Characteristics of Adult Outpatients and Inpatients with COVID-19 — 11 Academic Medical Centers, United States, March–May 2020

1Descriptions of coronavirus disease 2019 (COVID-19) in the United States have focused primarily on hospitalized patients. Reports documenting exposures to SARS-CoV-2, the virus that causes COVID-19, have generally been described within congregate settings, such as meat and poultry processing plants and long-term care facilities. Understanding individual behaviors and demographic characteristics of patients with COVID-19 and risks for severe illness requiring hospitalization can inform efforts to reduce transmission. During April 15–May 24, 2020, telephone interviews were conducted with a random sample of adults aged ≥18 years who had positive reverse transcription–polymerase chain reaction (RT-PCR) test results for SARS-CoV-2 in outpatient and inpatient settings at 11 U.S. academic medical centers in nine states. Respondents were contacted 14–21 days after SARS-CoV-2 testing and asked about their demographic characteristics, underlying chronic conditions, symptoms experienced on the date of testing, and potential exposures to SARS-CoV-2 during the 2 weeks before illness onset (or the date of testing among those who did not report symptoms at the time of testing). Among 350 interviewed patients (271 [77%] outpatients and 79 [23%] inpatients), inpatients were older, more likely to be Hispanic and to report dyspnea than outpatients. Fewer inpatients (39%, 20 of 51) reported a return to baseline level of health at 14–21 days than did outpatients (64%, 150 of 233) (p = 0.001). Overall, approximately one half (46%) of patients reported known close contact with someone with COVID-19 during the preceding 2 weeks. This was most commonly a family member (45%) or a work colleague (34%). Approximately two thirds (64%, 212 of 333) of participants were employed; only 35 of 209 (17%) were able to telework. These findings highlight the need for screening, case investigation, contact tracing, and isolation of infected persons to control transmission of SARS-CoV-2 infection during periods of community transmission. The need for enhanced measures to ensure workplace safety, including ensuring social distancing and more widespread use of cloth face coverings, are warranted.

2The Influenza Vaccine Effectiveness in the Critically Ill (IVY) Network is a collaboration of U.S. medical centers conducting research on vaccine effectiveness for and epidemiologic studies of influenza, and recently started conducting epidemiologic studies on COVID-19. To explore the spectrum of illness across health care settings and potential community SARS-CoV-2 exposures after issuance of national social distancing guidelines on March 16, 2020, 11 academic medical centers in nine states conducted telephone-based surveys of a sample of patients with positive SARS-COV-2 test results during April 15–May 24, 2020 (testing dates = March 31–May 10, 2020). Medical centers submitted lists of persons with SARS-CoV-2 infection to Vanderbilt University and identified location of testing (intensive care unit [ICU], non-ICU hospital admission, emergency department [ED] without admission during the encounter, and other outpatient settings). To achieve a broadly representative cohort, selection of patients was made using site-specific stratified random sampling by location of testing. The median proportions sampled were 67% of inpatients and 53% of outpatients. Personnel from CDC telephoned patients during intervals of 14–21 days (97%) or 28–35 days (3%) after testing; up to seven call attempts were made per patient for each period. Interviews were conducted in English, Spanish, French, Creole, Portuguese, Arabic, Burmese, and Somali. Respondents or their proxies were asked to provide patient demographic and socioeconomic information, clinical signs and symptoms on the date of testing, underlying chronic conditions, and potential exposures to SARS-CoV-2 during the 2 weeks preceding illness onset (or 2 weeks before test date in patients who did not report symptoms). This 14-day exposure period was selected to encompass the estimated COVID-19 incubation period for most persons. Patients who responded at 28–35 days were asked the same questions, with the exception of signs or symptoms at the time of testing because the delay between symptom onset and interview date increased the potential for introducing recall bias.

3To compare responses among patients who received inpatient and outpatient testing, descriptive statistics were analyzed, using Wilcoxon rank-sum testing for continuous variables and chi-squared or Fisher’s exact test for categorical variables. Patients with proxy respondents or who had died were excluded because details about symptoms, medical conditions, and exposure histories were frequently unknown. Statistical analyses were conducted using Stata software (version 16; StataCorp).

4At least one telephone call was attempted for 798 randomly selected patients 309 inpatients [98 ICU and 211 non-ICU] and 489 outpatients [144 ED and 345 non-ED]) across the 11 sites. Among these, 544 (68%) answered calls, and 398 (50%) completed interviews. Sixty-seven (8%) patients or proxies refused, 37 (5%) were unable to complete the interview because of a language barrier, 42 (5%) requested a callback but could not be reached on further call attempts; 20 (3%) were reported to have died within 21 days of testing (nine proxy respondents interviewed and 11 refused). A total of 48 proxy interviews were excluded, leaving 350 of 398 for analysis.*

5Among the 350 respondents with completed interviews, 271 (77%) were tested as outpatients (70 ED and 201 non-ED) and 79 (23%) as inpatients (17 ICU and 62 non-ICU) (Table 1). The median number of patient respondents by site was 20 (interquartile range = 11–46). The median respondent age was 43 years; 185 (53%) were female, 116 (33%) white, 73 (21%) non-Hispanic black (black), 43 (12%) non-Hispanic of another race, and 116 (33%) Hispanic. Nineteen patients reported another positive SARS-CoV-2 test result before the test date applicable to this study. Among outpatients, 8% (22 of 271) were later admitted to the hospital after having outpatient testing.

6Compared with outpatients, inpatients were older (median age = 54 versus 42 years; p<0.001) and differed by race/ethnicity (p = 0.008) and annual household income (p = 0.003). Inpatients were less likely to be white (19% versus 37%) and more likely to have annual household income <$25,000 (28% versus 13%). Inpatients also had more underlying chronic conditions (median = two) than did outpatients (median = one) (p<0.001), including cardiovascular conditions, chronic respiratory disease, and diabetes.

7Among 316 (90%) respondents who answered questions on symptoms and did not report a previous positive SARS-CoV-2 test result,† 292 (92%) reported one or more symptoms on the date of SARS-CoV-2 testing (Table 2), including 238 (96%) of 248 outpatients and 54 (79%) of 68 inpatients. Both inpatients and outpatients reported a similar number of symptoms, but inpatients were more likely to describe dyspnea (72% versus 32%; p<0.001) and less likely to report loss of smell or taste (43% versus 59%; p = 0.030). Fewer symptomatic inpatients (39%, 20 of 51) reported a return to baseline level of health at 14–21 days than did symptomatic outpatients (64%, 150 of 233) (p = 0.001).

8Among 339 (97%) participants who provided exposure histories, 46% (153 of 332) reported a close case contact, defined as being within 6 feet of someone with a diagnosis of COVID-19, during the 2 weeks preceding illness onset or the date of testing for asymptomatic patients (Table 3). This was most commonly a family member (45%, 69 of 153) or a work colleague (34%, 52 of 153). Seven of the 339 participants were missing data in their case contact histories.

9Approximately two thirds (64%, 212 of 333) of participants were employed; however, only 35 of 209 (17%) were able to telework. Outpatients were more likely to report being employed than were inpatients (70% versus 42%; p<0.001) and interacted with persons outside the home more frequently (p<0.001). Among employed participants, 53 (25%) reported working in health care.

10Few studies have systematically collected data on COVID-19 patients from varied health care settings in the United States. In this multistate telephone-based survey of 350 U.S. COVID-19 outpatients and inpatients, inpatients were typically older and had more underlying chronic conditions, findings that have been previously observed with both COVID-19 and influenza patients. Compared with outpatients, inpatients reported lower household incomes and were less likely to be white. Differences by race/ethnicity are consistent with those reported previously (e.g., 43% of inpatients were Hispanic, and 28% were black), although in this descriptive analysis no adjustment for other factors was made to evaluate any independent association between race/ethnicity and COVID-19 severity.

11Approximately one third of symptomatic outpatients reported that they had not returned to baseline health by the interview date 14–21 days after testing positive for SARS-CoV-2 infection. In comparison, almost all outpatient working adults with laboratory-confirmed influenza reported returning to normal activities within 14 days of illness onset during the 2012–13 influenza season.

12Fewer than one half of patients were aware of recent close contact with someone with COVID-19, highlighting a need for increased screening, case investigation, contact tracing, and isolation of infected persons during periods of community transmission. This finding suggests that ensuring social distancing and more widespread use of cloth face coverings are warranted. A majority of COVID-19 patients reported working during the 2 weeks preceding illness, and few had the ability to telework, underscoring the need for enhanced measures to ensure workplace safety.

13The findings in this report are subject to at least six limitations. First, given that the survey was telephone-based, some nonresponse bias is possible. Patients with more severe illnesses might have still been hospitalized at the time of the survey or might have died, resulting in a higher proportion of nonrespondents among patients with more severe illness. Estimates of the frequency of clinical characteristics should therefore be interpreted with caution. Second, patients were sampled from academic medical centers with differing numbers of respondents; therefore, patients in this study are not representative of cases nationwide. With limited testing capacity, some groups (e.g., health care and other essential workers) might also have been preferentially tested. Third, data were obtained by self-report and might be subject to recall bias. Fourth, this survey documented a cross-section of symptoms reported on the date of testing, and symptoms might have changed during the course of illness. In addition, a few patients reported an earlier positive test result, which might have led to misclassification of test setting; however, this was infrequent (5%). Fifth, no adjustment for other factors to determine whether variables were independently associated with illness severity was made. Finally, a small proportion of respondents were asymptomatic at the time of testing. However, comparisons including demographics and exposure histories were similar when the analysis was restricted to only patients who reported symptoms.

14This study provides insights into epidemiologic characteristics of patients with laboratory-confirmed COVID-19 during March–May 2020, documenting differences between patients with medically attended outpatient and inpatient illness regarding demographic characteristics, baseline underlying chronic conditions, symptoms, and exposures that could be used to target public health interventions. In addition, among symptomatic respondents, inpatients and outpatients with COVID-19 reported similar numbers of symptoms, but different types of symptoms as previously described.§ Thus, a range of symptoms should prompt testing for SARS-CoV-2. The wide range of symptoms reported, and the lack of known COVID-19 contact in 54% of patients, underscores the need for isolation of infected persons, contact tracing and testing during ongoing community transmission, and prevention measures including social distancing and use of cloth face coverings.

15Exposures to SARS-CoV-2 have commonly been described in congregate settings rather than broader community settings.

16In a multistate telephone survey of 350 adult inpatients and outpatients who tested positive for SARS-CoV-2 infection, only 46% reported recent contact with a COVID-19 patient. Most participants’ contacts were a family member (45%) or a work colleague (34%). Two thirds of participants were employed; only 17% were able to telework.

17Case investigation, contact tracing, and isolation of infected persons are needed to prevent ongoing community transmission, given the frequent lack of a known contact. Enhanced measures to ensure workplace safety, including social distancing and more widespread use of cloth face coverings, are warranted.