Insight

Patient-reported outcomes in AMNOG: What the EQ-5D VAS reveals about assessment practice

Patient-reported outcomes (PROs) capture information on health status, symptoms or health-related quality of life directly from patients. They reflect patients' own experience of their condition and treatment, without interpretation by healthcare professionals.

In Germany's AMNOG process, PROs can help capture patient-relevant treatment effects that clinical endpoints alone may not fully reflect.

Patient-reported endpoints can be defined and analysed in different ways in benefit assessment. The chosen operationalisation determines which aspect of a patient's reported trajectory is captured and which treatment effects become visible in the assessment.

Consider the EQ-5D Visual Analogue Scale (EQ-5D VAS), on which patients rate their current health status from 0 to 100. Alongside analyses of continuous changes or changes at specified time points, time-to-event analyses can be presented. These examine when a predefined clinically relevant change first occurs.

Depending on the event definition, several types of time-to-event analysis are possible. Common approaches are time to first deterioration (TTFD), time to confirmed deterioration (TTCD) and time to persistent deterioration (TTPD), where the deterioration must remain present at all subsequent assessments through the end of observation.

The event definition matters because it captures different patterns of deterioration. A temporary decline can count as an event in a TTFD analysis even if the patient's health subsequently improves. Confirmed or persistent deterioration incorporates information from later assessments and therefore requires more extensive follow-up data.

The chosen operationalisation can consequently affect the estimated magnitude of a treatment effect and whether a potential advantage over the comparator becomes apparent. For pharmaceutical companies, this makes it part of the evidence and dossier strategy. Reliable interpretation of confirmed or persistent deterioration, however, requires sufficiently comparable assessment conditions between study arms and the availability of the necessary follow-up observations.

EQ-5D VAS in AMNOG: Findings from 372 benefit assessment procedures

A retrospective analysis of 372 AMNOG procedures examined which EQ-5D VAS operationalisations pharmaceutical companies submitted, which were assessed by the Institute for Quality and Efficiency in Health Care (IQWiG), and which were ultimately considered by Germany's Federal Joint Committee (Gemeinsamer Bundesausschuss, G-BA).

Among time-to-event analyses, TTFD was considered by far the most frequently. First deterioration was included in the final assessment in 109 procedures, compared with 13 for TTPD and six for TTCD. Other approaches, including mixed models for repeated measures (MMRM), mean change from baseline and binary analyses, also played a role. The EQ-5D VAS was not ultimately considered in 94 procedures.

Particularly informative are 30 procedures in which the pharmaceutical company submitted both TTFD and at least one confirmed or persistent deterioration analysis with the original dossier.

TTFD was ultimately considered in 23 of these 30 procedures, including 20 in which it was the only time-to-event analysis considered. TTFD and TTPD were both considered in three procedures. TTCD alone and TTPD alone were each considered in one procedure. No time-to-event analysis was included in the final assessment in five procedures.

TTFD was therefore frequently considered even when alternative operationalisations were available. This does not, however, establish a general methodological preference for first deterioration. On the contrary, in individual procedures IQWiG regarded confirmed or persistent deterioration as a conceptually appropriate or more informative operationalisation, provided that the assessment schedule supported a reliable analysis.

The procedures involving daratumumab, ixazomib and dostarlimab illustrate the role of the assessment schedule.

EQ-5D VAS in G-BA benefit assessments: Three AMNOG case studies

Daratumumab (D-1241): G-BA considers confirmed persistent deterioration

In the daratumumab procedure, the original Module 4 included a TTFD analysis of EQ-5D VAS using a deterioration threshold of at least 15 points.

IQWiG initially considered TTFD but noted that confirmed or persistent deterioration might also be relevant because assessments continued beyond progression until the end of the study. Corresponding analyses were submitted during the written comments procedure.

In the final data cut submitted during that procedure, median follow-up was 61.9 and 59.2 months, respectively. The G-BA ultimately considered confirmed persistent deterioration.

The available follow-up made it possible to distinguish temporary changes from sustained deterioration. This allowed the assessment to use additional longitudinal information that a TTFD analysis does not fully capture.

The procedure illustrates how confirmed or persistent operationalisations can provide additional information about the course of deterioration when the data are suitable.

Ixazomib (D-753): TTFD remains decisive despite a subsequently submitted confirmation analysis

In the ixazomib procedure, EQ-5D VAS was also originally analysed as TTFD. The pharmaceutical company submitted an analysis of confirmed deterioration during the written comments procedure.

The methodological assessment differed from that for daratumumab. Opportunities to confirm deterioration were not sufficiently comparable between treatment arms. Differences in the number and distribution of assessment opportunities, together with declining questionnaire completion, limited the usability of the additional analysis.

TTFD therefore remained decisive for the final assessment.

The case demonstrates that the interpretability of confirmed or persistent deterioration depends heavily on the assessment schedule. Analysing such an operationalisation retrospectively cannot remedy pre-existing differences between arms in assessment opportunities and response rates.

Dostarlimab (D-996): EQ-5D VAS assessment schedule and observation time

The dostarlimab procedure illustrates another methodological issue. TTFD was considered in the assessment even though median observation times were comparable between treatment arms.

Although median observation times differed only moderately, at approximately 23 versus 20 months, the assessment schedule yielded around 18 versus 14 possible assessments over a median observation period. IQWiG explicitly stated that the difference in observation time alone did not render persistent deterioration unsuitable.

What matters is how many assessments are actually available after first deterioration, how far apart they occur and whether opportunities for confirmation are comparable between treatment arms.

Long or similar median follow-up therefore does not automatically mean that confirmed or persistent deterioration can be analysed reliably.

Patient-reported outcomes in the AMNOG dossier: Implications for evidence strategy

The retrospective analysis shows that TTFD is frequently considered in assessment practice. The three cases nevertheless demonstrate that the suitability of an operationalisation depends substantially on the assessment schedule and available follow-up.

Four conclusions follow for pharmaceutical companies.

TTFD, TTCD and TTPD can capture different treatment effects

TTFD, TTCD and TTPD capture different aspects of deterioration. For example, a treatment may be associated with more frequent short-term changes in health status while sustained deterioration occurs less often.

In such circumstances, confirmed or persistent analyses can provide additional information about the treatment effect. Where the data allow, submitting multiple operationalisations and examining differences between their results may therefore be appropriate.

Interpret differences between TTFD, TTCD and TTPD methodologically

Where several methodologically justified operationalisations are available, differences between TTFD, TTCD and TTPD can provide further information about the course of deterioration. Such analyses should preferably be prespecified, or at least justified on clinical and methodological grounds independently of the observed result.

PRO assessment schedules in study design: Requirements for TTCD and TTPD

Whether confirmed or persistent deterioration can subsequently be analysed reliably is largely determined by study design. If sufficient follow-up assessments are missing after initial deterioration, it is impossible to establish reliably whether the deterioration was confirmed or persisted over time.

Follow-up duration is not the only relevant factor. The frequency and timing of PRO assessments, and whether assessment conditions are comparable between arms, also matter. Different assessment intervals or measurements taken at different points in the treatment cycle can mean that patients in the two arms do not have the same opportunities to confirm deterioration or document subsequent recovery.

The pembrolizumab procedure (D-653) provides an example. Because treatment cycles differed, PRO assessments in the study arms sometimes took place at different points within the respective treatment cycle. The G-BA regarded this as a potential risk of bias that also affected EQ-5D VAS analyses.

Study planning should therefore consider potential event definitions at an early stage. Assessment intervals, the timing of measurements, follow-up after progression or treatment discontinuation, and data completeness all influence which operationalisations can later be analysed reliably in the AMNOG process.

Justify the operationalisation of patient-reported outcomes in the AMNOG dossier

The procedures analysed do not establish a universally applicable ranking of TTFD, TTCD and TTPD. Selection should reflect the clinical question and the available study information.

In the AMNOG dossier, this means transparently describing the event definition, comparability of follow-up assessments and handling of missing measurements.

The strategic objective is to represent clinically relevant treatment effects appropriately while ensuring that the selected analyses are methodologically sound. Daratumumab illustrates that confirmed persistent deterioration may be considered under suitable conditions. Ixazomib and dostarlimab illustrate the limitations when the necessary follow-up is not sufficiently comparable.

Methodological basis of the EQ-5D VAS analysis

This analysis was conducted by co.value, the market access consultancy of co.faktor – a Cytel company. It was based on a retrospective descriptive review of 372 identified AMNOG procedures with documented reference to EQ-5D VAS. The data cut-off was 1 September 2026.

The analysis examined which EQ-5D VAS operationalisations were submitted by the pharmaceutical company, assessed by IQWiG and actually considered by the G-BA in its final benefit assessment of the endpoint. Classification depended on whether the analysis was methodologically considered, irrespective of statistical significance, clinical relevance of the observed effect or contribution to an added-benefit conclusion.

Particular attention was paid to time-to-event analyses of deterioration in health status, especially first, confirmed and persistent deterioration. Analyses submitted only after the original dossier, for example during the written comments procedure, were recorded separately and were not classified as available at initial submission.

The review primarily used official procedure documents, including G-BA resolutions, their supporting reasons and summary documentation, IQWiG benefit assessments and the pharmaceutical company's Module 4. Amendments, addenda and documents submitted during the written comments procedure were consulted where necessary.

Frequently Asked Questions

What are patient-reported outcomes (PROs) in AMNOG benefit assessment?

PROs are reports made directly by patients about symptoms, health status or health-related quality of life. They can inform the assessment of patient-relevant treatment effects under AMNOG. The EQ-5D VAS is a tool for patients to rate their general health status.

What is the difference between TTFD, TTCD and TTPD for EQ-5D VAS?

TTFD measures time to the first clinically relevant deterioration. TTCD requires confirmation at a subsequent assessment. TTPD captures persistent deterioration without subsequent recovery under the applicable event definition. These approaches can reveal different treatment effects.

Which EQ-5D VAS operationalisation does the G-BA consider most frequently?

In the retrospective analysis of 372 AMNOG procedures, TTFD was considered in 109 procedures, compared with 13 for TTPD and six for TTCD. This does not establish a general G-BA preference. The event definition and the reliability of the study data remain decisive.

When can confirmed or persistent deterioration be useful in an AMNOG dossier?

TTCD and TTPD can add information when sufficient follow-up assessments are available and assessment conditions are comparable between treatment arms. The daratumumab procedure shows that confirmed persistent deterioration can be considered in the final assessment under suitable conditions.

What are the requirements for collecting PROs in clinical studies?

Key factors include the number and timing of assessments, comparable assessment intervals between treatment arms, data completeness and continued observation after progression or treatment discontinuation. These factors determine whether confirmed or persistent deterioration can subsequently be analysed reliably.

More insights