BACKGROUND/RATIONALE:
While much attention is appropriately focused on the acute phase of minimizing further spread of the COVID-19 virus, protecting frontline clinicians, and treatment of those who contract the illness, it is also important to plan for surveillance and monitoring sequelae of COVID-19 illness and treatment. Recent clinical reports have revealed that COVID-19 disease is associated with a high inflammatory response that can result in viral pneumonia, and cardiac symptoms and complications. The severity, extent, and short-term versus long-term respiratory, cardiovascular, and neurological effects of COVID-19, along with the effect of specific COVID-19 treatments (e.g., steroids, unlabeled drug use, new drugs, and therapies, etc.) are not yet known. Additional concerns remain about the long-term mental health impacts on those who recover, especially among those with pre-existing mental health conditions. Severity of symptoms and mortality have been highest among the elderly, males, and those with comorbidities including chronic lung disease, hypertension, diabetes mellitus, cardiovascular diseases, and cerebrovascular diseases. The health impacts on vulnerable populations are also coming to light.

Among the Veteran population, the VA has tracked 361,988 positive and 6,431 active COVID 19 cases as of October 27, 2021 (VA website: https://www.accesstocare.va.gov/Healthcare/COVID19NationalSummary). With knowledge that Veterans enrolled in the VA tend to have more comorbidities compared to the general population, for those who contract COVID-19, they may be hospitalized and ultimately recover, and they may experience exacerbation of preexisting respiratory, cardiac, and mental health conditions, as well as develop new conditions. The long-term toll of COVID-19 hospitalization among survivors is still being realized. Data from early in the pandemic suggest that Veterans hospitalized with COVID-19 may experience higher readmission rates than the general population. Further regarding the mental health (MH) impact of contracting COVID-19, there remain limited data available on how hospitalization affects MH distress. Much is unknown about whether COVID-19 hospitalization contributes to the development of new MH conditions. Veterans may be at elevated risk of negative MH outcomes due to their higher prevalence of trauma and other risk factors.

OBJECTIVE(S):
During the nine-month methods development study we sought to:
1)Identify VA patients who were treated for COVID-19 nationally using available data in our prior national cohort of VA primary care patients and additional national VA data sets and explore methods for outcome ascertainment with input from a Clinical and Epidemiology Advisory Committee (CEAC);

2)Among Veterans hospitalized for COVID-19, describe the patient-level health impacts through 2020 and beyond when available, including hospital readmissions, exacerbation of existing conditions, development of new COVID- related chronic conditions, and mortality; and

3)Among Veterans hospitalized for COVID-19, explore, and examine factors associated with patient-level health impacts including hospital readmissions, exacerbation of existing conditions, development of new COVID- related chronic conditions, and mortality, during the first six to nine months after diagnosis.

METHODS:
Objective 1: We convened a Clinical & Epidemiology Advisory Committee (CEAC) virtually for three meetings during the nine-month project (Oct 2020, Dec 2020 & Mar 2021). The CEAC reviewed and discussed the study team's initial criteria for identifying COVID-19 cases, criteria for identifying and classifying subsequent health conditions, and preliminary analyses. The CEAC advised to expand the study population to incorporate data from the new COVID-19 Shared Data Resource (CSDR) developed during the project. The CEAC also assisted in identifying information gaps that may require new routinely collected data (e.g., ventilator days) or new patient reported data (symptoms after recovery; treatment outside the VA) that could be a focus in a subsequent study.

We explored methods using the CSDR to identify Veterans hospitalized for COVID-19, approaches using the CDW for developing a comparison cohort and for determining 90 day and 6-month outcomes. We initially focused on pulmonary, diabetes and mental health conditions. While information for the COVID cases was available in the CSDR, appropriately matched comparison cases being drawn from the CDW proved more time consuming than the project period allowed. For efficiency we focused on those Veterans hospitalized for COVID-19 and conducted descriptive analyses on specific sub-populations to refine outcome ascertainment approaches, assess data completeness, and gauge trends in selected outcomes for exacerbation of conditions and new conditions, mortality, hospital discharge disposition, and hospital readmissions.

Objective 2: Using the CSDR, we identified Veterans hospitalized in a VA facility with a positive diagnosis of COVID-19, based on a positive polymerase chain reaction (PCR) test indicating the presence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and admitted between March 1, 2020 through July 31, 2020 (n=5,486). In addition to the CSDR, data sources included electronic health records from the VA Corporate Data Warehouse (CDW) to identify hospital admissions, outpatient encounters, comorbidities, and additional sociodemographic characteristics. Discharge disposition was assessed for the initial COVID-19 hospitalization and was consolidated from 23 to three categories (community, another hospital or long-term care (LTC) facility, and unknown) due to very small sample sizes for several categories. In addition to discharge disposition, discharge status, which indicates whether a patient was discharged against medical advice (AMA) was flagged. VA hospital readmission for acute care was assessed within 90 days of discharge for those patients who were discharged alive.

Objective 3: Using the CSDR and CDW we explored approaches for identifying new pulmonary and mental health conditions using ICD-10-CM-diagnosis codes and medications. Focused on mental health outcomes, we selected all Veteran patients who were hospitalized for COVID-19 between March through August 2020 using the CSDR. This time frame allowed evaluation of new MH conditions up to six months post-hospitalization through March 2021. We excluded patients who died during hospitalization (n = 1,158) or within 6 months of discharge (n = 715) and were missing covariate information (n = 95) for a final sample of 3,518. We operationalized development of new MH conditions based on whether a Veteran had been diagnosed with the condition within two years prior to hospitalization. We analyzed the odds ratio of new MH conditions using a logistic regression model adjusting for demographic and clinical measures, as well as engagement in VHA MH care before and after hospitalization to account for limited opportunities for diagnosis.


FINDINGS/RESULTS:
Objective 1: Preliminary findings indicated that among the 5,486 Veterans hospitalized in the VA with COVID-19 between March 1 through July 31, 2020, 52.4% had diabetes mellitus; among those who died during hospitalization 48% had diabetes with complications and 11% had uncomplicated diabetes. Among those discharged, 2% developed new onset diabetes within six months of discharge. Among those with pre-existing diabetes, 9% had a worsening of their diabetes. Whether these trends were due to COVID-19, treatment for COVID-19 or other factors warrants further research in a controlled observational study.

Objective 2: We explored and examined the initial COVID-19 admission, discharge disposition, and readmission in 90 days for Veterans who survived their initial COVID-19 hospitalization. Of the 5,486 Veterans hospitalized in the VA with COVID-19 between March 1 through July 31, 2020, 21% died in the hospital. Among those who survived and without missing data (n=4,038), most were discharged to the community (74%); 18% were transferred to another hospital/long term care, and 7% had an unknown discharge disposition, and 1% were discharged against medical advice (AMA). Among those discharged 10.8% were readmitted within 90 days, often presenting with ongoing COVID-19 issues (38%) or circulatory conditions (15%). A larger proportion of Veterans with an unknown discharge disposition were readmitted within 90 days (79/325, 24.3%) compared to either community discharges (10.9%) or transfers to another hospital/LTC (4%; p<0.001). In regression analysis readmissions were associated with having an unknown discharge disposition, shorter stay, or use of a mechanical ventilator during the initial COVID-19 hospitalization; more comorbidities and living in a rural location (all p < 0.001). These rates are similar to those reported in most other studies, except for one by Donnelly and colleagues (2020). Our sample of Veterans were younger (average=66.2 versus 70.2 years) and our cohort was twice as large, covered a longer period, and focused on readmission for acute care, likely accounting for the differences observed in the readmission rate between studies.

Objective 3: We explored and examined the development of new mental health diagnoses among Veterans hospitalized at a VA hospital for COVID-19 up to 6-months following hospitalization. Among those Veterans hospitalized during this period (N=3,518), prior to hospitalization, 61% of our sample had a mental health diagnosis. The most common, pre-hospitalization MH conditions were depressive disorders (40%), anxiety disorders (23%), and posttraumatic stress disorder (22%). Eight percent (N=277/3518) developed a new MH diagnosis with the common new MH diagnoses being depressive (28%), anxiety (28%), and adjustment disorders (24%). We found patients aged <45 years (OR = 1.56, p = .05) or 55-64 years (OR = 1.61, p < .01) compared to those aged >=65 and those who lived in rural areas (OR = 1.49, p < .05) compared to urban were more likely to develop new MH conditions. Women were less likely than men to develop a new MH condition (OR = 0.56, p < .05). As Elixhauser comorbidity scores increased, the odds of new MH conditions decreased (OR = 0.99, p < .001). We found no significant interaction effects between race, gender, and MH care engagement in our exploratory analyses. We found no significant associations with other demographics and clinical variables. When considering specific predictors, age and rurality were associated with increased likelihood of developing a new MH condition. It is unsurprising that younger age groups showed increased risk, as MH conditions generally onset prior to 50 years of age. In terms of rural dwelling patients, there may be more limited access to MH prevention resources increasing likelihood of distress. Clinicians may consider implementing additional MH prevention measures (e.g., screening) to mitigate MH morbidity post-COVID-19 hospitalization

IMPACT:
Continued negative outcomes of the COVID-19 virus affect a small but critical number of Veterans who were hospitalized at VA facilities. Our analyses of hospital discharge disposition and readmissions suggest that a pre-determined discharge disposition with additional support/follow-up after hospital discharge could be beneficial in reducing adverse outcomes in persons with COVID-19. Yet more complete ascertainment of discharge disposition is needed for it to be a useful surveillance measure.

The current study evaluated the longitudinal impact of COVID-19 hospitalization on the development of new MH conditions using a large, national sample. As this was an observational study, it is not possible to infer causation from our findings. Due to the data limitations, we were unable to evaluate additional variables of interest such as social support and changes in post-hospitalization or care sought outside the VA, which may inform MH, as well as other condition trajectories. Future research should consider exploring social and clinical variables and care sought outside the VA to identify modifiable risk factors to prevent post-hospitalization MH distress and/or promote resilience.

The knowledge gained in working with the CEAC, the extant VA data to explore approaches for outcomes ascertainment informed the proposal and ongoing work of the COVID-19 Observational Research Collaboratory (CORC) begun in April 2021. Methods that include an emulated clinical trial design with a contemporaneous comparison group and incorporation of non-VA data such as Medicare and Medicaid are needed and is a focus of CORC. Further, we provided feedback to CORC, CIPHER and CSDR on methods to incorporate VA community care and Fee-basis data for ascertaining healthcare use. We encouraged the CSDR team to expand outcomes beyond the initial 60-days available in the data repository. Continued work in CORC with outcomes ascertainment will also combine information from prospective surveys and extant VA electronic health record data. Understanding the longer-term health impacts on Veterans will better position the VA and public health authorities to address their future health needs.

PUBLICATIONS:

Journal Articles

1. Gardner J, Brown G, Vargas-Correa J, Weaver F, Rubinstein I, Gordon HS. An assessment of Veterans attitudes and willingness to receiving the COVID-19 vaccine: a mixed methods study. BMC infectious diseases. 2022 Mar 29; 22(1):308. [view]
2. Weaver FM, Niederhausen M, Hickok A, O'Neill AC, Gordon HS, Edwards ST, Govier DJ, Chen JI, Young R, Whooley M, Hynes DM. Hospital Readmissions Among Veterans Within 90 Days of Discharge Following Initial Hospitalization for COVID-19. Preventing chronic disease. 2022 Dec 1; 19:E80. [view]
3. Chen JI, Hickok A, O AC, Niederhausen M, Laliberte AZ, Govier DJ, Edwards ST, Gordon HS, Slatore CG, Weaver FM, Young R, Hynes DM. Psychiatric disorders newly diagnosed among veterans subsequent to hospitalization for COVID-19. Psychiatry Research. 2022 Jun 1; 312:114570. [view]

JPRE

1. Bajema KL, Berry K, Streja E, Rajeevan N, Li Y, Yan L, Cunningham F, Hynes DM, Rowneki M, Bohnert A, Boyko EJ, Iwashyna TJ, Maciejewski ML, Osborne TF, Viglianti EM, Aslan M, Huang GD, Ioannou GN. Effectiveness of COVID-19 treatment with nirmatrelvir-ritonavir or molnupiravir among U.S. Veterans: target trial emulation studies with one-month and six-month outcomes. medRxiv : the preprint server for health sciences [Preprint]. 2022 Dec 16. [view]