Factors Influencing Adherence in Hep-C Infected Patients
Factors Influencing Adherence in Hep-C Infected Patients
This systematic review was prepared according to the standards of the recommendations for systematic reviews of prognostic factors and reported according to MOOSE. A systematic literature search was performed in MEDLINE (via Pubmed) and Embase (via Embase) (TM). The search strategy combined various terms and medical subject headings related to adherence, hepatitis-C and ribavirin (the full search strategies for each database are available in Additional File 1). The search was performed on March 25th 2013. Study type, publication date and language were not limited in the electronic search strategy to maximize sensitivity.
To be eligible for this review the studies had to meet the following inclusion criteria:
No exclusion criteria were applied.
Two reviewers independently performed the study selection according to the inclusion criteria in a two-step procedure (DP; SA, TM). Firstly, the titles and abstracts of all hits in the electronic databases were screened. Secondly, the full-texts of all potentially relevant articles were obtained and screened. Any differences between the reviewers were discussed until consensus was reached. In addition, the reference lists of all included publications were hand-searched and a Google Scholar search was performed to identify grey literature (TM). The authors were contacted in case of any missing information regarding the inclusion criteria (TM).
According to established recommendations a differentiation between initiation, implementation and discontinuation adherence should be made. In clinical practice initiation and discontinuation of a medication regimen can be easily assessed. In contrast, implementation adherence is often not obvious. Furthermore, research indicates that the reasons for discontinuation and implementation adherence can differ. Thus, it was decided to focus on implementation adherence. Non-adherence can be intentional (e.g. conscious decision not to take) or non-intentional (e.g. forgetting). Furthermore, it has been shown that non-adherence is mostly non-intentional and that there are different influencing factors. Therefore, we excluded studies that only measured intentional non-adherence like surveys that gathered reasons for non-adherence stated by patients.
The risk of bias of included studies was assessed using the methodology checklist for prognostic studies provided by the National Institute for Health and Care Excellence (NICE) (evaluation questions for the instruments are available in Additional file 2). The study population was considered representative, if there were no specific inclusion criteria (e.g. psychiatric disorders) i.e. if the population was representative for the general hepatitis-C infected population in western countries. Confounding was rated appropriate if the potentially relevant factors were incorporated in the analysis. The risk of bias assessment was performed independently by two reviewers (TM, DP). Disagreements were resolved in a discussion or by involving a third person.
The data were extracted in pre-designed standardized tables. For each study the number of analyzed patients, the study inclusion criteria (demographic, socioeconomic, disease related, medication related), the country the study took place, the used adherence measure and mathematical operationalization, and the medications to which the adherence measure (ribavirin or ribavirin and interferon) refers to were extracted. With respect to the results, the influencing factor and the effect on adherence (effect direction or compared categories; effect size and measure) and the statistical significance (95%-CI or p-value) were extracted. All data in the tables on the influence refers to an increase of the respective factor, independent from whether the factor is positive (e.g. educational level) or negative (adverse events). Higher risk ratios (RR) and odds ratios (OR) mean lower adherence in the reference group. e.g. in a comparison of higher versus lower age, a higher OR indicates a higher adherence in the higher aged population. In case the studies used univariate as well as multivariate analysis methods, only the results of the multivariate analysis were extracted. Data extraction were performed by one reviewer (TM) and verified by a second (DP). If effect sizes and statistical significance were not reported in the publications, the OR with confidence intervals were calculated by the authors (Microsoft Excel 2010) using the double data entry method, provided that there were sufficient data (TM, DP).
A quantitative data synthesis using a meta-analysis was planned a priori but was not performed to avoid misleading results due to heterogeneity regarding the included patients, the adherence measurements and definitions/operationalization, the measurement of influencing factors, and the statistical analysis methods (e.g. adjustments, categorizations). Furthermore, in most studies there was a significant lack of reporting especially regarding values for not statistically significant results to allow a recalculation of data.
For all factors that were analyzed in at least two studies a summary estimation of the effect direction and the effect size was made. Two reviewers rated the evidence for an effect, considering the consistency of the effect direction (within and between studies), the effect size, the statistical significance, the sample size and the risk of bias of included studies that analyzed the respective factor (TM, DP). Discrepant ratings were discussed until consensus.
A p-level of <0.05 was considered statistically significant.
Methods
Sources
This systematic review was prepared according to the standards of the recommendations for systematic reviews of prognostic factors and reported according to MOOSE. A systematic literature search was performed in MEDLINE (via Pubmed) and Embase (via Embase) (TM). The search strategy combined various terms and medical subject headings related to adherence, hepatitis-C and ribavirin (the full search strategies for each database are available in Additional File 1). The search was performed on March 25th 2013. Study type, publication date and language were not limited in the electronic search strategy to maximize sensitivity.
Study Selection
To be eligible for this review the studies had to meet the following inclusion criteria:
Patients: Adult patients with hepatitis-C virus infection
Exposure: Potential adherence influencing factor/s (exposure [factor] is not controlled by the investigator, e.g. different dosages or therapies)
Medication: Regimes containing ribavirin
Outcome: Quantitative patient implementation adherence measure (not persistence, not exclusively intentional-adherence measures)
Region: Study conducted in WHO- mortality Stratum A (very low child mortality and low adult mortality)
Publication language: English or German
No exclusion criteria were applied.
Two reviewers independently performed the study selection according to the inclusion criteria in a two-step procedure (DP; SA, TM). Firstly, the titles and abstracts of all hits in the electronic databases were screened. Secondly, the full-texts of all potentially relevant articles were obtained and screened. Any differences between the reviewers were discussed until consensus was reached. In addition, the reference lists of all included publications were hand-searched and a Google Scholar search was performed to identify grey literature (TM). The authors were contacted in case of any missing information regarding the inclusion criteria (TM).
According to established recommendations a differentiation between initiation, implementation and discontinuation adherence should be made. In clinical practice initiation and discontinuation of a medication regimen can be easily assessed. In contrast, implementation adherence is often not obvious. Furthermore, research indicates that the reasons for discontinuation and implementation adherence can differ. Thus, it was decided to focus on implementation adherence. Non-adherence can be intentional (e.g. conscious decision not to take) or non-intentional (e.g. forgetting). Furthermore, it has been shown that non-adherence is mostly non-intentional and that there are different influencing factors. Therefore, we excluded studies that only measured intentional non-adherence like surveys that gathered reasons for non-adherence stated by patients.
Assessment of Risk of Bias
The risk of bias of included studies was assessed using the methodology checklist for prognostic studies provided by the National Institute for Health and Care Excellence (NICE) (evaluation questions for the instruments are available in Additional file 2). The study population was considered representative, if there were no specific inclusion criteria (e.g. psychiatric disorders) i.e. if the population was representative for the general hepatitis-C infected population in western countries. Confounding was rated appropriate if the potentially relevant factors were incorporated in the analysis. The risk of bias assessment was performed independently by two reviewers (TM, DP). Disagreements were resolved in a discussion or by involving a third person.
Data Extraction and Synthesis
The data were extracted in pre-designed standardized tables. For each study the number of analyzed patients, the study inclusion criteria (demographic, socioeconomic, disease related, medication related), the country the study took place, the used adherence measure and mathematical operationalization, and the medications to which the adherence measure (ribavirin or ribavirin and interferon) refers to were extracted. With respect to the results, the influencing factor and the effect on adherence (effect direction or compared categories; effect size and measure) and the statistical significance (95%-CI or p-value) were extracted. All data in the tables on the influence refers to an increase of the respective factor, independent from whether the factor is positive (e.g. educational level) or negative (adverse events). Higher risk ratios (RR) and odds ratios (OR) mean lower adherence in the reference group. e.g. in a comparison of higher versus lower age, a higher OR indicates a higher adherence in the higher aged population. In case the studies used univariate as well as multivariate analysis methods, only the results of the multivariate analysis were extracted. Data extraction were performed by one reviewer (TM) and verified by a second (DP). If effect sizes and statistical significance were not reported in the publications, the OR with confidence intervals were calculated by the authors (Microsoft Excel 2010) using the double data entry method, provided that there were sufficient data (TM, DP).
A quantitative data synthesis using a meta-analysis was planned a priori but was not performed to avoid misleading results due to heterogeneity regarding the included patients, the adherence measurements and definitions/operationalization, the measurement of influencing factors, and the statistical analysis methods (e.g. adjustments, categorizations). Furthermore, in most studies there was a significant lack of reporting especially regarding values for not statistically significant results to allow a recalculation of data.
For all factors that were analyzed in at least two studies a summary estimation of the effect direction and the effect size was made. Two reviewers rated the evidence for an effect, considering the consistency of the effect direction (within and between studies), the effect size, the statistical significance, the sample size and the risk of bias of included studies that analyzed the respective factor (TM, DP). Discrepant ratings were discussed until consensus.
A p-level of <0.05 was considered statistically significant.
Source...