Quality of Life Research

, Volume 17, Issue 5, pp 715–724

The impact of chronic hepatitis C and co-morbid illnesses on health-related quality of life

Authors

  • Jeffrey W. Kwan
    • Division of Gastroenterology and Hepatology, Department of MedicineVA Palo Alto Health Care System, (154C), 3801 Miranda Avenue
    • Stanford University School of Medicine
  • Ruth C. Cronkite
    • Stanford University School of Medicine
    • Center for Health Care EvaluationVA Palo Alto Health Care System
    • Department of SociologyStanford University
    • Center for Primary Care and Outcomes ResearchStanford University School of Medicine
  • Antony Yiu
    • Center for Health Care EvaluationVA Palo Alto Health Care System
    • Department of SociologyStanford University
  • Mary K. Goldstein
    • Geriatric Research, Education, and Clinical CenterVA Palo Alto Health Care System
    • Center for Primary Care and Outcomes ResearchStanford University School of Medicine
  • Lewis Kazis
    • Center for Health Quality, Outcomes, and Economic ResearchEdith Nourse Rogers Memorial Veterans Hospital
    • Division of Gastroenterology and Hepatology, Department of MedicineVA Palo Alto Health Care System, (154C), 3801 Miranda Avenue
    • Stanford University School of Medicine
Article

DOI: 10.1007/s11136-008-9344-3

Cite this article as:
Kwan, J.W., Cronkite, R.C., Yiu, A. et al. Qual Life Res (2008) 17: 715. doi:10.1007/s11136-008-9344-3

Abstract

Objectives

Determine the relative impact of chronic hepatitis C (CHC) and co-morbid illnesses on health-related quality of life (HRQoL) in 3023 randomly selected veterans with known hepatitis C virus antibody (anti-HCV) status who previously completed a veteran-specific HRQoL questionnaire (SF-36V).

Methods

Multiple regression analyses were performed to measure the relative contribution of anti-HCV status, four demographic variables, and ten common medical and six psychiatric co-morbidities to HRQoL between 303 anti-HCV(+) and 2720 anti-HCV(−) patients.

Results

Anti-HCV(+) veterans were younger, reported a lower HRQoL on seven of eight 36-Item Short Form Health Survey for Veterans (SF-36V) subscales (P ≤ 0.001) and the mental component summary (MCS) scale (P < 0.001). The ten medical and six psychiatric co-morbidities had variable impact on predicting lower HRQoL in both groups. After adjusting for demographic variables and co-morbid illnesses, we found that anti-HCV(+) patients reported a significantly lower MCS score (P < 0.001) and a trend toward a lower physical component summary (PCS) score (P < 0.07) compared to anti-HCV(−) veterans. Among the anti-HCV(+) veterans, co-morbid medical illnesses contributed to impaired PCS but not to MCS.

Conclusions

Veterans with CHC were younger than HCV(−) veterans and hence less likely to have other co-morbid medical illnesses. Medical co-morbidities seen in those veterans with CHC contribute to impaired PCS but not MCS. Anti-HCV(+) status negatively affects HRQoL, particularly MCS, independently of medical and psychiatric co-morbidities.

Keywords

Chronic hepatitis CCo-morbid conditionsQuality of lifeVeterans

Abbreviations

Anti-HCV

Antibody to hepatitis C virus

CHC

Chronic hepatitis C

COPD

Chronic obstructive pulmonary disease

EIA

Enzyme-linked immunoassay

HCV

Hepatitis C virus

HRQoL

Health-related quality of life

ICD-9

Ninth version of the Clinical Modification of the International Classification of Diseases

IRT

Item response theory

MCS

Mental component summary

PCS

Physical component summary

PTSD

Posttraumatic stress disorder

RP

Relative precision

SF-36

36-Item Short Form Health Survey

SF-36V

36-Item Short Form Health Survey for Veterans

VAPAHCS

VA Palo Alto Health Care System

VHA

Veterans Health Affairs

VISN

Veterans Integrated Service Network

Introduction

An estimated 3.2 million individuals in the USA are infected with chronic hepatitis C (CHC) [1], and CHC is the leading indication for liver transplantation [2]. Most patients are asymptomatic, and symptoms, if present, tend to be nonspecific. Patients often attribute symptoms such as fatigue to CHC rather than to other co-morbid illnesses. Health-related quality of life (HRQoL) in this patient population is impaired relative to that of healthy individuals or those with chronic hepatitis B [38]. These impairments are not associated with age, sex, mode of transmission, substance abuse history, social support, serum aminotransferase levels, or liver histology [7, 9, 10]. In contrast, co-morbid illnesses, particularly psychiatric illnesses, are common among patients with CHC [11, 12].

We have previously demonstrated the relative impact of CHC and co-morbid psychiatric illnesses on HRQoL in U.S. veterans [13], but much less is known about the effect of co-morbid medical illness on HRQoL [7, 14]. The presence and number of active medical illnesses (defined as a chronic somatic, non-psychiatric condition requiring treatment and monitoring) had strong correlations with 36-Item Short Form Health Survey (SF-36) summary and subscale scores among 107 interferon treatment failure patients screened for participation in a clinical trial [7]. In particular, patients with painful medical co-morbidities (e.g. arthritis and back pain) had lower modified SF-36 scores than did patients without these co-morbidities. Among the 220 CHC patients seen at their referral hepatology clinic of the same institution, the presence of co-morbid illness was the most important predictor of impaired HRQoL [14]. Among patients with two co-morbid medical illnesses, there were significant reductions in SF-36 scores in two subscales and the physical component summary (PCS) score. With three or more co-morbid medical illnesses, significant reductions in SF-36 scores were seen in six subscales and the PCS scale, compared to CHC patients with no co-morbid illness. Both of these studies were limited by referral and selection bias, the use of population controls instead of a true control group with negative hepatitis C virus antibody [anti-HCV(−)], and lack of determination of the relative contribution of each co-morbid illness.

The U.S. veteran population represents a unique cohort with a higher prevalence of CHC than the general population [1517]. Veterans receiving health care from Veterans Health Affairs (VHA) facilities have more complex medical and mental health problems than their counterparts using non-VHA facilities or the general population [18, 19], resulting in significantly reduced HRQoL [20, 21]. In 1999, a large-scale HRQoL survey not targeted to any specific medical or psychiatric illness was conducted on randomly selected veterans. While most published studies on HRQoL in CHC patients have been conducted in highly selected populations and/or in the context of a large clinical trial, the VHA Large Health Survey is based on a large sample of randomly selected VHA enrollees, not targeted to any specific illness or clinic patient population, thereby minimizing selection or referral bias. This affords us a valuable opportunity to investigate the impact of co-morbid illnesses on HRQoL in unselected veterans with CHC as compared to uninfected veterans. We hypothesized that not only is HRQoL impaired among veterans with CHC compared to those without CHC, but that co-morbid medical and psychiatric illnesses also contribute significantly to these impairments independent of CHC.

There is an increasing need to study the interaction of multiple diseases, partly due to an aging population with an increasing prevalence of comorbidities. Co-occurring conditions adversely affect function, QoL, and mortality [22], where the degree of impairment of the physical domains of QoL is associated with the number of medical comorbid illnesses [23]. However, studies frequently exclude psychiatric comorbidity and other confounding factors [22, 23]. In the study reported here, we evaluate the relative impact of comorbid somatic and mental conditions, controlling for demographics, on HRQoL. Since the majority of CHC veterans are not candidates for anti-viral therapy, insight into the impact of comorbid conditions on HRQoL will point to interventions other than anti-viral therapy that may help to improve the HRQoL of these veterans. This is especially important for gastroenterologists and hepatologists, who care for patients with CHC, since such insight may facilitate coordination with other health care providers in terms of treatment for comorbid conditions.

Methods

Study population

The study sample comprised 3023 American veterans enrolled in the VA Palo Alto Health Care System (VAPAHCS) who had participated in the 1999 Large Health Survey of Veterans and had been tested for hepatitis C antibodies (anti-HCV). The anti-HCV status was determined from laboratory records of more than 20,000 veterans at the VAPAHCS who were tested between July 1992 and December 2004, using either a second or third generation enzyme-linked immunoassay (EIA II or EIA III, respectively; Abbott Laboratories, Abbott Park, IL). The study protocol was approved by the Stanford University Human Subjects Panel.

Large health survey

The 1999 Large Health Survey of Veterans randomly sampled 1.4 million veteran enrollees nationally between July 1999 and January 2000 for self-reported demographic and health information using a mailed questionnaire [24]. The Survey included information on age, sex, ethnicity, education level, and marital status in addition to HRQoL. The overall response rate was 63.1% for the national sample and 62.4% for the VAPAHCS sample.

Health-related quality of life

The HRQoL was measured using the self-administered 36-Item Short Form Health Survey for Veterans (SF-36V), a veteran-specific version of the Medical Outcomes Study SF-36 [25], a widely used instrument for assessing disease non-specific HRQoL [26]. The SF-36 has two summary component scales [PCS and mental component summary (MCS) scores] that are the weighted sums of eight subscales [27]. The SF-36V version has been validated in the veteran population [25] and normalized to the general U.S. population [28].

Co-morbid medical illnesses

For the purpose of this study, the term medical illness refers to somatic illness only and does not include any psychiatric or mental illnesses. The co-morbid medical illnesses were obtained from the VHA administrative databases based on the diagnosis using the Ninth Version of the Clinical Modification of the International Classification of Diseases (ICD-9). Among all of the major medical diagnoses available in the VHA administrative databases, we selected only those diagnoses with a prevalence of at least 5% in the anti-HCV(+) cohort for analysis. Similar ICD-9 diagnoses were grouped for analysis. For example, osteoarthritis (715.9), other arthritis (V13.4, V17.7), and rheumatoid arthritis (714.0) were combined as “Arthritis” for the purpose of this study. Using these pre-defined criteria, the ten most common conditions in the anti-HCV(+) cohort included in our analysis were: hypertension (401.9), low back pain (724.2), arthritis (715.9, V13.4, V17.7, 714.0), diabetes mellitus (250.0, 648.0), chronic obstructive pulmonary disease (COPD)/asthma (490–496, 493.9, V17.5), cancer (199.1), enlarged prostate (600.0), cataract (366.9/366.1), anemia (285.9), and urinary tract infection (599.0, V13.02).

Co-morbid psychiatric illnesses

As with the co-morbid medical illnesses, several of the ICD-9 psychiatric diagnoses of a similar nature were grouped together for the purpose of our analysis. The six most common diagnoses included in the study were: anxiety (300.0, 300.2), depression (296.2, 296.3, 300.4, 311), bipolar disorder (296.0, 296.1, 296.4, 296.5, 296.7, 296.8, 296.9), schizophrenia (295), posttraumatic stress disorder (PTSD) (309.81), and alcohol dependence (303, 305.0).

Statistical analysis

Comparisons between the anti-HCV(+) and anti-HCV(−) subgroups

Anti-HCV(+) veterans were compared with anti-HCV(−) veterans on baseline demographic characteristics (age, sex, ethnicity, educational level, and marital status), the prevalence of each of ten co-morbid medical illnesses and six co-morbid psychiatric illnesses, and the eight subscales and two summary scales of the SF-36V survey. Chi-square tests of association were used to compare dichotomous variables, and independent group t tests were used to compare continuous variables.

Predictors of HRQoL

Multiple regression analyses (Ordinary Least Squares estimation) were estimated to examine the effect of anti-HCV status relative to each of ten co-morbid medical illnesses and six psychiatric illnesses on HRQoL. Two separate regression equations were estimated, one each for the PCS and MCS. The predictor variables included a set of four demographic characteristics (age, ethnicity, education level, and marital status), a dichotomous variable for anti-HCV status (1 = positive, 0 = negative), and a set of dichotomous variables for the presence of each of ten medical and six psychiatric illnesses.

Differential effects of co-morbid medical conditions on HRQoL

To investigate whether the medical illnesses might have a variable effect on HRQoL depending on a veteran’s anti-HCV status, separate regression equations were estimated for each anti-HCV status subgroup. Two regression models were estimated for each, one for the PCS and the other for the MCS. The same demographic, and medical and psychiatric illness variables were used as predictors in all models. The F statistic associated with the Chow test was used to determine whether the set of regression coefficients differed across the regression models for the anti-HCV(+) and the anti-HCV(−) subgroups [29]. All analyses were conducted using SPSS ver. 11.5. (SPSS, Chicago, IL)

Results

Comparison of demographic characteristics, co-morbid medical and psychiatric illnesses, and HRQoL based on anti-HCV status

Compared to the anti-HCV(−) individuals, the anti-HCV(+) patients were younger, significantly more likely to be male, nonwhite, unmarried, and less likely to have at least a high school education. The anti-HCV(+) individuals were significantly more likely to have co-morbid psychiatric illnesses (alcohol dependence, bipolar disorder, depression, PTSD, and schizophrenia) and low back pain, but less likely to have cancer, cataracts, enlarged prostate, or hypertension. Overall, the average number of medical conditions for the anti-HCV(+) group was significantly lower than that of the anti-HCV(−) group, but the average number of psychiatric conditions was significantly higher for the anti-HCV(+) group (Table 1).
Table 1

Demographic characteristics and prevalence of medical and psychiatric diagnoses in anti-HCV(+) and anti-HCV(−) veterans (n = 3023)

 

Anti-HCV (+) (n = 303)

Anti-HCV(−) (n = 2720)

P valuea

Demographic characteristics

   Male (%)

99.0

96.0

.021

   Nonwhite (%)

42.9

33.2

.001

   Unmarried (%)

67.3

50.3

<.001

   Education ≥ 12 years (%)

50.8

59.4

.004

   Mean age, years (SD)

52.8 (9.05)

60.8 (12.66)

<.001

Medical illnesses

   Anemia (%)

8.9

10.6

.355

   Arthritis (%)

30.0

31.5

.591

   Cancer (%)

12.5

19.7

.003

   Cataract (%)

10.2

18.8

<.001

   COPD (%)

12.9

16.8

.082

   Diabetes (%)

18.5

22.2

.137

   Enlarged prostate (%)

11.6

23.2

<.001

   Hypertension (%)

42.9

59.6

<.001

   Low Back Pain (%)

26.4

20.4

.014

   Urinary Tract Infection (%)

7.9

7.9

.990

   Mean no. of medical illnesses

1.8

2.3

<.001

Psychiatric illnesses

   Alcohol dependence (%)

39.6

16.0

<.001

   Anxiety (%)

13.9

11.2

.170

   Bipolar Disorder (%)

11.6

6.3

<.001

   Depression (%)

38.0

24.3

<.001

   PTSD (%)

22.4

13.2

<.001

   Schizophrenia (%)

12.2

7.2

.002

   Mean no. of psychiatric illnesses

1.4

0.8

<.001

Anti-HCV, Hepatitis C virus antibody status; COPD, chronic obstructive pulmonary disease; PTSD, postraumatic stress disorder

aP values based on χ2 for dichotomous variables and independent groups t tests for continuous variables (age, number of medical and psychiatric conditions)

Compared to the anti-HCV(−) veterans, the anti-HCV(+) veterans had significantly lower MCS scores as well as significantly lower mean SF-36V scores on seven of eight domains. The two groups had similar PCS scores. While the mean PCS and MCS scores for the anti-HCV(−) group were comparable to the VA national average and the mean for veterans from this region [Veterans Integrated Service Network (VISN) 21] and the VAPAHCS, the mean MCS score for the anti-HCV(+) group was considerably lower than the mean scores for each of these comparison groups (Table 2).
Table 2

Health related quality of life (HRQoL) scores in anti-HCV(+) and anti-HCV(−) veterans (n = 3023)

 

Anti-HCV (+) (n = 303)

Anti-HCV(−) (n = 2720)

P valuea

Domain (0–100), mean (SD)

   Physical functioning

55.3 (29.3)

56.5 (29.4)

.505

   Role physical

29.5 (40.5)

38.3 (41.7)

.001

   Bodily pain

41.8 (27.6)

49.5 (27.2)

<.001

   General health

41.1 (25.1)

49.6 (24.7)

<.001

   Vitality

37.2 (24.3)

44.7 (24.8)

<.001

   Social functioning

45.6 (31.0)

58.8 (31.1)

<.001

   Role emotional

36.3 (45.1)

54.4 (48.1)

<.001

   Mental health

51.2 (24.7)

64.4 (23.8)

<.001

Summary component scores (0–100)

   Physical component summary score, PCS (SD)

36.1 (11.7)

36.8 (11.9)

.385

   Mental component summary score, MCS (SD)

36.9 (14.0)

44.5 (14.1)

<.001

 

VA national average

VISN 21

Palo Alto

Comparison scores

   PCS

36.90

38.17

38.37

   MCS

45.08

45.55

45.52

aP values based on t-test for independent groups

Multiple regression analyses of predictors of HRQoL

The relative contribution of anti-HCV status, demographic characteristics, and co-morbid medical and psychiatric illnesses on HRQoL is shown in Table 3. The overall explained variance (adjusted R2) in the PCS score for the total sample was 0.20, signifying that 20% of the variation in the PCS scores was accounted for by the 21 predictor variables (anti-HCV status, four demographic, ten medical illnesses, and six psychiatric illnesses) in the model. Similarly, the 21 predictor variables accounted for 27% of the variation in the MCS scores.
Table 3

Effect of demographic characteristics, medical and psychiatric diagnoses, and anti-HCV status on physical and mental component scores

 

Total sample (n = 3023)

Anti-HCV(+) group (n = 303)

Anti-HCV(−) group (n = 2720)

 

PCS

MCS

PCS

MCS

PCS

MCS

Constant

43.47

39.33

40.14

40.18

43.42

39.23

Anti-HCV status

  Anti-HCV(+)

−1.21

−3.57***

    

Demographic characteristics

  Nonwhite

−0.61

−0.70

−0.44

−0.82

−0.60

−0.69

  Unmarried

2.16***

−0.79

2.34

0.46

2.15***

−0.91

  Education ≥ 12 years

1.99***

3.17***

1.12

3.80**

2.13***

3.12***

  Age

−0.03

0.11***

0.05

-0.02

−0.04

0.12***

Medical illnesses

  Anemia

−2.36***

−0.96

−3.63

−1.90

−2.26***

−0.83

  Arthritis

−5.31***

−0.56

−6.60***

−0.75

−5.15***

−0.55

  Cancer

0.16

0.43

0.75

4.30

0.08

0.11

  Cataract

−1.02

0.18

−1.62

1.60

−0.99

0.19

  COPD

−4.00***

−0.78

−5.82**

0.83

−3.87***

−0.88

  Diabetes

−2.53***

−1.22*

−3.82*

1.22

−2.39***

−1.41*

  Enlarged prostate

0.75

0.06

−0.56

0.35

−0.80

0.14

  Hypertension

−1.24**

0.56

−3.37**

2.01

−1.00*

0.37

  Low back pain

−5.35***

−2.00***

−3.64**

−0.23

−5.56***

−2.24***

  Urinary tract infection

−3.40***

0.56

−2.35

−1.72

−3.53***

0.84

Psychiatric illnesses

  Alcohol dependence

0.29

−2.05**

−0.25

0.10

0.45

−2.51***

  Anxiety

0.83

−3.23***

3.92*

−6.32**

0.51

−2.70***

  Bipolar disorder

0.42

0.11

−0.64

2.17

0.68

-0.37

  Depression

−0.03

−6.55***

0.12

−7.50***

0.02

−6.54***

  PTSD

−2.08**

−9.74***

−2.99

−7.35***

−1.94**

−10.21***

  Schizophrenia

0.85

−2.22*

1.50

−4.22

0.85

−1.83

Adjusted R2

.20***

.27***

.22***

.19***

.19***

.26***

 P ≤ 0.10, * P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001

The numbers in the table correspond to regression coefficients, except for the last row where the numbers represent the proportion of explained variance (adjusted for the number of predictors)

Effect of anti-HCV status

Anti-HCV(+) status was associated with a lower overall HRQoL. The anti-HCV(+) individuals scored 3.57 points lower on the MCS than anti-HCV(−) veterans (P ≤ 0.001) while the PCS score was 1.21 points lower among anti-HCV(+) veterans (P ≤ 0.07).

Effects of demographic characteristics

Having a high school education or above was associated with higher PCS and MCS scores in all groups except for one [PCS in the anti-HCV(+) group]. Being unmarried was associated with higher PCS scores for the total sample and the anti-HCV(−) subgroup. Age was associated with a higher MCS score for the total sample and the anti-HCV(−), group and for each increase of 1 year in age, the MCS score increased by 0.11 and 0.12, respectively.

Effects of co-morbid medical illnesses on PCS score

Total sample. Of the ten co-morbid medical illnesses evaluated, seven (anemia, arthritis, COPD, diabetes, hypertension, low back pain, and urinary tract infection) appeared to be associated with a significantly lower PCS score, with low back pain and arthritis having the strongest association, after controlling for anti-HCV status, demographics, and other co-morbid illnesses. Veterans with low back pain and arthritis had PCS scores that were 5.35 (P ≤ 0.01) and 5.31 points (P ≤ 0.01), respectively, lower than veterans without these illnesses. Overall, these seven medical illnesses demonstrated a stronger negative association with the PCS score than did anti-HCV status (−1.21).

Based on anti-HCV status. To evaluate whether co-morbid medical illnesses might have a differential effect on the PCS score depending on HCV status, separate regression models were estimated for the anti-HCV(+) and anti-HCV(−) subgroups. The coefficients associated with the demographic, medical and psychiatric illness variables remained essentially the same (i.e., the Chow test [29] for equality of the coefficients across both samples was not significant). Among the anti-HCV(+) group, five (arthritis, COPD, diabetes, hypertension, low back pain) of the ten medical illnesses had a significant negative effect on the PCS score. Among the anti-HCV(−) group, seven of the ten medical illnesses (except cancer, cataract and enlarged prostate) had a significant negative effect on the PCS score. Taken together, the predictors accounted for 22% of the variation observed in the PCS score for the anti-HCV(+) group, and 19% of the variation observed in the PCS score for the anti-HCV(−) group.

Effects of co-morbid medical illnesses on MCS score

Total sample. After controlling for anti-HCV status, demographics, medical and psychiatric illnesses, patients with diabetes and low back pain scored 1.22 (P = 0.03) and 2.00 points (P ≤ 0.01), respectively, lower than patients without these diagnoses. The negative effects of these two medical illnesses were not as strong as the effect of anti-HCV(+) status (−3.57) on the MCS score.

Based on anti-HCV status. Taken together, the set of regression coefficients were found to differ across subgroups when predicting the MCS score [F (21, 2981) = 1.98; P ≤ 0.01]. Among the anti-HCV(+) group, none of the medical illnesses had a significant effect on the MCS score. In the anti-HCV(−) group, two medical illnesses had small but significant negative effects on the MCS score: diabetes (−1.41; P ≤ 0.02) and low back pain (−2.24; P ≤ 0.01). Taken together, the predictors accounted for 19% of the variation in the MCS score for the anti-HCV(+) group and 26% of the variation in the MCS score for the anti-HCV(−) group.

Effects of co-morbid psychiatric illnesses on PCS score

Total sample. Of the six co-morbid psychiatric illnesses evaluated, only PTSD was associated with a significantly lower PCS score. After controlling for anti-HCV status, demographic characteristics, and co-morbid illnesses, veterans with PTSD scored 2.08 points lower than veterans without PTSD (P ≤ 0.01). This diagnosis appears to have a somewhat stronger negative effect than anti-HCV(+) status (−1.21) on the PCS score.

Based on anti-HCV status. In the anti-HCV(+) group, anxiety was associated with a higher PCS score (3.92; P ≤ 0.04), while PTSD was associated with a lower PCS score for both the anti-HCV(+) (−2.99, P ≤ 0.08) and anti-HCV(−) groups (−1.94; P ≤ 0.01).

Effects of co-morbid psychiatric illness on MCS score

Total sample. Five psychiatric illnesses had large negative effects on the MCS score: alcohol dependence (−2.05; P ≤ 0.01), anxiety (−3.23; P ≤ 0.01), depression (−6.55; P ≤ 0.01), PTSD (−9.74; P ≤ 0.001), and schizophrenia (−2.22; P ≤ 0.02). The negative effects of both depression and PTSD on the MCS score were stronger than the effect of anti-HCV(+) status (−6.55 and −9.74 vs. −3.57, respectively).

Based on anti-HCV status. After separate regression equations had been estimated for the two anti-HCV subgroups, the results were similar—with two exception. Three psychiatric illnesses (anxiety, depression, and PTSD) had significant negative effects on the MCS score across both anti-HCV status groups, but the effect of PTSD on MCS was somewhat stronger for the anti-HCV(−) group . Alcohol dependence was also associated with a lower MCS score in the anti-HCV(−) group only (−2.51; P ≤ 0.01).

Patient scenarios demonstrating effects of co-morbid illnesses on predicted PCS and MCS scores

In order to provide concrete examples of how co-morbid illnesses influence HRQoL, we used the coefficients from the regression analyses to estimate PCS and MCS scores for an “average” veteran with CHC (i.e. 53 year-old, white, unmarried, high school-educated male, Table 1), who may have various additional co-morbid medical and psychiatric illnesses (Table 4).
Table 4

Predicted PCS and MCS scores for various common conditions among anti-HCV(+) patients

Common medical conditions among anti-HCV(+) patients

Predicted PCSa

Predicted MCSa

Anti-HCV(+) with no co-morbid illnesses

44.35

43.74

Co-morbid medical illnesses

  Hypertension

40.71

45.75

  Anemia

40.72

41.84

  Diabetes

40.53

45.75

  Arthritis

37.75

42.99

  Hypertension and diabetes

37.75

46.97

  Anemia and diabetes

36.90

48.06

  Anemia and arthritis

34.12

43.74

  Hypertension, diabetes, and arthritis

30.56

46.22

  Anemia, hypertension, and diabetes

33.53

45.07

  Anemia, arthritis, hypertension, and diabetes

26.93

44.32

Co-morbid psychiatric illnesses

  Alcohol dependence

44.10

43.84

  Depression

44.47

36.24

  PTSD

41.36

36.39

  PTSD and alcohol dependence

41.11

36.49

  PTSD and depression

41.48

28.89

  PTSD, depression, and alcohol dependence

41.23

28.99

Co-morbid medical and psychiatric illnesses

  Hypertension, diabetes, and PTSD

34.17

38.87

  Hypertension, diabetes, and alcohol dependence

33.92

39.72

  Hypertension, diabetes, depression, and alcohol dependence

37.03

39.57

  Hypertension, diabetes, arthritis, PTSD, and depression

27.69

31.37

  Hypertension, diabetes, arthritis, PTSD, depression, and alcohol dependence

27.44

31.47

  Anemia, hypertension, diabetes, arthritis, PTSD, depression, and alcohol dependence

23.81

29.57

aAll predicted PCS and MCS scores were generated for a 53 year-old, unmarried patient with a high school education or more

Predicted PCS score

An average veteran who is anti-HCV(+) with no co-morbid medical or psychiatric conditions has a predicted PCS score of 44.4. An anti-HCV(+) veteran with the same baseline demographic characteristics, but with some additional medical co-morbidity (either hypertension, anemia, or diabetes) has a predicted PCS score of between 40.5 and 40.7. Arthritis has a stronger negative effect and would lead to a predicted PCS score of 37.8. If this individual were to have two medical co-morbidities, his predicted PCS score could vary between 34.1 (for anemia and arthritis) to 37.8 (for hypertension and diabetes), which is roughly similar to the average PCS score for the veterans in this sample and nationwide (36.1 and 36.9, respectively, see Table 2), and what would be expected given that the average number of medical co-morbidities for this sample is 1.8. As the number of co-morbid medical conditions increases, we see that the predicted PCS scores continue to drop, with a particularly low predicted score of 26.9 for a veteran with four co-morbid medical conditions (anemia, arthritis, hypertension and diabetes). If this individual were to have one co-morbid psychiatric condition, his predicted PCS score would range from 41.4 for someone with co-morbid PTSD to 44.1 and 44.5 for someone with alcohol dependence or depression, respectively. The addition of a second or even a third co-morbid psychiatric illness for someone with a diagnosis of PTSD has little additional effect on the predicted PCS scores, which range from 41.1 to 41.5. The combination of two co-morbid medical conditions and one co-morbid psychiatric condition results in predicted PCS scores of 33.9 (for hypertension, diabetes and alcohol dependence) to 34.2 (for hypertension, diabetes and PTSD), which is only slightly lower than the average PCS score for this study sample, which typically has on average 1.8 co-morbid medical conditions and 1.4 co-morbid psychiatric conditions (see Table 1). Additional predicted PCS scores are given in Table 4 for various combinations of co-morbid medical and psychiatric conditions. Overall, the pattern of predicted scores suggests that co-morbid medical conditions have more of a negative impact on predicted PCS scores than on predicted MCS scores.

Predicted MCS score

An average veteran who is anti-HCV(+) with no co-morbid medical or psychiatric conditions has a predicted MCS score of 43.7. Interestingly, the addition of co-morbid medical conditions results in predicted MCS scores that are similar, ranging from 41.8 for someone with co-morbid anemia to 45.8 for someone with hypertension or diabetes. The same individual with co-morbid depression or PTSD has a much lower predicted MCS score of 36.2–36.4, which is similar to the average MCS score for this study sample. A veteran with PTSD and depression has a particularly low predicted MCS score of 28.9. Overall, the pattern of predicted scores suggests that a co-morbid psychiatric condition has a much greater negative impact on predicted MCS scores than does a co-morbid medical condition.

Discussion

Our findings demonstrate that, compared to anti-HCV(−) controls, anti-HCV(+) veterans are more likely to have psychiatric illnesses. With regard to medical illnesses, they are more likely to have low back pain but less likely to have cancer, cataracts, hypertension, and prostate problems. The differences in prevalence of these medical co-morbid conditions are probably due to younger age (on average about 8 years younger) of the anti-HCV(+) group. Overall, the anti-HCV(+) and (−) groups have similar PCS scores, but the anti-HCV(+) group has a statistically lower MCS score. After controlling for four demographic variables as well as ten medical and six psychiatric co-morbidities, we found that positive anti-HCV status was associated with a relatively large and statistically significant negative effect on the MCS score (3.57 points) and a trend toward a smaller negative effect on the PCS score (1.21 points). The ten medical and six psychiatric co-morbidities had a variable impact in reducing the HRQoL in both groups. In the anti-HCV(+) group, co-morbid medical illness was more strongly associated with the PCS score, whereas co-morbid psychiatric illness was more strongly associated with the MCS score. Consistent with our previous findings [13], we found that anti-HCV(+) status had a stronger overall negative effect on the MCS than the PCS score, even following the adjustment for presence of co-morbid psychiatric diagnoses.

This study consisted of randomly selected veterans representing the full spectrum of CHC disease, ranging from asymptomatic disease to decompensated cirrhosis, and was not limited to those with liver biopsies and/or meeting eligibility criteria for a particular clinical trial. A chart review of the anti-HCV(+) group revealed that, at the time the survey was completed, 3% had documented compensated cirrhosis, and another 3% had decompensated cirrhosis (data not shown). The large sample size and use of internally verified anti-HCV(−) controls also serves to strengthen our findings.

Our anti-HCV(+) group was significantly younger and therefore less likely to have the medical problems more commonly found in an older population, such as cancer and prostate problems. Low back pain was the only medical diagnosis that was found to be more common among the anti-HCV(+) group. This finding is consistent with other disease prevalence studies of U.S. veterans [21]. Of the ten medical diagnoses included in our analysis, arthritis, COPD, diabetes, and low back pain had the strongest negative impact on the PCS score in the anti-HCV(+) group. This result is similar to that of a recent study, which determined that cancer history and the presence of two or more chronic medical conditions were associated with significantly lower PCS scores among CHC patients [30]. The anti-HCV(+) status resulted in a minor and statistically insignificant reduction in the PCS score. Our findings suggest that among CHC individuals, co-morbid medical illnesses contribute more to impaired PCS than MCS. Interestingly, none of the ten medical diagnoses had a statistically significant impact on the MCS score in the anti-HCV(+) group.

The results of this study demonstrate that anti-HCV(+) status in our study cohort was associated with significantly lower MCS scores, independent of medical or psychiatric co-morbid illnesses. Several studies have explored the proposed effect of “labeling” on patients’ perception of health [10, 31]. In one study, the majority of patients with the diagnosis of CHC perceived financial insecurity, internalized shame, and social rejection, regardless of the method of HCV acquisition or socioeconomic status [32]. Our study does not allow us to determine any labeling effect on HRQoL. If the hypothesis is supported that perception of financial insecurity, internalized shame, and social rejection are intermediators of the relation between CHC diagnosis and reduced mental health, then a possible intervention would be counseling and educating all CHC patients at the time of diagnosis to minimize perceptions of stigma associated with CHC. Such counseling could help alleviate the labeling and associated psychological distress that may negatively impact one’s HRQoL. The labeling effect and the benefit of intervening need to be examined in future research.

Our findings have several limitations. First, as this is a cross-sectional study, one must use caution in inferring causality between anti-HCV status, co-morbid illnesses, and HRQoL. The findings suggest associations among these factors instead of cause and effect. Second, the anti-HCV status was determined over an extended period of time, and approximately one-third of the anti-HCV(+) subjects and three-fourths of the anti-HCV(−) subjects had testing done after 1999, when the Survey began. Given the chronic nature of hepatitis C and the recent decline in newly infected cases, our presumption was that the HCV antibody was unchanged from the time the Survey was administered. We had only a few cases of acute hepatitis C during the study period (data not shown). We also do not know the proportion of veterans who were aware of their anti-HCV status at the time of the survey. It should be noted that not knowing the antibody status would minimize the “labeling” effect. Third, the anti-HCV antibody is used as a surrogate marker for CHC infection and, as such, a small proportion of these veterans could represent false-positives. Prior studies have shown that only 3.6% of veterans testing positive for anti-HCV represent false positives [33], which is similar to our previous findings [17]. Fourth, the determination of medical and psychiatric illnesses was derived from ICD-9 diagnoses in the VHA administrative databases, which can be subject to coding errors and may not be entirely accurate. However, the prevalence of many of the co-morbidities in our cohort is similar to what has been reported in other studies [21, 34]. In addition, the validity of grouping some of the ICD-9 diagnoses together has not been established. Fifth, we relied on summated subscale scores of ordinal items from the SF-36V to measure the dimensions of the HRQoL. These limitations stem from constraints on access to archival data from the veteran survey. Such summated scores are often limited by the lack of: (1) an ordered continuum of items that reflect a unidimensional construct and (2) interval-level items that can be algebraically added [35]. We acknowledge that alternative models of scoring, such as Rasch Item Response Theory (IRT) scaling models where raw scores are transformed into latent trait variables, may have yielded improved relative precision (RP) in discriminating between subgroups of patients, especially those at the extremes of the score distributions [3538]. Sixth, our subjects are predominantly male and chose to receive medical care from the VHA; consequently, they may differ somewhat from the general population [19, 39]. We also did not stratify patients according to the severity of their liver disease and medical or psychiatric co-morbid conditions. Previous studies have found a poor correlation between the HRQoL and liver test abnormalities [9, 40]. Finally, because the Survey was self-administered, patients who were very sick or hospitalized were perhaps less likely to return the Survey. However, the large sample size and random selection process increases the likelihood that the whole spectrum of disease was represented in our analysis.

In summary, anti-HCV(+) status is associated with a significantly impaired HRQoL among American veterans. The MCS score was statistically significantly lower while the PCS score was not. Our results suggest that a hepatitis C diagnosis impacts HRQoL negatively, independent of co-morbidities, but that the presence of co-morbid medical illness can further diminish HRQoL, particularly the physical components. Co-morbid psychiatric conditions affect mostly the MCS score, with depression and PTSD having a particularly large negative impact on HRQoL for both groups of veterans. We believe that one could improve the HRQoL in these individuals by properly diagnosing and optimizing the treatment of underlying co-morbid illnesses regardless of anti-viral therapy. Our findings are particularly relevant since less than 12% of veterans with CHC received anti-viral therapy; co-morbid medical (e.g., anemia) or psychiatric conditions (e.g., depression, alcohol dependence, schizophrenia, bipolar disorder) are predictive of non-treatment for CHC [41]. Lessons learned from the VA health care system can provide a model for monitoring the outcomes of care in other managed care systems. Future study should be performed to validate the predictive model in an independent sample and to evaluate the labeling effect.

Acknowledgements

This study is supported by Department of Veterans Affairs Health Services Research and Development Service research funds and VA grant LIP 62–096. Views expressed are those of the authors and not necessarily those of the Department of Veterans Affairs.

Copyright information

© Springer Science+Business Media B.V. 2008