Cost-Effectiveness Analysis in Randomised Controlled Trials
Introduction
- Cost-effectiveness analysis (CEA) in healthcare
- Importance in decision-making
- Integration with randomised controlled trials (RCTs)
Randomised Controlled Trials
- Gold standard for evaluating interventions
- Key features:
- Randomisation
- Control group
- Blinding (when possible)
Randomised Controlled Trials (cont’d)
- Measure efficacy/effectiveness
- Efficacy:
capacity
for producing a desired result or effect/effectiveness.
- Effectiveness:
adequate
to accomplish a purpose; producing the intended or expected result.
Cost-Effectiveness Analysis (CEA) or Cost-Utility Analysis (CUA)
- Compares costs and health outcomes of interventions
- Measures:
- Costs, e.g., direct and/or indirect
- Outcomes, e.g., life-year gained, avoided unintended pregnancy, quality-adjusted life years (QALYs)
- Result: Incremental cost-effectiveness ratio (ICER) or incremental cost-utility ration (ICUR)
Integrating CEA with RCTs
- Advantages:
- Reduced bias
- Concurrent data collection
- Timely results
- Challenges/limitations:
- Sample size considerations
- Generalisability
- Time horizon
Steps in Conducting CEA alongside RCT
- Define the research question/objective
- Choose comparator(s)
- Determine perspective (e.g., societal, healthcare payer, etc.)
- Identify costs and outcomes
- Collect data during the trial
- Analyse and interpret results
Measuring Costs
- Direct medical costs:
- Intervention costs
- Healthcare utilisation
- Direct non-medical costs:
- Patient transportation
- Caregiver time
- Indirect costs:
Measuring Outcomes
- Clinical endpoints
- Patient-reported outcomes
- Quality of life measures
- Conversion to QALYs
- Example: EQ-5D-3L, EQ-5D-5L, SF-6D questionnaire
Analysis and Interpretation
- Calculate ICER: \(ICER = \frac{\text{Cost}_A - \text{Cost}_B}{\text{Effect}_A - \text{Effect}_B}\)
- Willingness-to-pay (WTP) threshold
- Sensitivity analysis
- Cost-effectiveness plane (Figure 1)
Analysis and Interpretation (cont’d)
Figure 1: Cost-effectiveness plane (Bozic et al (2003). JBJS, 85(1), 129-142.)
Example: Technology-Enhanced Group-baded (TEG) Rehabilitation Therapy
- WTP threshold: $50,000 per QALY
- Standard treatment: Cost = $1000, QALYs = 0.049
- New treatment: Cost = $1150, QALYs = 0.050
Example (cont’d)
- Stata do-file that produces utility weights for Australia using EQ-5D-5L based on a DCE (Norman, 2023).
- Data should be in 5 columns mo, sc, ua, pd, ad; otherwise, replace names with columns in your data set.
gen modec = .
replace modec = 0 if mo==1
replace modec = ‐0.039 if mo==2
replace modec = ‐0.067 if mo==3
replace modec = ‐0.237 if mo==4
replace modec = ‐0.242 if mo==5
gen scdec = .
replace scdec = 0 if sc==1
replace scdec = ‐0.030 if sc==2
replace scdec = ‐0.058 if sc==3
replace scdec = ‐0.213 if sc==4
replace scdec = ‐0.221 if sc==5
gen uadec = .
replace uadec = 0 if ua==1
replace uadec = 0 if ua==2
replace uadec = ‐0.055 if ua==3
replace uadec = ‐0.162 if ua==4
replace uadec = ‐0.162 if ua==5
gen pddec = .
relace pddec = 0 if pd==1
replace pddec = ‐0.044 if pd==2
replace pddec = ‐0.081 if pd==3
replace pddec = ‐0.276 if pd==4
replace pddec = ‐0.285 if pd==5
gen addec = .
replace addec = 0 if ad==1
replace addec = ‐0.032 if ad==2
replace addec = ‐0.066 if ad==3
replace addec = ‐0.238 if ad==4
replace addec = ‐0.238 if ad==5
gen n5dec = 0
replace n5dec = ‐0.153 if mo==5 | sc==5 | ua==5 | pd==5 | ad==5
gen eq5d5lscore = 1 + modec + scdec + uadec + pddec + addec + n5dec
replace eq5d5lscore = . if mo==. | sc==. | ua==. | pd==. | ad==.
Calculate QALY
- What is QALY?
- A Quality-Adjusted Life Year (QALY) is a measure used in health economics to assess the value of medical interventions.
- It combines both the
length of life
and the quality of life
into a single figure.
- How to calculate QALYs?
- Preference-based measures (e.g., EQ-5D) or valuation techniques (e.g., time-trade-off, standard gamble, visual analogue scale)
- Area under the curve method (Mathews et al. 1990)
Calculate QALY (cont’d)
- Utility value: represents the quality of life, ranging from 0 (death) to 1 (perfect health). Utility Elicitation Calculator
- Time-trade-off (TTO): Choosing between ill health for a specific time or perfect health with a shorter life expectancy.
- Standard gamble (SG): Choosing between ill health or a medical intervention that could restore perfect health or kill them.
- Visual analog scale (VAS): Rating ill health on a scale from 0 to 100, where 0 represents death and 100 represents perfect health.
- Time in health states: amount of time people live in various health states.
- Calculation: Multiply the utility value by the time spent in a health state.
QALY
Figure 2: Possible lifecourses (Source: Utility Elicitation, p. 292.)
AUC
- Treatment group: 1.981 {[0.5(0.189 + 0.620)6 + 0.5(0.620 + 0.883]6)/ 12}
- Control group: 0.998 {[0.5(0.189 + 0.516)6 + 0.5(0.516 + 0.656)6]/ 12}
Figure 3: QALYs for two treatments (Source: Measuring and Valuing Health Benefits for Economic Evaluation, p. 268
Cost-Effectiveness Analysis (back to the example!)
Figure 4: Cost-Effectiveness of TEG vs routine care
Example (cont’d)
- Net Monetary Benefit (NMB):
- NMB = (WTP threshold × Effectiveness) - Cost
wtp <- 50000
eff_new <- 0.05
cost_new <- 1150
eff_old <- 0.049
cost_old <- 1000
standard_nmb <- (wtp * eff_old) - cost_old
new_nmb <- (wtp * eff_new) - cost_new
TEG rehabilitation NMB: 1350