In 20% of all traffic accidents in Sweden with a fatal outcome, a driver was under the influence of drugs. Statistically, every 500th driver on Swedish roads is under the influence of alcohol. However, these alcohol incidents, with their steady decrease over the years, are less of a problem than the significant increase in driving under the influence of other drugs.

The police's current method of initial detection consists of a pen test and gut feeling. Making a judgment is very difficult due to the small amount of evidence, especially in consideration of the laborious procedure that follows. Police officers must drive the suspect to the nearest hospital or police station to wait for trained personnel to conduct a urine or blood test. This procedure can take from one hour to four and the result follows weeks later from the laboratory. In addition, 20-25% of the tests come back negative, which means that thousands of innocent people have to undergo these procedures every year in Sweden.

We have developed substance:ID based on the Liquid Chromatgraphy Mass Spectrometry technology of a startup that enables the police to detect driving under the influence of drugs on the road. For this purpose, replaceable cartridges are used, which avoid cross contamination and contain the chemical components. The testing procedure itself is minimally invasive, as only a sweat sample needs to be taken from the fingertips, and the suspect does not even need to get out of the car for it. To perform the procedure, substance:ID assists both the police and the suspect in performing the test to ensure a consistent accurate result and a good user experience. Through a quick screen test that can be performed in two to three minutes, an initial suspicion can be investigated. If suspicion is confirmed, the device can perform a longer and more accurate test that has comparable precision to urine and blood tests, shortening the entire conviction to about 22 minutes instead of up to four hours. In addition, Substance:ID saves resources directly and indirectly through avoiding commutes, people, time spent, and accidents.

substance:ID enables testing for driving under the influence of drugs time efficient, accurate, portable, inexpensive, hygenic, non-cheatable and non-invasive, thus enabling greater safety on the roads.

Design Process

The entire project was carried out in very close cooperation with the Umeå Traffic Police and the Police Training Unit of the University of Umeå. In the process, we had the opportunity to interview the police and conduct user tests in total eleven times during the ten project weeks. In addition to this, we repeatedly validated our qualitative insights with the police with nationwide and worldwide research to verify the validity of the problem and our assumptions.

The detailed police experience helped us greatly in identifying problems, prioritizing them, and finding proposals to fix them. We were also able to test the many ideas we developed directly with the user. To do this, we put a strong focus on prototyping to make our ideas understandable and tangible to the police. This allowed us to make not only faster but also more justified decisions. In addition, we were very lucky to have an expert in the police unit who is particularly engaged in incidents of driving under the influence of drugs and could provide us with deep insights.

We divided the project into sprints, in which we iteratively reworked and improved our ideas, thereby learning a lot about the product and its environment. At the end of each week, we had tangible results through which we could make concrete decisions by incorporating feedback from the police. In addition, we were able to test new skills that we acquired during the project piece by piece and did not have to implement an idea from scratch at the end of the project.

Industrial Design

For the Industrial Design, especially in the earlier sprints, we put a lot of emphasis on very large quantity of sketches and foam models to explore a wide range. However, through the sprint format, we committed to making a decision at each end of the week on what outcomes we wanted to take into the next sprint. The quick decisions allowed us to move forward quickly and at the same time provided us with time to make corrections to previous decisions. In the later sprints, we transitioned into milled and 3D printed form prototypes that we could use to validate more detailed feature decisions from CAD.

Central to the industrial design was the external image that substance:ID should embody. In doing so, we placed value on a less intimidating aesthetics that looks more inviting to the suspect. In addition, we were faced with the fact that substance:ID should not hide its high-tech nature and be allowed to communicate the great value it has to the outside world, but nevertheless it should be and communicate robustness. It should fit into the police context and not be eye-catching for bystanders, but still communicate to the immediate suspect what the function of the device is and how to use it.

Interaction Design

In order to be able to prototype the user experience of the concept more precisely, we also developed the interaction design in addition to the industrial design. For this, we tested both digital and physical interactions with the device and aligned them with each other as best as possible to deliver a wholistic user experience for both user groups.

In the design process, we prototyped and tested different user interface concepts and tested different supports and formats of these for both users. Again, we took an iterative approach, starting very early with a rough cardboard prototype that could only provide a rough idea of the testing procedure with a click dummy on the smartphone. In order to include the physical interactions with the device, we built a breadboard prototype, which, however, was not very informative in user tests due to its large abstraction. Therefore, we refined this into a comparable format to cardboard prototypes in a wooden enclosure. At the end of the project, for the final meeting with the police, we built a high-fidelity prototype with working buttons, LEDs, haptic feedback and a display.

Furthermore, we defined the new test procedure in detail and developed user flows in several iterations that reflected the function of the device. Dependent on this, we then tested several user interfaces with animations and interactions. In the end, the prototype was able to reflect the entire range of functions we had defined and provided a realistic impression of substance:ID.

Technology

In the project, we decided quite early on to use the technology Liquid Chromatography Mass Spectrometry due to the great progress made in the last few years. For this, we took a role model from the startup Intelligent Fingerprinting, who have incorporated this technology into a desktop device. For the test to work, a sample of sweat is taken from the fingertips of the suspect. In order to get enough sample material for a precise test, all ten fingers are pressed for five seconds on a field that takes the sample. The sample is then mixed with a buffer solution, which is located in a container in the cartridge and can be activated by pressing a button. The sample then enters a flow phase with the buffer solution and moves to an optical window through which light is transmitted. This light is split and can then be analyzed for substances in the sweat. The test takes ten minutes to analyze today, but can be performed in as little as 30 seconds in the future, according to the startup.

The huge time, environmental and financial damage caused by accidents and testing methods today offers great latitude for more efficient solutions. The price of the startup's device is in the low thousands, and a test cartridge costs less than a fifth of the laboratory tests used today. Furthermore, the device could become even cheaper, as the scope of application is not limited to traffic police only.

Impact

The impact of the device in use can help society and the environment in many aspects. First, the police would save a lot of time as suspects can be convicted faster and no more trips to the hospital or police station are needed. This would also allow police to focus in their duties more on helping people. Furthermore, substance:ID saves more time because the nature of the test gives the ability to screen, but at the same time has the ability to perform a test with the same accuracy of today's laboratory tests. The portability of substance:ID makes it possible to perform the test quite dignified directly at the car window of the suspects in the field without hospitals and bathroom facilities. Thus, the device also saves long car trips for the police and other involved persons. The low cost of resources for sample collection and testing makes the test very affordable for the police. In addition, less plastic and gasoline is used for sample collection, which benefits the environment. Sanitation is also prevented by preventing cross-contamination and there is no hazardous waste that must be disposed of in a special manner. In addition, the test is not fraudulent, as is often the case with urine tests.

For the user experience of the device, we put a lot of emphasis on a good usability during the sample collection to make it as least intrusive as possible for police and suspects. Besides, the way of sample collection is much more dignified than blood or urine collection. In addition, the concept could improve police work by taking difficult decisions away by using technology. This not only increases society's trust in the police, but also eliminates racist and other prejudicial decisions for both sides. substance:ID makes it possible to conduct more tests and therefore ensure greater safety and fewer deaths on the streets.