Cost Effectiveness Analysis (CEA)

Cost effectiveness analysis (CEA) is a form of full economic evaluation that compares the costs and health outcomes of two or more interventions, expressing results as an incremental cost effectiveness ratio (ICER): the difference in costs divided by the difference in health effects. The health effect is measured in natural units, such as life years gained, cases detected, infections prevented, or points gained on a clinical scale. CEA is the most commonly used form of economic evaluation in health care because it directly links spending to health outcomes without requiring the monetization of health that cost benefit analysis demands.

Health care resources are finite, and decision makers must allocate them to maximize health gains. CEA provides the evidence base for these allocation decisions by quantifying how much additional health benefit each additional dollar of spending buys. Health technology assessment agencies worldwide, including NICE in England, Canada's Drug Agency (formerly CADTH), and PBAC in Australia, use CEA results to inform coverage and reimbursement decisions. For researchers, conducting and interpreting CEA is increasingly expected alongside clinical trials and observational studies, especially when studying interventions with significant cost implications.

Researchers estimate the total costs and health outcomes of the intervention and its comparator over a defined time horizon. Costs typically include direct medical costs (drugs, hospitalizations, physician visits), direct nonmedical costs (transportation, caregiving), and sometimes indirect costs (lost productivity). Health outcomes are measured in the same natural units for both alternatives. The ICER is calculated as: (Cost of intervention minus Cost of comparator) divided by (Effect of intervention minus Effect of comparator). When the intervention costs more and is more effective than its comparator, it is considered cost effective if the ICER falls below the decision maker's willingness to pay threshold (for example, $50,000 per life year gained). If it costs less and is more effective, it dominates the comparator; if it costs more and is less effective, it is dominated. Neither case needs a threshold comparison, and a negative ICER should not be compared with the threshold, because it can come from either case. When the intervention costs less but is also less effective, the question is whether the savings justify the health given up. Decision tree models, Markov models, and microsimulation models are commonly used to project costs and outcomes over longer time horizons. Sensitivity analyses test how the ICER changes when key assumptions are varied.

Use CEA when comparing interventions that produce outcomes measurable in the same natural health units and when the decision context involves allocating health care resources. It is standard practice for submissions to health technology assessment bodies and is frequently conducted alongside randomized controlled trials as a trial based economic evaluation. CEA is also used in model based evaluations that synthesize data from multiple sources to project long term costs and outcomes.

A common error is comparing an intervention to "do nothing" rather than to the most relevant active comparator, which may overestimate cost effectiveness. Another mistake is using an inappropriate time horizon that does not capture all relevant costs and consequences; a lifelong chronic disease intervention evaluated over only one year may miss important downstream savings or complications. Researchers sometimes fail to discount future costs and outcomes, which inflates the apparent value of long term benefits. Ignoring uncertainty by reporting a single ICER without probabilistic sensitivity analysis or confidence intervals understates the decision uncertainty. Finally, the choice of perspective (societal versus health system) significantly affects which costs are included and should be explicitly stated and justified.

Health economists evaluate a new screening test for colorectal cancer compared with the current fecal immunochemical test. Over a 20 year time horizon modeled for a cohort of 100,000 individuals aged 50 to 75, the new test costs an additional $15 million but detects 850 more cancers at an early stage, resulting in 2,400 additional life years gained. The ICER is $6,250 per life year gained, well below the commonly used $50,000 threshold. Probabilistic sensitivity analysis shows that the new test is cost effective at the $50,000 threshold in 92% of 10,000 Monte Carlo simulations. These results support adoption of the new screening approach.