The cost-effectiveness of using waking salivary cortisol in the diagnosis of adrenal sufficiency

Adrenal Insufficiency (AI) is a deficiency in the stress hormone cortisol, and when untreated it can lead to life‐threatening complications. Clinically, AI can present with non‐specific signs and symptoms such as fatigue, muscle aches and abdominal pain, often resulting in delayed diagnosis. Currently, the Short Synacthen test (SST) is the reference standard for diagnosing AI, but requires venepuncture and a 2‐3 hour hospital attendance. A new home‐based test is available that uses a patient’s waking salivary cortisone (WSC) levels as a surrogate marker of that collected by the SST. As well as being more convenient for patients, the test only costs £19 per patient compared to the £184 for the SST.

A recent diagnostic accuracy study undertaken in Sheffield compared SST and WSC and produced negative and positive prediction values for AI of 96% (95% CI; 90 to 99) and 95% (95% CI; 87 to 99), respectively. However, despite the potential costs saving of £165 per patient if WSC was to directly replace SST, this is not considered realistic given the slightly inferior diagnostic performance of WSC relative to the SST. Consequently, it is more likely to be adopted as part of a two‐stage diagnostic strategy, with WSC used to screen for AI and the hospital‐based SST used to test for AI in patients with an equivocal WSC result.

In this study, we report an economic evaluation of a two‐stage diagnostic strategy from the NHS perspective and with a time horizon bounded by the generation of a non‐equivocal biochemical diagnosis. This is undertaken by formulating a decision tree to look at costs and diagnostic performance for the two approaches, which is populated from the Sheffield diagnostic accuracy study and local unit costs.

The economic results for the primary analysis show that a two‐stage diagnostic strategy costs £103 per patient less than the current diagnostic strategy (or 58% less). This is associated with a worse diagnostic performance with 7 out of 220 patients (3.2%) receiving a false diagnosis (four false positive and three false negative diagnoses). When the secondary analyses are considered, the use of alternative diagnostic cutoffs for SST has little impact on the results. The use of a societal perspective increases cost savings to £127 per patient (or a 59% reduction relative to current care).

The two‐stage strategy is considered to be a valuable alternative to the current diagnostic process, due to its cost savings, that are associated with improved patient experience and quicker diagnosis for the vast majority of patients. Other benefits include the enhanced resilience of the WSC testing process due to the avoidance of hospital visits (which becomes even more important during health service crises, including pandemics) and the need to procure synacthen. Further studies should look to build on our findings, by adopting the proposed two‐stage diagnostic pathway in the routine clinical practice of other hospitals.