I received the following email, entitled “A research lead (potentially bigger than the opioid epidemic,” from someone who wishes to remain anonymous:
My research lead is related to the use of psychotropic medications in Alzheimer’s patients. I should note that strong cautions have already been issued with respect to the use of these medications in the elderly (e.g. https://www.mind.uci.edu/alzheimers-disease/articles-of-interest/medications-to-avoid-for-patients/). As a practical matter, however, at present agitated and aggressive behaviors are considered a common symptom of advanced Alzheimer’s and therefore the use of these drugs as chemical restraints is common even by the most conservative physicians. Furthermore, as the medical profession is very diverse many physicians have not updated their practice and are far from cautious in prescribing these medications to the elderly. Indeed, chemical restraints are the norm in nursing home practice, so it is my belief that psychotropic medications are ubiquitous in nursing homes.
A family member is an active medical professional (a physician assistant who works as a front-line health care provider in a family practice office in rural America), so I have some insight into how a conservative medical practitioner will behave. I have observed two things: first, when a symptom is in the list of potential symptoms for a disease that has been diagnosed in the patient, there is a strong presumption that the symptom is caused by the disease. Second, when a symptom is a side effect of a medication (particularly one prescribed by another doctor), there is a presumption that the need for the medication outweighs the side effect.
A combination of close familiarity with my mother’s symptoms/behavior on and off a variety of drugs and my knowledge of the ubiquitous bank gaming of regulatory controls in financial markets leads me to wonder whether the drug companies aren’t playing the same kind of game. In particular the addition of “behavioral and psychological symptoms of dementia” (BPSD) to the criteria for the diagnosis of late-stage Alzheimer’s is as far as I can tell of very recent vintage. (I believe they were introduced with the DSM-5 in 2013.) BPSD are symptoms treated by psychotropic medications. These same medications are also commonly used to treat mild sleep and anxiety disorders in the general population.
The problem with all the psychotropic medications is that they are used to treat the same behaviors that they can also cause as side effects, including irritability, anxiety, “disinhibition” which maps into a willingness to hit out at or behave abusively to others, aggressiveness, and self-harm. In sufficiently high doses, however, the patient is heavily sedated and they are very effective chemical restraints.
My suspicion is that the introduction of BPSD into the definition of common symptoms of Alzheimer’s has developed as a result of the ubiquitous use of psychotropic medications in this population. That is, as far as I can tell the studies that have found BPSD to be common in Alzheimer’s are population-based studies that did not control for the use of medications that have as side effects BPS behaviors. Successfully bringing BPSD into the clinical definition of Alzheimer’s is hugely profitable for the drug companies that now have physician’s biases – to attribute symptoms to the disease that has already been diagnosed, rather than to the drug that may cause it as a side effect – working on their side.
In short, what I would really like to see is a careful statistician’s review of the studies that find that BPSD are common symptoms of Alzheimer’s and an analysis of whether sufficient controls for the use of medicines that have as side effects the same symptoms have been implemented. (With psychotropic medications, the length of use is an important factor because they build up in the system. Thus long-term use has different effects from short-term use. Long-term prescriptions are not a good proxy for long-term use, since refill/renewal of shorter term prescriptions is common practice.)
I don’t know anything about this, but I ran it by an expert on eldercare who said that, yes, this is a big deal. So I’m posting this in the hope that someone will look into it more carefully. As you can see from the discussion above, there are statistical entry points to this one, and it touches on some interactions between causal inference and decision analysis.