Table 4 Summary table of the structure of the three economic modelling papers included in the review

Hadzijahic et al. (2000) [29], USA

To determine whether it is less costly to request CT or EUS first to identify advanced oesophageal cancer; to determine which variables most affect the overall cost of identifying advanced disease.

Oesophageal cancer patients who underwent both CT and EUS between July 1995 and April 1999 (n = 124, mean age = 62.7 years, 98 (79%) men, and 72 (58%) white).

Cost-minimisation study using decision tree model to compare which of the two initial staging strategies (EUS first or CT first strategy) would cost less to detect advanced disease in patients diagnosed endoscopically with oesophageal cancer.

Not stated specifically, the study took local referral centre perspective.

Price year: 1999

Currency: US dollars (USD$).

CT first strategy vs. EUS first strategy.

Decision analysis using decision tree model.

Not stated specifically.

Overall cost of identifying advanced disease of the two strategies: EUS first and CT first strategies.

Initial CT is the least costly strategy if the probability of finding advanced disease by initial CT is greater than 20%, if the probability of finding advanced disease by initial EUS is less than 30%, or if the cost of EUS is greater than 3.5 times the cost of CT. EUS found advanced disease more frequently than CT (44% vs. 13%; p < 0.0001) and initial EUS was the least costly strategy (Initial EUS strategy expected cost was US$804 (£824, 2017 price year) vs. initial CT strategy expected cost $844 (£867, 2017 price year)).

Harewood et al. (2002) [30], USA

To examine which staging/management technique was the least costly: EUS FNA, CT-guided FNA or surgical management of oesophageal tumours.

Patients with apparently “resectable” oesophageal cancer on CT (i.e. patients with non-metastatic oesophageal cancer).

Cost-minimisation study using decision tree model to determine which strategy is least costly among the different alternatives: CT-FNA, EUS-FNA and ‘proceed straight to surgery’ options.

Third party payer perspective.

Price year: 2001

Currency: US dollars (USD$).

CT-FNA vs. EUS-FNA vs. ‘proceed directly to surgery’.

Decision analysis using decision tree model.

Not stated specifically.

Least costly staging strategy among the three strategies (CT-FNA vs. EUS-FNA vs Surgery)

EUS FNA was the least costly strategy at $13,811 (£14,578, 2017 price year), followed by surgery at $13,992 (£14,768, 2017 price year) and CT-FNA at $14,350 (£15,147, 2017 price year).

EUS FNA remained the least costly option, provided that the prevalence of celiac lymph node (CLN) involvement was greater than 16%. Below this value, surgery became the least costly strategy.

The final outcome of the model was also sensitive to variation in the sensitivity of EUS FNA. Provided that the sensitivity of EUS-FNA was greater than 66%, EUS-FNA remained the least costly staging option in the management of oesophageal tumours.

Despite changing the values of two or three variables simultaneously in the two- and three-way sensitivity analyses, the result still showed that EUS FNA remained the least costly strategy.

Wallace et al. (2002) [31], USA

To compare the health care costs and effectiveness of multiple staging options for patients with oesophageal cancer.

All Medicare-eligible patients whose invasive oesophageal cancer was diagnosed between January 1991 and December 1996. Data were obtained retrospectively from the SEER–Medicare databases.

Cost-effectiveness study using decision tree model to compare the costs and effectiveness of six strategies (CT alone vs. CT + EUS vs. CT + TL vs. CT + EUS + TL vs CT + PET+EUS vs. PET+EUS).

Third-party payer perspective

Price year: 2000

Currency: US dollars (USD$).

The costs and effectiveness of the six strategies were compared – CT alone vs. CT + EUS vs. CT + TL vs. CT + EUS + TL vs CT + PET+EUS vs. PET+EUS.

Decision analysis using decision tree model.

Not stated specifically

Cost, QALYs and cost per QALY of the six strategies

Under baseline assumptions, CT + EUS-FNA was the least costly strategy and offered more QALYs, on average, than all other strategies with the exception of PET+EUS-FNA. The latter was slightly more effective but also more costly. The marginal cost-effectiveness ratio comparing PET+EUS-FNA with CT + EUS-FNA was $60,544 per QALY (£66,588 per QALY, 2017 price year). These findings were robust and changed very little in all of the sensitivity analyses.