fax: (33) 1 44 738 462
Chronic hepatitis C virus infection: Does it really impact health-related quality of life? A study in rural Egypt
Article first published online: 24 NOV 2004
Copyright © 2004 American Association for the Study of Liver Diseases
Volume 40, Issue 6, pages 1434–1441, December 2004
How to Cite
Schwarzinger, M., Dewedar, S., Rekacewicz, C., Elaziz, K. M. A., Fontanet, A., Carrat, F. and Mohamed, M. K. (2004), Chronic hepatitis C virus infection: Does it really impact health-related quality of life? A study in rural Egypt. Hepatology, 40: 1434–1441. doi: 10.1002/hep.20468
- Issue published online: 24 NOV 2004
- Article first published online: 24 NOV 2004
- Manuscript Accepted: 2 SEP 2004
- Manuscript Received: 6 APR 2004
- Agence Nationale de Recherche sur le SIDA–France (ANRS 12.79)
- European Community (INCOMED Programme: Viral Hepatitis Surveillance in Mediterranean Countries. Grant Number: ICA3-CT-2000-30011
Previous Western studies showed a consistent and marked reduction in health-related quality of life (HRQOL) in patients chronically infected with hepatitis C virus (HCV). However, these studies were conducted on patients whose knowledge of their serological status may have affected their HRQOL. This HRQOL survey conducted in the Egyptian rural population provides a unique opportunity to clarify this issue among a population whose serological status is unknown. HRQOL was assessed by an Arabic translation of the Short-Form 12, and a visual analog scale of the relative severity of one's health status. HCV chronic infection was defined by positive tests for anti-HCV antibody and HCV-RNA. HRQOL was compared according to HCV chronic infection status in linear mixed models adjusted for potential confounding factors, such as age, sex, education, and health care–related risk factors, and adjusted for interviewer as a random effect. One hundred forty-six Egyptians chronically infected with HCV had similar Short-Form 12 and visual analog scale scores, compared with 1,140 uninfected controls from the same rural community. In individuals chronically infected with HCV, serum aminotransferase levels did not correlate with HRQOL. In conclusion, this study did not find a significant reduction of HRQOL in patients chronically infected with HCV compared with uninfected, contemporaneous controls. This may be explained in part by a lower morbidity amongst patients chronically infected with HCV in rural Egypt and a higher morbidity amongst uninfected controls as compared with those of Western studies, as well as a lack of awareness of hepatitis C serological status. (HEPATOLOGY 2004;40:1434–1441.)
Approximately 15% of 59 million Egyptians in 1996 were estimated to have positive test results for anti–hepatitis C virus (HCV) antibody, and based on 60% viremia, more than five million Egyptians are chronically infected with HCV.1 The treatment of chronic hepatitis C (CHC) patients is considered a public health priority in Egypt to reduce both the burden of liver disease and the transmission of HCV. However, dramatic health care budget constraints limit access to the costly treatment recommended in Western countries.2–4 Decision to treat should depend on the expected benefits from treatment of CHC patients, who are mostly infected with the genotype 4, living in the Nile Delta rural areas, and generally unaware of their HCV serological status in the absence of systematic screening.1
Previous Western studies have reported a consistent and marked reduction in health-related quality of life (HRQOL) among CHC patients as compared with nationally representative samples of adults, particularly in physical health–related domains.5–16 The HRQOL of CHC patients declines even more during the 6 to 12 months of treatment, but it returns to the pretreatment level during the 6 months after treatment because of significant improvements in HRQOL of sustained virological responders.7, 10, 11, 15, 17 However, these studies were conducted on patients whose knowledge of their serological status, rather than the disease itself, may have affected their HRQOL.12, 18
The Egyptian rural population, with the highest prevalence of HCV in the world, provides a unique opportunity to clarify this issue. Our primary objective was to compare HRQOL in individuals chronically infected with HCV and unaware of their serological status with that of uninfected controls from the same Egyptian rural community. Our secondary objective was to compare HRQOL within individuals chronically infected with HCV across serum aminotransferase levels.
Patients and Methods
All adults and children from a village in the lower Nile Delta region (Zwayat-Razin, Meynoufeya Governorate), Egypt, were invited to participate in a cohort study on HCV risk factors between May and December 2002 (n = 5,130; response rate of 78%). The village of Zwayat-Razin was selected because it is an endemic area for Schistosomiasis mansoni and, as such, best represents the epidemic of HCV in rural Egypt.1 The HRQOL substudy was conducted among all adults of 5 of the 6 districts of the village included in the cohort study, because of budget constraints. The HRQOL interview occurred at a median time of 3 weeks (extremes: same day to 6 months) after the risk factors interview. Three interviewers blinded to HCV serological status of participants followed an initial 2-month training period with joint interviews and were then purposely matched to the interviewee's gender. Participants were not yet aware of their HCV serological status at the time of the HRQOL survey. There was no difference in chronic HCV infection prevalence between participants included in the HRQOL substudy (11.4%) and others from the risk factors study (11.9%). The study was approved by the institutional review board at the University of Ain Shams, Cairo, and informed consent was obtained from each participant.
HRQOL Data Collection.
HRQOL was evaluated by two instruments: Short-Form 12 (SF-12) and a visual analog scale (VAS). The SF-12 is a widely used generic HRQOL questionnaire19 developed by the Medical Outcome Trust to reduce the time of administering the SF-36.20, 21 The health concepts underlying the SF-12 and the SF-36 range from those reflecting predominantly physical wellness, including items on physical functioning, the ability to perform expected physical roles, the degree of bodily pain, and the overall sense of general health, to those reflecting predominantly social and emotional well-being, including items on the ability to perform expected emotional and social roles, overall sense of vitality, and overall sense of mental health. The 12 items of the SF-12 were selected from the SF-36 to best reflect its physical health summary score (PCS) and its mental health summary score (MCS).21 We performed a translation and cross-cultural adaptation of the SF-12 by a forward translation from US English followed by an independent backward translation into English, and adaptations were made to the context of rural Egypt, such as description of activities of a typical day. The Arabic SF-12 version showed satisfactory psychometric properties (i.e., reliability, construct validity, convergent validity) in our sample, except for the item Mental Health 2 about “”Have you felt downhearted and blue?“ which had a low sensitivity. We decided, however, to include all component items in the PCS and the MCS, using the Likert method of summated items.22 Presented results did not change after the removal of Mental Health 2 from the MCS (data not shown). We also developed a VAS to assess the relative severity of one's health state, which was shown to be feasible in developing countries,23 with a high sensitivity to change in CHC patients24 and in other chronic diseases of mild to moderate severity.25 The VAS was designed as a 100-point scale between death and the best imaginable health state, and participants were asked to rate their health state at different points in time, such as during the preceding month, 1 year ago, and then 5 years ago. The SF-12 PCS, SF-12 MCS, and VAS scores ranged from 0 to 100, with higher scores reflecting better HRQOL.
HCV chronic infection was defined by positive tests for anti-HCV antibody with two enzyme-linked immunosorbent assays (INNOTEST HCV Ab IV, 4th generation test, Innogenetics, Ghent, Belgium; and ABBOTT HCV EIA 3.0, 3rd generation test, Abbott Laboratories, Wiesbaden, Delknheim, Germany) and a detectable HCV viremia (reverse transcriptase polymerase chain reaction, in house assay using 5′UTR primers with modification)26 with or without elevated liver enzymes. All univariate and multivariate analyses of HRQOL were adjusted for interviewer as a random effect in linear mixed models. Mean HRQOL score in individuals chronically infected with HCV was compared with that of uninfected controls, adjusting for potential confounding factors such as risk factors for HCV chronic infection that decreased HRQOL. In individuals chronically infected with HCV, the association of HRQOL and serum aminotransferase (alanine aminotransferase [ALT], aspartate aminotransferase [AST]) level was tested by using Pearson correlation. Significance was attributed at the 5% level, and data were analyzed with SAS software (SAS 8.02, Cary, NC).
Overall, 1,286 adult Egyptians from the village of Zwayat Razin were interviewed to collect HRQOL data after the survey on HCV infection risk factors. Participants were 58.9% female and had a mean age of 33.1 (SD, 12.0) years, 60.3% were illiterate, and 73.6% were farmers. The prevalence of self-declared chronic diseases was 19.0% (12.4% high blood pressure, 6.7% rheumatic diseases, 1.9% diabetes, 0.5% liver diseases). Approximately 54.6% reported a previous admission to hospital or clinic, 6.1% had been transfused, and 6.5% had received parenteral antischistosomal therapy. None of the 146 (11.4%) individuals chronically infected with HCV were aware of their serological status.
We looked for potential confounding factors between HCV chronic infection and HRQOL scores. Age, sex, education, self-declared chronic disease, previous admission to hospital or clinic, history of transfusion, and past parenteral antischistosomal therapy were significant risk factors for HCV chronic infection (Table 1). Conversely, these sociodemographic or disease-related variables, except past parenteral antischistosomal therapy, were significant predictors of HRQOL scores in individuals chronically infected with HCV as well as uninfected controls (Table 2). Accordingly, further multivariate analyses on HRQOL included those six potential confounding factors—age, sex, education, self-declared chronic disease, previous admission to hospital or clinic, and history of transfusion.
|Infected by HCV (n = 146)||Uninfected Controls (n = 1,140)||P Value*|
|Mean (SD) age, yr||39.6 (12.3)||32.2 (11.8)||<.0001|
|Number of pregnancies (in women):|
|Read and write||24.6||15.3||<.01|
|Lived whole life here||87.7||89.2||NS|
|Self-declared chronic disease||28.8||17.7||<.01|
|High blood pressure||17.8||11.7||<.05|
|Previous admission to hospital or clinic||66.4||53.1||<.01|
|History of transfusion||12.3||5.4||<.001|
|Parenteral antischistosomal therapy||19.2||4.8||<.0001|
|Explanatory Variable of HRQOL Scores||Infected by HCV (n = 146)||Uninfected Controls (n = 1,140)|
|n||Physical Health Summary Score||P||Mental Health Summary Score||P||VAS During the Last Month||P||n||Physical Health Summary Score||P||Mental Health Summary Score||P||VAS During the Last Month||P|
|Female||63||32.5 (24.9–40.1)||*||45.3 (39.8–50.9)||NS||62.2 (58.4–66.0)||NS||694||45.8 (43.5–48.2)||**||48.6 (46.9–50.3)||*||64.9 (63.8–66.2)||*|
|Male||83||55.3 (48.4–62.3)||54.9 (49.8–60.1)||63.5 (60.0–67.1)||446||65.2 (62.2–68.2)||58.6 (56.4–60.8)||71.5 (70.0–73.1)|
|>38||89||39.7 (33.4–46.0)||48.1 (43.5–52.7)||58.4 (55.6–61.2)||368||31.2 (28.4–34.0)||42.8 (40.5–44.9)||56.9 (55.5–58.2)|
|26 to 38||33||43.4 (30.7–56.1)||*||50.3 (41.0–59.6)||*||68.3 (62.8–73.8)||**||397||56.3 (53.5–59.1)||***||54.5 (52.3–56.7)||***||69.4 (68.0–70.8)||***|
|<26||24||60.7 (47.0–74.3)||63.6 (53.7–73.6)||73.1 (67.1–79.0)||375||70.6 (67.8–73.4)||60.3 (58.1–62.5)||76.1 (74.7–77.5)|
|Illiterate||88||38.0 (31.9–44.2)||48.2 (43.7–52.8)||59.4 (56.6–62.2)||687||44.3 (42.1–46.6)||48.9 (47.2–50.5)||64.1 (62.9–65.2)|
|Read and write||36||53.5 (41.0–66.0)||***||54.8 (45.5–64.0)||*||64.7 (58.9–70.3)||**||174||59.9 (55.3–64.6)||***||57.7 (54.3–61.2)||**||69.0 (66.7–71.4)||***|
|Any education||22||68.3 (52.8–83.8)||61.7 (50.3–73.2)||76.5 (69.5–83.6)||279||71.0 (67.3–74.7)||59.1 (56.4–61.8)||75.5 (73.6–77.3)|
|Self-declared chronic disease|
|Yes||42||23.4 (14.9–31.9)||**||37.4 (31.4–43.5)||NS||54.3 (50.1–58.5)||*||202||21.9 (17.6–26.1)||***||36.6 (33.4–39.8)||**||54.3 (52.1–56.5)||**|
|No||104||56.1 (50.1–62.1)||55.7 (51.5–59.9)||66.6 (63.6–69.5)||938||60.3 (58.3–62.2)||56.4 (54.9–57.8)||70.3 (69.3–71.3)|
|Admission to hospital or clinic|
|Yes||97||42.5 (36.2–48.8)||*||51.3 (46.8–55.8)||NS||62.0 (58.9–65.0)||NS||605||47.8 (45.3–50.4)||**||48.8 (47.0–50.6)||***||64.3 (63.1–65.5)||*|
|No||49||52.9 (43.1–63.7)||51.1 (43.3–58.9)||67.1 (61.9–72.3)||535||59.0 (56.3–61.8)||57.3 (55.3–59.2)||71.3 (69.9–72.6)|
|History of transfusion|
|Yes||18||18.9 (3.8–33.9)||***||30.0 (19.4–40.6)||***||52.8 (45.5–60.1)||***||61||38.5 (30.5–46.5)||*||42.9 (37.3–48.6)||*||59.9 (56.0–63.9)||*|
|No||128||48.9 (43.5–54.3)||54.1 (50.3–57.9)||64.5 (61.9–67.1)||1079||53.8 (51.9–55.8)||53.3 (51.9–54.7)||68.0 (67.0–68.9)|
|Parenteral antischistosomal therapy|
|Yes||28||42.1 (28.5–55.7)||NS||52.3 (42.5–62.0)||NS||61.6 (55.1–68.1)||NS||55||44.4 (36.0–52.7)||NS||52.1 (46.1–58.0)||NS||62.4 (58.2–66.5)||NS|
|No||118||44.9 (38.8–51.0)||51.8 (47.4–56.2)||63.8 (60.9–66.8)||1085||53.5 (51.6–55.5)||52.8 (51.4–54.2)||67.9 (66.9–68.8)|
Multivariate analyses showed that HCV chronic infection was not associated with the SF-12 PCS, SF-12 MCS, and VAS scores, whereas age, sex, education, self-declared chronic disease, and previous admission to hospital or clinic remained independent predictors of HRQOL scores (Table 3). Further comparison of adjusted mean scores of the eight subscales of the SF-12 between individuals chronically infected with HCV and uninfected controls did not show any significant difference (Fig. 1). Using a VAS, participants assessed retrospectively the relative severity of their health status at 3 different points in time from the preceding month to 5 years ago. Adjusted mean VAS scores did not differ significantly at any time between the two groups.
|Explanatory Variable of HRQOL Scores||n||Physical Health Summary Score†||P||Mental Health Summary Score†||P||VAS During the Last Month†||P|
|HCV chronic infection|
|Yes||146||50.7 (45.5–55.8)||NS||51.9 (47.6–56.2)||NS||66.9 (64.4–69.6)||NS|
|No||1140||50.1 (46.7–53.4)||49.2 (46.4–51.9)||66.5 (64.8–68.2)|
|Female||757||43.2 (39.3–47.1)||*||46.7 (43.5–50.0)||*||64.7 (62.7–66.6)||NS|
|Male||529||57.5 (53.4–61.6)||54.3 (50.9–57.8)||68.8 (66.7–70.8)|
|>38||457||36.6 (32.7–40.5)||44.3 (41.0–47.5)||58.9 (56.9–60.8)|
|26 to 38||430||51.1 (46.8–55.4)||***||51.1 (47.6–54.7)||**||67.9 (65.7–70.1)||***|
|<26||399||63.5 (58.9–68.0)||56.2 (52.4–60.0)||73.4 (71.1–75.7)|
|Illiterate||775||44.6 (40.8–48.4)||48.5 (45.4–51.7)||64.2 (62.3–66.1)|
|Read and write||210||51.1 (46.2–56.0)||**||52.1 (47.9–56.2)||NS||66.6 (64.1–69.1)||*|
|Any education||301||55.4 (50.9–59.9)||51.0 (47.3–54.7)||69.4 (67.1–71.6)|
|Self-declared chronic disease|
|Yes||244||40.2 (35.5–44.9)||***||44.6 (40.7–48.5)||*||63.2 (60.8–65.6)||*|
|No||1042||60.6 (56.9–64.2)||56.4 (53.4–59.5)||70.3 (68.4–72.1)|
|Admission to hospital or clinic|
|Yes||702||47.2 (43.6–50.9)||*||48.1 (45.1–51.1)||*||64.7 (62.8–66.5)||NS|
|No||584||53.5 (49.3–57.8)||53.0 (49.4–56.5)||68.8 (66.6–70.9)|
|History of transfusion|
|Yes||79||49.7 (43.7–55.7)||NS||49.2 (44.1–54.2)||NS||66.5 (63.5–69.6)||NS|
|No||1207||51.1 (48.2–53.9)||51.9 (49.5–54.3)||66.9 (65.5–68.4)|
Of 146 individuals chronically infected with HCV, 115 (79%) came to the local hepatology clinic for the results of their HCV serological status at a median time of 7 weeks after the HRQOL interview (interquartile range, 4-10 weeks). ALT and AST levels were above normal values (40 IU/mL) in 32% and 23% of individuals, respectively. ALT and AST levels did not significantly correlate with PCS, MCS, and VAS scores during the preceding month.
In this study, 1,286 Egyptians unaware of their HCV serological status were prospectively interviewed on their HRQOL by interviewers blinded to the HCV serological status of the participants. An Arabic version of the SF-12 of satisfactory psychometric properties did not differ between the 146 individuals chronically infected with HCV and uninfected controls. The assessment of the relative severity of one's health status by a VAS was also similar between the two groups at different times from the preceding month to 5 years prior. In individuals chronically infected with HCV, serum ALT and AST levels did not correlate with HRQOL measures.
Contrary to previous Western studies, this study did not find a significant reduction of HRQOL in individuals chronically infected with HCV compared with uninfected, contemporaneous controls.5–16 This may be explained by a lower morbidity rate amongst patients chronically infected with HCV in rural Egypt and a higher morbidity rate amongst uninfected controls as compared with those of Western studies, a more precise handling of confounding factors in the comparison with population controls, as well as a lack of awareness of hepatitis C serological status.
The lack of reduction of HRQOL in individuals chronically infected with HCV could be related to better outcomes of HCV infection in Egypt. The epidemic of HCV is largely attributed to the parenteral antischistosomal therapy program, which from 1961 to 1986 exposed children and young adults of the Nile Delta areas to contaminated needles.1 First, other studies have shown consistently that subjects infected at younger ages have better outcomes.27–31 Second, HCV-related morbidity was apparently low in the cohort, as shown by the small proportion (32%) of chronically infected subjects with any elevation of ALT. However, it compares with that of a cohort of Irish women among whom 55% had elevated liver enzymes 17 years after contamination by anti-D immune globulin,28 and it may be attributed in part to the low alcohol consumption expected in the Egyptian community. Conversely, Western HRQOL studies included CHC patients from clinical trials5–8, 10, 11, 14, 15 or tertiary care centers,9, 12, 13, 16 and they may be subject to referral bias, because these centers attract primarily individuals with already established chronic liver disease. In addition, eligibility criteria excluded individuals with normal ALT levels. However, none of the Western studies showed an association between HRQOL and disease severity markers, as evidenced again in this study with ALT levels.
In this study, rural Egyptians appeared to have significantly lower HRQOL scores than comparable Western populations. If the SF-12 PCS and MCS scales had been scored using US norm-based methods—in which the mean is 50 (SD, 10) in the 1998 general US population32—then the SF-12 PCS and MCS of the uninfected rural Egyptians would have been significantly lower than in the United States (44.1 [SD, 11.9] and 44.7 [SD, 9.4], respectively; P < .0001). Conversely, the SF-12 PCS and MCS of the HCV chronically infected Egyptians would not differ significantly from the U.S. norm for liver disease including chronic hepatitis and cirrhosis (PCS, 41.6 [SD, 11.8] vs. 39.9 [SD, 11.9]; MCS, 45.6 [SD, 9.4] vs. 45.4 [SD, 10.7], respectively32(p120)). Accordingly, one might postulate that CHC infection might have little impact compared with the numerous other factors impacting HRQOL in this remote area of Egypt.
This study showed that several confounding factors could explain the association of HRQOL and HCV. Sociodemographic factors related to past parenteral antischistosomal therapy, such as older age, male sex, and illiteracy, and ongoing iatrogenic factors, such as self-declared chronic disease, previous admission to hospital or clinic, and history of transfusion, were significant risk factors for HCV. In addition, these factors were all significantly associated with a reduction in HRQOL, confounding the association between HCV infection and HRQOL. This contrasts with previous Western HRQOL studies that compared CHC patients with national population norms with adjustments made for age, sex, and sometimes comorbidity. One may argue that adjustment was performed on a too limited number of HRQOL explanatory factors when matching CHC patients to population norms. For instance, low education may have acted as a confounding factor because it was associated with HCV infection,33, 34 and it consistently decreased HRQOL.35–37
As compared with previous Western HRQOL studies, none of the 146 Egyptians chronically infected with HCV were aware of their serological status. The lack of a reduction of HRQOL in these patients underlines the HCV labeling effect on HRQOL suggested in previous studies.12, 18 This labeling effect of HCV may be related to the significant reduction of HRQOL associated with the perceived negative impact of HCV on health,38 emotional distress,38 and stigmatization experienced by CHC patients.39 In rural Egypt, further studies should evaluate the effect of HCV labeling on HRQOL. However, we may expect that a low education level, limited access to health care, and the current large burden of disease as compared with the potential distressing long-term outcomes related to HCV infection would limit the impact of HCV labeling on HRQOL.
These study results may be limited by the instruments used to assess HRQOL. No current HRQOL instrument has been validated in the Arabic language. Our choice of the SF-12 questionnaire was guided by its reduced time for administration as compared with the SF-36 or the World Health Organization quality of life assessment tool (WHOQOL-BREF) (26 items); its cross-cultural validation in many languages around the world, including developing countries40; its ability to predict SF-36 physical and mental health summary scores widely used in the field of hepatitis C.6–16, 18, 38, 41 Our Arabic translation and cross-cultural adaptation of the SF-12 showed satisfactory but poorer reliability than the original U.S. SF-12. However, it showed a good convergent validity against VAS (Pearson correlation with PCS and MCS of 0.65 and 0.49, respectively) and a good construct validation in extreme groups. Another limitation may be related to a lower sensitivity of SF-12 to changes in HRQOL as compared with the SF-36 or the SF-36 expanded with specific questions related to hepatitis infection,11 although findings were similar with VAS of proven sensitivity in HCV.24 Finally, we had to rely on interviews in an illiterate community instead of the self-assessment of HRQOL used in most previous Western studies. Possible interaction with interviewer was taken into account in all statistical analyses by adjusting for interviewer as a random effect in linear mixed models. However, two Western studies showed consistently that SF-36 scores collected by self-completion were lower than those obtained by interview administration, the largest differences being in the health concepts reflecting social and emotional well-being.42, 43 To the extent that the surveyed Egyptians as well as the interviewers were unaware of the HCV serological status of the interviewee, the difference in the mode of administration of HRQOL questionnaires should not undermine the presented results, but it may limit the comparability with previous Western studies.
The generalizability of the study results to Western settings is limited by the many differences observed in rural Egypt with regard to sociodemographic variables (e.g., 60% were illiterate, 78% were married or partnered, and 89% lived their whole life in the village), the higher morbidity, the risk factors for HCV infection (intravenous drug use is very rare in rural Egypt, and the attributable fraction of HCV infections related to past injections, including parenteral antischistosomal therapy, was estimated at approximately 50% [M.K.M, personal communication]), the viral genotype 4, and the unawareness of HCV serological status. Accordingly, it would be of interest to replicate the study results in Western settings by adjusting HRQOL on larger sets of explanatory variables when matching to a population norm, or alternatively by using a common source population.
In conclusion, this study shows that HRQOL was not significantly reduced in adults chronically infected with HCV and unaware of their HCV status in the long phase before clinical manifestations. Egypt is currently considering options to treat HCV. This study showed that public health policies should concentrate on their health benefits in the long-term: reduction in mortality2, 44, 45 and HCV transmission, and their opportunity costs compared with other options—for example, reinforcing safer procedures in the health care system.
- 13Comorbid illness is an important determinant of health-related quality of life in patients with chronic hepatitis C. Am J Gastroenterol 2001; 96: 2737–2744., , , , , .Direct Link:
- 14Comorbidities and quality of life in patients with interferon-refractory chronic hepatitis C. Am J Gastroenterol 2001; 96: 170–178., , , , , , et al.Direct Link:
- 16Health-related quality of life in chronic liver disease: the impact of type and severity of disease. Am J Gastroenterol 2001; 96: 2199–2205., , , , , .Direct Link:
- 22Psychometric Theory. 3rd ed. New York: McGraw-Hill, 1994., .
- 24Health-state utilities and quality of life in hepatitis C patients. Am J Gastroenterol 2003; 98: 630–638., , , , , , et al.Direct Link:
- 32How to Score Version 2 of the SF-12 Health Survey (With a Supplement Documenting Version 1). Lincoln, RI: QualityMetric Incorporated, 2002., , , .
- 34Infection with hepatitis B and C viruses, social class and cancer. IARC Sci Publ 1997: 319–324., , .