Friday, August 16, 2013

How I Became Sick



by Jim Wood

I am a patient.  I used to be a regular human being, but that was a long time ago.  Back in the 1980s, as a result of a routine office visit and blood work, I was found to have slightly elevated blood pressure and blood sugar.  Nothing to worry about, merely in the upper range of normal, but probably worth keeping an eye on in the future.  Then, a few years later, I was diagnosed with two conditions called prediabetes and prehypertension, neither of which I had ever heard of before; so I was put on a lifetime’s worth of medication.  A few years after that, I was diagnosed with full-blown type 2 diabetes and essential (if mild) hypertension.  So my meds were augmented and the dosages upped.  By now I was a definite patient, requiring frequent monitoring and more aggressive treatment.

Sphygmomanometer; Wikimedia Commons
 But here’s the curious thing: my numbers (blood pressure and blood sugar) had not become worse.  In fact, they had improved a bit, which presumably had something to do with the meds.  Moreover, I had not developed any symptoms.  I did not have peripheral neuropathy, I had never experienced a coma, my feet and kidneys were just dandy, I had never had a stroke or any sign of heart disease.  In fact, the only major health problem I had turned out to be the result of one of the medications I was on, which actually caused serious hypotension, at least in me.  The scrip was withdrawn and now I’m fine.  I remain a patient, but a patient without a disease.

If there was no deterioration in my physical condition, how did I go from having minor not-quite-symptoms to having two life-threatening illnesses?  The answer is simple: the magic trick was performed by a series of Expert Medical Panels – well-known physicians and epidemiologists who were aided by representatives from the pharmaceutical industry.  These worthies decided on several occasions to lower the threshold values used to diagnose hypertension and type 2 diabetes.  Were these changes inspired by new evidence that even mild elevations in blood pressure or blood sugar caused serious disease and a greater risk of death?  No.  They were inspired by a longstanding commitment to “preventive medicine” – the well-intentioned idea that we can intervene in people’s lives before they develop overt disease (i.e. before they display real symptoms) and delay or prevent those diseases from ever happening.  I remember that, when I was a callow youth, the idea of preventive medicine was fairly new and the public reaction to it was overwhelmingly positive.  This was something that just plain made sense.  And, in fact, some preventive measures worked really well: don’t smoke, get some exercise, watch your diet, and get a bunch of vaccinations.  But other preventive measures had effects that were more equivocal, improving people’s health only at the margin and mostly turning a whole bunch of perfectly normal people into patients.

Take hypertension as an example.  The thresholds (both systolic and diastolic) for diagnosing hypertension were reduced in 1997 as a result of the Sixth Report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure, which appeared in Archives of Internal Medicine.  It has been estimated that this change increased the number of people in the U.S. diagnosed each year with the disease by about 13 million.  (This figure comes from the excellent book by Dr. Gilbert Welch et al., Overdiagnosed, Beacon 2011.)  That’s a lot of new customers, er, patients.  Did it help any of those new patients?  Well, some no doubt, but while the number of people being treated for hypertension skyrocketed, the mortality rate from causes of death associated with hypertension changed pretty much not at all.  What the change in diagnostic criteria did was greatly increase the number of normal people who have to be treated without benefit in order to increase the chance that you help just one person who really needs it (given that you can’t empirically distinguish the normals from the ones needing help).  In the past, when the diagnostic criteria were set high, nearly everyone who was diagnosed and treated probably benefitted to some degree (if they didn’t die from some other cause or indeed from hypertension meds first).  Using the current standards, however, a large fraction, perhaps a majority, of people being diagnosed and treated probably have no underlying disease and are unlikely to develop it in the future even without treatment.  Am I one of those people?  I have no way of knowing, short of going off all my medications and waiting to see if I die in short order.  Do I want to do that?  Not really (being both a moral and a physical coward).  Do I suspect that I may be receiving treatment that either is unnecessary or at best will have only trivial benefits?  Well, yes.

The basic underlying epistemological questions here are (1) can you ever make valid predictions about future disease risks; (2) how can you turn continuous biological variation in, say, blood pressure into a simple, dichotomous (yes/no) criterion for clinical decision-making; and (3) how can you capture inter-individual physiological variation with a single set of numbers?  The answers are: (1) no, (2) you can’t – at least not without injecting a big dose of arbitrariness and error into the process – and (3) you can’t, period.  In assessing the value of some diagnostic criterion like a blood pressure cut-off, we speak of sensitivity and specificity.  Sensitivity is the probability that your criterion correctly identifies everyone with the condition of interest, including people who are not yet displaying clear-cut pathologies (one minus the probability of a false negative).  Specificity is the probability of correctly diagnosing only the people with the condition of interest and not those with some other condition or no condition at all (one minus the probability of a false positive).  Both sensitivity and specificity are virtues, and you’d like to design your diagnostic methods so as to maximize both of them simultaneously.  But, generally speaking, you can’t because they compete with each other.  To guarantee that you catch all criminals, jail everyone (a rather Javert-like solution, I admit).  By the same token, to guarantee that you treat all cases or potential cases of a disease, diagnose and treat everyone.  Sure, most people won’t actually have the specific disease you’re trying to diagnose (the specificity will be as low as it could ever get), but that’s the price you pay for perfect sensitivity.  Although what I’ve just described is, of course, a gross exaggeration of current clinical practice, there’s no doubt that things have been moving in that general direction.  The whole goal of preventive care is to maximize sensitivity (in this case the ability to detect a disease even when there’s no overt sign of it) at the expense of specificity in order to “improve” prediction.  But that’s a mug’s game because a prediction with a high false positive rate is not a good prediction.

Is all this the result of an evil conspiracy of greedy drug manufacturers and health-care providers?  I honestly don’t know (even if I sometimes have my suspicions).  I think it’s entirely possible that it’s simply the result of misdirected good intentions and a fervent if wrongheaded faith in preventive medicine.  I genuinely like and respect my primary care physician, who assures me that I’m going to die if I go off my meds.  (I’m pretty sure I’m going to die anyway, but that’s another story.)  I truly think she has my best interests at heart.  But do I think she’s bought into a mind-set that seriously needs to be rethought?  Absolutely.  Do I enjoy her treating me like a patient in the absence of any discernible disease?  Not a bit.  Do I have the courage to tell her so?  See the comment about moral cowardice above.

There are many reasons that unnecessary patienthood is annoying and costly (to me, never mind the general public) and maybe sometimes even dangerous.  As already noted, I’ve had a non-trivial medical emergency (which landed me in the hospital) caused by one of my medications.  And I have to do “co-pays” for all my exams, treatments, and prescriptions – even though, by U.S. standards, I’m ridiculously well-insured.  And God knows I’ve experienced my fair share of annoyance scheduling frequent physical exams and blood tests and taking the time to refill my many prescriptions.  But my real objection to being a patient runs deeper than any of that – being treated like a patient is infantilizing.  I bristle at that subtle hint of moral disapproval when my doctor tells me my numbers have “worsened” (often within the range of measurement error).  I bristle even more at that slight tone of “What a good boy!” that accompanies any “improvement” in my numbers.  I was treated like a child for the first several years of my life, and I got thoroughly fed up with it.  Why is it happening to me again?

There are those in our society who need health care and are not getting it.  I am not one of them.  If we want to improve the health of the U.S. population in general, the single best thing we can do is eliminate disparities in access to health care.  My being treated is not furthering that goal at all, although it is increasing the costs of health care for everyone.  Let’s face it: the lovely idea of preventive medicine comes in a poisoned chalice.

9 comments:

Ken Weiss said...

It's ridiculous how we try to handle this, and after my most recent visit to the doctor, and his frightening diagnosis, I have changed my view.

Doc told me I had 'premortis sapiens'. "What the ... is that?" I asked. It's the multitudinous progress towards inevitable death of a wise (hence sapiens) organism.

The solution is this: For every known cause of disease and death, there is a risk. For most of them there are at least some treatments. So, with the pharmaceutical industry's approval, my doc has now put me on a coctail of meds for all of the causes, with dose proportional to estimated risk.

That approach should settle all these issues for once and all.

Ken Weiss said...

And, hell, then there was more! I went to my doctor to be poked and prodded because otherwise I wouldn't be able to get Penn State's affordable health care insurance. I had my phenome typed, I had my genome typed, I had my PET, CAT, and MRI reports, I had my nutrition evaluated, and I don't even smoke (any longer).

Then, damn! I was leaving the doctor's office, feeling safe and satisfied as I wended my way to Rite-Aid with my fat sheath of prescription orders to treat my premortis, when I saw this story:
http://www.bbc.co.uk/news/technology-23709009

Now, it turns out that using _Facebook_ is bad for your health! Double damn!

What do we need to do next??

Anonymous said...

Jim: I resemble your remarks, thank you.

The BP goalpost-shifting was briefly a topic of controversy at the turn of the century. For views harmonising with yours, see:

Port et al. Systolic blood pressure and mortality. Lancet (2000) 355:175-180.

www.ncbi.nlm.nih.gov/pubmed/10675116?dopt=Abstract

Port et al. There is a non-linear relationship between mortality and blood pressure. Eur Heart J, (2000) 21:1635-1638.

www.ncbi.nlm.nih.gov/pubmed/11032685?dopt=Abstract

Ken Weiss said...

There is also the truly cogent but overlooked point that if BP really is predictive and you lower it, you very directly risk surviving a quick cardiovascular demise and increasing risk of a slow and nastier one from cancer and other degenerative disorders. Public health policy nor medicine have really dealt with such issues; they are too inconvenient to think about.

Jim Wood said...

Congratulations, Ken. You've achieved immortality. Oh, wait, some day you'll get struck by lightening or run over by a Mack truck. Best never to leave your house. Wait, then you'll be exposed to asbestos or radon and...

Jim Wood said...

Your comment raises the important issue of competing causes of death: if one thing don't getcha, another thing will. Competing causes (and especially competing and interacting causes) are, as you suggest, difficult to think about, let alone analyze using real data. But basically they mean that if you were to completely eliminate a disease that currently causes, say, 40 percent of all deaths in a population, that would achieve nothing even close to a 40 percent reduction in mortality. People would just die of something else.

Jim Wood said...

Thanks, Anonymous. I look forward to reading them.

Ken Weiss said...

Yes, and then there's the Catch 23 that being immortal will be very boring perhaps and depressing. So that means some more anticipatory meds.

Anonymous said...

Untreated hypertension and diabetes were associated with Alzheimer's disease. Based on that I'd say people with increased genetic risk for Alzheimer's Disease should be treated more aggressively for even slight elevation in blood pressure and blood sugar. Those without such risk - less so.