The cost-effectiveness of screening for oral cancer in primary care

Abstract

OBJECTIVES: To use a decision-analytic model to determine the incremental costs and outcomes of alternative oral cancer screening programmes conducted in a primary care environment. DESIGN: The cost-effectiveness of oral cancer screening programmes in a number of primary care environments was simulated using a decision analysis model. Primary data on actual resource use and costs were collected by case note review in two hospitals. Additional data needed to inform the model were obtained from published costs, from systematic reviews and by expert opinion using the Trial Roulette approach. The value of future research was determined using expected value of perfect information (EVPI) for the decision to screen and for each of the model inputs. SETTING: Hypothetical screening programmes conducted in a number of primary care settings. Eight strategies were compared: (A) no screen; (B) invitational screen--general medical practice; (C) invitational screen--general dental practice; (D) opportunistic screen--general medical practice; (E) opportunistic screen--general dental practice; (F) opportunistic high-risk screen--general medical practice; (G) opportunistic high-risk screen--general dental practice; and (H) invitational screen--specialist. PARTICIPANTS: A hypothetical population over the age of 40 years was studied. MAIN OUTCOME MEASURES: The main measures were mean lifetime costs and quality-adjusted life-years (QALYs) of each alternative screening scenario and incremental cost-effectiveness ratios (ICERs) to determine the additional costs and benefits of each strategy over another. RESULTS: No screening (strategy A) was always the cheapest option. Strategies B, C, E and H were never cost-effective and were ruled out by dominance or extended dominance. Of the remaining strategies, the ICER for the whole population (age 49-79 years) ranged from pound 15,790 to pound 25,961 per QALY. Modelling a 20% reduction in disease progression always gave the lowest ICERs. Cost-effectiveness acceptability curves showed that there is considerable uncertainty in the optimal decision identified by the ICER, depending on both the maximum amount that the NHS may be prepared to pay and the impact that treatment has on the annual malignancy transformation rate. Overall, however, high-risk opportunistic screening by a general dental or medical practitioner (strategies F and G) may be cost-effective. EVPIs were high for all parameters with population values ranging from pound 8 million to pound 462 million. However, the values were significantly higher in males than females but also varied depending on malignant transformation rate, effects of treatment and willingness to pay. Partial EVPIs showed the highest values for malignant transformation rate, disease progression, self-referral and costs of cancer treatment. CONCLUSIONS: Opportunistic high-risk screening, particularly in general dental practice, may be cost-effective. This screening may more effectively be targeted to younger age groups, particularly 40-60 year olds. However, there is considerable uncertainty in the parameters used in the model, particularly malignant transformation rate, disease progression, patterns of self-referral and costs. Further study is needed on malignant transformation rates of oral potentially malignant lesions and to determine the outcome of treatment of oral potentially malignant lesions. Evidence has been published to suggest that intervention has no greater benefit than 'watch and wait'. Hence a properly planned randomised controlled trial may be justified. Research is also needed into the rates of progression of oral cancer and on referral pathways from primary to secondary care and their effects on delay and stage of presentation.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 210 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.