Why do insulin pumps have a single sensor?

Question

Sometimes a sensor can malfunction, or the area surrounding it can be infected, leading to wrong glucose measurements. Even just the fact to rely on a single sensor seems to psychologically limit the adoption of such devices, both by physician and patients.

Naively, using multiple sensors located at different places looks like a reasonable solution.

Why do most (all ?) insulin pumps use a single glucose sensor ? What are the reasons not to use multiple sensors ? Anticipating on the "sensors are expensive" answer -> They are mostly because of R&D costs. So once developed, using several ones should not be a lot more expensive than using one.

Answer 1

There have been few studies (on animals and humans) with multiple CGM sensors showing a more accurate estimation of the BG. [1][2]

The site of application of the sensor, the discordance between multiple data sources and errors in the calibration of the sensors, are open challenges to be precisely addressed before having a system of this kind. Moreover, unless we're strictly speaking of subcutaneous sensors, from a patient perspective there is considerable trouble in maintaining multiple sensors. (e.g. in terms of site changes, application in relevant areas, etc..)

Current sensors achieve a precision around 9% MARD, while the benefit of a system with multiple sensors would be a slight improvement on this number, due to the intrinsic limitations of the CGM, upfront a significant more expensive system and more complexity for the patient. (e.g. delay in the estimation, the noise of the measurement,..)

[1] https://ieeexplore.ieee.org/abstract/document/6615958

[2] https://www.liebertpub.com/doi/abs/10.1089/152091503322640980