Methods
The Office of Clinical Systems Development and Evaluation (CSDE–10E2A) produces a weekly CAN Score Report to help identify the highest-risk patients in a primary care panel or cohort. CAN scores range from 0 (lowest risk) to 99 (highest risk), indicating how likely a patient is to experience hospitalization or death compared with other VA patients. CAN scores are calculated with statistical prediction models that use data elements from the following Corporate Data Warehouse (CDW) domains: demographics, health care utilization, laboratory tests, medical conditions, medications, and vital signs (eAppendix available online at 10.12788/fp.0385).
The CAN Score Report is generated weekly and stored on a CDW server. A patient will receive all 6 distinct CAN scores if they are: (1) assigned to a primary care PACT on the risk date; (2) a veteran; (3) not hospitalized in a VA facility on the risk date; and (4) alive as of the risk date. New to CAN 2.5 is that patients who meet criteria 1, 2, and 4 but are hospitalized in a VA facility on the risk date will receive CAN scores for the 1-year and 90-day mortality models.
Utilizing VA Informatics and Computing Infrastructure (VA HSR RES 13-457, US Department of Veterans Affairs [2008]), we obtained 2 lists of veterans aged 70 to 75 years on February 8, 2019, with a calculated CAN score of ≥ 75 for 1-year mortality and 1-year hospitalization on that date. A veteran with a CAN score of ≥ 75 is likely to be prefrail or frail.9,10 Veterans who did not have a corresponding calculated CAN score on February 7, 2020, and February 12, 2021, were excluded. COVID-19 was declared a public health emergency in the United States on January 31, 2020, and the World Health Organization declared COVID-19 a pandemic on March 11, 2020.17 We picked February 7, 2020, within this time frame and without any other special significance. We picked additional CAN score calculation dates approximately 1 year prior and 1 year after this date. Veterans had to be alive on February 12, 2021, (the last date of the CAN score) to be included in the cohorts.
Statistical Analyses
The difference in CAN score from one year to the next was calculated for each patient. The difference between 2019 and 2020 was compared with the difference between 2020 to 2021 using a paired t test. Yearly CAN score values were analyzed using repeated measures analysis of variance. The number of patients that showed an increase in CAN score (ie, increased risk of either mortality or hospitalization within the year) or a decrease (lower risk) was compared using the χ2 test. IBM SPSS v26 and GraphPad Prism v18 were used for statistical analysis. P < .05 was considered statistically significant.
Results
There were 3538 veterans at PVAHCS who met the inclusion criteria and had a 1-year mortality CAN score ≥ 75 on February 8, 2019.
We excluded 6 veterans from the final analysis due to lack of 1-year mortality CAN score for 2020 or 2021. The final number included in the analysis was 3532 (Figure 1). The mean (SD) age was 71.8 (1.3) years. There were 3488 male (98.8%) and 44 female (1.2%) veterans represented (Table 1). Our data show a decrease in mean 1-year mortality CAN score in this subset of older frail veterans by 4.9 (95% CI, -5.3 to -4.5) in the year preceding the COVID-19 pandemic (Table 2). The 1-year mean mortality CAN score increased significantly by 0.2 (95% CI, -0.3 to 0.6; P < .0001 vs pre-COVID) in this same subset of patients after the first year of the COVID-19 pandemic (Figure 2). Mean CAN scores for 1-year mortality were 81.5 (95% CI, 81.2 to 81.7), 76.5 (95% CI, 76.1 to 77.0), and 76.7 (95% CI, 76.2 to 77.2) for 2019, 2020, and 2021, respectively.