Subgroup analyses of clinical effectiveness to support health technology assessments

Abstract

dSubgroup analysis is an integral part of access and reimbursement dossiers, in particular health technology assessment (HTA), and their HTA recommendations are often limited to subpopulations. HTA recommendations for subpopulations are not always clear and without controversies. In this paper, we review several HTA guidelines regarding subgroup analyses. We describe good statistical principles for subgroup analyses of clinical effectiveness to support HTAs and include case examples where HTA recommendations were given to subpopulations only. Unlike regulatory submissions, pharmaceutical statisticians in most companies have had limited involvement in the planning, design and preparation of HTA/payers submissions. We hope to change this by highlighting how pharmaceutical statisticians should contribute to payers' submissions. This includes early engagement in reimbursement strategy discussions to influence the design, analysis and interpretation of phase III randomized clinical trials as well as meta-analyses/network meta-analyses. The focus on this paper is on subgroup analyses relating to clinical effectiveness as we believe this is the first key step of statistical involvement and influence in the preparation of HTA and reimbursement submissions.

Serum Chloride and Sodium Interplay in Patients With Acute Myocardial Infarction and Heart Failure With Reduced Ejection Fraction

“…9,43 Of note, these acute HF cohorts had a much smaller sample size than the present series, limiting the power for the assessment of interaction analysis, which is a statistically weak test depending on sample size and event rate. 24 In this regard, our data supplement those assessed by Grodin et al, thus providing further insight into sodium/chloride interplay. Chloride is likely a stronger prognosticator than sodium (as also demonstrated herein), although the interpretation of these electrolytes should be considered together.…”

Implementation of AMNOG: An industry perspective

“…We support the position of ICH E9 and the considerations advocated by Paget et al. (). While we understand the desire to conduct subgroup analysis in support of the search for personalized medicine, we are concerned that subgroup analyses conducted in the manner by IQWiG and G‐BA could lead to unintended consequences.…”

“…Paget et al. () wrote about subgroup analysis of clinical effectiveness in support of HTA. Paget et al.…”

Concepts of ‘Personalization’ in Personalized Medicine: Implications for Economic Evaluation

“…It is also desirable for the evidence base identifying between-patient heterogeneity to come from pre-specified analyses. Statistical methods used to identify and investigate subgroup effects for HTA purposes have been reviewed recently [25], and a methodological framework to determine the optimal number (and specification) of subgroups in cost-effectiveness analysis for decision making is beginning to emerge [24]. …”