Applying GRADE Criteria to Clinical Inputs to Cost-Effectiveness Modeling Studies

Area of Science:

• Health Economics
• Clinical Epidemiology
• Health Technology Assessment

Background:

• Concerns exist regarding the quality of clinical data used in cost-effectiveness models.
• Economic evaluations rely heavily on clinical effectiveness data, necessitating robust data quality.

Purpose of the Study:

• To evaluate the appropriate use of clinical effectiveness data in cost-effectiveness modeling studies published between 2001 and 2015.
• To assess the quality of clinical evidence underpinning economic evaluations.

Main Methods:

• A systematic assessment of 72 modeling studies across three therapeutic areas.
• Application of Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria for clinical evidence quality: risk of bias, publication bias, indirectness, inconsistency, and study limitations.
• Analysis of trends in evidence quality over time.

Main Results:

• Over 60% of studies used data with high/unclear risk of bias; over 80% showed publication bias.
• Around 30% relied on indirect clinical evidence, with this increasing over time.
• Inconsistency and imprecision were low (<10%), but study limitations were poorly described in over a third of studies.

Conclusions:

• Despite using precise and consistent data (often RCTs), cost-effectiveness models are frequently impacted by methodological shortcomings in clinical efficacy data.
• Modelers and assessors must enhance their awareness and application of GRADE criteria for clinical evidence quality.
• Improving the assessment of clinical evidence quality is vital for the validity of economic evaluations.