Clinical resources

Treatment Value Discussion Compass

An educational framework for discussing benefit, harm, quality of life, evidence certainty, patient goals, treatment burden, cost, and uncertainty in thoracic oncology evidence. It is a discussion aid, not a QALY counter and not a recommendation engine.

Purpose

This page makes the trade-offs visible when a treatment looks promising, expensive, toxic, uncertain, burdensome, or poorly matched to a patient goal.

Benefit Harm QoL Evidence Patient goals Cost Not formal HTA

Educational use only

The score is a simplified teaching exercise. It supports structured reflection on trial evidence, absolute benefit, toxicity, patient burden, and resource consequences. It does not replace national reimbursement assessment, cost-effectiveness modelling, trial interpretation, patient preference, clinical judgment, or MDT decision-making.

The optional QALY and ICER fields are recorded as context only. They are not included in the score.

Why this fills a gap

Formal value frameworks and health technology assessments exist, but bedside and MDT discussions often need a simpler transparent structure.

Traditional trial reading often separates relative effect, absolute benefit, toxicity, quality of life, cost, and uncertainty into different parts of a paper or reimbursement report. This tool puts those domains on one page so the assumptions can be challenged.

It should not be presented as a validated value framework. Its function is to make the reasoning explicit before a formal guideline, reimbursement, or clinical decision is made.

Compass method

The total score is scaled to 100 points. Higher scores mean a stronger discussion signal, not a treatment recommendation.

Formula

ntog_treatment_value_discussion_score =
clinical_benefit_score + safety_tolerability_score + evidence_confidence_score + patient_implementation_fit_score + resource_reasonableness_score

Maximum score = 100. QALY gain, ICER, and local reimbursement notes are contextual fields only.

Domain weights

  • Clinical benefit: 35 points
  • Safety and tolerability: 20 points
  • Evidence confidence: 20 points
  • Patient and implementation fit: 15 points
  • Resource reasonableness: 10 points
Domain What it asks Why it matters
Clinical benefit How large is the absolute and relative benefit? Prevents overreading a small absolute gain with an impressive hazard ratio.
Safety and tolerability How frequent, severe, chronic, reversible, or fatal are the harms? Separates manageable toxicity from irreversible or hospitalising toxicity.
Evidence confidence How mature, relevant, and externally valid is the evidence? Weak comparators, immature data, crossover, or exploratory subgroups reduce confidence.
Patient and implementation fit Does the treatment match the patient goal and practical situation? A population benefit may not fit frailty, travel burden, symptoms, or care goals.
Resource reasonableness Is the cost and delivery burden proportionate to the likely benefit? Cost should be explicit, but not reduced to a QALY counter in this educational tool.

Score exercise

Enter simplified trial or treatment data. Blank numeric fields are treated as unavailable and will reduce certainty or trigger caution flags.

Interpretation bands

  • 0 to 34: weak or highly uncertain value signal.
  • 35 to 54: mixed value signal. Benefits, harms, cost, or uncertainty need close review.
  • 55 to 74: favourable discussion signal if patient fit and evidence validity are acceptable.
  • 75 to 100: strong discussion signal. Still not an automatic treatment recommendation.

Example loaders

Synthetic examples for teaching the logic.

1. Treatment context

2. Clinical benefit

Use absolute OS gain where available.
Relative effect. Interpreted with absolute benefit.
At a specified time point. Used instead of NNT if stronger.
At the same time point as the absolute benefit.

3. Harm and tolerability

4. Evidence confidence

5. Patient and implementation fit

6. Resource and cost context

Drug, administration, monitoring, and expected toxicity management if known.

7. Optional economic context, not scored

NTOG treatment value discussion score

Not calculated

The score is only a structured discussion aid. It is not a recommendation, formal health-economic assessment, or QALY calculator.

Structured export


              

Example matrix

The examples are illustrative. They are not treatment recommendations.

Scenario Benefit pattern Harm pattern Evidence pattern Cost pattern Typical discussion signal
Clear benefit OS gain, favourable HR, good absolute effect Manageable and reversible toxicity Phase III, mature, active comparator Moderate cost or favourable access Favourable to strong
Relative-only signal Impressive HR but small absolute gain Moderate toxicity May be statistically strong but clinically narrow Variable Mixed, requires absolute-benefit discussion
Costly uncertain option Small or surrogate benefit Meaningful toxicity or burden Immature or indirect evidence High cost and high resource pressure Weak to mixed

Relationship to existing frameworks

This NTOG page is intentionally simpler than formal value frameworks and HTA methods.

  • ASCO Value Framework: clinical benefit, side effects, symptoms or quality of life, and cost context. ASCO value in cancer care
  • ESMO-MCBS: graded magnitude of clinical benefit for cancer medicines. ESMO-MCBS
  • NICE health technology evaluation: economic evaluation, resource use, costs, uncertainty, subgroup analysis, and benefits not captured in QALY calculations. NICE economic evaluation manual

Limitations

A number can hide more than it reveals unless the assumptions are made explicit.

  • Median OS benefit may not reflect long-term survivors, delayed separation of curves, or cure fraction.
  • Hazard ratios do not show absolute benefit and may mislead when hazards are non-proportional.
  • NNT and NNH depend on time point, baseline risk, endpoint choice, censoring, and follow-up duration.
  • QoL data can be missing, selectively reported, or collected after dropout from treatment.
  • Cost thresholds differ between countries, payers, reimbursement systems, and negotiated prices.
  • Subgroup results may be underpowered, exploratory, or biologically plausible but not proven.
  • Patient goals, frailty, symptoms, travel burden, and family context may override a population-level value signal.