Costs and absence of HCV-infected employees by disease stage

Abstract

Objectives:

Quantify the costs and absenteeism associated with stages of the Hepatitis C virus (HCV).

Study design:

Retrospective analysis of the HCMS integrated database from multiple geographically diverse, US-based employers with employee information on medical, prescription, and absenteeism claims.

Methods:

Employee data were extracted from July 2001–March 2013. Employees with HCV were identified by ICD-9-CM codes and classified into disease severity cohorts using diagnosis/procedure codes assigning the first date of most severe claim as the index date. Non-HCV employees (controls) were assigned random index dates. Inclusion required 6-month pre-/post-index eligibility. Medical, prescription, and absenteeism cost and time were analyzed using two-part regression (logistic/generalized linear) models, controlling for potentially confounding factors. Costs were inflation adjusted to September 2013.

Results:

All direct costs comparisons were statistically significant (p ≤ 0.05) with mean medical costs of $1813 [SE = $3] for controls (n = 727,588), $4611 [SE = $211] for non-cirrhotic (n = 1007), $4646 [SE = $721] for compensated cirrhosis (CC, n = 87), $12,384 [SE = $1122] for decompensated cirrhosis (DCC, n = 256), $33,494 [SE = $11,753] for hepatocellular carcinoma (HCC, n = 17) and $97,724 [SE = $32,437] for liver transplant (LT, n = 19) cohorts. Mean short-term disability days/costs were significantly greater for the non-cirrhotic (days = 2.03 [SE = 0.36]; $299 [SE = $53]), DCC (days = 6.20 [SE = 1.36]; $763 [SE = $169]), and LT cohorts (days = 21.98 [SE = 8.21]; $2537 [SE = $972]) compared to controls (days = 1.19 [SE = 0.01]; $155 [SE = $1]). Mean sick leave costs were significantly greater for non-cirrhotic ($373 [SE = $22]) and DCC ($460 [SE = $54]) compared to controls ($327 [SE = $1]).

Conclusions:

Employees with HCV were shown to have greater direct and indirect costs compared to non-HCV employee controls. Costs progressively increased in the more severe HCV disease categories. Slowing or preventing disease progression may avert the costs of more severe liver disease stages and enable employees with HCV to continue as productive members of the workforce.

Keywords::

• Hepatitis
• Absenteeism
• Presenteeism
• Costs of care/Healthcare expenditures
• Cirrhosis
• Liver-transplant
• Hepatocellular carcinoma

Introduction

Hepatitis C virus (HCV) is the most common blood-borne chronic infection in the US and is most prevalent among those born during 1945–19651. According to CDC and NHANES, 3.2–5.2 million persons in the US are chronically infected2^,3. Owing to high disease prevalence and because the HCV-afflicted cohort comprises an age band that is well represented in the US workforce (55-years-and-older workers estimated at 19.5% of total and rising to 25.2% in 2020)4, disease burden and treatment decisions will have significant implications with respect to overall healthcare costs captured in employee benefit structures. Employers, as payers of health benefits for HCV-infected employees, will require information that permits informed decision-making regarding optimized disease management.

It has been reported that ∼50–75% of those infected are not aware of their serostatus5^,6, and infection is often detected during routine lab testing or upon investigation of non-specific symptomology experienced with disease progression. The most common means of HCV transmission is injection drug use in some populations; however, transfusions before screening was implemented, tattoo procedures, dental procedures, medical occupation hazards, and sexual contact are all alternative etiologies in the at-risk, employed population5. Although HCV infection can be asymptomatic for many years, complications once the disease progresses can be severe.

HCV is a progressive disease, and 60–70% of infected persons will go on to develop chronic liver disease; 5–20% develop cirrhosis over 20–30 years. Common complications of decompensated cirrhosis include portal hypertension, variceal bleeding, ascites, and hepatic encephalopathy7. Furthermore, decompensation progresses to hepatocellular cancer (HCC) at a rate of 3% per year8. In the US, chronic HCV infection is the leading cause for liver transplantation1. Approximately 1–5% will die from the consequences of chronic HCV infection due to liver cancer and/or cirrhosis.

The economic burden of HCV infection on the healthcare system is well documented in the literature. HCV-related healthcare utilization and costs increase significantly with increasing patient age9 and increasing disease severity10^,11. Looking to the future, the total long-term annual cost associated with chronic HCV infection is expected to rise dramatically from $6.5 billion in 2012 to an estimated $9.1 billion by 20241 2. Therefore, the impact of chronic HCV in the employed US population is of considerable social and financial importance, with implications on timing of treatment initiation. Recent research has shown that successful anti-viral treatment during the early stages of disease can significantly reduce liver disease progression and the costs and events associated with advanced liver disease9^,11^,13^,14.

Because chronic HCV often progresses slowly, those infected with HCV can be expected to maintain role function in the workplace, both prior to and after diagnosis. Participation in the workforce might be expected over many years, at any stage of HCV involvement, depending on the rate of disease progression. Although the effects and consequences of HCV infection in the workplace have been described in the literature, the data were often obtained from self-reports, not compared against controls, or lack specific end points, such as absence days broken down by type, time, and cost. To date, only a few studies have truly measured the impact of the stages of HCV infection on the workforce. A common finding among these studies is that HCV infection causes a substantial indirect burden of illness by negatively impacting productivity, absenteeism, and healthcare benefit costs. A large study by Su et al.15 reported HCV-infected employees exhibited significantly more lost work days, and higher healthcare costs than non-infected controls. HCV-infected workers annually experienced 4.15 more absence days (sick leave, short- and long-term disability), and had significantly higher annual healthcare costs, resulting in a total incremental (direct plus indirect) difference of $8352 per year. Research by Brook et al.16 found that workers receiving interferon-based treatment for HCV infection experienced more monthly absence days (0.52) and incurred greater cost ($31.31 in additional monthly absence payments) when compared to non-treated HCV-infected workers. Neither the Su et al.15 or Brook et al.16 studies evaluated outcomes with respect to disease severity.

Research evaluating HCV disease severity in an employed population and its proportional impact on direct and indirect costs to the employer has not been reported in the literature. Since chronically HCV infected patients can spend years or decades in the workforce, it is imperative to evaluate how the different stages of chronic HCV infection affect employee absence and healthcare costs. This would allow employers to make informed decisions regarding health benefits, treatment initiatives, and formulary placement of therapies. The aim of this study was to examine health benefit costs and absenteeism among employees with HCV infection in various stages of liver disease compared to employees without HCV.

Patients and methods

Design

The current study is a retrospective analysis of administrative claims and employer payroll data of employees from the Human Capital Management Services (HCMS) integrated database, extracted from January 2001 to September 2013. This database represents multiple geographically diverse, US-based employers, totaling more than 1.9 million employees. Demographic (race, age, gender, region), payroll (employment status and salary), healthcare claims (medical and prescription), disability and absence data (short-term disability [STD], long-term disability [LTD], workers’ compensation [WC], and sick leave [SL]) were obtained from retail, service, manufacturing, and financial industries. All employees in the database were eligible for the study provided they had continuous benefit eligibility for the study duration. Information on employees’ STD, LTD, WC, and SL claims was collected from payroll, attendance, and claims records and included hours or days missed and payments made to the employees. The WC claims’ costs also include medical claims paid under the workers’ compensation benefit.

The data were de-identified to comply with the Health Insurance Portability and Accountability Act and the contractual obligations between HCMS and its employer-contributors. This database has been used in prior HCV research15^,16.

Employees with HCV were identified by International Classification of Diseases 9^th revision Clinical Modification (ICD-9-CM) codes (Table 1) and classified into one of the following disease cohorts (in order of increasing severity): non-cirrhotics, compensated cirrhosis (CC), decompensated cirrhosis (DCC), hepatocellular carcinoma (HCC), and liver transplant (LT). A description of ICD-9-CM codes describing disease cohorts is available in Table 2. Initial patient selection was based on identification of any primary, secondary, or tertiary ICD-9-CM code related to HCV; disease stages were classified according to the presence or absence of specific codes. Cohorts for the different stages of HCV were identified using a combination of ICD-9-CM (diagnosis and procedure) and Current Procedure Terminology codes found in published literature11^,17. Employees were excluded from all cohorts if ICD-9-CM codes indicated a diagnosis of Human Immunodeficiency Virus or Hepatitis B. For the LT cohort, the index date was the date of the LT procedure. For the remaining HCV cohorts, the index date was defined as the first date of the most severe HCV disease claim for any given employee. Non-HCV employees, without ICD-9-CM codes for HCV, served as a control cohort with a randomly selected index date. Inclusion required 6-month pre-index baseline and 6-month post-index continuous follow-up periods. The pre-index period was used to assess population baseline characteristics and pre-index modified Charlson Comorbidity Index values (mCCI; modified by removing the mild and severe liver disease components).

Table 1. HCV ICD-9-CM codes.

ICD-9-CM code	Description
070.44	Chronic HCV with hepatic coma
070.54	Chronic HCV without mention of hepatic coma
070.70	Unspecified viral hepatitis C without hepatic coma
070.71	Unspecified viral hepatitis C with hepatic coma

HCV, hepatitis C virus; iCD-9-CM, international Classification of Diseases, 9th revision, Clinical Modification.

Table 2. ICD-9-CM codes for stages of HCV.

Stages of HCV	ICD-9-CM codes for HCV (see Table 1) plus code(s)11^,17
HCV without cirrhosis	Without an ICD-9-CM or CPT code classifying them into a more severe category (see below)
Cirrhosis	ICD-9-CM code(s) 571.2, 571.5, 571.6
Compensated cirrhosis	With a cirrhosis code, but with no code indicating DCC events
Decompensated cirrhosis (by presence of indicator events)	ICD-9-CM Dx codes 456.0–456.2x, 348.3x, 572.2, 572.3, 572.4, 782.4, 789.59, 572.8, 530.7, 530.82, 578.x, Or CPT codes 49080, 49081, 37140, 37160, 37180, 37181, 37182, 37183, 43204, 43205, 43243, 43244, 43400, 43401, or ICD-9 proc codes 54.9, 42.91, 44.91, or 96.06
These codes also specifically identify DCC	ICD-9-CM code(s) 070.44, 070.71
Liver transplant	ICD-9-CM diagnosis code of V21.7X, V42.7x or CPT codes 47135 or 47136 or ICD-9 proc code 50.5x
Hepatocellular carcinoma	ICD-9-CM code 155.xx

HCV, hepatitis C virus; iCD-9-CM, international Classification of Diseases, 9th revision, Clinical Modification; CPT, Current Procedural Terminology; Dx, diagnostic; proc, procedure; DCC, decompensated cirrhosis.

Outcome measures

Outcome measures were evaluated for each cohort over the 6 months post-index date and included adjusted direct health benefit costs (medical and prescription, employer- and employee-paid portions combined), absence days due to SL, STD, LTD, and WC, and payments for indirect costs associated with these health-related absence days. The HCMS database records the number of hours absent, including fractional hours. Days absent were obtained from employee payroll records or disability/workers’ compensation claims data and were not imputed from medical claims data. Indirect costs for absence days were taken from actual employee cost data, rather than having been imputed from average salary values. Employees will typically use SL to cover health-related work absences for up to 2 weeks and usually will receive 100% of their salary during that time; however, the number of SL days offered per year may vary by employer, employee level, and tenure18. Illnesses lasting from 1 or 2 weeks and up to 6 months will generally be covered by STD benefits and employees generally receive between 60–100% of their salary during that time. Long-term disability will cover work absence due to illness lasting longer than 6 months and compensates at a rate of 50–70% of salary19. Workers’ compensation payments in this study include salary replacement payments during the absence (usually 66–80% of salary)20 and medical claims paid under the workers’ compensation benefit. Absence benefits vary by employer and employee; even within employers, employees may receive additional benefits over time. For example, some employees may become eligible for short-term disability only after they complete a minimum term of employment. Absence outcomes were only measured for employees who were eligible for each specific absence benefit during their entire post-index period.

Results are available for pre- and post-index mCCI values, HCV drug use, contraindications to HCV drug use, and individual co-morbid conditions (insomnia, anxiety, depression, psychoses, fatigue, anemia, thrombocytopenia, white blood cell conditions, and diabetes). Only comorbidities which have been shown to have a high prevalence with HCV patients and HCV treatment were analyzed. Unadjusted results are also provided for 6-month post-index health benefit costs (medical, prescription, SL, STD, LTD, and WC), and post-index absence days (SL, STD, LTD, and WC). In addition, descriptive measures, such as age, gender, marital status, race, salary, years with employer (tenure), exempt status (exempt = not eligible for overtime), full-time vs part-time status, job type (Sales and Related Occupations, Healthcare Practitioners & Support, Unknown, Office and Admin Support, Transportation/Material Moving, and Management Occupations, and ‘Other’), index year, region (defined by the first digit of the employee’s Zip code) are reported. Medical, prescription, and absence costs were inflation adjusted to September 2013 dollars using Consumer Price Index values for medical services, prescription drugs, and all consumer goods and services, respectively.

Statistical analysis

Continuous descriptive measures (age, salary, etc.) were compared between each HCV cohort and the non-HCV cohort using t-tests. Dichotomous descriptive measures (female, Hispanic, etc.) were compared between each HCV cohort and the non-HCV cohort using chi-squared tests.

Each direct and indirect outcome (each dependent variable) was modeled separately using two-part regression models. For example, in the cost or days absence time models, logistic regression was first used to predict the likelihood of subjects having any costs or days absence time (part 1). The second part used generalized linear models with a gamma (γ) distribution and a log link function to model costs or days absence time for subjects with more than zero costs or days absences. In each model, the main independent variables included indicator (dummy) variables to differentiate between the different cohorts. There were a total of five indicator variables in each model, one for each HCV cohort, leaving the non-HCV cohort as the reference group. The models were adjusted for covariates including age, weekly salary, exempt status, female gender, marital status, race indicators, region (defined by the first digit of the employee’s Zip code), year of the index date (minus 2000), pre-index mCCI, and job type indicators. The cohort variables age and salary were forced to remain in the regression models. All other independent variables were selected based on a stepwise selection process. All models and statistics were generated using version 9.3 of the SAS System for Windows (Copyright SAS Institute, Inc., Cary, NC), and results were considered statistically significant when p ≤ 0.05.

Results

The final study population consisted of 1386 HCV-infected employees, the majority of whom were in the non-cirrhotic stage (1007), and 727,588 non-HCV-infected employees in the control cohort. The impact of the inclusion and exclusion criteria can be seen in Figure 1. Employees in the study were well distributed throughout the US. The average ages in all five HCV cohorts (ranging from 47–56 years) were greater than in the non-HCV control cohort (40 years). Employee tenure was also significantly greater in the HCV cohorts than in the control cohort. Employees with a LT had significantly higher mean annual salary ($74,075; SE = $9442) than the control cohort ($54,343; SE = $46), and employees in the non-cirrhotic HCV cohort had lower annual salaries ($51,213; SE = $984) than the control cohort. A wide variation existed in the percentage of employees in each cohort who were female, ranging from 10.5% in the LT cohort to 46.7% in the control cohort. Employees with DCC and non-cirrhotics were less likely to have exempt status than those in the control cohort (25.4% and 23.8%, respectively, vs 31.6%). Complete descriptive statistics are shown in Tables 3–5. Information on job type and region of the country is available online in the journal's supplemental information.

Figure 1. Population selection.

Table 3. Demographic comparisons of the study cohorts.^a.

Metric	Non-cirrhotic HCV		Compensated cirrhosis		Decompensated cirrhosis		Hepatocellular carcinoma		Liver transplant		No chronic HCV (Control cohort)
	Mean or %	SE	Mean or %	SE	Mean or %	SE	Mean or %	SE	Mean or %	SE	Mean or %	SE
n^b	1007		87		256		17		19		727,588
Demographics
Age	47.10§	0.30	51.90§	0.69	49.36§	0.56	56.32§	0.92	51.03§	1.72	40.45	0.01
Percentage female	42.6%‡	1.6%	37.9%	5.2%	37.1%‡	3.0%	29.4%	11.4%	10.5%‡	7.2%	46.7%	0.1%
Married	35.1%‡	1.5%	27.6%	4.8%	33.2%	2.9%	35.3%	11.9%	47.4%	11.8%	30.6%	0.1%
Not married	24.5%	1.4%	21.8%	4.5%	21.5%	2.6%	0.0%‡	0.0%	15.8%	8.6%	26.0%	0.1%
Missing Marital status	40.4%	1.5%	50.6%	5.4%	45.3%	3.1%	64.7%	11.9%	36.8%	11.4%	43.4%	0.1%
White	36.7%‡	1.5%	28.7%	4.9%	35.9%	3.0%	17.6%	9.5%	47.4%	11.8%	31.6%	0.1%
Black	9.8%	0.9%	8.0%	2.9%	10.9%	2.0%	5.9%	5.9%	0.0%	0.0%	8.4%	0.0%
Hispanic	6.8%	0.8%	5.7%	2.5%	10.2%‡	1.9%	23.5%	10.6%‡	10.5%	7.2%	5.6%	0.0%
Other race	2.2%	0.5%	1.1%	1.1%	3.5%	1.2%	0.0%	0.0%	0.0%	0.0%	3.0%	0.0%
Race missing	44.5%§	1.6%	56.3%	5.3%	39.5%§	3.1%	52.9%	12.5%	42.1%	11.6%	51.5%	0.1%
Job-related variables
Tenure	9.05§	0.27	9.53‡	0.87	10.96§	0.62	18.12§	2.44	13.66‡	2.12	7.77	0.01
Annual salary	$51,213‡	$984	$53,173	$3912	$53,204	$2,126	$58,793	$7445	$74,075†	$9,442	$54,343	$46
Exempt	23.8%§	1.3%	25.3%	4.7%	25.4%†	2.7%	29.4%	11.4%	47.4%	11.8%	31.6%	0.1%
Full-time	96.3%‡	0.6%	98.9%	1.1%	97.3%†	1.0%	100.0%	0.0%	94.7%	5.3%	94.4%	0.0%

Salary costs are adjusted for inflation to September 2013 dollars.

^aAll employees had at least 6 months of health plan enrollment before and after their index date. p Values were calculated using t-tests for continuous variables and chi-squared tests for dichotomous variables.

^bDue to missing data, n’s for the Non-cirrhotic HCV, Compensated Cirrhosis, Decompensated Cirrhosis, Hepatocellular Carcinoma, Liver Transplant, and no HCV cohorts are 1007, 87, 256, 17, 19, and 727,570 for age, and 985, 85, 256, 17, 17, and 698,686 for salary, respectively.

†p ≤ 0.05 vs the Control Cohort; ‡p ≤ 0.01 vs the Control Cohort; §p ≤ 0.0001 vs the Control Cohort.

Table 4. Co-morbid conditions within the study cohorts.

Metric	Non-cirrhotic HCV		Compensated cirrhosis		Decompensated cirrhosis		Hepatocellular carcinoma		Liver transplant		No chronic HCV (Control cohort)
	Mean or %	SE	Mean or %	SE	Mean or %	SE	Mean or %	SE	Mean or %	SE	Mean or %	SE
n	1007		87		256		17		19		727,588
Pre-index non-regression-adjusted results
Modified^a Charlson Index	0.269§	0.023	0.345‡	0.075	0.512§	0.059	0.765	0.466	1.684‡	0.530	0.133	0.001
Insomnia Diagnosis	2.8%§	0.5%	5.7%‡	2.5%	3.1%‡	1.1%	11.8%§	8.1%	0.0%	0.0%	1.3%	0.0%
Anxiety Diagnosis	4.6%	0.7%	3.4%	2.0%	5.9%†	1.5%	5.9%	5.9%	5.3%	5.3%	3.5%	0.0%
Depression Diagnosis	6.1%§	0.8%	6.9%	2.7%	10.9%§	2.0%	17.6%‡	9.5%	10.5%	7.2%	3.7%	0.0%
Psychoses Diagnosis	4.0%§	0.6%	2.3%	1.6%	9.4%§	1.8%	17.6%§	9.5%	10.5%‡	7.2%	2.0%	0.0%
Fatigue Diagnosis	7.9%§	0.9%	6.9%	2.7%	10.9%§	2.0%	17.6%‡	9.5%	5.3%	5.3%	3.8%	0.0%
Anemia Diagnosis	5.7%§	0.7%	4.6%	2.3%	10.2%§	1.9%	23.5%§	10.6%	31.6%§	11.0%	1.9%	0.0%
Thrombocytopenia Diagnosis	1.0%§	0.3%	6.9%§	2.7%	3.9%§	1.2%	11.8%§	8.1%	0.0%	0.0%	0.1%	0.0%
White Blood Cell Condition	0.7%	0.3%	3.4%§	2.0%	3.9%§	1.2%	5.9%§	5.9%	5.3%§	5.3%	0.3%	0.0%
Diabetes Diagnosis	9.6%§	0.9%	12.6%§	3.6%	11.3%§	2.0%	11.8%	8.1%	21.1%§	9.6%	3.7%	0.0%
Post-index non-regression-adjusted results
Modified^a Charlson Index	0.304§	0.024	0.276†	0.063	0.617§	0.079	3.176‡	0.908	1.737‡	0.373	0.146	0.001
Insomnia Diagnosis	2.6%‡	0.5%	3.4%	2.0%	5.1%§	1.4%	0.0%	0.0%	0.0%	0.0%	1.4%	0.0%
Anxiety Diagnosis	5.4%†	0.7%	3.4%	2.0%	7.4%‡	1.6%	11.8%	8.1%	0.0%	0.0%	3.9%	0.0%
Depression Diagnosis	6.4%‡	0.8%	5.7%	2.5%	12.9%§	2.1%	17.6%‡	9.5%	5.3%	5.3%	4.0%	0.0%
Psychoses Diagnosis	3.7%‡	0.6%	3.4%	2.0%	7.8%§	1.7%	11.8%‡	8.1%	5.3%	5.3%	2.2%	0.0%
Fatigue Diagnosis	6.8%§	0.8%	8.0%†	2.9%	12.9%§	2.1%	0.0%	0.0%	10.5%	7.2%	3.9%	0.0%
Anemia Diagnosis	5.8%§	0.7%	5.7%†	2.5%	18.8%§	2.4%	17.6%§	9.5%	36.8%§	11.4%	2.0%	0.0%
Thrombocytopenia Diagnosis	1.0%§	0.3%	3.4%§	2.0%	4.3%§	1.3%	11.8%§	8.1%	5.3%§	5.3%	0.1%	0.0%
White Blood Cell Condition	2.2%§	0.5%	3.4%§	2.0%	4.7%§	1.3%	5.9%§	5.9%	5.3%§	5.3%	0.3%	0.0%
Diabetes Diagnosis	9.3%§	0.9%	10.3%‡	3.3%	9.0%§	1.8%	11.8%	8.1%	26.3%§	10.4%	4.0%	0.0%

All employees had at least 6 months of health plan enrollment before and after their index date. p Values are calculated using t-tests for continuous variables and chi-squared tests for dichotomous variables.

^aMild and severe liver condition components have been removed from the Charlson Index.

†p ≤ 0.05 vs the Control Cohort; ‡p ≤ 0.01 vs the Control Cohort; §p ≤ 0.0001 vs the Control Cohort.

Table 5. Use of HCV medications; direct and indirect costs, and absence time within the study cohorts.

Metric	Non-cirrhotic HCV		Compensated cirrhosis		Decompensated cirrhosis		Hepatocellular carcinoma		Liver transplant		No chronic HCV (Control cohort)
	Mean or %	SE	Mean or %	SE	Mean or %	SE	Mean or %	SE	Mean or %	SE	Mean or %	SE
n	1007		87		256		17		19		727,588
Percentage of cohort with contraindicating ICD-9-CM codes (available as Supplemental Table 3) at any time in the database	32.9%§	1.5%	32.2%‡	5.0%	65.2%§	3.0%	88.2%§	7.8%	84.2%§	8.4%	17.8%	0.0%
Unadjusted use of HCV medications
% using Ribavirin	3.38%§	0.57%	9.20%§	3.12%	5.47%§	1.42%	0.00%	0.00%	0.00%	0.00%	0.0018%	0.0005%
% using Peginterferon	14.20%§	1.10%	26.44%§	4.76%	14.45%§	2.20%	5.88%§	5.88%	0.00%	0.00%	0.0131%	0.0013%
% using Telaprevir	0.99%§	0.31%	1.15%§	1.15%	1.56%§	0.78%	0.00%	0.00%	0.00%	0.00%	0.0003%	0.0002%
% using Boceprevir	0.20%§	0.14%	0.00%	0.00%	0.78%§	0.55%	0.00%	0.00%	0.00%	0.00%	0.0007%	0.0003%
Unadjusted indirect (absence) costs (by type)
STD Costs^a	$374	$94	$47‡	$47	$1642†	$614	$2504	$1629	$6094†	$2701	$232	$3
LTD Costs^b	$126	$94	$0§	$0	$119	$87	$0§	$0	$1812	$1432	$35	$3
WC Costs^c	$756	$337	$486	$466	$618	$397	$0§	$0	$0§	$0	$256	$9
SL Costs^d	$537†	$64	$644	$263	$774‡	$139	$3316	$2162	$386	$333	$406	$2
Unadjusted absence days (by type)
STD Days^a	2.67†	0.55	0.67	0.67	9.13‡	2.27	22.58	14.51	22.69	11.75	1.47	0.02
LTD Days^b	1.35	0.90	0.00§	0.00	1.38	1.19	0.00§	0.00	63.54	43.78	0.39	0.03
WC Days^c	1.17	0.56	0.25	0.23	0.96	0.71	0.00§	0.00	0.00§	0.00	0.44	0.02
SLk Days^d	2.18	0.22	2.47	0.82	3.16‡	0.52	9.99	6.32	0.85	0.61	1.86	0.01
Unadjusted direct costs
Employee Medical Costs	$5562§	$430	$6021§	$976	$17,418§	$2394	$48,031‡	$13,053	$142,045‡	$47,440	$1926	$11
Employee Rx Costs	$2951§	$240	$5019§	$1064	$4,075§	$527	$7,344	$3,847	$10,846‡	$2,631	$437	$2

Costs are adjusted for inflation to September 2013 dollars.

All employees had at least 6 months of health plan enrollment before and after their index date. p Values are calculated using t-tests for continuous variables and chi-squared tests for dichotomous variables.

Eligible employee n’ for the Non-cirrhotic HCV, Compensated Cirrhosis, Decompensated Cirrhosis, Hepatocellular Carcinoma, Liver Transplant. and no HCV cohorts are:

^aSTD ns for non-cirrhotic, CC, DCC, HCC, LT, and controls are 623, 49, 168, 12, 13, and 456,097, respectively.

^bLTD ns for non-cirrhotic, CC, DCC, HCC, LT, and controls are 726, 61, 189, 14, 13, and 455,351, respectively.

^cWC ns for non-cirrhotic, CC, DCC, HCC, LT, and controls are 914, 79, 238, 17, 17, and 644,692, respectively.

^dSL ns for non-cirrhotic, CC, DCC, HCC, LT, and controls are 460, 28, 123, 11, 9, and 343,143, respectively.

†p ≤ 0.05 vs the Control cohort; ‡p ≤ 0.01 vs the Control cohort; §p ≤ 0.0001 vs the Control cohort.

CC, Compensated cirrhosis; DCC, decompensated cirrhosis; HCC, Hepatocellular carcinoma; LT, Liver transplant; HCV, hepatitis C virus.

Significant differences were observed for co-morbid conditions. Pre- and post-index mean mCCI values were generally significantly greater in the HCV cohorts than in the control cohort. This was particularly the case in the post-index period. In the pre-index period, all comorbidities (except anxiety) were significantly more prevalent in most HCV cohorts than in the control cohort. In the post-index period, this was the case for all comorbidities except anxiety and insomnia. Many comorbidities were 2–10-times more likely in employees with HCV, and thrombocytopenia in particular was 10–120-times more prevalent (Table 4).

Unadjusted medical and drug costs were highest in the LT cohort and lowest in the control cohort for the 6-month follow-up period (Table 5). Short-term disability and LTD costs were highest in the LT cohort, and the HCC cohort had the highest SL costs and days. Due to the small sample size, some of these values were not significantly different than in the control cohort.

All five HCV cohorts had significantly higher regression-adjusted 6-month direct medical and drug costs than the control cohort (Table 6). Direct costs increased by disease severity and were lowest in those with non-cirrhotic HCV, followed in order by the CC, DCC, HCC, and LT cohorts. The combined direct costs were significantly higher (p < 0.01) in the HCV cohorts than in the control cohort: the incremental (HCV cohort minus control cohort) direct costs per cohort were $5274 for non-cirrhotic HCV, $6746 for CC, $13,590 for DCC, $35,939 for HCC, and $102,529 for LT. Indirect components, such as SL and STD costs, were significantly higher in the DCC and non-cirrhotic HCV cohorts than in the control cohort, whereas the indirect costs varied by benefit type, with some costs for the HCV cohorts lower than the control cohort. In the LT cohort, SL costs were significantly lower than the control cohort. Overall, the total incremental indirect costs for each HCV cohort compared to the non-HCV control cohort were $350 for the non-cirrhotic HCV, $79 for the CC, $702 for those with DCC, $2941 for those with HCC, and $2148 for those with a LT.

Table 6. Regression-adjusted 6-month health benefit costs and absence days (during 6 months post-index).^a

Metric	Non-cirrhotic HCV			Compensated cirrhosis			Decompensated cirrhosis			Hepatocellular carcinoma			Liver transplant			No chronic HCV (Control cohort)
	n	Adj Mean	SE	n	Adj Mean	SE	n	Adj Mean	SE	n	Adj Mean	SE	n	Adj Mean	SE	n	Adj Mean	SE
Employee medical costs	1007	$4611§	$211	87	$4646§	$721	256	$12,384§	$1122	17	$33,494‡	$11,753	19	$97,724§	$32,437	727,588	$1813	$3
Employee Rx costs	1007	$2875§	$123	87	$4311§	$619	256	$3,416§	$283	17	$4,657‡	$1,527	19	$7,017§	$2,127	727,588	$398	$1
Subtotal direct costs		$7486§			$8957§			$15,801§			$38,151‡			$104,740§			$2211
Sick leave costs	460	$373	$22	28	$418	$106	123	$460‡	$54	11	$909	$446	9	$128‡	$59	343,143	$327	$1
STD costs	623	$299‡	$53	49	$76	$72	168	$763‡	$169	12	$2,685	$1,486	13	$2,537‡	$972	456,097	$155	$1
LTD costs	726	$57	$41	61	$0	NA	189	$17	$22	14	$0	NA	13	$135	$232	455,351	$15	$1
WC costs	914	$274†	$61	79	$238	$216	238	$114	$50	17	$0	NA	17	$0	NA	644,692	$154	$2
Subtotal indirect costs		$1003			$732			$1,354			$3,594			$2,800			$652
Total costs		$8488			$9689			$17,155			$41,744			$107,541			$2864
Sick leave days	460	2.00	0.12	28	2.13	0.54	123	2.49‡	0.29	11	4.79	2.34	9	0.83‡	0.38	343,143	1.77	0.00
STD days	623	2.03†	0.36	49	0.82	0.78	168	6.20‡	1.36	12	17.32	9.46	13	21.98‡	8.21	456,097	1.19	0.01
LTD days	726	0.63	0.45	61	0.00	NA	189	0.34	0.44	14	0.00	NA	13	8.46	14.55	455,351	0.23	0.01
WC days	914	0.47	0.21	79	0.09	0.09	238	0.31	0.33	17	0.00	NA	17	0.00	NA	644,692	0.22	0.00
Total days		5.12			3.05			9.34			22.11			31.26			3.41

Costs are adjusted for inflation to September 2013 dollars.

n, number of eligible employees; Adj, Adjusted; SE, standard error; NA, Statistics not produced, because no one in the cohort had a claim during the post-index study period.

†p ≤ 0.05 vs the Control cohort; ‡p ≤ 0.01 vs the Control cohort; §p ≤ 0.0001 vs the Control cohort.

^aAll employees had at least 6 months of health plan enrollment before and after their index date.

To produce and compare the regression-adjusted costs and days outcomes, two-part regression methods were used (separate pairs of models for each of the 10 cost and absence outcome variables). In each model, the main independent variables included indicator (dummy) variables to differentiate between the different cohorts. There were a total of five indicator variables in each model, one for each HCV cohort, leaving the non-HCV cohort as the comparison group. Other independent variables included age, weekly salary, exempt status (not eligible vs eligible for overtime pay), female gender, marital status, race indicators, region (defined by the first digit of the employee’s Zip code), year of the index date (minus 2000), pre-index modified Charlson Comorbidity Index (mild and severe liver disease components removed), and the following job type (comparisons between cohorts are available online in the journal's Supplemental Table 3) indicators: Sales and Related Occupations, Healthcare Practitioners & Support, Unknown, Office and Admin Support, Transportation/Material Moving, and Management Occupations (Other Job Type was the omitted reference category). The cohort variables, age, and weekly salary were forced to remain in the regression models. All other independent variables were selected based on a stepwise selection process in SAS.

The total costs for the non-cirrhotic HCV, CC, DCC, HCC, and LT cohorts were 3.0-, 3.4-, 6.0-, 14.6-, and 37.6-times greater than the control cohort costs, respectively. The total 6-month costs across all health benefit types for the six study cohorts are shown in Figure 2.

Figure 2. Regression-adjusted 6-month health benefit costs per employee. Sample sizes and significant differences for the different categories are presented in Table 6. *Significant difference compared to the control, p values and sample sizes are presented in table 6. LT, Liver transplant; HCC, Hepatocellular carcinoma; DCC, decompensated cirrhosis; CC, Compensated cirrhosis; HCV, hepatitis C virus.

Among the HCV subjects with employer absenteeism benefits, the STD absence days were significantly higher than the control cohort for the non-cirrhotic HCV, DCC, and LT cohorts (Table 6). The SL days for the DCC and non-cirrhotic HCV cohorts were higher than for the control cohort; however, only the difference between the DCC and control cohorts reached statistical significance (p < 0.05), while the difference between the non-cirrhotic HCV and control cohorts approached significance (p = 0.0555). The SL days between the control and LT cohorts were significantly different; however, they were lower for those with LTs. When combining all absence days, there were 1.71 more absence days for those with non-cirrhotic HCV, 5.93 more days for those with DCC, 18.7 more days for those with HCC, and 27.85 more days for those with a LT compared to the control cohort. The CC cohort had 0.36 fewer total absence days than the control cohort. Total absence days for the non-cirrhotic HCV, CC, DCC, HCC, and LT cohorts were 1.5-, 0.9-, 2.7-, 6.5-, and 9.2-times the control cohort, respectively, and are shown in Figure 3.

Figure 3. Regression-adjusted 6-month health benefit absence days per eligible employee. Sample sizes and significant differences for the different categories are presented in Table 6. *Significant difference compared to the control, p values and sample sizes are presented in table 6. LT, Liver transplant; HCC, Hepatocellular carcinoma; DCC, decompensated cirrhosis; CC, Compensated cirrhosis; HCV, hepatitis C virus.

Discussion

The social and economic burden of HCV infection is well documented in the literature10–12^,15^,21^,22. While anti-viral treatment improves the prognosis of even advanced liver disease, recent findings suggest that successful early treatment before the onset of cirrhosis may result in better long-term outcomes, both clinically and economically10^,11^,13. In the present study, direct costs (medical and prescription) of HCV cohorts were all significantly higher than those of the control cohort. Direct costs over the 6-month study period were higher with each progressive stage of liver damage, initially starting below $9000 for the non-cirrhotic chronic HCV cohort and reaching over $100,000 for the LT cohort. The increases in medical costs were disproportionately greater as the disease worsened. Prescription costs followed the same pattern (except for the CC and DCC cohorts), but increases were not as pronounced as those seen in medical costs. Co-morbid conditions such as anemia and white blood cell disorders were also 10–20-times higher in all HCV cohorts compared to the control cohort. Early treatment may, therefore, not only result in a reduction in associated direct and indirect costs, but also positively impact clinical outcomes by reducing the risk of further complications.

To date, few studies have evaluated the impact of HCV treatment on the employee and employer in the workplace. In this study, significant differences in SL and STD costs were evident between LT, DCC, and non-cirrhotic HCV cohorts compared to the control cohort. Although yet to be studied, it is reasonable to conclude that earlier diagnosis and treatment would have a positive impact on reducing health benefit costs and absenteeism associated with the more severe stages. Although the LT and HCC stages had low prevalence within the 728,974 employees, their impact in terms of costs and absences from work was quite large. Adjusted 6-month total costs (direct plus indirect) of these two stages averaged $107,541 and $41,744, respectively, and these employees had absent rates of 16–25% during the 6-month study period. Other stages of hepatitis C were more prevalent and still had significant impacts on costs and absences.

The newest CDC guidelines recommend that all persons born from 1945–1965 are screened at least once for HCV23, recognizing that up to 75% of HCV-infected individuals are unaware of their health status5. The barriers to effective HCV treatment have been historically challenging and can involve the patient, healthcare provider, and the healthcare system. Although HCV treatment options in recent years have greatly improved with respect to achieving a better clinical outcome, barriers to starting and completing therapy remain significant. Traditional HCV treatments have shown limited efficacy and are associated with significant toxicity, prompting many HCV-diagnosed patients to be non-compliant or choose to postpone treatment until more effective and less toxic alternatives become available24. In this study, the use of medications (ribavirin, peginterferon, telaprevir, and boceprevir) was low across the five HCV cohorts, with the highest prevalence of ribavirin (9%) and peginterferon (26%) use being in the CC cohort. Approximately 1.6% of the DCC cohort used telaprevir, and 0.8% used boceprevir. Another possible reason for the low utilization of these medications in the HCV cohorts may be due to contraindicated comborbidities (available online in the journal's Supplemental Table 3). Using ICD-9-CM codes, 84% of the LT cohort, 88% of the HCC cohort, 65% of the DCC cohort, 32% of the CC cohort, and 33% of the non-cirrhotic HCV cohort exhibited contraindications to interferon-based therapy (Table 5).

In this study, both costs and absences increased overall as the stages of HCV progressed from no cirrhosis, CC, DCC, HCC, and LT. The fact that the LT cohort had significantly more STD costs and days, but significantly fewer SL costs and days than the control cohort may have indicated previously expended SL banks that were used during early disease complications or while awaiting a transplant.

All costs and absences identified for subjects after their index date were included, regardless of whether or not hepatitis-C was listed on the claim; this allowed for comprehensive comparisons between cohorts. The objective measurement of absence time and costs are strengths of this study. Absence days were not subject to employee recall bias and were differentiated by type by obtaining this data from employee payroll records or disability/workers’ compensation claims. Indirect absenteeism costs were obtained from employee payroll records or disability/workers’ compensation claims data and were not imputed from average salary values. While this research on direct and indirect costs and the time associated with absenteeism was comprehensive, the focus was on absenteeism, not presenteeism. Prior work in this database documented trends toward reduced presenteeism associated with HCV compared to non-HCV controls15.

Although the rate of new infection with HCV declined by 90% from 1994 to 20062 3, the prevalence of more advanced liver disease is expected to rise sharply as the currently infected population ages14. Thus, the direct and indirect economic burden of HCV infection in the workplace is expected to further increase, as the cohort of HCV-infected individuals progress in their disease but remain of working age9^,11^,22^,25. To our knowledge, this is the first study to explore the indirect economic burden associated with varying stages of HCV liver disease and the only study of HCV-infected employees categorized by disease stage. Previous studies compared uncategorized HCV patients to controls or only addressed the effects on direct healthcare costs.

This research also used two-part stepwise regression methodology. The regression portion allowed the models to control for differences between the cohorts, and the stepwise procedure effectively incorporates the significances of the various parameters relative to each other. This effectively reduces the problems associated with the correlations of some variables including, for instance, age and tenure (years employed), which tend to be highly correlated—younger employees might not have been in the work pool long enough for their tenure to grow. The regression methodology requires independence between observations which precluded the inclusion of persons in more than one stage.

Limitations

As a recent study found only 35% of HCV patients to have private health insurance, the results of this study may not be generalizable to the entire HCV population25. Data for this study were obtained retrospectively from an employer database and did not include non-employed HCV-infected individuals, nor those who were employed but not covered by insurance. Because of the diminishing population with more severe disease, the protocol required people be included for their most severe stage only. The requirement of employment may have resulted in a less severe HCV population than otherwise observed. The continuous eligibility requirement may have influenced the generalizability of this study, as some stages of HCV are often too severe (LT and HCC) for the patient to continue employment and was the reason this study was limited to a 6-month time period. As a result, some cohorts’ small sample sizes were likely to impact the statistical significance, and the included subjects might not be representative of all subjects with these conditions. The patients in this study were identified based on ICD-9-CM codes, which failed to include those HCV-infected subjects that have not been diagnosed. Finally, because reimbursement is tied to accurate coding, clinician offices seek to code at the highest level that will be reimbursed, which may introduce a bias to classify subjects into more advanced stages of HCV infection.

Conclusion

HCV infection is associated with substantial direct and indirect costs. Costs are significantly greater in more advanced-stage HCV. In this study of HCV-infected employees, direct and indirect medical costs for each disease stage progressively contributed to total healthcare costs. The disease burden associated with HCV will continue to increase as more patients are diagnosed due to new testing recommendations and through the natural disease progression of the current HCV population. The at-risk HCV population is likely well represented by those who are active in the workforce; disease burden and treatment decisions will have significant implications with respect to overall health benefit costs. It is the payers of these health benefits who require information such as presented in this study to facilitate appropriate decision-making regarding the optimized management of HCV infection in their employed population.

Transparency

Declaration of funding

This research was sponsored by AbbVie. RAB, JES, RW, DRW, JCS, and NLK designed the study and developed the protocol. NLK and JWY conducted the analysis. All authors participated in the interpretation of the data. RAB drafted the initial manuscript with the assistance of Conny Burkett. All authors participated in the revisions to this manuscript. RAB and JES secured the funding for this manuscript with the assistance of JCS, DRW and RWB from AbbVie.

Declaration of financial/other relationships

JCS, DRW, and RWB are AbbVie employees and may hold AbbVie stock or options. RAB and JES are employees of the JeSTARx Group., NLK and JWY at the time of this research were employees of the HCMS Group. Both the JeSTARx and HCMS Groups received funding from AbbVie to conduct this research.

Acknowledgments

Additionally, the team thanks Conny Burkett, RPh, of Paradigm Consulting, Southlake, TX, for her contributions to interpretation of the data and for assistance in drafting the manuscript. Paradigm Consulting received funding from the JeSTARx Group for their assistance.

Previous presentations

Presented at Academy Health’s 2014 Annual Research Meeting, San Diego, CA, June 10, 2014. Baran R, Samp J, Walker D, et al. The direct and indirect costs of employees with hepatitis C vary by stage of disease. Available at: http://academyhealth.org/files/ARM/photos/2014%20ARM%20Posters2.pdf (p 277–8), accessed August 14, 2014.