EVENTS | VIEW CALENDAR
The Internet of medical things
In recent months there has been much buzz about mobile health, relating to smart phones and Big Data statistics. Likewise, there are expectations for self-reporting data for drug trials. While the concept is intriguing, I’m not convinced that such data will be of sufficient quality to make a difference. There are very few transducers that can provide reliable chemistry data, even in the laboratory setting. It may be more realistic to first improve the hospital environment for data acquisition and analysis and have that trickle down to consumers in the way military spending made the Internet and smart phones possible.
It is possible to integrate various modules sourcing local data for critical/intensive care. The integration can be physical or virtual in a variety of combinations near patient. A challenge is to bring together engineering talents that don’t normally work together with the ultimate user base of nurses and clinicians. Such a consortium is not without precedent when other challenges were defined. Could that be possible here? Some elements are beginning to come together, but make no mistake, this would take a decade.
We are challenged by prospects for more evidence to support medical decisions, while being unable to employ the new tools efficiently at the level of an individual patient. How will we synchronously integrate physical and chemical data to optimize therapeutic interventions and monitor them? How will we avoid providing caregivers with data that may be interesting, but not actionable?
I see analogies with notions of the “Internet of things” as played out in the popular press, as well as the vision of wearable medical technologies coupled to smartphones. Those are interesting ideas, but are very limited by the poor quality of the accessible numbers. There are no sensors that can do much more than estimate oxygen saturation or glucose. In critical care and intensive care, the possibilities are far better in terms of both the quality and variety of the data available to assist with acute decision-making. When the term “smart” is used with a device, there is both imbedded processing and connectivity that transmits data, while also maintaining measures of performance and preventive maintenance.
Seven Smart Tools:
All of the above have been developed separately, demonstrated at the proof-of-concept level or integrated into practice. None of them speak to the others. Time-tracked information will move step-by-step from only physics to physics and chemistry.
Virtual integration with wireless will come most easily. Physical integration is more a matter of electromechanical and human factors designed to reduce duplication of displays, barcode readers, power supplies, batteries and transportable carts. These utilities add weight, clutter and take space. Central informatics will simplify and integrate data management in forms that caregivers can use.
Integrating the decision making from multiple data streams presents opportunity for improved care with reduced labor. The bottom line is to have chemical diagnostics coordinated with physical diagnostics, enabling better phenotypic pharmacology. This is about information—acquiring it efficiently to make better decisions, enabling more focus on the patient and less on costly labor for data acquisition. Trust will take much longer.
Analogy: Bringing a phone to a camera and a camera to a phone, completely bypassing developing film to send a picture to your sister. What once took two weeks now can be accomplished in less than a minute globally. Camera performance, data storage, transmission speed all improved a thousand-fold or more in 15 years. Now music, television and books are in the mix with better interoperability every year. The same is possible with medical informatics.
Intensive care rooms are crowded with noisy incompatible devices tangled in cables and tubes. We can do much better. Consider that we had no means of regularly assessing electrolytes until awkward flame photometers spread in the 1950s. Blood gas analyzers of the modern variety didn’t exist prior to 1980 and became substantially improved in the last five years. They now have been reduced in size and increased in capability tenfold since the ‘80s. “Gases” are now only one of their capabilities. These benchtop instruments can be smaller yet. Costs can be reduced substantially with volume. A feature of this future vision is that the needed measurement science is already accomplished, although data management remains very awkward. The next step is integrated human factors design both for hardware and software, focusing on safety, quality data and visualization.
To deliver on the promise of precision (a.k.a. personal) medicine, we must make more measurements—not only the static measurements of genomics, but the dynamic measurements of the phenotypic response to disease and drugs, often in multiple combinations. Genomics without phenotype is like the sound of one hand clapping. The patients are waiting as individuals, not as averages. Physicians are inundated with information and inhibited by time. Several decades ago, a routine physical exam did not require a blood sample or much beyond a stethoscope, a blood pressure cuff, a finger and a little light to peek into things. Today we look at chemistry as well, every year or two. The exotic in 1990 is now a routine parameter to watch. In critical care, time matters much more. Minutes count. We are closer to using those minutes to make better decisions. Of course all of this would enhance drug trials, supplying more information one patient at a time. I’ve been prescribed drugs for seven decades. Not once has the circulating drug concentration been measured. That can’t be the future. You don’t add peppers to chili without tasting it.
Readers may be interested to know that I recently founded No-Links, the social network for antisocial people. No one joins and no one is distracted from serious work. Our members do not want to hear about birthdays or work anniversaries. Our members want time to think. Enjoy your summer.