Mir M. Ali, Ph.D., Health Economist
Office of Behavioral Health, Disability, and Aging Policy
Office of the Assistant Secretary for Planning and Evaluation
April 6, 2021
Link to Printer Friendly Version in PDF Format (11 PDF pages)
A significantly greater percentage (2.7%) of patients with a COVID-19 diagnosis experienced a new onset of mental health condition following their COVID-19 diagnosis compared to patients who tested negative for COVID-19 (1.0%) or had a COVID-19 symptom (1.4%).
Among patients who experienced a new onset of mental health condition, the most common types were anxiety (70-75%) and major depression (31-33%).
The odds of developing a mental health condition after COVID-19 diagnosis were significantly higher among individuals with health-related social needs associated with childhood upbringing (odds ratio [OR] 13.3), education (OR 8.3), employment (OR 4.5), or housing (OR 3.5).
Experts have raised concerns about COVID-19’s intersection with mental health since the start of the pandemic, especially among individuals with pre-existing mental health conditions. Individuals with mental health conditions have more chronic medical conditions that increase their risk for contracting severe COVID-19 compared with individuals without any mental health condition. Although researchers have predicted adverse mental health outcomes following a COVID-19 diagnosis,[3, 4] few studies quantify the mental health sequelae of COVID-19. Taquet et al. (2020) documented a bidirectional relationship between a COVID-19 diagnosis and mental health conditions but cautioned that their findings could be due to confounding socioeconomic factors that they were unable to take into account.
The pandemic has also raised concerns that social inequalities in health could unevenly impact COVID-19 related mortality and morbidity, with individuals and families living in poverty and disadvantaged conditions being disproportionately impacted. During the COVID-19 pandemic, many low-wage workers have been deemed as essential including agricultural workers, home care workers, grocery store workers, and restaurant workers. In addition, evidence suggests that social determinants of health (SDOH) indicators such as low income, social exclusion, unemployment, adverse childhood experiences, and food and housing insecurity can lead to poor mental health outcomes. Little is known about how the SDOH indicators are correlated with onset of new mental health conditions among individuals who contracted COVID-19. This is an important omission in the literature, as policy responses to the mental health impact of the pandemic could be more effectively formulated if the underlying factors that contribute to adverse mental health outcomes are better understood. Thus, the dual purposes of this brief are: (1) to estimate the rate of new onset of mental health conditions among individuals with a COVID-19 diagnosis who had no mental health conditions or mental health-related health services utilization in the 12 months prior to their COVID-19 diagnosis; and (2) to quantify the contribution of SDOH indicators to the development of the mental health condition.
The data for the study were drawn from December 2018-January 2021 IQVIA US Open Source Claims, a multi-payer pre-adjudicated health insurance claims database covering all 50 states and Washington, D.C. IQVIA US Open Source Claims includes professional claims generated by office-based physicians (CMS-1500), institutional claims generated by hospitals and other institutions (UB-04), and prescription claims. Analysis was completed using the E360™ Platform and E360™ Analytics Workbench. The version of the data that was accessible to ASPE contains information only on patients who were diagnosed with COVID-19, had a test for COVID-19, or exhibited COVID-19 symptoms such as fever, fatigue, shortness of breath, or cough. Descriptive statistics and logistic regression models were used to estimate the probability of, and risk factors for, developing a mental health condition after a COVID-19 diagnosis. The sample used for analysis includes 14,870,194 individuals who were diagnosed with COVID-19, had a test for COVID-19, or exhibited COVID-19 symptoms and who had no mental health diagnoses or mental health-related services utilization in the past 12 months. To examine the rate of new onset of mental health condition after COVID-19 and to see how this rate differed in comparison to patients who were not diagnosed with COVID-19, the sample was categorized into three mutually exclusive cohorts: (1) those who were diagnosed with COVID-19; (2) those who tested negative for COVID-19 (the "Negative Test cohort"); and (3) those who exhibited symptoms but were not diagnosed with COVID-19 or tested for COVID-19 (the "Symptoms cohort"). Patients with COVID-19 symptoms were examined on the theory that they could represent patients with COVID-19 lacking a diagnosis. The data set available to ASPE did not allow the construction of a matched control cohort to be compared with the patients diagnosed with COVID-19.
Psychiatric conditions were measured by whether the patient had any of the following mental health-related diagnoses after either their COVID-19 diagnosis or COVID-19 test or COVID-19 related symptoms: psychotic disorder, major depression, other mood disorder, anxiety and related disorder, and other mental health conditions. The primary variable of interest related to whether persons had indicators of SDOH as measured by the International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) Z codes. The codes relevant for the analysis were whether persons were indicated to have problems related to education and literacy (Z55), employment and unemployment (Z56), occupational risk (Z57), social environment (Z60), childhood upbringing (Z62), primary support including family (Z63), and psychosocial environment (Z64-Z65). Other variables controlled for in the analysis include indicators of hospitalization due to COVID-19; underlying medical conditions associated with higher risk of illness from COVID-19 (asthma, bronchitis, cancer, diabetes, heart disease, hypertension, kidney disease, liver cirrhosis, nicotine dependence and obesity); and demographic characteristics (age and gender).
Patients with a COVID-19 diagnosis were more likely to experience new onset of a mental health condition compared to patients with either a negative test for COVID-19 or a COVID-19 symptom.
Table 1 displays the percentage of new onset of mental health condition among each cohort and the percentage of the specific type of mental health condition among those who have a new onset of any mental health condition. The incidence of any new onset of mental health condition was higher among the COVID-19 cohort compared to the Symptoms cohort (2.7% vs. 1.4%; p < 0.001) and the Negative Test cohort (2.7% vs. 1.0%; p < 0.001). The median days to a new mental health diagnosis was lower among the COVID-19 cohort compared to the Negative Test cohort (26 days vs. 56 days; p < 0.001) and the Symptoms cohort (26 days vs. 123 days; p < 0.001). Among those with any mental health condition, the most common type was anxiety and related disorders for the COVID-19 cohort (75.2%), the Negative Test cohort (71.4%), and the Symptoms cohort (69.5%). Major depression was the next most common mental health condition, with prevalence of 33.1% among the COVID-19 cohort, 31.2% among the Negative Test cohort, and 32.3% among the Symptoms cohort.
Patients with a COVID-19 diagnosis who experienced a new onset of mental health condition were more likely to have severe COVID-19-related risk factors, hospitalizations and SDOH indicators.
Table 2 shows that 27.5% of patients with a new onset of mental health diagnosis following a COVID-19 diagnosis experienced a COVID-19 related hospitalization. The hospitalization rate among patients who did not have a mental health diagnosis was 14.7%. Among patients who were diagnosed with a mental health condition following a COVID-19 diagnosis, rates of medical conditions associated with severe COVID-19 were higher compared to those who did not have a mental health diagnosis. Utilization of Z codes related to SDOH was low across the sample population; however, those with a new onset of mental health condition had a higher rate of any SDOH-related Z code utilization compared to those who did not experience a new onset of mental health condition (3.1% vs 0.6%).
Multivariable logistic regression estimates (Table 3) show that the odds of developing a mental health condition following COVID-19 were higher among individuals with a SDOH Z code--specifically, among patients with SDOH Z codes related to childhood upbringing (OR 13.3, 95% CI 11.8-15.0), housing (OR 3.5, 95% CI 3.3-3.7), employment (OR 4.5, 95% CI 4.0-5.0), or education (OR 8.3, 95% CI 7.3-9.3). Hospitalization due to COVID-19 also was correlated with higher odds of developing a mental health condition (OR 1.6, 95% CI 1.5-1.6). To check whether SDOH Z codes were more strongly correlated with COVID-19 diagnosis than non-diagnosis, logistic regression models were estimated for the Negative Test cohort and the COVID-19 Symptom cohort. The results reported in Appendix Table A1 shows that the SDOH Z codes were also statistically significant among these cohorts, but the magnitudes were larger among the COVID-19 patients.
Compared to patients with negative COVID-19 test or COVID-19 symptoms, patients with a COVID-19 diagnosis were more likely to experience a new onset of a mental health condition. Anxiety and major depression were the most common type of mental health condition among patients, regardless of whether they had a COVID-19 diagnosis or not. The findings from the study are consistent with previous research findings that a COVID-19 diagnosis was associated with the development of a mental health condition. However, the incidence rate identified here (2.7%) is lower than what has been reported in the literature. Specifically, Taquet et al. (2020) using electronic health records reported a 5.8% incidence rate of mental health condition among individuals who contracted COVID-19 but had no mental health diagnosis in the past 12 months. The difference between the incidence rate reported here and Taquet et al. (2020) might be attributable to differences in how the 12-month clean period was defined. In the present study, the 12-month clean period was based on either a diagnosis or services utilization related to mental health, whereas the Taquet et al. (2020) based the clean period on diagnosis only. It is possible that Taquet et al. (2020) might have included individuals who utilized health services for their mental health condition without a mental health diagnosis appearing in their electronic health records during the clean period. When a clean period of mental health diagnosis only was applied to this study, the number of patients with a COVID-19 diagnosis in the study becomes 4,916,785 among whom 232,786 (or 4.7%) experienced a new onset of mental health condition--an estimate closer to what was reported in Taquet et al. (2020). Thus, the present study imparts a conservative bias to the incidence rate of new onset of a mental health condition after COVID-19.
The underlying mechanism between COVID-19 and mental health conditions is not well understood. Biological factors directly related to COVID-19 (compromised immune system, loss of functioning, prolonged hospitalization) and/or social factors (economic instability, housing insecurity, lack of social support) might be associated with a new onset of mental health condition. Future studies could unpack these underlying mechanisms behind new mental health conditions among patients with COVID-19.
SDOH indicators or health-related social needs were a significant predictor of new onset of mental health conditions among patients with COVID-19 after accounting for pre-existing medical conditions associated with severe COVID-19, hospitalization due to COVID-19, and demographic characteristics. Despite the low rate of utilization of the Z codes related to SDOH (see Appendix A for more details), its impact was highly significant and large in magnitude. This is consistent with research that documented underutilization of SDOH Z codes,[9, 10, 11] even though there is a growing consensus that addressing SDOH is crucial to the overall objective of improving population health.
Research has shown that income inequality is associated with depression prevalence and that adverse childhood experiences along with food and housing insecurity are correlated with anxiety disorder.[13, 14] The findings from this study add to this growing body of literature and show that economic insecurity related to employment, education, and housing, along with problems related to social environment and childhood upbringing, can be risk factors in developing mental health conditions among patients with COVID-19, even among those with no recent history of mental illness.
Health systems are beginning to adopt patient-level screening for health-related social needs. For example, many states, health care plans, and providers have recognized the importance of addressing these health-related social needs for Medicaid and CHIP beneficiaries to lower health care costs, improve health outcomes, and increase the cost-effectiveness of health care services and interventions. The mental health profession is advocating for implementation of policies that helps improve education and employment opportunities and reduce adverse childhood experiences, along with development of new mental health treatment regimens. It is estimated that the impact of the COVID-19 pandemic on disadvantaged communities will be long lasting. Thus, integrating health-related social needs into policy efforts designed to tackle the COVID-19 pandemic may help not only to reduce the disparities in health outcomes that are being observed, but also to reduce the onset of mental health conditions that individuals are likely to experience subsequent to their COVID-19 diagnosis.
|Table 1: New Onset of Mental Health Condition after COVID-19 Diagnosis compared with other Health Events among Individuals with no Mental Health Related Claims in the Past 12 Months|
(N = 4,556,217)
|Negative COVID-19 Test2
(N = 8,593,827)
(N = 1,720,105)
|Any Mental Health Condition*||122,201 (2.7%)||85,182 (1.0%)||24,424 (1.4%)|
|Median Days to Mental Health Diagnosis*||26||56||123|
|Type of Mental Health Condition4|
|Anxiety and Related Disorder*||75.2%||71.4%||69.5%|
|Other Mood Disorder*||5.5%||6.9%||6.9%|
|Other Mental Health Condition*||15.4%||10.5%||11.8%|
* p < 0.001
|Table 2: Descriptive Statistics of Patients with a COVID-19 Diagnosis by Mental Health Condition|
(N = 4,556,217)
|Mental Health Condition
(N = 122,201)
|No Mental Health Condition
(N = 4,434,016)
|0 - 17*||11.01%||9.08%||11.13%|
|18 - 24*||10.35%||10.14%||10.36%|
|25 - 64*||57.38%||53.55%||57.62%|
|Any SDHOH Measure*||0.75%||3.10%||0.61%|
|Z55 (Education & Literacy)||0.03%||0.21%||0.02%|
|Z56 (Employment & Unemployment)||0.05%||0.23%||0.04%|
|Z57 (Occupational Risk)||0.03%||0.04%||0.03%|
|Z59 (Housing & Economic Circumstances)||0.17%||0.77%||0.13%|
|Z60 (Social Environment)||0.07%||0.24%||0.06%|
|Z62 (Childhood Upbringing)||0.03%||0.29%||0.02%|
|Z63 (Primary Support)||0.15%||1.01%||0.09%|
|Z64/Z65 (Psychosocial Event)||0.12%||0.38%||0.11%|
|* p < 0.001|
|Table 3: Logistic Regression Model of New Onset of Mental Health Condition|
|Variable||Odds Ratio||95% Confidence Interval||P-value|
|0 - 17 (reference)|
|18 - 24||1.23||1.20; 1.25||<0.001|
|25 - 64||0.93||0.92; 0.95||<0.001|
|Kidney Disease||1.17||1.15; 1.18||<0.001|
|Liver Cirrhosis||1.22||1.16; 1.29||<0.001|
|Nicotine Dependence||1.78||1.76; 1.81||<0.001|
|Z55 (Education & Literacy)||8.26||7.31; 9.34||<0.001|
|Z56 (Employment & Unemployment)||4.49||4.02; 5.01||<0.001|
|Z57 (Occupational Risk)||1.51||0.92; 1.44||0.219|
|Z59 (Housing & Economic Circumstances)||3.50||3.30; 3.72||<0.001|
|Z60 (Social Environment)||2.52||2.28; 2.79||<0.001|
|Z62 (Childhood Upbringing)||13.34||11.84; 15.03||<0.001|
|Z63 (Primary Support)||8.82||8.33; 9.33||<0.001|
|Z64/Z65 (Psychosocial Event)||2.45||2.27; 2.65||<0.001|
|Z64/Z65 (Psychosocial Event)||0.12%||0.38%||0.11%|
|Appendix Table 1: Logistic Regression Model of New Onset of Mental Health Condition|
|SDOH Measures||Negative COVID-19 Test||COVID-19 Mortality Rate|
|Z55 (Education & Literacy)||4.35
|Z56 (Employment & Unemployment)||5.01
|Z57 (Occupational Risk)||1.33
|Z59 (Housing & Economic Circumstances)||4.35
|Z60 (Social Environment)||2.03
|Z62 (Childhood Upbringing)||6.68
|Z63 (Primary Support)||5.52
|Z64/Z65 (Psychosocial Event)||1.83
|Models controls for demographic characteristics, COVID-19 related hospitalization and health conditions.|
Utilization of SDOH Z codes in less than 1% of the patient identified in this study is consistent to what was observed in prior research.[9, 11, 17] A recent study of Medicare Fee-for beneficiaries found that less than 1% of the patients had an SDOH-related Z codes in their claims. The study also reported that patient with SDOH Z codes were younger; more likely to be male; more likely to be Black; and were more likely to have a behavioral health diagnosis, especially alcoholism, drug disorders, and mental health conditions. The data source utilized here is limited in the number of patient demographic characteristics, and the study was not able to explore how the role of COVID-19 and SDOH measures on mental health may differ by these characteristics. Research has documented substantial health disparities by race/ethnicity related to the COVID-19 pandemic, and it is possible that communities of color are more likely to develop mental health conditions as a result of a COVID-19 diagnosis. More research is needed to explore racial disparities in new onset of mental health condition after a COVID-19 diagnosis. This study was also unable to identify the providers who might be utilizing the SDOH Z codes, but the literature documents mental health providers and large private non-for-profit urban teaching hospitals to be more likely to utilize them.[9, 17] It is likely that the patients assigned the SDOH Z codes may have more severe health conditions that might have prompted the providers to assign those Z codes. As such, the underutilization of SDOH Z codes here imparts a conservation bias to the estimates. However, a reason for the low utilization of the SDOH-related Z codes may be that no current national financial incentives exist to code the SDOH indicators. They are not included in any risk adjustment models for health care organization performance evaluation or federal value-based payment programs, and therefore may not be the focus of documentation improvement efforts by billing departments and coding experts to the degree that clinical comorbidities have been.
Xiang YT, Yang Y, Li W, Zhang L, Zhang Q, Cheung T, Ng CH. (2020). "Timely mental health care for the 2019 novel coronavirus outbreak is urgently needed." Lancet Psychiatry, 7(3): 228-229. doi: 10.1016/S2215-0366(20)30046-8.
Novak P, Sanmartin MX, Ali MM, Chen J. (2020). "Health conditions associated with severe illness from COVID-19 among individuals with serious mental illness." Psychiatr Serv, appips202000300. doi: 10.1176/appi.ps.202000300.
Holmes EA, O'Connor RC, Perry VH, Tracey I, Wessely S, Arseneault L, Ballard C, Christensen H, Silver RC, Everall I, Ford T, John A, Kabir T, King K, Madan I, Michie S, Przybylski AK, Shafran R, Sweeney A, Worthman CM, Yardley L, Cowan K, Cope C, Hotopf M, Bullmore E. (2020). "Multidisciplinary research priorities for the COVID-19 pandemic: A call for action for mental health science." Lancet Psychiatry, 7(6): 547-560. doi: 10.1016/S2215-0366(20)30168-1.
Yao H, Chen JH, Xu YF. (2020). "Patients with mental health disorders in the COVID-19 epidemic." Lancet Psychiatry, 7(4): e21. doi: 10.1016/S2215-0366(20)30090-0.
Abrams EM, Szefler SJ. (2020). "COVID-19 and the impact of social determinants of health." Lancet Respir Med, 8(7): 659-661. doi: 10.1016/S2213-2600(20)30234-4.
Centers for Disease Control and Prevention. COVID-19 in Racial and Ethnic Minority Groups. Accessed March 23, 2021 at https://www.cdc.gov/coronavirus/2019-ncov/need-extra-precautions/racial-ethnic-minorities.html.
Shim RS, Compton MT. (2018). "Addressing the social determinants of mental health: If not now, when? If not us, who?" Psychiatr Serv., 69(8): 844-846. doi: 10.1176/appi.ps.201800060. Epub 2018 Jun 1. PMID: 29852822.
Taquet M, Luciano S, Geddes JR, Harrison PJ. (2021). "Bidirectional associations between COVID-19 and psychiatric disorder: Retrospective cohort studies of 62 354 COVID-19 cases in the USA." Lancet Psychiatry, 8(2): 130-140. doi: 10.1016/S2215-0366(20)30462-4.
Weeks WB, Cao SY, Lester CM, Weinstein JN, Morden NE. (2020). "Use of Z-codes to record social determinants of health among fee-for-service Medicare beneficiaries in 2017." J Gen Intern Med., 35(3): 952-955. doi:10.1007/s11606-019-05199-w.
Chaiyachati KH, Grande DT, Aysola J. (2016). "Health systems tackling social determinants of health: Promises, pitfalls, and opportunities of current policies." Am J Manag Care, 22(11): e393-e394.
Mathew J, Hodge C, Khau, M. (2019). Z Codes Utilization among Medicare Fee-for-Service (FFS) Beneficiaries in 2017. CMS OMH Data Highlight No. 17. Baltimore, MD: Centers for Medicare & Medicaid Services, Office of Minority Health.
Gottlieb LM, Wing H, Adler NE. (2017). "A systematic review of interventions on patients' social and economic needs." Am J Prev Med., 53(5): 719-729. doi: 10.1016/j.amepre.2017.05.011.
Dube SR, Anda RF, Felitti VJ, Chapman DP, Williamson DF, Giles WH. (2001). "Childhood abuse, household dysfunction, and the risk of attempted suicide throughout the life span: Findings from the Adverse Childhood Experiences Study." JAMA, 286(24): 3089-96. doi: 10.1001/jama.286.24.3089.
Messias E, Eaton WW, Grooms AN. (2011). "Economic grand rounds: Income inequality and depression prevalence across the United States: An ecological study." Psychiatr Serv., 62(7): 710-2. doi: 10.1176/ps.62.7.pss6207_0710.
Horwitz LI, Chang C, Arcilla HN, Knickman JR. (2020). "Quantifying health systems' investment in social determinants of health, by sector, 2017-19." Health Aff (Millwood), 39(2): 192-198. doi: 10.1377/hlthaff.2019.01246.
Lipson D. (2017). Medicaid’s Role in Improving the Social Determinants of Health: Opportunities for States. National Academy of Social Insurance. Available at https://www.nasi.org/research/2017/medicaid%E2%80%99s-role-improving-social-determinants-health.
Truong HP, Luke AA, Hammond G, Wadhera RK, Reidhead M, Joynt Maddox KE. (2020). "Utilization of social determinants of health ICD-10 Z-codes among hospitalized patients in the United States, 2016-2017." Med Care, 58(12): 1037-1043. doi: 10.1097/MLR.0000000000001418.
U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES
Office of the Assistant Secretary for Planning and Evaluation
200 Independence Avenue SW, Mailstop 447D
Washington, D.C. 20201
For more ASPE briefs and other publications, visit: aspe.hhs.gov/reports
ABOUT THE AUTHORS
Mir M. Ali, Ph.D. is a Health Economist in the Office of Behavioral Health, Disability, and Aging Policy in the Office of the Assistant Secretary for Planning and Evaluation.
Mir M. Ali, Ph.D., Health Economist. Mental Health Consequences of COVID-19: The Role of Social Determinants of Health (Issue Brief). Washington, DC: Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services. April 9, 2021.
All material appearing in this report is in the public domain and may be reproduced or copied without permission; citation as to source, however, is appreciated.
Subscribe to ASPE mailing list to receive email updates on new publications: aspe.hhs.gov/join-mailing-list
For general questions or general information about ASPE: aspe.hhs.gov/about