Innovation does not exist in a vacuum. Many meaningful innovations rely on multidisciplinary technological advancements and proper market conditions—being in the right place at the right time. Right now, we are on the brink of a revolution in how consumer medical devices are designed.

It may be possible to fundamentally alter how we design consumer medical devices, moving away from a process led by consumer insights and usability testing, to one utilizing real world data to improve patient outcomes. A combination of industry trends and technological improvements makes this possible for the first time, creating a new opportunity for medical device companies and product designers.

A Long Time in the Making

In 1999, during a speech to President Clinton at an event called “Millennium Evening at the White House,” Dr. Vinton Cerf talked about the future—a place where it was possible to connect everyday objects to the Internet. This future was a world where devices like your bathroom scale could send your weight to your doctor and have that information added to your medical record. At the time, there were about 60 million Internet-connected devices, almost all were hardwired computers.

Only a few years later, in October 2004, an article titled “The Internet of Things” was published in Scientific American, stating that “giving everyday objects the ability to connect to a data network would have a range of benefits.” It is hard to imagine that connected devices were still just an idea, with no real understanding of the extent to which they could be used or how they would affect our culture and industries.

Today, a little more than 12 years later, these ideas are not only feasible, but connected devices are becoming pervasive in society and will only continue to become more widespread. In a 2015 report titled “Worldwide Internet of Things Forecast, 2015–2020,” the International Data Corporation estimated the number of Internet of Things (IoT) devices at 9 billion—and that excludes devices such as smartphones, tablets, and computers. By 2020, that number is forecasted to jump to an astounding 28.1 billion.

It seems as if every new consumer device is now a “smart” device and anything that can become “connected” already is or soon will be. Everything from fitness trackers and hairbrushes, to coffee machines and thermostats, the list of connected devices includes anything and everything.

Medical Devices Are Not What They Used to Be

Following closely behind consumer products, connected medical devices are entering the market and becoming increasingly more common. Realizing the prediction Dr. Cerf made 18 years ago, we are uploading our weight to the Internet; however, medical devices are also doing much more. Connected devices can be used to track the blood glucose levels and insulin injections of a diabetic, inhaler usage for patients with asthma or COPD (Chronic Obstructive Pulmonary Disease), and even monitor vital signs during the course of a pregnancy.

While many consumer IoT applications are used to create convenience in our lives, connected and smart medical devices serve a very different purpose. They are part of the growing trend in healthcare called Digital Health, the idea of turning health data into valuable information that can be utilized by multiple stakeholders. Primarily, these stakeholders are the patients and doctors who use the data as a means to improving patient outcomes. By tracking device usage and medication administered, patients can keep themselves accountable for tracking their own health status while doctors are given the ability to track a patient’s progress and compliance to prescribed usage remotely. There are also commercial benefits for stakeholders such as pharmaceutical and insurance companies. And there is an additional stakeholder that can benefit from this information that has not yet been realized: the device designer. One of the most valuable benefits of connected medical devices will be using the data the devices collect to help designers create even better products to improve patient outcomes.

Market Trends Create Room for Cross-Discipline Innovation

The reason we will be able to design better products relies on a growing trend within consumer devices in general—one that has recently been spreading to medical devices. Devices are becoming easier to use and more intuitive with the help of technology. They are labeled as “smart devices” and are designed to reduce potential sources of error by automating the use process and guiding users through the steps as much as possible. Picture a standard inhaler; most people think of something that looks like the inhaler shown in Figure 1.

However, this is an inhaler of the past. These devices are usually mechanical and come with complex instructions while the devices currently being developed are often electro-mechanical with screens that guide the user through the process step by step. Many of these devices have alarms, notifications, and applications to connect to your smartphone. The inhalers of today look more like those shown in Figure 2.

These smart inhalers are designed to help guide the user through the use process via sensors on key components to track where the user is in the process. If the user deviates from the standard procedure a warning can be provided to let the user know something might have been done incorrectly and inform them which step to perform next.

There are currently no standards or best practices for what types of sensors are necessary for these devices. The type and quantity of sensors used depends on the design intent, though conceivably there could be sensors on all key components. For example, the device could recognize when the cap is removed, indicating the device is about to be used to administer the medication. An accelerometer and gyroscope could be used to recognize when the user is shaking the device and if the device is in the correct orientation. The canister could detect force pressure, indicating the user is taking their dose or priming the cartridge, a process in which medication is expelled into the air to ensure any old medication is removed and the inhaler is ready to deliver a dose. Skin sensors can be used to detect when the user puts their mouth on the mouthpiece, while another sensor can be used to detect the length and strength of the user’s inhalation. Many of these, if not all, have been implemented in devices on the market or devices currently in development.

This is not just happening with inhalers. A similar transformation is happening with reusable auto-injectors. Auto-injectors are a type of syringe in which the needle is automatically inserted into the skin and the medication is delivered all by simply pushing a button. This differs from conventional syringes or pen-injectors where the user is required to manually insert the needle and depress the plunger to deliver the medication. The typical auto-injector looks similar to the device shown in Figure 3.

Auto-injectors were primarily fixed-dose devices that used pre-filled syringes for the medication. However, many of the new auto-injectors use a disposable needle that connects to a reusable medication cartridge. The auto-injectors being developed today are vastly different and look more like those shown in Figure 4.

Like inhalers, auto-injectors can have sensors on various parts of the device. Different sensors on the device could be used to detect any or all of the following:

• How much medication is in the cartridge
• How long the cartridge has been in the device
• The temperature of the medication
• If a needle is loaded and how often the needle is changed
• If the device is correctly placed on the user’s skin prior to initiating the injection
• How long the device is held on the skin during the injection, as this is necessary to ensure that the full dose of medication is delivered.

This is not an aspirational list of features or technology that is not currently feasible. Devices currently available on the market or that are being developed are implementing sensors like these.

Where Does this Fit in With the Current Design Process?

Because of the move toward smart devices to improve patients’ experiences, there is valuable usage data available from the device sensors. To understand why this data is useful and how it can be utilized, it is important to understand how medical devices are designed.

Before a new medical device can go on the market, the US Food and Drug Administration (FDA) requires human factors testing to ensure the devices are safe and effective. As described by the FDA for Premarket Notification, often called 510(k) approval, this testing is to “demonstrate that the device to be marketed is at least as safe and effective, that is, substantially equivalent, to a legally marketed device.” One aspect of this involves conducting validation testing, a final user study where a production-ready device is tested with representative users doing representative tasks.

Additionally, even before validation testing, it is necessary to conduct formative studies. These studies are part of an iterative design process in which users are observed interacting with the device at multiple stages of development to improve the design and address any usability issues that arise during development.

However, even with the most thorough human factors plan, there are inherent limitations. One drawback is that the studies are conducted using simulated scenarios in a representative environment and are not being used naturally. Additionally, since participants are paid for their time and are being observed, there are biases that can affect their behavior, potentially altering the results of the study.

Furthermore, we are only able to ask participants to perform a task a couple of times, so there is no accurate information about the long-term use of the device. Specifically, how they will use the device on the tenth, hundredth, or thousandth use. For these reasons, observed behavior does not always match real world usage.

Second, as defined by the FDA, the goal is to create a safe and effective product.  However, that doesn’t necessarily mean it will be utilized effectively. A device can be safe and easy to use when users are asked to complete a task in a study, but in real life if the user misses their dose, they will not achieve a better clinical outcome.

Perhaps the process to administer the medication takes too long. During a user study, this could be a minor annoyance, but in the real world this could mean the difference between taking the correct dose on time or skipping it for the day. While human factors testing is incredibly useful in ensuring safe and effective products, there are certain limitations that are impossible to overcome without understanding what is happening during normal, everyday, use.

The potential for data available from device sensors, which can be automatically uploaded, made anonymous, and analyzed, is immense. One of the primary goals of Digital Health is “Nudging,” a way to purposefully change or alter someone’s decision-making process and behavior. The data allows the designer to see what behaviors or actions might need a nudge, a modification to the design to help encourage the user to make the correct decisions about their medication.

For example, the data might suggest that patients who delay their dose by more than two minutes are more likely to end up missing their dose. This might indicate a need for a design change to help ensure patients take their dose within the prescribed time frame. If the data suggests that patients are not leaving their auto-injector on their skin long enough, designers may need to create a nudge to increase compliance with the length of time the device is left on skin.

Where Does This Leave Us?

Correcting these design problems is the easy part, relatively speaking. Until recently, collecting this type of data has been impossible. The data has the potential to transform how products are designed. This will facilitate the design and creation of the next generation of smart devices, devices where the user requirements and design choices are not only formed through user research and best practices, but are also based on real use data.

Unlike consumer devices, where the product life cycle can be as little as 12-18 months, medical devices generally move at a much slower rate. So, while this first round of smart devices is entering the market or being developed, it may take some time before the second generation devices enter development and we are able to implement this data as a design input.

This is just the beginning. It is easy to imagine that once this data can be collected and analyzed, it can be used in conjunction with other data sources, such as electronic health records, fitness trackers, and public data. This is beneficial because effective management of these conditions are much more than the medication you take; it’s the lifestyle you live. For an asthmatic, the relevant data points could be the combination of temperature, humidity, pollen count, and exercise. For a diabetic, it could be the combination of nutrition, sleep, and exercise. Once we can combine all of this data and track it against patient outcomes, we can start to close the loop on disease management. The Digital Health movement will not only make it possible to design better devices, but will allow us to draw more meaningful conclusions and create lifestyle changes to improve outcomes.

While the benefit of IoT will not be fully realized for years and the design implications not even fully understood yet (think smart cities and driverless cars), there are benefits that can be gained even at this early stage of the technology. Though connected medical devices are being used effectively today, there is unrealized potential to utilize the use data to refine the design of consumer medical devices and improve patient outcomes.