Introduction National Institute for Health and Clinical Excellence approved the use of relatively expensive capsule endoscopy (CE) to investigate mid gastrointestinal haemorrhage (MGIH) and small bowel Crohn's disease (CD) in 2004. It gives a high diagnostic yield (31–76% MGIH, 43–71% CD), does not use radiation and causes few complications. In the past cheaper contrast fluoroscopy was the investigation of choice. Robust data that compare the two are limited. Which minimises costs when used in a public general hospital?

Methods As part of our service audit we used electronic records to identify all patients who had had CE±fluoroscopy to investigate MGIB or CD at our institution from January 2007 to May 2008. (Patients referred from elsewhere for CE were excluded.) We established the indication for investigation, other tests that had been done and the positive diagnostic yield for each. We used activity based costing to determine procedure costs. We then constructed a cost minimising pathway for patient investigation for MGIB and CD and subjected our conclusions to sensitivity analyses.

Results For CD investigation, 136 patients had fluoroscopy, 21 CE and 3 both. For MGIH 47 had fluoroscopy, 59 CE and 8 had both. The positive diagnostic yield for CD was 23% with fluoroscopy, approximate half that of CE (57%) and for MGIH, 2% with fluoroscopy and 56% with CE. Observation of actual practice suggests that both positive and negative fluoroscopy results may help clinicians conclude investigation of suspected CD in that few of the patients we studied had had both investigations. It appeared that the investigations in themselves were rarely definitive and were interpreted in the clinical context of individual patients. Relative costs for CE in the UK range from £1075 to 1368 and fluoroscopy £146–288. Fluoroscopy detects approximately half the CD cases that CE would at less than a quarter of the cost of CE.

Conclusion To investigate CD, our analysis suggests that fluoroscopy should be used initially and only if negative and clinical suspicion of CD remains should CE then be done. The very low diagnostic yield for fluoroscopy vs. CE for MGIH demands that CE should be the imaging modality of choice and fluoroscopy abandoned as a diagnostic option. To determine the place for new, often expensive, investigation techniques cost minimisation analysis is effective. We estimate that, for our patient cohort, the use of CE as the first investigation for MGIH and only after a negative fluoroscopy study for suspected CD, our institution could have saved ∼£7000. Extrapolating these figures nationally the judicious use of CE has the potential to save the NHS £1.3 million per annum. The analysis does not factor in other CE benefits such as faster diagnosis of MGIH and less radiation exposure.