Compliance as an Explanatory Variable in Hepatitis-C

  • Carmen Mak
  • Samuel M. Heft
  • Harold Amkraut
  • Mei-Hsiu Ling


A Hepatitis C clinical trial program for patients infected with Hepatitis C virus of genotype I demonstrated that the sustained virologic response rate is higher with 48 weeks of Ribavirin in combination with Interferon (I/R) than with 48 weeks of interferon and placebo (UP). The intent-to-treat (ITT) response rate for the I/R group is 29%, versus 9% for the UP group. However, since not all patients completed their treatment, it is likely that the benefit would have been greater if all the patients had taken their full dose for 48 weeks. Assessing the impact of taking the full dose is not trivial; clearly the ITT approach underestimates the response, whereas subsetting the patients who took a large majority of their full dose (eg, 80% of the total dose) likely overestimates the response. Using methodology developed by Efron and Feldman (I), we develop a statistical model that incorporates compliance as a covariate, thereby allowing for an estimate of the fully compliant response rate. It is estimated that the treatment benefit as measured by the difference in sustained virologic response rates of I/R versus UP at full compliance ranged from 23% to 26% compared to 20% for the ITT analysis and 27% for the 80% compliant subgroup.

Key Words

Compliance Dose-response Hepatitis C Interferon Ribavirin 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Efron B, Feldman D. Compliance as an explanatory variable in clinical trials (with discussion). J Am Stat Assoc. 1991;86:9–26.CrossRefGoogle Scholar
  2. 2.
    Alter MJ. Epidemiology of hepatitis C. Hepatol. 1997;26:62S–65S.CrossRefGoogle Scholar
  3. 3.
    Poynard T, Leroy V, Cohard M, et al. Meta-analysis of interferon randomised trials in the treatment of viral hepatitis: effects of dose and duration. Hepatol. 1996;24:778–789.CrossRefGoogle Scholar
  4. 4.
    McHutchison JG, Gordon SC, Schiff ER, et al. Interferon alfa-2b alone or in combination with Ribavirin as initial treatment for chronic hepatitis C. New Engl J Med. 1998;339:1485–1492.CrossRefGoogle Scholar
  5. 5.
    Poynard T, Marcellin P, Lee SS, et al. Randomised trial of Interferon alpha-2b plus Ribavirin for 48 weeks or for 24 weeks versus interferon alpha-2b plus placebo for 48 weeks for treatment of chronic infection with hepatitis C virus. Lancet. 1998;352: 1426–1432.CrossRefGoogle Scholar
  6. 6.
    Sheiner LB, Rubin DB. Intention-to-treat analysis and the goals of clinical trials. Clinical Pharmacol Therapeutics. 1995;57:6–15.CrossRefGoogle Scholar
  7. 7.
    Heitjan DF. Ignorability and bias in clinical trials. Stat Med. 1999;18:2421–2434.CrossRefGoogle Scholar
  8. 8.
    Pocock SJ, Abdalla M. The hope and the hazards of using compliance data in randomized controlled trials. Stat Med. 1998;17:303–317.CrossRefGoogle Scholar
  9. 9.
    Lipid Research Clinic Program. The lipid research clinics coronary primary prevention trial results, Parts I and II. JAMA. 1984;251:351–374.CrossRefGoogle Scholar
  10. 10.
    McCullagh P, Neider JA. Generalized Linear Models. 2nd Edition. London: Chapman & Hall/CRC; 1989.CrossRefGoogle Scholar
  11. 11.
    Rubin DB. Estimating causal effects of treatments in randomized and nonrandomized studies. J Education Psychol. 1974;66:658–701.CrossRefGoogle Scholar
  12. 12.
    Holland P. Statistics and causal inference (with discussion). J Am Stat Assoc. 1986;81:945–970.CrossRefGoogle Scholar
  13. 13.
    Hastie TJ, Tibshirani RJ. Generalized Additive Models. London: Chapman & Hall/CRC; 1990.Google Scholar
  14. 14.
    Knight K. Mathematical Statistics. London: Chapman & Hall/CRC; 2000.Google Scholar
  15. 15.
    Chambers JM, Hastie TJ. Statistical Models in S. London: Chapman & Hall; 1991.Google Scholar

Copyright information

© Drug Information Association, Inc 2001

Authors and Affiliations

  • Carmen Mak
    • 1
  • Samuel M. Heft
    • 1
  • Harold Amkraut
    • 1
  • Mei-Hsiu Ling
    • 1
  1. 1.Statistics, Schering-Plough Research InstituteKenilworthUSA

Personalised recommendations