Methods: In this single-center retrospective study, we identified all patients admitted to the internal medicine service at UCSF Medical Center between June 1, 2014 and September 30, 2016 and obtained the timing of all computerized orders for those patients during the hospitalization. UCSF has an open-ICU where the internal medicine resident teams enter all orders for ICU patients (except for respiratory care and sedation). For each patient day, we determined the level of care based on location: intensive care unit (ICU), “step-down” (for patients with greater monitoring needs than a general medicine unit), or general medicine unit. The 3M APR-DRG Severity of Illness grouper was used to characterize discharge severity of illness for each patient which was categorized as Low, Moderate, Major, and Extreme. We used linear regression to identify statistically significant differences in order frequency based on level of care and severity of illness.
Results: We identified 2,047,576 orders placed during the study period for 15,349 patients. The mean order frequency was 34.2 orders per day for an ICU patient, 20 for a step-down patient, and 22.8 for a general medicine unit patient, and the difference in mean daily orders was statistically significant when comparing general medicine patients to patients in the ICU (p < 0.0001) and to patients in step-down units (p < 0.0001). The total orders placed during hospitalization were significantly associated with a patient’s discharge severity of illness. Patients with a low severity of illness received on average 95.3 orders during hospitalization, patients with a moderate severity of illness received on average 119.6 orders, patients with major severity of illness received on average 167.2 orders, and patients with an extreme severity of illness received on average 365.8 orders per hospitalization. Comparisons between each of these severity of illness categories, compared to low severity of illness, were significant (p<0.0001).
Conclusions: We found that on average, ICU patients require over twice as many daily orders than general medicine patients, with step-down patients somewhere in between. Furthermore, patients with a higher severity of illness require significantly more daily orders than those with lower illness severity. These findings suggest that the frequency and volume of electronic orders may serve as at least a partial proxy for patient complexity.
To cite this abstract:Van Groningen, N; Fang, MC; Najafi, N; Rajkomar, AR; Prasad, P; Khanna, R . COMPUTERIZED ORDERS AS A PROXY FOR PATIENT CLINICAL ACUITY. Abstract published at Hospital Medicine 2017, May 1-4, 2017; Las Vegas, Nev. Abstract 126. Journal of Hospital Medicine. 2017; 12 (suppl 2). https://www.shmabstracts.com/abstract/computerized-orders-as-a-proxy-for-patient-clinical-acuity/. Accessed May 24, 2019.