Technology Fails. I know this, so do you. Somehow, I am always surprised when low-end technology fails, such as a refrigerator. In the case of my organization, it was an expensive failure. The refrigerator of note was one in which we stored vaccines. The failure was a simple one. The unit was left open and there was no automated temperature monitoring to alert the staff. It failed over a weekend, so the problem was not discovered until Monday morning and we lost over $30,000.00 worth of medication and inconvenienced many of our patients. This was the catalyst that started our journey into IoT.
We looked at several commercially available temperature alarming solutions and found them to be both expensive, and limited to a single task. Concomitantly, I was speaking with a vendor partner who spoke of some IoT work they were doing with monitoring food carts, and we came upon the idea of exploring if temperature monitoring would be a possible solution for my issue. I really liked this idea, because it opened up an opportunity to not only build out an IoT infrastructure for my organization, but also set the stage for many other possibilities, such as; patient monitoring of vital signs, some real-time location tracking, integrating wearable devices into our care plans, and the potential to transmit patient-specific information to the clinical staff, which would enhance patient safety, provide for even higher quality of care, as well and contribute to improved staff satisfaction when using technology.
The challenge, of course, was to find a solution that could support rapid deployment, be in line with our digital strategy, and provide the appropriate cybersecurity healthcare would require. Yes, I had high expectations but believed the effort was worthwhile. We knew we would have to invent the workflow, and “product” and thus sought a partner that would complement the skills we had in-house. We wanted the solution to be cloud-based (I'm trying to make my organization as cloudy as possible), along with being ‘Network agnostic’, built on an infrastructure we can grow with relatively inexpensive sensors, and a reporting counsel (dashboard), which was visually interesting, and avoided “non-essential data noise”. It had to integrate into our organization’s Emergency Alerting System and be able to deliver an alert using cellular service. Also, we had to get smart on IoT technology.
Having the fundamental idea of what we wanted to accomplish with an IoT capability, we built several use cases and evaluated network topologies and technologies that would enable rapid deployment, provide reliable communications and be at a competitive cost
Having this fundamental idea of what we wanted to accomplish with an IoT capability, we built several use cases and evaluated network topologies and technologies that would enable rapid deployment, provide reliable communications and be at a competitive cost. Our criteria considered: Data Rates, Range, Battery Life Device and Sensor Cost. The three categories of network technologies we evaluated were: LAN (Bluetooth, WiFi, Zigbee), Cellular (3G, 4GLTE, NB-IoT) and LPWAN, (Low Powered Wide Area Network). Of the three, we choose LPWAN, because our use cases did not require high data rates. The range was more than adequate, and the battery life and device costs were the best.
LoRa & LPWAN are a communication protocol and long range, low power wireless devices that seems to have become the de facto technology for Internet of Things (IoT). Other aspects contributing to a greater advantage was that LPWAN is great for single-building applications and gave us the greatest flexibility in setting up our own network. That was critical in controlling costs from a POC perspective, especially since we had a ‘large inventory’ of use cases.
Perhaps one of the more challenging aspects of our journey was the availability of LoRaWAN devices that had the necessary healthcare certifications required by our Lab and Pharmacy departments. While we had to adjust our initial deployments, we were able to work through our partner to address device certifications and availabilities. However, we feel that this will be an ongoing issue regardless of network technology.
Operationally, we needed to deliver access to the devices on the network in case of adverse conditions, assets moving out of secure areas and for maintenance activities. That required an administrative platform that could not only provision the devices from one interface but also allow for automated alerting and escalation, including corrective action capture for compliance reporting. Our partner was able to provide an administrative platform that met our needs with minimal customization from a UI perspective.
Our deployment model was simple: 1) Connect & Monitor [establishment of an IoT network and delivering the data into a repository}; 2) Analysis & Feedback [implementation of an administration system with data visualization & insights], and 3) Provide Expansion Network and Device expansion to include asset tracking, as well as other use cases.
Our project was a success. We currently have 252 sensors, monitoring Pharmacy and Laboratory temperatures throughout the Organization. We are also monitoring temperatures in the Network and Telecommunications closets. Our dashboard visually presents the status of the devices and highlights temperature sensors that are out of range. These provide an alert to appropriate staff, and if unanswered, escalates to the Department Manager or Director, as necessary. The dashboard also shows devices that need to be re-calibrated within the next 2-weeks (a regulatory requirement), as well as the battery condition of the sensors. It is a solution which was accepted and praised by leaders in our various departments. Our next step is to expand the temperature sensors to all of our physician practices, and to start the work on asset tracking for some high cost equipment. One of our lessons learned was to calibrate the sensors and set the alert level early in the process. When we first turned on the system, the team was receiving about 3,000 messages per hour, that our sensors were out of range. Oops! But all in all, we feel as if we met our requirements and have delivered a solution to the Organization that is effective in both a functional and financial model.