Weighted mixed-comparator cost

Signature

C_mix = sum_(j=1)^J (w_j * C_j)

• w_j Proportion of the comparator population receiving current-care option j (proportion)
• C_j Expected relevant cost of current-care option j (currency per defined population or person)
• J Number of mutually exclusive current-care options represented in the mixed comparator (count)

Function

Mixed-comparator construction function

Implementations

• Excel

  Calculate weighted comparator cost

  Excel multiplies each option cost by its treatment share and adds the products.

  =SUMPRODUCT(TreatmentShares,OptionCosts)

Assumptions

• Valid treatment shares

  Treatment shares are non-negative and sum to 1 across the mutually exclusive comparator options.

• Representative current practice

  Treatment shares reflect the expected comparator mix for the relevant population setting and period.

• Consistent costing basis for mixed comparator

  Option costs use the same perspective, population, time horizon, currency and price year.

Worked examples

• Weighted comparator cost

  Current care consists of Option A for 70% of patients at £10,000 each and Option B for 30% at £15,000 each. Mixed-comparator cost is £11,500 per patient.

  C_mix = (0.70*10000) + (0.30*15000) = 11500

Common errors

• Using outdated treatment shares

  A mixed comparator based on historical or irrelevant treatment shares may not represent the decision-maker's expected current practice.

Sources

• NICE manual for mixed-comparator cost

  NICE. Health technology evaluations: the manual. Section 4.6 Economic evaluation.

  View source →

Canonical Identity