What struck me most when I first took my CPAP machine home was the large degree to which my sleep became “medicalized.” That is, it became understood in terms of the assumptions, values, and desires of medical professionals, not my own. The “MyAir” app associated my machine only tells me how long I’ve kept my mask on, whether it has leaked, and how many apnea episodes I have had every hour. Sleep quality is not measured or represented anywhere. Ironically, I get pretty good numbers when I lie awake for hours on end, wishing that the panicked feeling of suffocation would subside just enough for me to fall asleep. Even nights that I do sleep, I awake four to five times per night, never reaching the deepest level of sleep. My slumber can be nearly as unhealthy as before despite the “good” numbers sent to my doctor by the machine.
Most telling is the way that my usage of the machine is talked about. The primary concern of my doctor and insurance company is “compliance” – so much so that a respiratory technician was made to show me a scary four figure number that I would be responsible for paying if I do not wear my mask the required four hours per night. Unfortunately, there is no equally threatening monetary incentive for my doctor to ensure that I am actually asleep and sleeping well for the night. I can be totally compliant while being completely miserable.
The tendency be overly enthralled with seemingly objective but unrepresentative measures and take too little care in understanding how people interact with their technologies is tragically common. Robert Pool calls this the “machine centered philosophy of engineering.” Under the spell of this philosophy, whatever machine technologists come up with is framed as ideal. The only imaginable problem then becomes the failure of people to adapt themselves to the machine, not that designers failed to give empathic consideration of what people can reasonably do. A classic example of this machine-centered view was the control room in nuclear plants like Three Mile Island. Operators were blamed for mistakes made in the run up to a near meltdown at the plant, but one of the underlying causes was that the array of dial and gauges in the room were not set up to be comprehensible to operators but easier for the designers and builders to lay out.
Once one notices that CPAP machines are a machine-centered approach to treating sleep apnea, their status as the “gold standard” treatment begins to appear much less certain. Indeed, nearly 50% percent of diagnosed apnea sufferers never adapt to their machines and stop using them. “Gold standard” status perhaps makes sense in the simplified environment of the clinical study but not in real life. Yet alternative treatments to the CPAP machine receive little attention from sleep doctors, perhaps because they do not reliably get patients’ AHI (average incidents per hour) down to the sought after 5 or less. However, consider that a “compliant” CPAP patient only need wear their mask 4 hours a night. Their actual nightly AHI may actually be little different than people using these alternative treatments. Someone managing to wear a CPAP mask 5 hours per night with an AHI of 4, but going back to an AHI of 25 for the remaining 3 hours, has a nightly AHI of almost 12—which would classify them as suffering from moderate sleep apnea and is no different than what alternative treatments accomplish. However, under the spell of machine-centered thinking, this would be seen solely the patient’s fault for being insufficiently diligent rather than a failure of the CPAP approach more broadly.
Looking at other cases of machine-centered failure, however, provides lessons regarding how sleep apnea treatment could be more person-centered. For instance, autopilot leads to new kinds of plane crashes because trying to completely delegate the process of flying to an algorithm deskills pilots, leading them to make elementary mistakes when the autopilot shuts off in unusual circumstances. The alternative is to “informate,” which involves using automation technologies to help pilots become better at their jobs: help them maintain attention, periodically test their skills, provide feedback on performance, etc. Informating takes the cognitive aspects of pilots and the human-machine interface as part of the design, rather than expect users to be superhuman. The challenge for sleep apnea researchers is learn to think out of the machine-centered box. Rather than simply delegate the holding open of patients’ throats to a machine, how could patients be better empowered to do that themselves?
This alternative approach is mostly undone science. While there are a few studies looking into how physical therapy exercises, playing the didgeridoo, and a cannabinoid could reduce the frequency of apnea incidents by up to 50 percent, there are few follow-up studies, much less any research attempting these treatment options in combination. Little to no energy has been spent by my doctor to try to diagnose exactly why my airway collapses. Given that breathing is a semi-voluntary act, what reason is there to believe that I could not retrain my respiratory system to have a more suitable level of musculature?
Insofar as today’s paradigm of compliance to CPAP reigns, apnea sufferers like myself are left in the dark, trying to piece together sparse information on the Internet in order to design our own alternative and complementary treatment pathways. This need not be the case. I could use the help of a trained medical professional, rather than go it alone. Absent a less machine-centered, more person-centered paradigm of apnea treatment, I do not have any other options.