Comprehensive Code List Associated with Underlying Medical Conditions Identified by the CDC as High-Risk Factors for Progression to Severe COVID-19 Outcomes

Abstract

A list of diagnosis codes mapped to CDC-defined high-risk conditions for severe COVID-19 outcomes is currently not available in the literature. We reviewed the CDC list of underlying conditions associated with severe COVID-19 and a coding expert and two clinicians mapped the relevant high-risk conditions to the appropriate ICD-10-CM codes. We additionally assessed the prevalence of these conditions in the Optum de-identified-Clinformatics^® Data Mart Database and the Optum de-identified Electronic Health Record dataset. A comprehensive list of approximately 8200 codes were mapped to the CDC-defined high-risk underlying conditions; these ICD-10-CM codes were stratified into three groups corresponding with the CDC strength of evidence category (conclusive, suggestive, or mixed evidence). Applying these codes to administrative claims and EHR datasets demonstrated a consistent prevalence of high-risk conditions over four years (2018–2021). These findings present a comprehensive list of codes that can be used by clinicians and researchers to identify and characterize patients at high risk for severe COVID-19 outcomes.

1. Introduction

As the coronavirus disease 2019 (COVID-19) virus continues to evolve, a need remains to understand underlying conditions that represent a high risk for progression to severe COVID-19 illness and death. The CDC has identified age over 50 years and several underlying medical conditions as high-risk factors for progression to severe illness [1]. Not being vaccinated or up to date with COVID-19 vaccination also increases the risk of severe illness [1]. In addition, patients aged 50–64 years had a death risk ratio of 4.3, and patients aged 65–74 had a death risk ratio of 6.7 [2]. Racial and ethnic minority groups are also at disproportionate risk of being affected by COVID-19 [3]. Furthermore, patients with one underlying condition had a death risk ratio of 1.5, while those with 2–5 conditions had a death risk ratio of 2.6 [4].

The availability of a systematically generated list of ICD-10-CM codes will enable health services researchers to implement a consistent definition of patients at high risk of severe COVID-19 outcomes in real-world data sources. Additionally, the use of this code list to identify high-risk cohorts may support population health decision-making in developing informed vaccination and treatment strategies in a real-world setting.

The CDC has developed a list of high-risk conditions with the first list published on 31 May 2021. The CDC periodically updates this list as new evidence emerges [3]. This study was conducted to develop the first comprehensive list of ICD-10-CM diagnosis codes for underlying medical conditions associated with high risk of developing severe COVID-19.

2. Materials and Methods

In this study, the CDC’s high-risk underlying conditions were obtained from CDC publications and corresponding references [3]. These conditions were mapped to ICD-10-CM codes by an ICD-10 coding expert (LH) [5]. Additional ICD-10-CM codes that the coding expert deemed as related to the conditions were further examined by two clinicians (JA and FD) who determined if the codes were clinically aligned with the literature and if the inclusion of those codes was warranted.

The resulting comprehensive list of codes for CDC-defined high-risk conditions was applied to Optum’s de-identified Clinformatics^® Data Mart Database and Optum’s^® de-identified Electronic Health Record dataset from 1 January 2018 to 31 December 2021, on a yearly basis (2018, 2019, 2020, 2021) to determine performance and consistency across four years of U.S. data. The Clinformatics^® Database consists of medical, pharmacy, and death data from 70M+ lives across the United States (US). The Optum^® EHR data encompass patient-level electronic health records from 105M+ lives across the US. Since the CDC considers age ≥ 50 as a risk factor, data were stratified across the following age groups: 12–49 and ≥50 years old. Data provided by Optum to third parties are de-identified and provisions are in place to prevent re-identification to protect patients’ confidentiality. This study is exempt from additional IRB approval because it is retrospective and non-interventional, and it used anonymized data from The Clinformatics^® Database.

3. Results

Using CDC references, along with clinical and coding expert assessments, a total of 8211 codes were mapped to the CDC list of high-risk conditions (

Table 1. ICD-10 code list for underlying conditions/characteristics associated with higher risk for severe COVID-19.

Conditions/Characteristics	Number of ICD-10 Codes
Conclusive High Risk
Pregnancy (females aged < 44)	2520
Cancer-active or history	1464
Disability	1234
Stroke and cerebrovascular disease	686
Diabetes Mellitus	470
Heart conditions a	417
Chronic lung diseases b	313
Immunocompromised state c	163
Mood disorders	132
Chronic kidney disease (CKD) at any stage	129
Solid organ or blood stem transplant (includes bone marrow transplants)	83
Dementia	82
Tuberculosis	66
Chronic liver disease d	22
Smoking; current or former	12
HIV infection	8
Suggestive Higher Risk
Substance use disorders e	265
Sickle cell disease	37
Overweight and obesity	34
Mixed Evidence
Hypertension	66
Thalassemia	8

^a Heart failure, coronary artery disease, or cardiomyopathies. ^b Moderate to severe asthma, bronchiectasis, bronchopulmonary dysplasia, chronic obstructive pulmonary disease, emphysema and chronic bronchitis, interstitial lung disease, pulmonary fibrosis, cystic fibrosis, and pulmonary embolism. ^c Primary caused by genetic defects, secondary/acquired from prolonged use of corticosteroids or other immune weakening medicines. ^d Cirrhosis; non-alcoholic fatty liver disease; alcoholic liver disease; autoimmune hepatitis. ^e Alcohol, opioid, or cocaine use disorder.

). The complete list of codes is described in Supplementary Materials Table S1.

Despite age being considered a high-risk factor, our analysis did not include ≥50 years as a characteristic associated with high risk for severe COVID-19. In the 2021 calendar year, a total of 12,809,403 (93.1%) patients ages 12 years and older were identified in The Clinformatics^® Database, with 4,661,642 patients aged 12–49 years, and 8,147,761 patients aged ≥ 50 years. Among these, 1,792,771 (38.5%) and 6,245,388 (76.7%) had at least one characteristic or condition associated with high risk for severe COVID-19 across ages 12–49 years and ≥ 50 years, respectively. This was similar and consistent with the years 2018 to 2020 (Supplementary Materials Tables S2 and S3).

Among the 12–49 years cohort, the five most prevalent conditions reported in the claims database (

Table 2. Prevalence of CDC-defined underlying medical conditions associated with higher risks for severe COVID-19 in 2021 (The Clinformatics^® Database).

Cohort	N (%)
Ages 12+ years and continuously enrolled in The Clinformatics® Database during calendar year 2021	12,809,403
	12–49 Years Old	≥50 Years Old
Total number of patients	4,661,642	8,147,761
	N (%)	N (%)
At least 1 characteristic or condition during calendar year associated with a higher risk for severe COVID	1,792,771 (38.45)	6,245,388 (76.65)
Obesity and Overweight	698,827 (14.99)	4,640,392 (56.95)
Mood Disorders	558,107 (11.97)	2,603,997 (31.96)
Hypertension	404,603 (8.68)	2,253,393 (27.66)
Disability	355,326 (7.62)	1,993,866 (24.47)
Cancer-Active and History	184,403 (3.96)	1,799,700 (22.09)
Diabetes Mellitus	171,975 (3.69)	1,316,315 (16.16)
Smoking	165,492 (3.55)	1,291,967 (15.86)
Pregnancy (Females aged < 44)	164,102 (3.52)	1,243,098 (15.26)
Heart Conditions a	149,376 (3.20)	1,169,069 (14.35)
Immunocompromised State b	110,014 (2.36)	1,164,822 (14.30)
Lung Disease c	81,714 (1.75)	738,042 (9.06)
Substance Use Disorder d	65,755 (1.41)	705,150 (8.65)
Chronic Liver Disease e	62,182 (1.33)	292,602 (3.59)
Stroke and Cerebrovascular Disease	34,648 (0.74)	213,127 (2.62)
Chronic Kidney Disease at Any Stage	29,728 (0.64)	79,510 (0.98)
HIV	9011 (0.19)	29,722 (0.36)
Sickle Cell Disease and Thalassemia	7204 (0.15)	21,361 (0.26)
Tuberculosis	6276 (0.13)	15,681 (0.19)
Solid Organ or Stem Cell Transplant	5541 (0.12)	12,956 (0.16)
Dementia	324 (0.01)	0 (0.00)

^a Heart failure, coronary artery disease, or cardiomyopathies; ^b primary caused by genetic defects, secondary/acquired from prolonged use of corticosteroids or other immune weakening medicines; ^c moderate to severe asthma, bronchiectasis, bronchopulmonary dysplasia, chronic obstructive pulmonary disease, emphysema and chronic bronchitis, interstitial lung disease, pulmonary fibrosis, cystic fibrosis, pulmonary embolism; ^d alcohol, opioid, or cocaine use disorder; ^e cirrhosis; non-alcoholic fatty liver disease; alcoholic liver disease; or autoimmune hepatitis.

) were overweight and obesity (15.0%), mood disorders (12.0%), hypertension (8.9%), disabilities (7.6%), and active or history of cancer (4.0%). Similarly, among those ≥50 years, the most prevalent conditions (Table 2) observed were hypertension (57.0%), heart conditions (32.0%), obesity and overweight (27.7%), diabetes mellitus (24.5%), and active or history of cancer (22.1%). This was similar and consistent with the years 2018 to 2020 (Supplementary Materials Tables S2 and S3).

Similar results were observed in the Optum^® EHR dataset and within other stratified age groups (12–17, 18–49, 50–64, and 65+ years old) (Supplementary Materials Tables S4–S7).

4. Discussion

This study developed a list of ICD-10-CM codes for use by clinicians and researchers to define cohorts and identify patients at risk for severe COVID-19 outcomes. The list of codes was developed systematically and identified consistent disease trends across multiple years of claims and EHR data (Supplementary Materials Table S2). The list of codes is aligned with CDC references to ensure the relationship between the codes and underlying conditions related to severe COVID-19 is appropriate.

Understanding high-risk conditions among age groups allows the identification of patients who may be at high risk for progression to severe illness and supports appropriate clinical decision-making. In a previous study, several coexisting medical conditions were associated with an elevated risk of serious outcomes from COVID-19, including death. These comorbid conditions were obesity, diabetes with complications, chronic kidney disease, chronic obstructive pulmonary disease/bronchiectasis, neurocognitive disorders, and coronary atherosclerosis [4]. Identification of these conditions by applying our code list to a large administrative claims database has demonstrated similar results. When applying our code list to claims data, we observed a greater percentage of patients ≥50 years old with at least one high-risk condition compared to other ages. The stratification of age groups in our datasets determined the most prevalent conditions among patients 12–49 years and ≥50 years, considering age ≥ 50 years is a single risk factor for developing severe COVID-19 [3].

This study sought to develop an explorative comprehensive list of ICD-10 CM codes for underlying conditions identified by the CDC as being associated with a high risk of developing severe COVID-19. This code list can be implemented as is or adapted to be fit for purpose depending on the research question or patient population. It is up to the researcher/clinician to choose the most practical and relevant codes for a given health system or database. The performance of this code list was not validated, given the explorative nature of our study. Our study is limited given the variability of documentation and coding/billing practices across different institutions [5]. Furthermore, based on differing institutional practices, certain ICD-10 codes for the identified comorbidities may be under and/or over-estimated, resulting in misclassification bias [5]. Consistent and precise coding of comorbidities is needed to appropriately classify conditions based on ICD-10 codes. Our study focused on ICD-10 codes in EHR and claims data, additional limitations exist given that our study did not consider lab test results and the type of medication prescribed that may be related to certain conditions not captured using ICD-10 codes alone. Within our list, there are overlapping codes for more than one condition. While appropriate when assessing an individual disease condition, this needs to be taken into consideration when assessing the prevalence of all high-risk conditions. This is especially true for patients that have an immunocompromised state or cancer. Additionally, during the COVID-19 pandemic older adults experienced declines in healthcare resource utilization such as outpatient visits, which may have resulted in a variation in prevalence for the conditions observed in this study [6]. The list of codes is up-to-date as of 1 April 2023 (based on the 9 February 2023 CDC update). However, it will require revision when new ICD-10 codes are introduced, or as the CDC guidelines are refreshed.

Lastly, while hypertension was reported as one of the most prevalent high-risk conditions among all ages, it is important to recognize that the CDC places this in the mixed evidence category in relation to developing severe COVID-19 outcomes [3]. Due to the evolving nature of COVID-19 and the CDC’s list of high-risk conditions, additional codes may be added or deleted in future updates to this code list as new findings become available. Also, per CDC the list of underlying medical conditions is not exhaustive and may be updated with new available evidence.

5. Conclusions

This study systematically summarizes a comprehensive list of ICD-10-CM codes that can be used to identify and characterize patients with underlying conditions that put them at risk for severe COVID-19 outcomes. In conclusion, the CDC states that the list of underlying medical conditions is not all-inclusive and may undergo updates based on emerging evidence.