scoring sheet which was used in our pre-intervention data collection we collected an additional 10 days of ICU patient documentation for our post-intervention comparison. We saw an overall improvement in the score in post vs pre-intervention (Average pre-intervention score =7.8, average post-intervention score = 8.4). This was due to the better documentation of patient’s illnesses. Documentation of hyponatremia showed an improvement of 44%. Encephalopathy and malnutrition were still not adequately captured in provider documentation.
Conclusions: Given that Mortality data is comprised of a three year rolling average, our quality improvement aim was targeted at documentation, which is considered a process outcome, rather than actual mortality scores. We saw an improvement in documentation of one factor that better captures illness severity but will need to determine if that improvement is sustained as well as implement additional interventions to improve provider documentation of malnutrition and encephalopathy.