Electronic Health Records (EHR) Significantly Under-Capture Patient Co-Morbidity
Event Type
Workshop
Applications
Computational Science
Education and Training and Outreach
HPC Community Collaboration
HPC Training and Education
Machine Learning and Artificial Intelligence
Performance
Workforce
W
TimeSunday, 14 November 20215pm - 5:15pm CST
Location223
DescriptionBackground:
The concept of a digital twin in healthcare is predicated upon mimicking, as closely as possible, the clinical state of the patient. However, EHR implementations might suffer in data quality, as they are incumbent upon accurate/complete data entry (typically) into discrete data fields. Unfortunately, the well-documented problem of healthcare provider “click fatigue” might preclude full capture of data.
One of the most important predictors of cancer outcome is performance status, which also conceptually incorporates the patient’s comorbid illnesses. Does the EHR potentially underestimate patients’ disease burden? In this analysis, we compared ICD-10 capture of patient co-morbidities from the EHR’s Diagnoses with those from the billing system (coded by billers from each clinic note).
We hypothesized that discrete comorbidity data from the billing system would be statistically larger in number and scope than those captured within the EHR.
Materials/Methods:
We utilized a cohort of lung cancer patients treated with radiotherapy who were covered under a retrospective research protocol. Comorbidity information was separately pulled from the EHR (PowerChart, Cerner Medical Systems, Kansas City, MO) and the billing (B) system (Soarian, Cerner Medical Systems, Kansas City, MO). Cohorts were standardized, utilizing only patients with data available from both systems. ICD-9 codes, SNOMED codes and invalid ICD-10 codes were removed. Because chronic conditions were recapitulated at each patient visit in B, all duplicates were removed. By doing so, we generated a “maximal” comorbidity list for each patient from each system for purposes of comparison.
Charlson Comorbidity Index (CCI) and Elixhauser (EL) scores (validated instruments of comorbidity) were generated for each patient from these ICD-10 codes (CCI-EHR, CCI-B, EL-EHR, and EL-B, respectively). The CCI captures morbidity across 17 specific health sub-domains and the EL captures 31. Mean EHR and B scores across each sub-domain were compared for both CCI and EL using Welch’s t-test. Lastly, the EHR and B data sets were merged to generate CCI-Combined and EL-Combined scores, which were statistically compared across sub-domains to the CCI-B and EL-B results in turn. All analyses were performed in R version 4.03 using the ‘comorbidity’ package to generate the CCI/EL scores.
Results:
After cleaning, EHR and B contained 1929 and 4179 distinct ICD-10 codes, respectively, across 2059 patients. 2582 of these codes were exclusive to B. Mean scores across the CCI sub-domains were typically greater by an order of magnitude or more in B. Interestingly “cancer” was documented more frequently in the EHR (p=0.19) and “AIDS” was numerically more common in B, but not significantly so (p=0.058). EL scores were vastly significantly higher in B compared with EHR, except for blood loss anemia (p=0.083). Comparing CCI-Combined and EL-Combined to the CCI-B and EL-B data sets statistically improved four sub-domain scores -- cerebrovascular disease (p< 2.2e-16) and renal disease (p<2.2e-16) in CCI and hypothyroidism (p<2.2e-16) and solid tumor diagnosis (p=9.6e-9) in EL.
Conclusion:
The EHR significantly underestimates patient co-morbidity. If billing data are available, they should be incorporated into patient co-morbidity calculations.
The concept of a digital twin in healthcare is predicated upon mimicking, as closely as possible, the clinical state of the patient. However, EHR implementations might suffer in data quality, as they are incumbent upon accurate/complete data entry (typically) into discrete data fields. Unfortunately, the well-documented problem of healthcare provider “click fatigue” might preclude full capture of data.
One of the most important predictors of cancer outcome is performance status, which also conceptually incorporates the patient’s comorbid illnesses. Does the EHR potentially underestimate patients’ disease burden? In this analysis, we compared ICD-10 capture of patient co-morbidities from the EHR’s Diagnoses with those from the billing system (coded by billers from each clinic note).
We hypothesized that discrete comorbidity data from the billing system would be statistically larger in number and scope than those captured within the EHR.
Materials/Methods:
We utilized a cohort of lung cancer patients treated with radiotherapy who were covered under a retrospective research protocol. Comorbidity information was separately pulled from the EHR (PowerChart, Cerner Medical Systems, Kansas City, MO) and the billing (B) system (Soarian, Cerner Medical Systems, Kansas City, MO). Cohorts were standardized, utilizing only patients with data available from both systems. ICD-9 codes, SNOMED codes and invalid ICD-10 codes were removed. Because chronic conditions were recapitulated at each patient visit in B, all duplicates were removed. By doing so, we generated a “maximal” comorbidity list for each patient from each system for purposes of comparison.
Charlson Comorbidity Index (CCI) and Elixhauser (EL) scores (validated instruments of comorbidity) were generated for each patient from these ICD-10 codes (CCI-EHR, CCI-B, EL-EHR, and EL-B, respectively). The CCI captures morbidity across 17 specific health sub-domains and the EL captures 31. Mean EHR and B scores across each sub-domain were compared for both CCI and EL using Welch’s t-test. Lastly, the EHR and B data sets were merged to generate CCI-Combined and EL-Combined scores, which were statistically compared across sub-domains to the CCI-B and EL-B results in turn. All analyses were performed in R version 4.03 using the ‘comorbidity’ package to generate the CCI/EL scores.
Results:
After cleaning, EHR and B contained 1929 and 4179 distinct ICD-10 codes, respectively, across 2059 patients. 2582 of these codes were exclusive to B. Mean scores across the CCI sub-domains were typically greater by an order of magnitude or more in B. Interestingly “cancer” was documented more frequently in the EHR (p=0.19) and “AIDS” was numerically more common in B, but not significantly so (p=0.058). EL scores were vastly significantly higher in B compared with EHR, except for blood loss anemia (p=0.083). Comparing CCI-Combined and EL-Combined to the CCI-B and EL-B data sets statistically improved four sub-domain scores -- cerebrovascular disease (p< 2.2e-16) and renal disease (p<2.2e-16) in CCI and hypothyroidism (p<2.2e-16) and solid tumor diagnosis (p=9.6e-9) in EL.
Conclusion:
The EHR significantly underestimates patient co-morbidity. If billing data are available, they should be incorporated into patient co-morbidity calculations.