Comparison of reimbursement decisions before and after real-world evidence review in France

When assessing new medicines for reimbursement decisions, the Haute Autorité de Santé (HAS) Transparency Committee may request the collection of real-life data in the form of real-life post-registration studies to be conducted by the pharmaceutical company if questions remain unanswered about a medicine’s effect or if there is an uncertainty about the generalisability of RCT data to real-life practice. The medicines are then reassessed by the Transparency Committee using results from these studies alongside other existing RCT data on the medicine.

This review studied the impact of the real-life study results on two ratings determining:

  • the medicine’s eligibility for reimbursement and the reimbursement rate (the actual therapeutic benefit [ACB])
    • takes into consideration severity of the disease, efficacy, side-effects, positioning in the therapeutic strategy among other available therapies and public health benefits
    • a medicine can be given either an insufficient ACB (resulting in not being reimbursed) or one of 3 ‘acceptable’ ACB reimbursement levels (important – 65%; moderate – 30% and mild – 15%)
    • Note: In France, real-life study results are only requested for medicines given an ‘acceptable’ ACB reimbursement level.
  • the comparative effect of the medicine over standard care which impacts on the price (the clinical added value [CAV])
    • takes into consideration efficacy and tolerance compared to standard care
    • a medicine can be given one of 5 possible levels (major, important, moderate, minor, no improvement)

In-depth analysis of specific cases where the reimbursement rate or comparative effect rate was downgraded (resulting in less or no coverage) was conducted in order to understand the discrepancies between RCT and real-life post-registration studies. For more details about the case study analysis, see here.

What was examined?

All medicines reassessed by the Transparency Committee for which real-life post-registration studies results – including interim analysis – had been taken into account between 2011 and 2014 were included.

The objectives of the post-registration studies were classified as follows:

  1. Terms of use and prescription requirements – where additional data was requested on how the medicine is actually used in usual practice and could include any of the following: dosages used, treatment duration, observance and/or the identification and description of the included patients and prescribers
  2. Impact on morbidity and mortality – typically where additional data was requested on the benefits of the medicine
  3. Safety – adverse events from the medicine observed in usual practice
  4. Impact on the healthcare system – usually in response to a request for data on the impact of the medicine on the use of other health products and services
  5. Quality of life – refers to the consequences of the medicine on the disease in physical (autonomy, physical abilities), psychological (emotionalism, anxiety, depression for example) and social fields (relationship to the family, friends or professional environment).

For each objective, results from the post-registration studies were ranked into 3 categories, based on the information stated within the transparency committee report:

  • comparable – when results aligned with RCT data
  • uninterpretable – when no conclusion could be drawn
  • not comparable – when results were different from RCT data.

What were the findings and conclusions?

Findings

  • Fifty different medicines, assessed in 68 different indications were identified.
  • Real-life study data included ad-hoc studies, observational studies, registries and databases.
  • Overall, 63% (43/68) of the indications maintained a similar ACB and CAV rating from the initial assessment; 37% (25/68) of the indications were downgraded.
  • The reasons for requesting real-life studies were multiple for most indications but a request for data on the use of the medication and on its effectiveness were the most common (see table 1 and 2 below).

Table 1. Downgrading of ACB and CAV rating by main real-life study objective and comparability of results to the RCT data

ACB

Abbreviations and headings: PRS – post-registration studies; MM – impact on morbidity and mortality; terms of use – terms of use and prescription requirements; safety – adverse events; QoL – quality of life; healthcare system – impact on the healthcare system.

Table 2. Downgrading of ACB and CAV rating by main real-life study objective and comparability of results to the RCT data, excluding growth hormones, slow-acting anti-rheumatic drugs (SAARDs) and drugs with a marketing authorization granted before 2000

CAV

Abbreviations and headings: PRS – post-registration studies; MM – impact on morbidity and mortality; terms of use – terms of use and prescription requirements; safety – adverse events; QoL – quality of life; healthcare system – impact on the healthcare system.

  • When no additional data was submitted or when real-life study results were considered to have evidence gaps, the results were considered uninterpretable. Evidence gaps usually concerned missing data (patients lost to follow-up), lack of representativeness of the practitioners and patients to those in real-life, or methodological issues. Other parameters that differed from RCTs included how medicines were prescribed (for example, dosage, intake frequency or treatment duration) or problems with adherence.
  • Real-life study results were not comparable to RCT data usually because the real-world studies used the medicine in a different patient population, showed an important efficacy-effectiveness gap, or demonstrated poorer tolerability.
  • Sensitivity analysis showed no difference in the reason for requesting real-life studies: by removing medicines with marketing authorisation before 2000, medicines classed as either slow-acting anti-rheumatic drugs (SAARDs) or growth hormones had a similar number of requests for morbidity and mortality data as for the terms of use of the medicine.

Conclusions

  • A downgraded ACB or CAV are most often observed when real-world study results are not comparable to RCT data than when they are considered uninterpretable.
  • Downgrading of the ACB or CAV was less likely to happen for real-world studies ranked comparable to RCT data.
  • Since there may be additional studies considered in the reassessment (such as RCTs published later), it is difficult to single out the main drivers of the Transparency Committee’s reassessments and make accurate conclusions on the impact of real-life post-registration studies on the ACB or CAV level.

What are the limitations of this study?

  • Evidence requirements have evolved over time: this study included medicines that were first assessed more than 15 years ago when evidence requirements on comparative efficacy and effectiveness to determine reimbursement decisions were lower.
  • The severity of the disease targeted by the medicine had a bigger impact on the ACB years ago, but now the medicine’s efficacy is more important and the effect size required to achieve a given CAV has increased substantially.

Key contributors

Milon Waththuhewa, Sophie Stamenkovic, Houria Mouas, Muriel Vray, Mira Pavlovic and François Meyer, Haute Autorité de Santé (HAS)