Malfunctions and errors can be significantly reduced by integrating such technologies as MEMS, energy-harvesting RFID, and low-power CMOS process technologies and IP.
Recently, the FDA reported that “of the 56,000 medical device reports relating to the use of infusion pumps received by the FDA over a five-year period, approximately 1% (560) were related to deaths, 34% (19,040) to serious injuries, and the remainder to system malfunctions.” As a result, the FDA issued a guidance directed specifically toward infusion pumps. It states that when paired with the proliferation of ambulatory infusion pumps, auto-injectors, and syringe pumps, as well as the desirability of a disposable infusion pump, there’s a driving need for miniature, low-power, cost-effective, precise methods of monitoring dose size and drug flow rate. This is to eliminate, or at least reduce, dosing errors, and therefore the number of deaths and injuries, caused by these systems. The integration of emerging MEMS, energy-harvesting RFID, and low-power CMOS process technologies and intellectual property (IP) has the potential to address these concerns and provide additional benefits in the form of several different devices in the near future.
Current system shortcomings
Current infusion pumps and auto-injectors lack true closed-loop feedback, like a flow sensor, to ensure that the correct dose size and flow rate are administered for any particular drug, relying instead on calculations for dose-size and flow-rate data derived from drip sensors/counters for gravity-based pumps and screw/piston position sensors or load cells used in volumetric pumps. These flow rate and dose size calculation methods are subject to inherent errors introduced by volumetric chamber irregularities, temperature, fluid viscosity, atmospheric pressure, and back pressure variability. In addition, and perhaps most critically, there’s the potentially fatal human error of administering the wrong drug.
Ideally, you’d like to have a cost-effective, highly accurate biocompatible fluid flow meter with an integrated temperature sensor at the point of administration (subcutaneously). It should be capable of harvesting ambient energy of some type (heat, kinetic, solar, RF, etc.) to power itself and communicate via an RF protocol (Zigbee, BTLE, RFID, etc.) with the main system microcontroller (MCU). This would ensure the most accurate flow rates and dose sizes. For an infusion pump, this would be at the end of the infusion set. But in an auto-injector or syringe pump, it could be in the drug cartridge volumetric chamber, where a unique identifier communicated over the RF protocol to the main MCU would confirm that the correct drug was being administered and enable. If this wasn’t the case, it would disable the operation of the pump/injector accordingly.
The added benefit of this mechanism would be to eliminate the possibility of competitors cloning the cassette or patients unintentionally using counterfeit drugs. The data could also be encrypted, as is currently implemented in NFC readers/tags, to ensure communication integrity. As it would be in constant communication with the main microcontroller, the data received could be logged for later download. It could also be used for alarms, such as free-flow/improper flow of fluid, occlusion, air-in-line, dose limit/Bolus limit exceeded, empty reservoir, no reservoir, and drug mismatch to reduce or eliminate harmful errors. The temperature data from the temp sensor would be used for flow rate compensation calculations. Critically, this solution would survive all forms of sterilization, be it steam or gamma radiation.
Most of the technology and IP required to implement this type of solution already exists, with the remainder currently in development. Taking circuit blocks from an energy harvesting dual-interface EEPROM, a low power MCU, a temperature sensor, and the MEMS structure from a pressure sensor could easily be integrated into a single-chip device that addresses the many FDA concerns. Complimentary electronics-on-plastic technology provides a biocompatible platform for strain gauges and flow sensors to be paired with the silicon to provide a complete system-in-package while keeping costs low by utilizing printed electronics and metals not subject to the high mask prices and expensive fabs that traditional monolithic circuits depend on. This same technology has the added benefit of easily integrating the RFID antenna onto the same piece of plastic.
One could implement an energy harvesting sensor system today by architecting a system based on currently available products. An energy harvesting RFID IC, an ultra low-power MCU, a temp sensor, and a capacitive pressure sensor, flow sensor, or strain gauge would enable you to effectively build a remote sensor that could communicate with and run off of an RFID reader’s E-field. This would allow for direct contact with the fluid of interest (assuming effective packaging techniques), increasing system accuracy over current indirect systems. As seen in the figure, several strain gauges or pressure sensors used in conjunction with a temperature sensor would directly provide data from which flow rate could be calculated.
An energy-harvesting sensor system can be designed using currently available products.
There are two key components of this intermediate solution. The first is the energy harvesting dual interface EEPROM (ST model M24LR16E-R), which has the ability to power itself off of the E-field from a standard NFC/RFID interface. In addition, it can provide power to an external device that it harvests from the RF interface, specifically the second key component of the solution, the ultra low-power MCU. These two devices, used in conjunction with a RFID reader on the main pump unit, would enable seamless data transfer from the sensor MCU to the main controller via the dual I2C and RFID interfaces.
This sensor system, utilized in the presence of an RFID reader on the pump or handheld unit, would be completely battery-less, avoiding all of the compliance, bio-compatibility, and drug compatibility issues associated with batteries. Upon the initial use of the hose or drug container to be used with the infusion pump, the RF interface would provide power to the sensor system. It would authenticate that the drug is compatible with the pump by reading a pre-stored unique ID. It would then sync the sensor with the RTC from the pump by writing the initial timestamp to the EEPROM. This, in turn, would power up the ultra low-power MCU, start the internal RTC, and then put the system to sleep. After the initial sync process, any time an injection is performed, the pump would wake up the sensor over the RF interface via interrupt on the dual-interface EEPROM. The microcontroller would interrogate the pressure sensor/flow sensor/strain gauge, and provide real-time feedback over RFID to the main controller, allowing for more accurate dosing.
If you assume constant flow throughout the system, this solution could be implemented in a prepackaged cartridge for both auto-injectors and infusion pumps to increase total pump accuracy and FDA compliance. AT the same time, it would reduce the harmful incident rate and increase pharma revenue by preventing the cloning of delivered drugs via specific devices.
Nick Trombly is a market development engineer at STMicroelectronics, based in Schaumburg, IL, where he’s responsible for developing system and application specific solutions for the Healthcare sector. He received his MSEE and BSEE from Michigan State University.