Impact of COVID-19 on work loss in the United States- A retrospective database analysis

Abstract

Objectives

This study investigates the utilization of work absence benefits among United States (US) employees diagnosed with COVID-19, examining frequency, duration, cost, and types of work loss benefits used.

Methods

This retrospective analysis of the Workpartners Research Reference Database (RRDb) included employees eligible for short- and long-term disability (STD and LTD employer-sponsored benefits, respectively), and other paid work absence benefits from 2018 to 2022. Workpartners RRDb includes over 3.5 million employees from over 500 self-insured employers across the US. Employees were identified by codes from adjudicated medical and disability claims for COVID-19 (2020–2022) and influenza, as well as prescription claims for COVID-19 treatments. Associated payments were quantified for each absence reason.

Results

Approximately 1 million employees were eligible for employer-sponsored paid leave benefits between January 2018 and December 2022. The mean age was 37 years (22% >50 years), and 49.4% were females. COVID-19 was the 2nd most common reason for an STD claim (6.9% of all STD claims) and 13th for an LTD claim (1.7% of all LTD claims) from 2020–2022. The mean duration for COVID-19 STD claims was 24 days (N = 3,731, mean claim=$3,477) versus 10 days for influenza (N = 283, mean claim=$1,721). The mean duration for an LTD claim for COVID-19 was 153 days (N = 11, mean claim=$19,254). Only 21.5% of employees with STD claims in the COVID-19 cohort had prior COVID-19-associated medical or pharmacy claims; over half (range 53%–61%) had documented high risk factors for severe COVID-19.

Conclusion

COVID-19 and influenza have the potential to cause work loss in otherwise healthy employees. In this analysis, COVID-19 was the second most frequent reason for an STD claim at the start of the pandemic and remained high (ranked 5th) in 2022. These results highlight the impact of COVID-19 on work loss beyond the acute phase. Comprehensively evaluating work loss implications may help employers prioritize strategies, such as vaccinations and timely treatments, to mitigate the impact of COVID-19 on employees and their companies.

PLAIN LANGUAGE SUMMARY

COVID-19 results in short- and long-term symptoms that may affect employees’ ability to work. Short- and long-term disability (STD and LTD, respectively), other work absences, and medical and pharmacy claims from the Workpartners Research Reference Database were analyzed for US adult (≥18 years) employees. COVID-19 claims were identified using the Center for Disease Control and Prevention recommended International Classification of Diseases codes during the analysis from 2020 to 2022. During 2020 to 2022, COVID-19 ranked as the second most frequent reason for STD claims and 13th most frequent among LTD claims. Influenza ranked 58th overall with no LTD claims (2018–2022). The average COVID-19 STD claim lasted 24 days and cost employers $3,477 per claim, and LTD claims averaged 153 days, costing $19,254. Only 21.5% of employees with STD claims in the COVID-19 cohort had prior COVID-19-associated medical or pharmacy claims, and over half (range 53%–61%) had a documented high-risk factor for severe COVID-19. Our results highlight the ongoing and substantial impact of COVID-19 on work absence benefit utilization beyond the acute phase. This analysis demonstrates the need for employers and researchers to review all available medical, pharmacy, and disability claims to assess the acute and long-term impact of COVID-19 on employees and prioritize mitigation strategies to reduce the burden of the virus to their employees.

Keywords:

• COVID-19
• long COVID-19
• influenza
• work absence
• disability claims
• employment

JEL CLASSIFICATION CODES:

• I10
• J32

Introduction

Since Coronavirus disease 2019, or COVID-19, an infectious disease caused by the SARS-CoV-2 virus, was declared to be a global pandemic in March 2020, over 774 million COVID-19 cases, including 103.4 million cases in the United States (US) have been reported to the World Health Organization (WHO)¹. According to the Centers for Disease Control and Prevention (CDC), 59.7% of the US population has had COVID-19, and 6.8% of adults reported being affected by continued symptoms of COVID-19² that significantly impacted their quality of life and ability to work^3–5.

Evidence from systematic reviews examining long-term COVID symptoms during the initial waves of the pandemic revealed that 38%–72% of individuals have at least one symptom for more than two months from the onset of COVID-19, and up to 54% experienced persistent symptoms for six months or more^6–9. Both acute and long-term COVID-related health challenges can lead to utilization of disability insurance, workplace absence, and decreased productivity, with only 27.3% able to return to work at the same capacity as before COVID-19 and 45.2% requiring a reduced work schedule¹⁰.

A comprehensive analysis of work absences claims data in the US describing the changes in benefit utilization since the start of the pandemic had not been conducted. A recent targeted literature review identified a gap in information regarding the impact of COVID-19 across various job-types¹¹. The current analysis aims to explore the evolving impact of COVID-19 on paid work absence benefit utilization, and quantify the frequency and duration of work absences due to acute and long-term effects of COVID-19 among US employees. The assessment is based on work loss claims and payments associated with short-term disability (STD), long-term disability (LTD), workers’ compensation (WC), family medical leave (FML), and discretionary/incidental time off (DITO). Secondary aims include assessing medical and prescription costs for employees identified with COVID-19, and unadjusted comparisons of COVID-19 work-related absences to an influenza cohort. Influenza was chosen as a benchmark for comparison based on the similarities of acute symptoms, as well as the opportunity to leverage data associated with an existing endemic disease state as COVID-19 exits the pandemic stage.

Methods

Study design

This retrospective descriptive database analysis of health utilization trends and work loss examined data from 2018 to 2022 medical, pharmacy, and work absence claims from de-identified US employees and a subset of employees with COVID-19 and influenza in the Workpartners Research Reference Database (RRDb). The Workpartners RRDb captures data on almost 4 million employees working for self-funded, employers across the US. Patient characteristics, employment information, and work absences were extracted for analysis.

Patient selection

Outcomes were evaluated for three cohorts: any employee in the RRDb with eligibility for any of the absence benefits (overall population), and the subsets of these eligible employees who had COVID-19 or influenza, as identified through medical and/or work loss claims. Employees with COVID-19 were also identified by medication prescription claims for COVID-19 medications. Prescription data were not used to identify employees with influenza because prescription medications used to treat influenza can also be used for prophylaxis following exposure. Employees with COVID-19 and influenza were identified by applicable International Classification of Diseases (ICD) codes from medical, STD, or LTD claims and by COVID-19 prescription treatments once they became available (2021 to 2022). The date of initial COVID-19 or influenza medical, STD, or LTD claim or COVID-19 prescription medication claim was defined as the employee’s index date. Employees were required to be 18 years or older, eligible and enrolled in the specified benefit. Employees were excluded from analyses if benefit claims data were not provided by the employer, or if the date of birth was missing (<0.01%) to ensure that all individuals were at least 18 years of age. Disability claims where the benefit programs provided either non-specific or non-standard reasons were reported as “Unknown.”

Cohort selection involved evaluating all patients with COVID-19 (2020–2022) or influenza (2018–2022), regardless of the length of health plan enrollment surrounding the index date. Examining data from 2018 and 2019 established a baseline of work absence claims data before the onset of the COVID-19 pandemic. This approach provided the necessary context to assess the impact of COVID-19 on work absence outcomes relative to influenza. Additionally, data from 2018 and 2019 were evaluated to examine influenza trends prior to the COVID-19 pandemic, since mitigation measures aimed at limiting the spread of COVID-19, such as mask-wearing, hand hygiene, and social distancing, resulted in reduced rates of influenza during the pandemic¹².

The WHO established U07.1 as the ICD-10 code to identify COVID-19 in April 2020, with reporting guidelines from the CDC available shortly after. The CDC had provided interim guidance on the coding of COVID-19 cases prior to April 2020¹³. U09.9 (post COVID-19 condition, unspecified) became effective in October of 2021 (Supplementary Table 1A). Following standard practice, these codes (U07.1 and U09.9) were utilized in medical claims to identify employees with COVID-19.

Disability insurers began to use several non-specific coronavirus codes to identify COVID-19 disability claims and continued to utilize these even after the introduction of U07.1 in 2020. There were six ICD-9 and ICD-10 diagnosis codes (sourced only from “Disability claims” in Supplementary Table 1A), which had not been utilized in disability claims prior to March 2020, and exhibited a marked increase after (Supplementary Figure 1). Therefore, all eight ICD diagnosis codes listed in Supplementary Table 1A were utilized in STD and LTD claims to identify employees with COVID-19. Similarly, due to the utilization of ICD-9 codes in disability claims data, both ICD-9 and ICD-10 codes were utilized for identifying influenza disability claims (Supplementary Table 1B).

Outcomes measured

The following outcomes were measured annually for each STD, LTD, (non-work related illness) WC (work related illness or injury), and FML leave reason: the number of claims in the calendar year after the index date, claims per 10,000 full-time-equivalent employee-year (FTEEY), the percentage of employees eligible for the benefit associated with a claim, the percentage of all claims from that benefit for the given reason. Additionally, the number of closed claims (claims no longer incurring absence days or paid amounts) were quantified, as well as the duration and payments associated with these closed claims. Non-working days such as weekends or holidays are not included in the length of the closed claim. The same measures for each leave reason were analyzed for the combined-year study periods of 2020–2022 for each population and 2018–2022 for the overall population. Additionally, the number of medical or pharmacy claims from employees with a diagnosis or prescription for the treatment of COVID-19 on or before the start of the absence claim was evaluated, along with the time from the start of the most recent episode of COVID-19 to the start of the absence claim.

To provide more granular detail on these categories of interest for STD and LTD, leave reasons based on individual ICD diagnosis codes were categorized using the 283 Agency for Healthcare Research and Quality’s (AHRQ) categories¹⁴ in addition to the COVID-19 and influenza cohort definitions. WC leave reasons were categorized using 64 standard WC nature of injury codes¹⁵. FML reasons were reported by commonly used employer database categories including: Own (illnesses of the employee), Family (to care for a family member), Pregnancy (for employee pregnancy), Military (for military leave), and Other¹⁶.

For DITO, the following outcomes were measured for each leave reason (DITO values are limited to administrative, sick or paid time off): the total number of lost work days, lost work days per FTEEY, DITO payments per FTEEY, the percentage of employees eligible for DITO that had a DITO leave, the percentage of all DITO lost days for the given reason, and the percentage of employees using DITO one week prior to two weeks after the start of the employee’s most recent episode of COVID-19. Time (days) from the start of the most recent medical/pharmacy of COVID-19 claim to the start of the absence claim were calculated for employees in the COVID-19 cohort but not in the overall employee population. DITO reasons were limited to those that potentially could be used during an illness per employer policies and included Sick (for an employee illness), Paid Time Off (PTO) (for any reason), and Administrative (recommended by some employers for use during a COVID-19 episode).

Absence benefits

Various employee work absence benefits were examined, including STD, LTD, WC, DITO, and FML. Based on typical policies found in the RRDb, STD is generally utilized by full-time employees for non-work-related illnesses or injuries for up to six months duration. LTD is also usually accessible to full-time employees and is used for non-work-related illnesses or injuries, following the STD portion of the claim. WC is available to all employees and is designated for work-related illnesses and injuries. Both LTD and WC benefits can span over multiple years. DITO is often limited to full-time employees, encompassing sick leave for non-work-related illnesses, PTO for any reason, and administrative leave for jury duty, bereavement leave, or other reasons. FML is a federally mandated non-paid benefit that is generally used to protect employment during the employee’s illness or the illness of a family member, care for a newborn or adopted child, or military leave. Additional details are provided in Supplementary Table 2.

Direct costs

Direct overall and COVID-19-specific medical and prescription costs were evaluated for those in the COVID-19 cohort with six months of continuous eligibility before and after their index date. Costs were inflation-adjusted to December 2022 dollars using components of the Consumer Price Index (CPI) from the US Bureau of Labor Statistics. Medical costs used the medical CPI, prescription costs prescription CPI, and salary and payments for absences used the general CPI.

Statistical analysis

Categorical and binary variables were summarized using frequencies and percentages. Continuous variables were summarized using means and standard errors (SE). Student’s t-tests evaluated differences in continuous variables using SAS version 9.4 (SAS Institute, Inc, Cary, North Carolina).

Results

Overall, 1,266,323 employees were identified between January 2018 and December 2022, with 757,523 employees eligible and enrolled in medical and pharmacy reimbursement plans, and 989,098 were eligible and enrolled in work absence benefits. Of those, 96.3% of employees were eligible for WC, 77.7% FML, 43.3% STD, 32.8% LTD, and only 9.5% DITO (Table 1). The mean age for the overall 2018–2022 population calculated on December 31 of each employee’s first year in the RRDb was 37.0 (SE 0.01) years, and 49.4% were females. The mean age for the COVID-19 cohort at their index date was 43.7 (SE 0.05) years, and 52.6% were females. The mean age for the influenza cohort at their index date was 41.6 (SE 0.08) years, and 50.4% were females. More employees in the overall population live in the East North Central region of the United States (17.8%) than the other nine regions. Retail Trade (28.1%) and Health Care and Social Assistance (26.4%) were the industries with the most employees in the database. In the overall population, the average annual salary was $63,429, and 58.0% were full time employees. Further baseline characteristics for the overall population and the COVID-19 cohort by year are listed in Table 2. Clinical characteristics available through medical and pharmacy claims data for severe COVID-19 risk factors are reported in Supplementary Table 3 and show that over half (range 53%–61%) had a documented high-risk factor. The percent of employees in the influenza cohort with at least one risk factor for severe COVID-19 ranged from 53% − 57%.

Table 1. Attrition criteria.

Overall employee population	(2018–2022)	2020	2021	2022
	N	N	N	N
All employees from 01 January 2018 to 31 December 2022	1,266,323	691,840	719,936	735,305
Age > =18	1,255,180	687,054	715,184	729,348
Eligible for any work loss benefit	989,098	531,129	555,985	579,630
Eligible for medical/pharmacy benefits	757,523	436,030	427,140	417,358
Eligible for STD	428,756	262,103	269,663	279,523
Eligible for LTD	323,982	195,413	207,113	208,749
Eligible for WC	952,562	505,479	527,647	557,970
Eligible for FML	768,184	414,742	441,387	449,315
Eligible for DITO	94,408	64,852	69,906	71,719
COVID-19 cohort	(2020–2022)	2020	2021	2022
	N	N	N	N
Identified through any source	72,539	14,560	23,815	34,164
Identified through medical claim	64,421	13,043	22,844	28,534
Identified through pharmacy claim	5,005	0	0	5,005
Identified through STD claim	3,105	1,515	965	625
Identified through LTD claim	8	2	6	0
COVID-19 with work loss eligibility	(2020–2022)	2020	2021	2022
	N	N	N	N
COVID-19 and eligible for any work loss benefit	56,447	11,147	28,144	52,292
COVID-19 and eligible for STD	37,634	7,332	18,336	34,932
COVID-19 and eligible for LTD	26,919	4,555	12,599	25,165
COVID-19 and eligible for WC	54,210	10,913	26,734	50,223
COVID-19 and eligible for FML	45,353	8,529	22,952	42,572
COVID-19 and eligible for DITO	10,085	1,279	4,250	9,909

Abbreviations. STD, short-term disability; LTD, long-term disability; WC, workers' compensation; FML, family medical leave; DITO, discretionary/incidental time off.

Table 2. Cohort demographics and baseline information.

Characteristics		2018–2022		2020–2022
		Influenza	Overall	COVID-19	Overall
Employee count (N)		24,199	1,255,180	72,539	1,010,051
Age continuous
	Mean (SE)	41.6 (0.08)	37.0 (0.01)	43.7 (0.05)	37.9 (0.01)
Age
	18–49 years	71.8%	78.3%	64.4%	74.9%
	50–64 years	26.2%	19.7%	32.6%	21.3%
	65+ years	2.0%	1.9%	3.0%	2.5%
Sex
	Male	49.4%	49.3%	47.3%	48.9%
	Female	50.4%	49.4%	52.6%	50.0%
	Unknown	0.2%	1.3%	0.14%	1.1%
Marital status
	Married	18.2%	12.4%	13.1%	12.2%
	Single	18.9%	27.3%	15.2%	25.0%
	Other/unknown	62.9%	60.3%	71.7%	62.8%
Race/ethnicity
	White	14.5%	10.4%	12.8%	10.6%
	Black	2.9%	2.6%	3.1%	2.8%
	Asian	2.1%	1.1%	1.4%	1.3%
	Hispanic	6.5%	3.6%	6.5%	4.1%
	Other	1.2%	0.7%	1.2%	0.9%
	Unknown	72.9%	81.5%	75.0%	80.4%
Geographic region (first zip code digit)
	New England (0)	6.0%	7.7%	6.8%	7.7%
	East North Central (1)	14.1%	17.8%	21.9%	19.1%
	Mid-Atlantic (2)	15.7%	12.2%	11.1%	11.8%
	South-East (3)	11.7%	8.3%	10.1%	8.5%
	East Central (4)	5.5%	6.1%	7.2%	6.1%
	North Central (5)	2.3%	3.9%	2.4%	3.7%
	Central (6)	5.0%	5.4%	4.2%	5.2%
	South Central (7)	19.3%	10.4%	13.0%	10.3%
	Mountain (8)	11.4%	13.1%	12.0%	12.7%
	Pacific (9)	8.6%	13.4%	10.6%	13.6%
	Unknown	0.4%	1.7%	0.8%	1.5%
Tenure (years)
	Mean (SE)	7.3 (0.05)	3.9 (0.01)	8.1 (0.03)	4.8 (0.01)
Status
	Full time	74.5%	58.0%	75.9%	60.2%
	Part time	1.1%	20.1%	1.2%	18.7%
	Unknown	24.3%	21.9%	22.9%	21.0%
Compensation (annual salary)
	Mean (SE)	$87,354 ($451)	$63,429 ($67)	$86,331 ($278)	$66,621 ($76)
Pay grade
	<40K	10.4%	33.9%	9.4%	31.2%
	40–89.9K	38.3%	26.4%	40.7%	28.2%
	90 + K	27.1%	15.8%	25.8%	17.5%
	Unknown	24.1%	23.9%	24.1%	23.1%
Job type
	Exempt	26.8%	16.5%	26.1%	17.9%
	Non-exempt	45.9%	58.1%	46.7%	56.7%
	Unknown	27.3%	25.4%	27.2%	25.4%
Industry type (NAICS code)
	Utilities (22)	4.9%	3.0%	6.1%	3.4%
	Construction (23)	0.6%	0.2%	0.4%	0.3%
	Manufacturing (31–33)	20.5%	13.0%	16.8%	12.5%
	Retail trade (44–45)	13.8%	28.1%	12.4%	27.6%
	Transportation and warehousing (48–49)	1.4%	1.9%	1.7%	1.7%
	Information (51)	0.3%	0.2%	0.2%	0.3%
	Finance and insurance (52)	10.9%	8.9%	9.8%	10.2%
	Professional, scientific, and technical services (54)	1.0%	1.4%	0.4%	0.8%
	Administration and support and waste management and remediation services (56)	9.0%	7.6%	9.2%	7.8%
	Educational services (61)	5.4%	3.3%	6.5%	3.4%
	Health care and social assistance (62)	25.1%	26.4%	31.4%	26.9%
	Other services (except public administration) (81)	0.1%	0.2%	0.2%	0.2%
	Other	7.0%	5.9%	5.1%	4.9%

Overall population

Short-term disability (STD)

There were 56,695 STD claims filed during the period of 2020–2022. COVID-19 was the second most frequent reason for STD claims among employees during that period, and in 2020 (9.7% of all STD claims). COVID-19 accounted for 6.7% of all STD claims in 2021, and 4.3% in 2022. Influenza had declining percentages of the total STD claims for each individual year analyzed, particularly during the pandemic years, with influenza accounting for 0.6% of total claims in 2018, 0.4% in 2019, 0.3% in 2020, 0.1% in 2021, and 0.2% in 2022. COVID-19 STD claims (identified during the 2020–2022 period) were more frequent than influenza claims (2018–2022) (4.4% vs 0.3% of all STD claims) and 2020–2022 period (6.9% vs 0.2%). Compared to the mean duration of 43 days from 56,361 all-cause closed STD claims during 2020–2022, the mean duration of COVID-19 STD claims was 24 days and 10 days for influenza. The mean cost of the 55,013 closed all-cause STD claims with payment data during 2020–2022 was $6,784. The mean cost for COVID-19 STD claims was $3,477, and influenza was $1,354. Other top STD claim reasons are listed in Table 3.

Table 3. Top five reasons for STD claims in the overall population.

Year	STD reason	Rank	No. of claims	Claims per 10,000 FTEEY	% of total STD claims	Mean duration of closed claims (days)	Mean cost of closed claims^a ($)
2020–2022	Other pregnancy and delivery including normal	1	7,440	102.2	13.1%	37	$6,773
	COVID-19	2	3,916	53.8	6.9%	24	$3,477
	Other complications of birth; puerperium affecting management of mother	3	3,031	41.6	5.3%	49	$9,032
	Mood disorders	4	2,600	35.7	4.6%	53	$9,814
	Spondylosis; intervertebral disc disorders; other back problems	5	2,519	34.6	4.4%	57	$7,973
2020	Other pregnancy and delivery including normal	1	2,417	102.0	13.0%	39	$7,123
	COVID-19	2	1,814	76.6	9.7%	21	$3,075
	Other complications of birth; puerperium affecting management of mother	3	1,025	43.3	5.5%	51	$8,840
	Spondylosis; intervertebral disc disorders; other back problems	4	872	36.8	4.7%	52	$7,436
	Mood disorders	5	809	34.2	4.3%	53	$8,862
2021	Other pregnancy and delivery including normal	1	2,521	104.7	12.7%	36	$6,636
	COVID-19	2	1,322	54.9	6.7%	28	$3,816
	Other complications of birth; puerperium affecting management of mother	3	982	40.8	4.9%	49	$8,459
	Mood disorders	4	965	40.1	4.9%	53	$9,874
	Anxiety disorders	5	890	37.0	4.5%	52	$10,189
2022	Other pregnancy and delivery including normal	1	2,502	99.8	13.7%	36	$6,578
	Other complications of birth; puerperium affecting management of mother	2	1,024	40.9	5.6%	48	$9,763
	Mood disorders	3	826	33.0	4.5%	54	$10,662
	Spondylosis; intervertebral disc disorders; other back problems	4	793	31.6	4.4%	59	$8,369
	COVID-19	5	780	31.1	4.3%	24	$3,828

Abbreviations. STD, short-term disability; FTEEY, full-time-equivalent employee-year.

^aExcludes claims with missing payment information.

Long-term disability (LTD)

Compared to STD claims, there were 1,436 LTD claims from 252,069 LTD eligible employees recorded during the period of 2020–2022. Of the 760 closed LTD claims, the mean duration was 187 days. The mean cost of for 720 LTD claims with available payment data was $20,638. The most common diagnosis for LTD claims was Spondylosis; Intervertebral Disc Disorders; Other Back Problems, with 156 claims, a mean duration of 213 days, and mean costs of $17,036. Other prevalent LTD diagnoses observed were musculoskeletal and behavioral health diagnoses. COVID-19 was the most common respiratory diagnosis, and the 13th most common overall cause for LTD claims (N = 24) (Table 4). Notably, COVID-19 had a shorter average duration than the most common reason: Spondylosis; intervertebral Disc Disorders; Other Back Problems with a difference of 60 days), and COVID-19 claims incurred a higher mean cost ($2,218 difference). Additionally, COVID-19 ranked as the 20th most common LTD claims diagnosis in 2020 but moved to 11th in 2021 and 12th in 2022. Of note, a diagnosis of influenza did not result in any LTD claims.

Table 4. COVID-19 LTD and WC claims in the overall population.

Year	COVID-19 rank among claim reasons	No. of claims	Claims per 10,000 FTEEY	% of total claims	Mean duration of closed claims (days)	Mean cost of closed claims^a ($)
LTD
2020–2022	13	24	0.4	1.7%	153	$19,254
2020	20	5	0.3	1.0%	296	$25,358
2021	11	13	0.7	2.6%	53	$6,204
2022	12	6	0.3	1.4%	119	$39,673
WC
2020–2022	16	36	0.3	0.5%	46	$29,981
2020	9	23	0.6	1.1%	53	$29,163
2021	16	12	0.3	0.5%	38	$34,311
2022	28	1	0.0	0.0%	2	$375

Abbreviations. LTD, long-term disability; WC, workers’ compensation; FTEEY, full-time-equivalent employee-year.

^aExcludes claims with missing payment information.

Workers’ compensation (WC)

There was a total of 6,598 WC claims from 769,450 WC eligible employees during 2020–2022. COVID-19 ranked 16th overall (Table 4), with a steady decrease in counts over time (23, 12, and 1 claim for 2020, 2021, and 2022, respectively). The mean duration was 46 days and mean cost of $29,981.

Family medical leave (FML)

It is not possible to discern if FML leaves were due to COVID-19. The distribution of FML reasons was stable across the 2020 to 2022 time period. FML claims resulted in 111,010 total claims during 2020–2022 from 634,522 distinct FML eligible employees. Of the total claims, 57.8% of FML claims were for the employee’s own illness, with a mean of 23 days absent.

Discretionary/incidental time off (DITO)

Exact reasons for the changes seen in “Sick day” utilization are typically not available because of the lack of detailed information provided by employers and because of differences in policies between employers. Overall, employers paid $2.29 billion to 84,049 distinct employees for 5,118,830 work days associated with DITO from 2020 to 2022. PTO was used by over 93% of employees overall, (3,698,657 work days, $1.79 billion/20.1 work days and $9,751 per FTEEY). DITO for illness or “Sick days” were used by 52% of eligible employees and resulted in 1,023,926 lost work days, $335.4 million total costs, or 5.6 days and $1,825 per FTEEY. The percent of employees taking DITO for “Administrative leaves” increased from 28% to 46% from 2019 to 2020. Relatively few Administrative days were taken prior to 2020 (0.4–0.5 per FTEEY), compared to 3.8 Administrative days per FTEEY taken in 2020. Notably, payments for administrative leave increased eight-fold from 2019 to 2020 ($203 to $1,611 per FTEEY).

COVID-19 cohort

Of the 72,539 employees within the COVID-19 cohort, 56,447 were eligible for any employer sponsored work loss benefits in the 2020–2022 period. The majority (∼90%) were indexed based on a medical claim; however, 10.4% of the cohort were identified by an STD claim without a prior medical claim for COVID-19 in 2020, 4.1% in 2021 and 1.8% in 2022. In 2022, COVID-19 cohort members were older, had higher annual salaries, and were less likely to be eligible for overtime pay (non-exempt status) than in 2020 and 2021. Relative to the overall employee population, employees with COVID-19 were most concentrated in select geographic regions (Table 2).

Short-term disability (STD)

There were 37,634 employees identified as having COVID-19 who were also eligible for STD benefits. Almost 10% (3,728) filed 3,916 COVID-19-related STD claims in 2020–2022 accounting for 43.8% of the total STD claims in the COVID-19 cohort. In comparison 1% of the influenza cohort had an STD claim for influenza in 2020–2022 and 2% from 2018 to 2022. Out of 3,916 COVID-19 STD claims reported within this cohort, only 842 (21.5%) claims had a prior COVID-19-associated medical or pharmacy claim. Additionally, among the 842 COVID-19 STD claims linked to a prior COVID-19 medical or pharmacy claim, the median STD claim began within one day (75% started within six days) of the start of the employee’s most recent COVID-19 diagnosis. The mean duration of closed COVID-19 STD claims increased from 21 days in 2020 to 28 days in 2021, and then decreased to 24 days in 2022. Of note, the duration of closed claims exhibited a wide range from 1 day to 260 days. Furthermore, the mean duration of closed STD claims varied by ICD code, ranging from a mean of 18 days (B972) to 65 days (U09.9). The mean cost from closed claims with pay data (N = 3,731) was $3,477.

Long-term disability (LTD)

Compared to 8,932 STD claims, there were 175 LTD claims from 26,919 distinct employees (0.65%) in the COVID-19 cohort. COVID-19 was the most common LTD diagnosis claim for the 2020–2022 period, with 13.7% of total LTD claims. There were 24 total LTD claims during the study period; 13 claims (54%) were still open at the time of this analysis. The mean duration of closed claims was 153 days, ranging from 1 day to 709 days. Of the 11 closed claims with pay data, the mean cost was $19,254. COVID-19 was also the top LTD claim diagnosis for 2020 and 2021, and ranked third in 2022. Notably, although COVID-19 ranked as the third in 2022, it had the highest mean cost at $39,673 from two closed claims with pay data compared to $25,358 from four claims in 2020 and $6,204 from five claims in 2021. The ICD codes B97.29 and U07.1 are the most frequent codes for COVID-19 LTD claims. Similar to STD claims, 29.17% of COVID-19 LTD claims had a prior COVID-19-associated medical or pharmacy claim. As reported with the overall population, there were no LTD claims for influenza in the COVID-19 cohort.

Workers’ compensation (WC)

There was a total of 383 WC claims from 2020 to 2022. Only three WC indemnity claims in 2020–2022 cited COVID-19 as the reason (mean duration of 38 days, mean cost $5,642). All three were initiated within six days of the start of their most recent COVID-19 infection identified medical claims.

Family medical leave (FML)

Over 10.2% of employees in the COVID-19 cohort had an FML for their own illness during 2020–2022, and 4.0% had an FML absence due to the illness of a family member. The median time from the onset of the most recent COVID-19 episode to the initiation of the FML was at least 124 days across all FML reasons in the 2020–2022 time period. A clinical diagnosis for a corresponding leave is unknown.

Discretionary/incidental time off (DITO)

Throughout the 2020–2022 period, among 10,085 eligible employees, absences taken one week prior to two weeks after the most recent COVID-19 episode resulted in 44,685 lost workdays at a cost of $19.3 million ($432/day). Notably, 41% of those days were classified as Administrative, while 35% were designated as PTO. PTO and Sick day absences were used by an increasing percentage of employees each year from 2020 to 2022, while the percentage of employees taking administrative leaves decreased from 56% to 47%.

Medical and prescription medication costs

Following a diagnosis of COVID-19 (post-index), 34% of all medical and prescription medication costs were COVID-19 related (as defined in Supplementary Table 1A) for employees in 2020, 39% in 2021, and 11% in 2022. Among the subset of the COVID-19 cohort with six months of pre- and post-index health plan coverage, medical and prescription medication costs are shown in Table 5. In 2020, individuals with COVID-19 incurred mean direct medical costs in the six-month post-index period of $7,566 (SE $395), and a median of $1,128. This increased to $8,540 (SE $431, median $1,240) in 2021, and decreased to a mean of $5,907 (SE $169) and a median of $1,097 in 2022. The direct prescription costs increased each year from 2020 to 2022 with a mean of $1,177 (SE $50, median $82) in 2020, $1,358 (SE $66, median $84) in 2021, and $1,958 (SE $53, median $126) in 2022.

Table 5. COVID-19 cohort six-month post-period direct medical and prescription costs.

	2020	2021	2022
Employee count, N	11,442	18,790	27,516
All cause total health costs, mean (SE)	$8,743 ($403)	$9,898 ($439)	$7,865 ($181)
All cause medical costs, mean (SE)	$7,566 ($395)	$8,540 ($431)	$5,907 ($169)
All cause pharmacy costs, mean (SE)	$1,177 ($50)	$1,358 ($66)	$1,958 ($53)
Covid-related total health costs, mean (SE)	$2,952 ($336)	$3,866 ($291)	$853 ($63)
Covid-related medical costs, mean (SE)	$2,952 ($336)	$3,866 ($291)	$849 ($63)
Covid-related pharmacy costs, mean (SE)	$0 ($0)	$0.01 ($0.00)	$3.47 ($0.04)

Note: Clinician-administered medications, such as remdesivir are included in the costs associated with the medical claims.

Discussion

In this analysis, COVID-19 resulted in higher rates of work absence claims, longer mean duration of work absences, and higher mean costs of claims during both the 2018–2022 and 2020–2022 study periods relative to influenza. Since the recognition of COVID-19 on 4 February 2020, COVID-19 ranked second (9.7%) among the top diagnoses for STD claims, alongside claims for pregnancy, low back pain, and mood disorders. An increase in COVID-19 LTD claims were observed after long COVID was added as a recognized condition that could result in disability under the Americans with Disability Act (ADA) in July 2021^17,18. COVID-19 ranked as the 20th most common LTD claims diagnosis in 2020 but increased to 11th in 2021 and 12th in 2022. The ICD-10 codes U09.9 (Post COVID-19 condition, unspecified) and U07.1 (COVID-19) were the only two LTD claim diagnoses in 2022.

While pandemic mitigation measures decreased the transmission rates of influenza from 2020 to 2022^19,20, influenza accounted for less than 1% of the total STD claims for each individual year analyzed in the overall population, both preceding and during the COVID-19 pandemic consistent with previously published trends²¹. Additionally, COVID-19 STD claims also had a higher mean cost than influenza STD claims ($3,477 vs $1,354). Notably, there were no LTD claims for an influenza diagnosis in the overall population or the COVID-19 cohort.

To the best of our knowledge, this is the first peer-reviewed publication to use integrated group health and paid absence claims data to describe the utilization of paid leave benefits by employees with COVID-19. It may also be the first publication to describe the utilization patterns of ICD codes for COVID-19 within a disability claims database. Medical claims are utilized for reimbursement and therefore healthcare providers are required to utilize appropriate ICD-10 codes in order for them to receive compensation for the care provided²². Absence vendors do have these same coding requirements for claims adjudication. We found that disability insurers used non-specific viral infection and pneumonia AHRQ category codes to document extended leave related to COVID-19 for years after the CDC recommended the use of U07.1 for COVID-19²³. Including ICD-9 and the six expanded ICD-10 codes that were used to identify COVID-19 through STD and LTD claims, increased the number of disability claims in our analysis from 411 to 1,814 within the COVID-19 cohort in 2020. The increased availability and utilization of rapid antigen tests may have enabled absence claims to be processed without an associated medical claim during the pandemic when access to in-person healthcare was limited^24,25. Based on this exploratory analyses, COVID-19 cases identified through both medical and disability claims might be a more accurate representation of the prevalence of symptomatic COVID-19 diagnoses impacting work loss within an employed population.

The most common industry represented in the RRDb was the Retail Trade industry (28.1%); however, nearly a third of employees (31.4%) in the COVID-19 cohort were in the Health Care and Social Assistance job industry. Employees working in the Health Care and Social Assistance sector may be at higher risk for exposure to COVID-19 relative to other industries but may be more likely to utilize mitigation measures such as mask wearing while at work than other industry segments^26–28. Those within the healthcare industry may be also more likely to be eligible for STD and LTD benefits than employees in other industries where employees are not employed full time or have a higher turnover rate. The mean salary was $86,331 in the COVID-19 cohort, with an increase from approximately $80,000 in 2020 and 2021 to a mean of $93,725 in 2022. The increase in average annual salary of over $10,000 may have been due to remote employees with relatively higher salaries compared to other employees in the COVID-19 cohort, returning to an office setting with potentially increased exposure and resulting illness from COVID-19²⁹. Although different job types were identified in this analysis, extrapolating the impact of COVID-19 on individual job types is not appropriate because the definition of job types varied between employers within the RRDb.

The impact of COVID-19 on work absences in this study provides objective support for previous studies that utilized internet-based surveys with open-ended questions to evaluate the impact of COVID-19 on work participation and labor shortages^10,30,31. Stelson et al. found that many participants expressed a strong desire to return to work for financial purposes, but symptoms of long COVID interfered with work and home life³⁰. Another study, conducted through online surveys distributed through COVID-19 support groups and social media, reported that the recovery time for COVID-19 exceeded 35 weeks for more than 91% of respondents. Among them, 45.2% required a reduced work schedule, and an additional 22.3% were not working at the time of the survey due to COVID-19¹⁰. Using responses from 2022 the US National Health Interview Survey (NHIS) to assess the impact of long COVID on work loss productivity, Bonner et al. reported that Long COVID affected work attendance and disability risk for those with chronic health conditions. For example, people with diabetes and Long COVID had a 130% increase in missed days compared to those with diabetes alone (21.3 missed days vs 9.3 missed days)³¹.

The impact of vaccination on work benefit utilization was limited due to inconsistent reporting of vaccination claims data within the US because COVID-19 vaccinations are administered at physician offices, pharmacies, or other vaccination locations. The safety and efficacy of mRNA COVID-19 vaccinations is well established^32–34. A nationwide study involving symptomatic outpatients boosted with the original monovalent BNT162b2 in the US was associated with a decrease in the prevalence and duration of acute and long-term symptoms leading to improved health-related quality of life (HRQoL) and work performance. Employed patients with COVID-19 completed the Work Productivity and Activity Impairment General Health (WPAI-GH) questionnaire, and the BNT162b2 vaccinated cohort was acutely associated with less decreases in WPAI-GH scores. The mean overall work productivity time loss (absenteeism and presenteeism combined) in the week following a COVID-19 diagnosis was 65.0% for the unvaccinated cohort and 53.8% for those workers who received a vaccination. The mean work hours lost was significantly lower for those boosted with BNT162b2 and highest for the unvaccinated cohort (15.85 and 28.53, respectively)³⁵. A follow up study evaluated the changes in HRQoL and work productivity six months after COVID-19 infection, and found that unvaccinated patients had the highest number of work hours lost compared to those boosted with BNT162b2³⁶. Similarly, compared to an unvaccinated/not up-to-date cohort, vaccination with bivalent BA.4/5 BNT162b2 was associated with better work performance, less absenteeism, and fewer work hours lost³⁷. Notably, these studies relied on patient self-reports in the week prior to the assessment, and did not analyze the costs associated with the COVID-19-associated work loss or validate the absences with payroll records.

Work absence due to COVID-19 is an under-explored area, emphasizing the need for increased knowledge both in the US and on a global scale. Medical care in the US differs from other countries with universal healthcare and a national registry database. The US has a diverse landscape with commercial healthcare providers and a healthcare system that combines both commercial and government-funded Medicare insurance. In this unique context, the analysis of medical and disability claim data allowed for valuable and objective insight within the workforce, including a detailed analysis of the duration of work absences and associated cost of COVID-19.

The current analysis highlights that COVID-19 continues to be a debilitating disease that poses a substantial employer burden in terms of medical, absenteeism, and disability costs. Strengths of the current study include the unique evaluation of clinical and economic impact of COVID-19 on employed patients using objectively measured claims data. Including identified ICD-9 and ICD-10 codes in addition to the recommended ICD-10 codes captured a more inclusive number of COVID-19 cases, given the coding patterns of some disability insurance providers when categorizing STD and LTD claims. With the addition of disability leaves in addition to medical and prescription drug claims to identify patients with COVID-19, the study captured a substantial number of patients who might have otherwise been missed.

Future studies have the potential to more comprehensively capture the burden on employees and offer additional contextual insights into the results in the current analysis. Key areas for future research include comparisons between individuals with COVID-19 and controls or other respiratory infections, identification of and comparisons between high- and low-risk patients, COVID-19 severity, and evaluating the impact of COVID-19-associated physical and mental conditions on work absence. For example, anxiety was the eighth most frequent reason for STD in 2018 and 2019 but increased to fifth or sixth in each year after COVID-19 from 2020 to 2022. Additionally, with longer time periods, the impact of COVID-19 on LTD claims can be better understood. Finally, administrative claims data can also be used to evaluate the impact of COVID-19 on employee dependents, including spouses and children, considering the employees’ roles as caregivers and the potential for disease transmission within the household.

Limitations

There are several inherit limitations related to the use of administrative claims data including the risk of clerical inaccuracies, recording bias secondary to financial incentives, and temporal changes in billing codes. The analysis was restricted to information available in an integrated database of health claims, work absences, and employee human resource information. This may account for some missing demographic information, such as race or ethnicity as some employees choose not to specify, and others are not uniquely described. The population within the Workpartners RRDb is representative of self-insured employers that provide absence benefits and may not be representative of all US employees. Employees working for relatively smaller employers may have utilized unpaid leave for a COVID-19 illness or may have lost their job if unpaid leave was available from their employer. Additionally, the findings do not reflect the full burden of patients with COVID-19 as the claims data only reflect interactions with the healthcare system and/or paid leave benefit claims. Employees with minor COVID-19 symptoms may not have sought or received a COVID-19 diagnosis from a healthcare provider. Other unmeasured factors, such as presenteeism are not included in this analysis, and therefore the current results may be viewed as conservative.

It is assumed that influenza data in 2018 and 2019 is representative of baseline influenza data before the COVID-19 pandemic^19,20. However, the medical claims data of COVID-19 and influenza and prescription claims data of COVID-19 may be underrepresented, as many healthcare providers were restricting access to healthcare to limit viral transmission during 2020–2022. Because prescription medications used to treat influenza may also be used for prophylaxis following exposure, prescription claims data were not used to identify influenza cases, potentially leading to inconsistent and biased comparisons. A direct association cannot be made to the impact of acquired or natural immunity to STD claims from this analysis. This analysis did not include the impact of vaccination on work absences, because the records of vaccinations are not consistently identifiable in claims databases and may be administered in multiple settings, thus limiting the ability to evaluate an association of vaccination on STD rates. Also, new variants of COVID-19 have emerged since this analysis and may have different impacts on future health outcomes than reflected in these data.

The Workpartners RRDb is focused on self-insured employers with absence benefits, these data do not represent unemployed individuals, or those working for employers that do not provide absence benefits, and those in job-types or industries not included in the RRDb. Lastly, individuals over the age of 65 years in the US are eligible for Medicare, thus claims for employees in this age segment are not comprehensive in this analysis.

Conclusions

Although the overall morbidity of COVID-19 disease is lower secondary to vaccinations, medical treatments, and natural immunity, COVID-19 continues to impact peoples’ ability to work. COVID-19 ranked as the 2nd most frequent STD reason, and 13th most frequent LTD reason during 2020–2022. Influenza ranked 28th in 2018 and 49th in 2019 for STD, with no associated LTD claims. Integration of group health and disability claims offered a comprehensive understanding of the impact of COVID-19-related work loss for an employed population. This research can enable employers to conduct a similar evaluation of COVID-19 trends within their population to appropriately prioritize resources such as vaccinations and timely treatment needed to mitigate the impact to the employee and the company.

Transparency

Declaration of financial/other relationships

RB, TG, and NK are employees of Better Health Worldwide. Better Health Worldwide received funding from Pfizer for the completion of the research. AD, ER, and DW are employees of Workpartners. Workpartners received funding through Better Health Worldwide for the research. JJ, AY, JGP, FD, JME, HS are employees and shareholders of Pfizer Inc.

Acknowledgements

None stated.

Additional information

Funding

This study was funded by Pfizer Inc.