In 1921, Frederick Banting discovered the hormone insulin, which saves the lives of patients with insulin-dependent diabetes [type 1 diabetes (T1D)] every day. One hundred years later, modern researchers are on the brink of putting on the market a smart device that can completely automate the complex life-sustaining insulin-dosing regimen required to manage this chronic disease. Although not a cure, Internet-of-Things (IoT) technology helps to ease the burden of disease management for these patients.
There are many life-changing technologies that have advanced medical treatments, such as artificial organs, prosthetics, and robotic surgery equipment, to assist medical providers in treating patients. However, this article focuses on “smart” medical devices that assist patients diagnosed with chronic conditions in the everyday care and management of their disease, which, in turn, improves quality of life and increases peace of mind.
First we need to answer the question: What makes any device smart? In other words, when the term smart is added to a product, what does that mean?
Generally, the term smart device should refer to an electronic device that meets certain criteria or has a certain architecture. By some definitions, a smart device is simply one that has an embedded sensor, such as the accelerometer or fingerprint sensors embedded in mobile phones. If, along with the sensing, the device includes data collection and analysis, is connected to a network (for example, Bluetooth or Wi-Fi), and performs some type of actuation based on the data, we can also categorize it as an IoT device.
To bring clarity and understanding to this topic of smart/IoT technology criteria, the National Institute of Standards and Technology (NIST) provides a special publication (SP) defining the building blocks/primitives of the IoT. In NIST SP 800-183,11 five primitives in IoT devices are defined:
1) a sensor (something measuring a physical property),
2) an aggregator (a software algorithm to transform the collected sensor data into information),
3) a communication channel (a medium to transmit the data),
4) an external utility/eUtility (hardware to process the data flow), and
5) a decision trigger (a condition to execute an action/transaction).Not all five primitives are necessary for a device to be considered smart.
Given this explanation and for the purposes of this article, we put devices into two categories: 1) those based on simple monitoring and 2) those that perform complex interactive tasks. The goal is to understand how the IoT empowers the medical field using the following two categories: 1) smart devices that collect data and provide information to aid in care decisions and 2) more sophisticated devices that are able to automate patient care based on real-time data. The remainder of the article will provide examples of both types.
This category includes medical devices that have the capability to aggregate, analyze, and store data. These types of medical devices collect and analyze real-time patient data to provide the patient, provider, or caregiver with more than a snapshot of information to make medical decisions.
For example, Parkinson’s disease is a brain disorder that leads to many physical movement problems. For patients with this disease, researchers developed a watch-like, movement-tracking device using a motion sensor to track abnormal movements. Along with the tracking aspect, the patient uses the device to record when medication is taken as well. Data from the sensor are recorded every 2 min and, coupled with the diary of medications, can ultimately help with adjustments of medication timing.9 Without a device like this, a provider makes medication decisions based on a snapshot patient exam every three months.
Another device in this category is a smart thermometer to track body temperature. Consider if the user lives alone and is not feeling well—it may be difficult for that person to track temperature and medication times. In another scenario, a user may need to accurately track his or her temperature over a few days. A smart thermometer is a device that, once it has taken the user’s temperature, offers guidance based on age and temperature history (to determine if the user is getting better or worse) and provides medication tracking. Some thermometers also have storage capabilities to store baseline temperatures for multiple users.
Smart asthma monitoring can be accomplished with a patch device (containing sensors) worn by the patient to detect symptoms such as cough rate, wheezing, respiration pattern, heartbeat, and temperature. This can help a patient/caregiver by using notifications and reminders for medication dosing. With the amount of data tracked, the patient may be able to discern asthma trigger patterns.
Heart disease is the leading cause of death worldwide. There is a shortage of heart donors; thus, care modalities while a patient is waiting for a heart transplant are a major research focus. Today, artificial hearts are only a temporary solution until a donor is found. Devices are also useful for the early discovery of abnormalities to avoid long-term damage or predict the risk of cardiac arrest.
Thus, to care for heart patients and those at risk for heart disease, wearable devices have been created and improved using the IoT. For example, researchers are trying to develop a “wear-and-forget” device with heart-monitoring sensors that could be inserted into clothing fabric; the goal is a dedicated system that monitors, stores, and sends data to a server. The data are collected from heart-activity monitors using inductive sensors, a mobile electrocardiograph device, and other peripheral devices and then analyzed by a medical professional to provide alerts and a diagnosis.1 Other researchers are working on a multisensory system using the IoT to collect and analyze body area sensor data to predict cardiac arrest. This system uses smartphone-based heart-rate detection and remote supervision to detect a health crisis.
Another disease that afflicts millions of people is diabetes. There are two types of diabetes: insulin-dependent diabetes, or T1D, in which the body does not make insulin, and insulin-resistant (type 2) diabetes. T1D is an autoimmune disease in which the pancreas stops making insulin, and it typically manifests at a young age (average age 4–14). Type 2 diabetes, in which the body does not use insulin efficiently, is usually caused by lifestyle and can occur at any age (average 45). T1D requires a lifetime of round-the-clock care. It can be managed manually with at least six finger pricks a day to check glucose levels and at least six needles of dosed insulin (two different types) a day. The manual care is complex and burdensome, which makes it a prime candidate for care automation.
Insulin is an essential hormone produced in the pancreas, the organ that regulates the amount of glucose in the blood. Without insulin, cells cannot absorb glucose for energy. As noted, persons with T1D need to continually monitor their blood sugar, food intake, and activity to determine the amounts of insulin required to keep their blood sugar in a normal range.
An automated way to check blood sugar did not appear on the market until 1999. Some researchers are also exploring the use of smart contact lenses to measure glucose levels.7 Currently, patients who choose to forgo finger pricks use a wearable device called a continuous glucose monitor (CGM), which is a system that provides a glucose reading to the patient every 5 min. The device includes a sensor, transmitter, and receiver. The sensor and transmitter are electrically connected. The sensor is inserted under the patient’s skin to measure the interstitial glucose level.
The sensor is typically a thin wire or filament whose end is coated with glucose oxidase. The glucose oxidase reacts with the glucose in the interstitial fluid, generating an electrical signal. The signal is passed along the wire, translated via the electronics on the CGM, and then transmitted via Bluetooth to a phone or receiver. If the blood sugar level (high or low) needs correcting, an alarm will sound. In addition, the continuous glucose data can be stored on the cloud to allow a patient or caregiver to perform a visual analysis of patterns to consider insulin amount adjustments for the patient.
Although less painful than multiple daily finger pricks, CGM devices are not pleasant to wear and can be difficult to use for an active person playing sports, for example. Although the devices are waterproof, they can get knocked off or become loose in the presence of excess sweat. However, given the correction alarm, the upside is peace of mind—especially while sleeping—which leads to a better management of blood sugar and a healthier outcome for the patient.
The challenge is that the amount of insulin one needs throughout a day varies and depends on many factors.