Unraveling COVID-19: A Large-Scale Characterization of 4.5 Million COVID-19 Cases Using CHARYBDIS.

Purpose: Routinely collected real world data (RWD) have great utility in aiding the novel coronavirus disease (COVID-19) pandemic response. Here we present the international Observational Health Data Sciences and Informatics (OHDSI) Characterizing Health Associated Risks and Your Baseline Disease In SARS-COV-2 (CHARYBDIS) framework for standardisation and analysis of COVID-19 RWD. Patients and
Methods: We conducted a descriptive retrospective database study using a federated network of data partners in the United States, Europe (the Netherlands, Spain, the UK, Germany, France and Italy) and Asia (South Korea and China). The study protocol and analytical package were released on 11th June 2020 and are iteratively updated via GitHub. We identified three non-mutually exclusive cohorts of 4,537,153 individuals with a clinical COVID-19 diagnosis or positive test, 886,193 hospitalized with COVID-19, and 113,627 hospitalized with COVID-19 requiring intensive services.
Results: We aggregated over 22,000 unique characteristics describing patients with COVID-19. All comorbidities, symptoms, medications, and outcomes are described by cohort in aggregate counts and are readily available online. Globally, we observed similarities in the USA and Europe: more women diagnosed than men but more men hospitalized than women, most diagnosed cases between 25 and 60 years of age versus most hospitalized cases between 60 and 80 years of age. South Korea differed with more women than men hospitalized. Common comorbidities included type 2 diabetes, hypertension, chronic kidney disease and heart disease. Common presenting symptoms were dyspnea, cough and fever. Symptom data availability was more common in hospitalized cohorts than diagnosed.
Conclusion: We constructed a global, multi-centre view to describe trends in COVID-19 progression, management and evolution over time. By characterising baseline variability in patients and geography, our work provides critical context that may otherwise be misconstrued as data quality issues. This is important as we perform studies on adverse events of special interest in COVID-19 vaccine surveillance.

Introduction

The World Health Organization (WHO) declared the coronavirus disease 2019 (COVID-19) pandemic on 11 March 2020 after 118,000 reported cases in over 110 countries.5 By the end of 2021, the number of COVID-19 cases increased to over 278 million cases globally, and the death toll exceeded 5 million.6 Thousands of publications have attempted to aid our scientific understanding of this public health emergency.7,8

Characterisation studies, called descriptive epidemiology, provide an important context into our understanding of disease by describing the basic attributes of who gets sick and in what context. The initial body of COVID-19 characterisation work gave researchers information on the stark difference in the perception of the novel coronavirus compared to flu-like illnesses: patients were male, younger, and with fewer concurrent comorbidities and less documented prior medication use.9

Utilising routinely collected real world data (RWD) can be a powerful asset for understanding an evolving pandemic response.1,2 Each data source provides novel information, be it the geographic variability of COVID-19, the impact of varying government strategies to contain spread or the evolution of treatment protocols. With extensive heterogeneity in public health strategies and clinical care across the world,10 a large repeated multi-center study to describe disease across locations, practices, and populations, but that holds data analysis constant would go far in determining what factors impact observed differences.

RWD networks are vital in helping to understand the magnitude of the problem, and developing possibly mitigating strategies both globally and locally.11,12 Here we present the global Observational Health Data Sciences and Informatics (OHDSI) community, an international open-science initiative of more than 3500 collaborators from 34 countries, response to the COVID-19 pandemic.3 Founded in 2015, the OHDSI data network enabled a rapid baseline understanding of COVID-19 in emerging hotspots (United States of America [USA], Spain and South Korea).9 Our work evolved into a systematic framework for analysing and reporting COVID-19 RWD that we call Characterizing Health Associated Risks, and Your Baseline Disease In SARS-COV-2 (CHARYBDIS).

CHARYBDIS offers multiple insights into COVID-19 clinical presentations, management and progression. Herein we aim to describe baseline demographics, clinical characteristics, treatments received, and outcomes among individuals diagnosed and hospitalized with COVID-19 in actual practice settings in nine countries from three continents. These data reflect an international community of research collaborators who are working to advance retrospective database research in RWD for COVID-19. Our body of research is freely available, foundational result set that can provide benchmarks in how COVID-19 manifests over time including its inevitable evolution as we roll-out additional vaccines and treatments.

Methods

Study Design, Setting and Data Sources

We conducted a descriptive retrospective database study using a federated network of data partners in the USA, Europe (the Netherlands, Spain, the UK, Germany, France and Italy) and Asia (South Korea and China). Each data partner mapped their source system to the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM).13–15 The use of a CDM ensured shared conventions, including consistent representation of clinical terms across coding systems. We assessed the plausibility, conformance and completeness of each contributing database using a common data quality tool for repeated assessment and monitoring the adherence to conventions across the network.16,17 We ensured technical reproducibility by using the same package of analytical code for all contributing data partners.18

The study protocol and analytical package were released on 11 June 2020 and iterative updates have continued to be released via GitHub: .4 23 real world healthcare databases contributed to the CHARYBDIS study (). Contributing institutes ranged from major academic medical centers to small community hospitals from across three continents. Date capture ranged from December 2019 to as recent as January 2021 (site specific dates in ). Prior to performing these analyses, all the data partners obtained Institutional Review Board (IRB) or equivalent governance approval. Each data partner executed the study package locally on their OMOP CDM. Only aggregate results from each database were publicly shared. Minimum cell sizes were determined by institutional protocols. All data partners consented to the external sharing of the result set on data.ohdsi.org.

Study Population and Follow-Up

We focused on three non-mutually exclusive COVID-19 cohorts: i) diagnosed with COVID-19 (a positive SARS-CoV-2 laboratory test or clinical diagnosis code documenting COVID-19 - earliest event served as the index date); ii) hospitalized with COVID-19 and; iii) hospitalized with COVID-19 and requiring intensive services. Due to variability in access to diagnostic testing, we specifically looked for the presence of a PCR or antigen laboratory test OR the use of clinical diagnosis codes documenting COVID-19 presentation.19 The codes used to identify cohorts and more detail on the definitions of the above cohorts can be found in . These cohorts were generated both with a requirement of at least 365 days of data availability prior to the index date, and without any requirement for prior observation time. Databases created specifically for COVID-19 tracking may be unable to support extensive lookback periods and thus, we used multiple definitions to ensure inclusiveness in our approach. Cohorts were followed from their cohort-specific index date to the earliest of death, end of the observation period, and up to 30 days post-index.

Stratifications

Each cohort was analyzed by the overall study population and stratified by additional available characteristics including: follow-up time; socio-demographics, baseline comorbidities, pregnancy status (yes/no), and flu-like symptom episodes (yes/no). Detailed definitions of each stratification are available in .

Baseline Characteristics, Symptoms, Medication Use and Outcomes of Interest

Information on socio-demographics was identified at or before baseline (index date). All conditions, symptoms and medications were identified and described at four different time intervals (1 year prior, 30 days prior, at index and up to 30 days after index). The definition of each symptom and outcome is provided in .

Statistical Analysis

We built this analysis using Health Analytics Data-to-Evidence Suite (HADES), a set of open source R packages for large scale analytics.20 Proportions, standard deviations (SD), and standardized mean differences (SMD) within each subgroup were tabulated as pre-specified in our study protocol. This analysis was descriptive in nature with the explicit intention of building an initial, repeatable framework for constructing prevalent rates of disease. Only cohorts or stratified sub-cohorts with a minimum sample size of 140 subjects were characterized. This cut-off was deemed necessary to estimate with sufficient precision the prevalence of a previous condition or 30-day risk of an outcome affecting ≥10% of the study population. SMDs were plotted in Manhattan-style plots, a type of scatter plot designed to visualize large data with a distribution of higher-magnitude values. Scatter plots were also created to compare the described conditions, symptoms and demographics of patients diagnosed (Y axis) to those hospitalized (X axis) with COVID-19.

Results

Patient Characteristics

Overall, we identified three non-mutually exclusive cohorts of 4,537,153 individuals with a clinical COVID-19 diagnosis or positive test, 886,193 hospitalized with COVID-19, and 113,627 hospitalized with COVID-19 requiring intensive services (Figure 1). Of these, the cohorts including patients with the requirement of at least of 365 days before index: 3,279,518 with a clinical COVID-19 diagnosis or laboratory positive test, 636,810 hospitalized with COVID-19, and 63,636 hospitalized with COVID-19 requiring intensive services ( and ).

Figure 1

COVID-19 cases across the OHDSI COVID-19 network.

Geographic Distribution

The USA data partners contributed 96% of the diagnosed with COVID-19 cohorts, including the single largest diagnosed cohort from IQVIA Open Claims (n=2,785,812). Europe contributed 4% of the diagnosed with COVID-19 cohorts, owing the single largest regional diagnosed cohort to SIDIAP-Spain (n=124,305). Asia contributed less than 1% of diagnosed with COVID-19 cohorts, with the single largest regional diagnosed cohort contributed from Daegu Catholic University Medical Center (n=599).

Demographic Distribution

In the USA, the proportion of diagnosed cases generally decreased with age, with most diagnosed cases being within the 25 to 60 age group. The proportion of cases hospitalized and intensive services increased with age, with the highest proportions of cases of hospitalized, or intensive cases in the 60 to 80 year age group (Figure 2). A slightly higher proportion of women were diagnosed than men but a greater proportion of men were hospitalized (and where available, required intensive services) than women in the USA databases. In Europe, databases captured diagnosed or hospitalized cohorts but had limited information on intensive services. In Europe, databases capturing hospitalized cases (HMAR, HM-Hospitales, SIDIAP, and SIDIAP-H) showed a similar trend to the USA databases in that there was a higher proportion of men were hospitalized than women (). Unlike the USA and European databases, there was also a higher proportion of women in hospitalized cases in the South Korean database (HIRA). Age-wise trends in the European and Asian databases were similar to those in the USA databases, in that the bulk of the diagnosed cases were in the 25 to 60 year age group, whilst the majority of the hospitalized cases were in the 60 to 80 year age group ().

Figure 2

Distribution of diagnosed, hospitalized and requiring intensive services COVID-19 cases by age and sex across the OHDSI COVID-19 network in the United States.

Comorbidities

Overall, the proportion of patients with type 2 diabetes mellitus, hypertension, chronic kidney disease, end stage renal disease, heart disease, malignant neoplasm, obesity, dementia, auto-immune condition, chronic obstructive pulmonary disease (COPD), and asthma was higher in the hospitalized cohort as compared to the diagnosed (Tables 1 and 2). Data on tuberculosis, human immunodeficiency viruses (HIV), and hepatitis C infections were sparse, and where available the proportions were generally low (≤1%). In the US databases, the proportion of pregnant women was generally higher in the hospitalized cohort than in the diagnosed, but not so in two European databases (HM and SIDIAP). The remaining five European and one of the Asian databases had data on pregnant women only in the hospitalized cohort, the proportion of which was < 2%.

Table 1

Characteristics of Persons with a COVID-19 Diagnosis or SARS-CoV-2 Positive Test Across the OHDSI COVID-19 Network*

	Asia		United States													Europe
	DCMC	NFHCRD	HealthVerity	Premier	OPTUM-EHR	OPTUM-SES	STARR-OMOP	TRDW	VA-OMOP	IQVIA-OpenClaims	IQVIA Hospital CDM	CUIMC	CU-AMC-HDC	UWM-CRD	OHSU	SIDIAP	IPCI	CPRD	IQVIA LPD France	IQVIA DA Germany	IQVIA LPD Italy
COVID-19 Cases (N)	559	403	587,683	66,132	160,613	7863	4788	1250	57,937	2,785,812	153,477	10,437	9481	3245	11,187	124,305	3306	3864	23,592	11,500	4816
Persons Tested	NR	397	3,898,593	219,230	1,025,584	41,673	56,881	6950	521,814	6,520,151	719,596	22,094	120,661	83,921	109,434	173,957	NR	5551	NR	NR	NR
Tested Positive, n (%)*	NR	392 (97.3)	425,610 (72.4)	NR	73,113 (45.5)	NR	1880 (39.3)	1035 (82.8)	32,847 (56.7)	NR	NR	6959 (66.7)	NR	3,140 (96.8)	8764 (78.3)	39,047 (31.4)	NR	2098 (54.3)	NR	NR	NR
Full 30-day follow up	162 (29.0)	276 (68.5)	67,071 (11.4)	3902 (5.9)	84,073 (52.3)	1269 (16.1)	2703 (56.5)	641 (51.3)	44,661 (77.1)	1,882,950 (67.6)	21,145 (13.8)	2008 (19.2)	8755 (92.3)	1,199 (36.9)	3760 (33.6)	81,914 (65.9)	2601 (78.7)	2723 (70.5)	9819 (41.6)	5588 (48.6)	3570 (74.1)
< 30-day follow up	397 (71.0)	127 (31.5)	520,612 (88.6)	62,230 (94.1)	76,540 (47.7)	6594 (83.9)	2085 (43.5)	609 (48.7)	13,272 (22.9)	902,862 (32.4)	132,332 (86.2)	8429 (80.8)	706 (7.4)	2046 (63.1)	7427 (66.4)	42,391 (34.1)	705 (21.3)	1141 (29.5)	13,773 (58.4)	5912 (51.4)	1246 (25.9)
Comorbidities, n (%)**
Type 2 Diabetes Mellitus	108 (19.3)	9 (2.2)	20,922 (3.6)	10,783 (16.3)	26,897 (16.7)	2673 (34.0)	555 (11.6)	179 (14.3)	19,083 (32.9)	724,991 (26.0)	35,576 (23.2)	1977 (18.9)	1396 (14.7)	391 (12.0)	603 (5.4)	9941 (8.0)	500 (15.1)	545 (14.1)	1318 (5.6)	1089 (9.5)	452 (9.4)
Hypertension	154 (27.5)	19 (4.7)	34,090 (5.8)	19,008 (28.7)	54,678 (34.0)	4393 (55.9)	1319 (27.5)	307 (24.6)	34,357 (59.3)	1,260,816 (45.3)	60,495 (39.4)	3771 (36.1)	2708 (28.6)	735 (22.7)	1065 (9.5)	21,337 (17.2)	688 (20.8)	779 (20.2)	3522 (14.9)	2611 (22.7)	1659 (34.4)
Heart disease	106 (19.0)	7 (1.7)	19,016 (3.2)	11,533 (17.4)	39,510 (24.6)	3726 (47.4)	977 (20.4)	245 (19.6)	24,699 (42.6)	936,271 (33.6)	33,846 (22.1)	3236 (31.0)	1871 (19.7)	440 (13.6)	778 (7.0)	17,759 (14.3)	470 (14.2)	722 (18.7)	1213 (5.1)	2007 (17.5)	1013 (21.0)
History of cancer	32 (5.7)	NR	6107 (1.0)	3157 (4.8)	18,536 (11.5)	1491 (19.0)	887 (18.5)	106 (8.5)	10,792 (18.6)	317,479 (11.4)	11,237 (7.3)	1480 (14.2)	843 (8.9)	184 (5.7)	469 (4.2)	8872 (7.1)	262 (7.9)	296 (7.7)	674 (2.9)	661 (5.7)	547 (11.4)
Hepatitis C	NR	NR	740 (0.1)	410 (0.6)	1395 (0.9)	112 (1.4)	61 (1.3)	35 (2.8)	3075 (5.3)	40,101 (1.4)	1966 (1.3)	144 (1.4)	90 (0.9)	54 (1.7)	88 (0.8)	648 (0.5)	NR	NR	40 (0.2)	31 (0.3)	53 (1.1)
Obesity	29 (5.2)	NR	15,072 (2.6)	7298 (11.0)	71,076 (44.3)	2468 (31.4)	1246 (26.0)	325 (26.0)	25,128 (43.4)	740,430 (26.6)	28,757 (18.7)	3729 (35.7)	3136 (33.1)	233 (7.2)	945 (8.4)	36,557 (29.4)	629 (19.0)	1428 (37.0)	2287 (9.7)	1345 (11.7)	674 (14.0)
Dementia	6 (1.1)	NR	4255 (0.7)	3697 (5.6)	5360 (3.3)	851 (10.8)	38 (0.8)	29 (2.3)	4019 (6.9)	219,062 (7.9)	7776 (5.1)	483 (4.6)	235 (2.5)	116 (3.6)	97 (0.9)	6013 (4.8)	64 (1.9)	327 (8.5)	55 (0.2)	339 (2.9)	81 (1.7)
Autoimmune condition	49 (8.8)	NR	7291 (1.2)	1678 (2.5)	13,396 (8.3)	1464 (18.6)	418 (8.7)	133 (10.6)	10,103 (17.4)	433,259 (15.6)	8965 (5.8)	1388 (13.3)	720 (7.6)	140 (4.3)	409 (3.7)	8260 (6.6)	476 (14.4)	394 (10.2)	1467 (6.2)	1183 (10.3)	636 (13.2)
Chronic obstructive pulmonary disease (COPD) without asthma	NR	NR	8160 (1.4)	3335 (5.0)	12,067 (7.5)	1449 (18.4)	231 (4.8)	89 (7.1)	12,665 (21.9)	297,269 (10.7)	12,008 (7.8)	809 (7.8)	733 (7.7)	112 (3.5)	249 (2.2)	15,819 (12.7)	213 (6.4)	294 (7.6)	696 (3.0)	868 (7.5)	350 (7.3)
Asthma without COPD	17 (3.0)	NR	10,458 (1.8)	3972 (6.0)	21,076 (13.1)	1125 (14.3)	521 (10.9)	112 (9.0)	6278 (10.8)	438,892 (15.8)	12,936 (8.4)	1376 (13.2)	1100 (11.6)	176 (5.4)	567 (5.1)	7567 (6.1)	322 (9.7)	494 (12.8)	2327 (9.9)	1097 (9.5)	420 (8.7)
Pregnant women	NR	NR	3543 (0.6)	1192 (1.8)	3917 (2.4)	109 (1.4)	52 (1.1)	27 (2.2)	86 (0.1)	41,329 (1.5)	2944 (1.9)	382 (3.7)	212 (2.2)	32 (1.0)	156 (1.4)	689 (0.6)	32 (1.0)	11 (0.3)	212 (0.9)	39 (0.3)	63 (1.3)
Chronic kidney disease broad	156 (27.9)	NR	7535 (1.3)	5711 (8.6)	17,531 (10.9)	1829 (23.3)	398 (8.3)	NR	10,239 (17.7)	364,857 (13.1)	16,250 (10.6)	1181 (11.3)	723 (7.6)	213 (6.6)	277 (2.5)	8144 (6.6)	197 (6.0)	478 (12.4)	194 (0.8)	562 (4.9)	192 (4.0)
End stage renal disease	155 (27.7)	NR	1683 (0.3)	1062 (1.6)	3008 (1.9)	359 (4.6)	122 (2.5)	NR	3273 (5.6)	96,555 (3.5)	5155 (3.4)	600 (5.7)	166 (1.8)	51 (1.6)	52 (0.5)	8 (0.0)	NR	17 (0.4)	NR	27 (0.2)	NR
Human immunodeficiency virus infection	NR	NR	829 (0.1)	357 (0.5)	763 (0.5)	67 (0.9)	20 (0.4)	NR	817 (1.4)	24,808 (0.9)	1309 (0.9)	163 (1.6)	56 (0.6)	45 (1.4)	43 (0.4)	290 (0.2)	NR	NR	83 (0.4)	18 (0.2)	19 (0.4)

Notes: *Proportions presented among diagnosed patients with a COVID-19 diagnosis or SARS-CoV-2 positive test by database (column percentage); since SIDIAP_H includes a subset of SIDIAP, results were not included in this table; - data not available or below the minimum cell count required (5 individuals); no prior observation time was required. **Prevalent conditions at index date.

Abbreviations: CU-AMC-HDC, U of Colorado Anschutz Medical Campus Health Data Compass; CUIMC, Columbia University Irving Medical Center; IQVIAHospitalCDM, IQVIA Hospital Charge Data Master; OHSU, Oregon Health and Science University; OPTUM-EHR, Optum© de-identified Electronic Health Record Dataset; OPTUM-SES, Optum® De-Identified Clinformatics® Data Mart Database – Socio-Economic Status (SES); STARR-OMOP, Stanford Medicine Research Data Repository; TRDW, Tufts Research Data Warehouse; UWM-CRD, UW Medicine COVID Research Dataset; VA-OMOP, Department of Veterans Affairs; NR, not reported by data partner.

Table 2

Characteristics of Persons Hospitalized with a COVID-19 Diagnosis or SARS-CoV-2 Positive Test Across the OHDSI COVID-19 Network*

	Asia		United States													Europe
	HIRA	NFHCRD	HealthVerity	Premier	OPTUM-EHR	OPTUM-SES	STARR-OMOP	TRDW	VA-OMOP	IQVIA OpenClaims	IQVIA Hospital CDM	CUIMC	CU-AMC-HDC	UWM-CRD	OHSU	HM Hospitales	SIDIAP	HMAR
COVID-19 Cases (N)	7599	304	22,887	36,019	29,061	4336	744	326	10,951	533,997	57,062	3439	1874	733	627	2544	18,369	2686
Hospitalized with positive test, n (%)	NR	125 (41.1)	13,262 (57.9)	NR	13,817 (47.5)	NR	128 (17.2)	232 (71.2)	8623 (78.7)	NR	NR	3075 (89.4)	NR	676 (92.2)	344 (54.9)	NR	13,685 (74.5)	773 (28.8)
Full 30-day follow up	7359 (96.8)	284 (93.4)	10,333 (45.1)	2361 (6.6)	18,555 (63.8)	851 (19.6)	657 (88.3)	NR	8548 (78.1)	412,537 (77.3)	11,876 (20.8)	943 (27.4)	1810 (96.6)	400 (54.6)	484 (77.2)	109 (4.3)	12,290 (66.9)	1254 (46.7)
< 30-day follow up	240 (3.2)	20 (6.6)	12,554 (54.9)	33,658 (93.4)	10,506 (36.2)	3485 (80.4)	87 (11.7)	NR	2400 (21.9)	121,460 (22.7)	45,186 (79.2)	2496 (72.6)	64 (3.4)	333 (45.4)	143 (22.8)	2435 (95.7)	6079 (33.1)	1432 (53.3)
Comorbidities, n (%)**
Type 2 Diabetes Mellitus	1760 (23.2)	NR	3880 (17.0)	8899 (24.7)	9531 (32.8)	1844 (42.5)	157 (21.1)	83 (25.5)	5839 (53.3)	254,505 (47.7)	16,480 (28.9)	1120 (32.6)	677 (36.1)	226 (30.8)	177 (28.2)	428 (16.8)	3295 (17.9)	294 (10.9)
Hypertension	1943 (25.6)	NR	6410 (28.0)	15,216 (42.2)	16,427 (56.5)	2977 (68.7)	389 (52.3)	123 (37.7)	9087 (83.0)	390,171 (73.1)	26,262 (46.0)	1770 (51.5)	1073 (57.3)	398 (54.3)	283 (45.1)	1139 (44.8)	5645 (30.7)	653 (24.3)
Heart disease	1271 (16.7)	NR	5178 (22.6)	10,384 (28.8)	13,274 (45.7)	2634 (60.7)	297 (39.9)	109 (33.4)	7421 (67.8)	319,842 (59.9)	16,165 (28.3)	1534 (44.6)	802 (42.8)	286 (39.0)	258 (41.1)	606 (23.8)	5148 (28.0)	362 (13.5)
History of cancer	410 (5.4)	NR	1132 (4.9)	2811 (7.8)	4939 (17.0)	1065 (24.6)	277 (37.2)	47 (14.4)	3401 (31.1)	106,805 (20.0)	5524 (9.7)	588 (17.1)	300 (16.0)	90 (12.3)	154 (24.6)	286 (11.2)	2616 (14.2)	179 (6.7)
Hepatitis C	61 (0.8)	NR	134 (0.6)	394 (1.1)	469 (1.6)	77 (1.8)	17 (2.3)	13 (4.0)	1037 (9.5)	14,408 (2.7)	1050 (1.8)	81 (2.4)	37 (2.0)	25 (3.4)	37 (5.9)	15 (0.6)	135 (0.7)	38 (1.4)
Obesity	16 (0.2)	NR	2238 (9.8)	6678 (18.5)	15,497 (53.3)	1626 (37.5)	312 (41.9)	126 (38.7)	5677 (51.8)	191,071 (35.8)	10,735 (18.8)	1651 (48.0)	988 (52.7)	138 (18.8)	167 (26.6)	149 (5.9)	8428 (45.9)	259 (9.6)
Dementia	436 (5.7)	NR	1815 (7.9)	3428 (9.5)	2376 (8.2)	637 (14.7)	17 (2.3)	17 (5.2)	2087 (19.1)	81,638 (15.3)	4044 (7.1)	373 (10.8)	140 (7.5)	95 (13.0)	25 (4.0)	108 (4.2)	1102 (6.0)	75 (2.8)
Autoimmune condition	813 (10.7)	NR	1215 (5.3)	1432 (4.0)	3320 (11.4)	931 (21.5)	89 (12.0)	54 (16.6)	3156 (28.8)	136,735 (25.6)	4205 (7.4)	570 (16.6)	226 (12.1)	67 (9.1)	83 (13.2)	121 (4.8)	1706 (9.3)	93 (3.5)
Chronic obstructive pulmonary disease (COPD) without asthma	145 (1.9)	NR	2213 (9.7)	3016 (8.4)	5176 (17.8)	1066 (24.6)	102 (13.7)	52 (16.0)	4641 (42.4)	118,421 (22.2)	7071 (12.4)	469 (13.6)	333 (17.8)	77 (10.5)	90 (14.4)	173 (6.8)	4848 (26.4)	138 (5.1)
Asthma without COPD	1560 (20.5)	NR	1004 (4.4)	2677 (7.4)	3746 (12.9)	628 (14.5)	127 (17.1)	39 (12.0)	1153 (10.5)	82,087 (15.4)	3825 (6.7)	498 (14.5)	245 (13.1)	58 (7.9)	101 (16.1)	112 (4.4)	957 (5.2)	99 (3.7)
Pregnant women	121 (1.6)	NR	341 (1.5)	682 (1.9)	1550 (5.3)	30 (0.7)	18 (2.4)	13 (4.0)	NR	12,748 (2.4)	2029 (3.6)	158 (4.6)	111 (5.9)	22 (3.0)	73 (11.6)	7 (0.3)	108 (0.6)	20 (0.7)
Chronic kidney disease broad	421 (5.5)	NR	2622 (11.5)	5339 (14.8)	6596 (22.7)	1357 (31.3)	162 (21.8)	NR	3958 (36.1)	164,710 (30.8)	8827 (15.5)	691 (20.1)	375 (20.0)	152 (20.7)	112 (17.9)	157 (6.2)	2658 (14.5)	186 (6.9)
End stage renal disease	30 (0.4)	NR	826 (3.6)	948 (2.6)	1506 (5.2)	296 (6.8)	31 (4.2)	NR	1520 (13.9)	53,747 (10.1)	3333 (5.8)	371 (10.8)	101 (5.4)	43 (5.9)	29 (4.6)	7 (0.3)	NR	91 (3.4)
Human immuno-deficiency virus infection	NR	NR	96 (0.4)	275 (0.8)	222 (0.8)	33 (0.8)	NR	NR	239 (2.2)	7009 (1.3)	516 (0.9)	73 (2.1)	14 (0.7)	11 (1.5)	11 (1.8)	NR	47 (0.3)	14 (0.5)

Notes: *Proportions presented among diagnosed patients with a COVID-19 diagnosis or SARS-CoV-2 positive test by database (column percentage); - data not available or below the minimum cell count required (5 individuals); no prior observation time was required. **Prevalent conditions at index date.

Abbreviations: CU-AMC-HDC, U of Colorado Anschutz Medical Campus Health Data Compass; CUIMC, Columbia University Irving Medical Center; IQVIAHospitalCDM, IQVIA Hospital Charge Data Master; OHSU, Oregon Health and Science University; OPTUM-EHR, Optum© de-identified Electronic Health Record Dataset; OPTUM-SES, Optum® De-Identified Clinformatics® Data Mart Database – Socio-Economic Status (SES); STARR-OMOP, Stanford Medicine Research Data Repository; TRDW, Tufts Research Data Warehouse; UWM-CRD, UW Medicine COVID Research Dataset; VA-OMOP, Department of Veterans Affairs; HM-Hospitales, HM-Hospitales Madrid; SIDIAP, Information System for Research in Primary Care; HMAR, Hospital del Mar; NR, not reported by data partner.

Other Analyses

Dyspnea, cough, and fever were the most common symptoms in diagnosed and hospitalized cohorts globally (). Where recorded, the proportion of dyspnea and malaise/fatigue was consistently higher in the hospitalized cohort as compared to the diagnosed. Anosmia/hyposmia/dysgeusia was present in less than 1% individuals in all but one database and more common in the diagnosed than the hospitalized cohorts ().

We further described a total of 19,222 conditions and 2973 medications registered during the year prior to the index date (). The same information is also described for 30 days prior to the index date, at index date, or during the first 30 days after index date (–) The full result set of comorbidities, presenting symptoms, medications and outcomes are reported by each cohort in aggregate counts, and are available in an interactive website: .

Discussion

CHARYBDIS is the world’s largest open science aggregate result set aimed at describing the baseline demographics, clinical characteristics, treatments received, and outcomes among individuals diagnosed and hospitalized with COVID-19. To accomplish this, we aggregated over 22,000 unique characteristics creating a multi-centre view to describe trends in COVID-19 progression, management and evolution over time. Globally, we observed similarities in the USA and Europe in gender (more women diagnosed than men but more men hospitalized than women) and age (most diagnosed cases between 25–60 years of age versus most hospitalized cases between 60–80 years of age) distributions. Similar to previous studies, we observed South Korea differed with more women than men hospitalized. We found similarities in comorbidities and presenting symptoms. The large, diverse sample size allows also for the identification of populations of great interest, including children and adolescents,25 pregnant women,26 patients with a history of cancer,27 patients with a history of autoimmune disorders,28 or patterns of drug utilization in COVID-19 treatment,21 and which were the focus of additional in-depth investigations.

Summary of Key Findings

We described characteristics of 4,537,153 individuals with a clinical COVID-19 diagnosis or positive test, 886,193 hospitalized with COVID-19, and 113,627 hospitalized with COVID-19 requiring intensive services from 9 countries. Up to 22,200 unique aggregate characteristics have been produced across databases, with all made publicly available in an accompanying website. The evidence framework is a method for systematically understanding cohort-level differences in COVID-19 from different regions and different points in the pandemic. In the months since we started this effort, our network has already aided in rapid study for coagulopathy and adverse of events of special interest for COVID-19 vaccines to inform regulatory bodies.22 This research community can be a public health utility to guide in 1) better patient characterization and stratification, 2) identifying areas of gap in knowledge/evidence, and 3) generating hypotheses for future research.

Comparison to Other Multi-Centre COVID-19 Consortia

We began our deep phenotyping work through an initial investigation of persons hospitalized with COVID-19 compared to prior flu seasons in our global federated network.9

The National COVID Cohort Collaborative (N3C) is a NIH NCATS funded initiative collecting centralizing patient-level data to study patterns in COVID-19 patients.23 This effort has over 80 participating institutions contributing 4.5M COVID-19 patients to date to a centralized harmonized repository. The consortia has enabled many US institutions in adoption of common data models in COVID-19 research. 4CE is another multi-site data-sharing collaborative of 342 hospitals in the US and in Europe, utilizing i2b2 or OMOP data models.24 The hospitalization cohorts presented in 4CE cohorts remain smaller than the scope of CHARYBDIS with only 36,447 hospitalized patients with COVID-19 as of August 2020.24 Even when adjusting for cohort overlap, our work to date with CHARYBDIS is nearly triple the diagnosis and double the hospitalized cohorts represented in prior research. Our results also have more international representation across the cascade of hotspots over the course of the pandemic’s spread. As we continue our research, we are working with researchers to create inpatient-outpatient linkages and understand COVID-19 patient trajectories across care settings.

Study Strengths

Our study has several strengths. This study is unique in its approach to characterizing COVID-19 cases across an international network of healthcare systems with varied policies enacted to combat this pandemic. This allows better understanding of the implications of the pandemic for different countries and regions, in the context of an international comparison. Particularly, it provides visibility into the variability of patient characteristics across healthcare settings. This study is the most comprehensive federated network of healthcare sites in the world, creating the single largest cohort study on diagnosed and hospitalized COVID-19 cases to date. The large, diverse sample size allows for extensive investigation on subgroups of interest. CHARYBDIS is the framework for additional in-depth investigations on children and adolescents,25 pregnant women,26 patients with a history of cancer,27 patients with a history of autoimmune disorders,21 or patterns of drug utilization in COVID-19 treatment.21 The size of these results are so large, we have hundreds combinations of subgroups of interest that remain unreported. There is significant opportunity for this framework to inform additional research.

Study Limitations

We recognize there are limitations in our approach. First, this study is descriptive in nature. Further analyses are needed to utilize these findings in clinical application. The observed differences between groups (eg diagnosed versus hospitalized) should therefore not be interpreted as causal effects without further statistical scrutiny. Answering causal questions is especially difficult in COVID-19 because of the varying processes by which patients were screened, tested, admitted, and treated; the critical importance of knowing the exact timing of treatments and outcomes in severe cases; and the lack of appropriate comparison groups. Simple multivariable models by themselves will not sufficiently address bias for multiple questions and were purposely not applied here. This study was carried out using data recorded in routine clinical practice and based on electronic health records (EHRs) and/or claims data. The analysed data are therefore expected to be incomplete in some respects and may have erroneous entries, leading to potential misclassification. We have selectively reported database-specific outcomes to minimise the impact of incompleteness. We are aware that this may mean the network assembled is not inherently valuable for every follow-on analysis as each data partner may have different elements missing. Hospital encounters may be unable to ascertain outcomes experienced in an outpatient data. Our EHR partners rely on structured data and may be missing key findings from clinical notes. Additionally, the under-reporting of symptoms observed in these data is a key finding of this study, and should be taken into consideration in previous and future similar reports from “real world” cohorts. Differential reporting in different databases is likely a function of differential coding practice as well as of variability in disease severity, with milder/less symptomatic cases more likely presenting in outpatient and primary care EHR, and more severe ones in hospital databases. Finally, the current result submissions are prejudiced to data in the initial wave of COVID-19 cases. Further analysis using this network requires stratification by calendar month. Lastly, we currently lack data partners in low to middle income countries and recognize these data are lacking representation of some of the hardest hit areas in the world (eg Brazil, India). As data are accumulated over time, future updates of the results will provide the opportunity to study more recent cohorts of COVID-19 patients, who seem to have a better prognosis overall compared to those diagnosed in the first half of the pandemic.

Conclusion

We constructed a global, multi-centre view to describe trends in COVID-19 progression, management and evolution over time. By characterising baseline variability in demographics across geography, our work provides critical context to the reliability of the insight we generate. In retrospective database studies, one can struggle to identify whether heterogeneity occurs because of patient variability or because of the variability in source systems we use to capture patient data. Here we use a network of retrospective databases standardised to the same data model adhering to a shared ontology and data quality processes. Our study provides a comprehensive view into the first year of the pandemic at a scale unlike most retrospective research. Our work sheds light on the natural history of millions of COVID-19 patients from the USA, 6 European countries and 2 Asian countries. This framework is open source and available for re-use enabling a repeatable, reproducible method to capture the evolving natural history of this novel coronavirus and can be extended to other disease of international interest. We believe it is critically important to repeat and reproduce the findings we produce in real world studies. Leveraging this global federated network to corroborate single center findings can provide context to national database findings in the presence of regional variability in COVID-19 management including vaccine rollout and treatments.