HAL for Hand Hygiene: The Computers are Watching

by Erica Mitchell | August 30 2017 | 0 Comments

Electronic monitoring HAL-01.jpgIn the sci-fi classic 2001: A Space Odyssey, an omnipresent computer named HAL monitors all activity aboard a spacecraft on a critical mission. The single-mindedness of this artificial intelligence makes HAL helpful and life-saving, but misses the mark on some decisions that require a more human touch. While our year 2017 has not yet brought us interstellar craft equipped with AI, we are living in a time when technology is supporting almost every field, including healthcare. This post will explore how technology is helping us with hand-hygiene compliance, and how, like HAL, there are some clear advantages as well as some disadvantages. (Thankfully, no hand hygiene technology is able to eject non-compliers out the airlock. Yet.)

The gold standard for monitoring hand hygiene is direct observation by a trained (human) healthcare worker. The World Health Organization recognizes the critical role of observation in its"Multimodal Hand Hygiene Improvement Strategy," noting that only direct observation by a healthcare worker can accurately both recognize most opportunities for hand hygiene as well as record the compliance with appropriate hand hygiene practices. So many procedures and patient interactions do not follow a predictable chain of events, they argue, so opportunities for hand hygiene do not always occur at the same places or moments. Only a direct observation can recognize these opportunities.

However, direct observation by a trained healthcare worker is difficult. It's expensive - observers have to be trained, validated by yet another set of observers, and paid for their time. The data collected must be collated, analyzed, and reviewed, requiring additional personnel. In order for that data to be meaningful, there has to be a lot of it, and many facilities have a hard time collecting a sufficiently large sample size to have meaningful outcomes to report. And finally, a point HAL would wholeheartedly agree with, there is the human element: Even with well-crafted, standardized observation practices, each individual observer will have their own interpretation and implementation of those practices, leading to potential problems with the data. And even if all the observations are performed perfectly, another human element can play a significant role: The Hawthorne effect, that is, people perform differently when they know someone is watching (just ask Dave).

As a way around some of these disadvantages, healthcare facilities have turned to technology to assist them in monitoring hand hygiene compliance. One simple technology comes in the form of electronic dispenser counters that simply count how many times a soap or gel dispenser has been accessed (by anyone; no data about who access it is collected). This provides data that can be combined with other information to provide a better picture of hand hygiene compliance. This data does not account for hand hygiene opportunities, however, and any malfunctions can skew data. In addition, installation and maintenance are expensive.

Another form of technological support is electronic compliance monitoring systems. There are three major types of systems based on where they are installed: Door minders, electronic badges worn by healthcare workers, and camera-based systems. Those installed on doors record entry and exit data (and sometimes activity in the room) along with use of soap/gel dispensers. Badges accomplish the same paring of entry/exit and access, but further individualize that data based on the person wearing the badge. Camera-based systems record entry and exit as well, but also record all the room activity which can then be reviewed by an auditor at a later time. All three systems generate lots and lots of data, but some of that data may not be helpful. The positioning of all the sensors is vital for accurate readings, and upkeep of all the sensors is time-consuming and expensive. It is difficult to separate the good data from the compromised data, leading to inaccurate assessments of a department's or individual's hand hygiene compliance. And with job performance on the line, bad data can have catastrophic consequences.

One of the biggest hurdles faced by a system that places technology in the position of monitoring and reporting on humans in any field is trust. Health care workers are no different - they will need to trust that the system being implemented, the one that will report data on their performance, is accurate, fair, and reliable. For health care workers whose job performance includes hand hygiene compliance based on data coming from an electronic monitoring system, it feels a little like getting into a driverless car and heading out onto the highway. Not many people would be willing to do that without some significant trust that the entire system will work, from the car itself to the infrastructure, policy, and training necessary to enable it to operate safely.

Thankfully, computers and artificial intelligence are not making all the decisions. Unlike Dave and his ill-fated crew, we do not have to rely solely on algorithms and programming to make decisions, plan, or improve. However, we do have to find new and creative ways to help us remember to wash our hands if our compliance is going to improve from just around 50%, so these technologies are here to stay. (Just don't ask them to open the pod bay doors!)