An essay on screening diagnostic processes and their effect on individuals, and implications for overdiagnosis. By Dr Owen P Dempsey MSc MRCGP
“(medical) man’s passion for ignorance dehumanises man in the name of a Cartesian humanism”
(Owen Dempsey 2015)
“In my view, this activity …….. reveals both a libidinal dynamism that has hitherto remained problematic and an ontological structure of the human world that fits in with my reflections on paranoiac knowledge.”
(Jacques Lacan, talking about child development, in The Mirror Stage as Formative of the ‘I’ Function as Revealed in Psychoanalytic Experience, 1949, Ecrits p93)
This essay explores the effects of cut-offs (chosen for screening and preventive diagnostic tests, called ‘diachronic‘ tests here) on overdiagnosis. The need to do such tests is driven by a need for medical knowledge, they become an ideological tool perpetuating overdiagnosis, require a disavowal of ‘not knowing’ and objectify the individual transforming him or her into a subject-of-medicine.
It is acknowledged by those working within EBM and specifically in relation to diagnostic testing that language, the names given to the different aspects of testing and their properties is a substantial problem, but what is the problem exactly?
First of all a strange counter-intuitive but real example.
Following a discussion on the Evidence Based Medicine mail base JISC list, about the meaning and effects of sensitivity, specificity, prevalence and predictive values, Teresa came up with the following comment:
“”Oh dear! such a complicated topic. Going back to the RASTER study on MammaPrint, can you speak to the following clinician rationales and comment?
“I’m going to order a MammaPrint test for my patient, because if it shows Low Risk, we can be 97% sure my patient won’t develop metastasis and I can confidently recommend no chemotherapy.”
“I don’t think MammaPrint would be a good test for me to order for my patient, because even if she is likely to develop metastasis, the MammaPrint test would have a 30% chance of missing that and showing a false-negative Low Risk result instead—potentially misleading.”” (cite)
The MammaPrint test …… a way of determining the likelihood of developing metastases in the future.
My (brief) response was:
“The shift of the low-high risk cut off point to favour low risk is odd. It seems to favour (as in promote) specificity rather than sensitivity. This is odd because usually cut-offs have tended to promote medicalisation, as in most screening tests. But to decide on any particular cut off might be presupposing that there is a piece of knowledge, here called ‘risk’, that the test can fully characterise or obtain. Maybe it can’t. But maybe the testers don’t acknowledge that some things aren’t necessarily measurable and perhaps never will be. It would be interesting to know the confidence intervals for the findings of the study, it sounds as though chance may play a role.“
Interestingly, perhaps, there was no further response from the list, so far. (There has now .. 24/2/2015 and I will return to this) But this has intrigued me and made me wonder if it had something to offer. What could I turn up if I attempted to ‘go to the root of things’?
Given a population with a significant (possibly) high prevalence of the target condition (in this case so called ‘high risk’ of metastases in the future), then, a high specificity at the expense of sensitivity can be misleading. Whereas, on the one hand specificity would protect the already low risk low prevalence population from false negatives, on the other hand, in the face of an important high prevalence diagnosis, (high risk of metastases here), then the low sensitivity will miss, in this example, 30% of those that might benefit from chemotherapy. This suggests that the cut off between low and high risk , should depend upon what is known of the prevalence of the risks, low and high, in the test population.
The more the prevalence of metastases, then the more sensitive the test should be, and the less specific, the rationale being that the more the prevalence of metastases then the more important it is not to miss them, simply because there are increasing numbers of them and the more forgivable it is to be less specific and to over diagnose those that would not benefit from chemotherapy as there are proportionately less of them. This is complicated. The example given suggests that the cut off inordinately favours specificity and I wonder how the authors justified their choice of cut off.
Of course for an individual the risk might be an unknown, if only because studies have not been done to measure risks of having or not having metastases without treatments, i.e. without a non-treatment arm. In such a case there can be no rationale for the cut off. In the example given it looks as if the population levels of risks have been estimated (but how accurately?) and a decision made for a cut off between low and high risk but the cut off has for some reason been inordinately skewed towards specificity. This could be because the tests cut offs were worked out on a low prevalence population but then inappropriately applied to a higher prevalence population. Does this sound possible?
I argue here that, amongst many difficulties, the root origin of the ongoing unstoppable epidemic of overdiagnosis is Modern Medicine’s desire to name that which it doesn’t understand, and to assume it can know that which it cannot know. Modern Medicine, for the ‘diachronic’ test, manipulates the cut off points for diagnostic tests to favour sensitivity, so that the borderline must always be assumed to be pathological. Note here however that paradoxically the cut-off for the MammaPrint test favoured specificity, which I cannot explain. It is used as the EBM cognoscenti would say as a ‘Rule Out’ test, fairly reliable at confirming you don’t have a condition, not so reliable if you do have the condition, best used for screening type tests where falsely ‘ruling in’ has serious consequences, that outweigh falsely ‘ruling out’. Here you might expect them to what to have a more sensitive ‘rule in’ type of cut off. Why would they do this? It had the effect of reassuring people they did not run a risk of metastases when they perhaps did, it would also reduce the number receiving prophylactic chemotherapy. However, putting this strangeness to one side for the moment I will persevere with screening diachronic testing.
Medical Technology is part of a diachronic process when it tries to predict the future, or the risk of future harm. It is diachronic because it is providing information that is applicable to a time period as opposed to synchronic, which applies to the state of the tissue today. This applies to most screening and preventive investigations, including private health care screening and cancer screening programmes. This is a huge profitable industry. It relies upon a sacred belief that tissues must reveal their futures up to the test, and more testing will provide more certainty. Medicine remains obsessed with the classification of tissue representations and now molecular appearances. The aporia of medicine is disavowed because they are the unanswerable questions posed by the borderline tissue of the breast biopsy: “What does this result mean for me, the individual in front of you?” Answer: “We cannot say, (because so much cancer is determined by random chance) which is why you must have the mastectomy, we cannot take the chance of not removing this indeterminable aporia.”
When a test is used to determine risk it functions to performatively name an individual and transforms him/her into a feminised subject-of-medicine with a dis-ease requiring treatment now or, if not, at least further investigation now or later. The test and tester and the tested have two issues to confront in general a) the dependence of the predictive value of the test on the sensitivity and prevalence of the condition and b) the dependence of the sensitivity of the test upon the choice of cut off between low risk (treatment should be avoided) and high risk (treatment can be recommended). The effect of the test upon treatment recommendations is critically dependent upon the choice of cut off. This choice, for conditions where the problem is diachronic, i.e. not synchronic with the timing of the test, has two further issues to resolve: a) the uncertainty of the level of population risk, b) the uncertainty of the individual risk even if the population risk is known, and c) the ability of the test to provide reliable information about the risk in question (e.g. of a person developing metastases in the future, or in a de novo screening situation (on a person with no previous symptomatology or signs of the disease in question) as in breast cancer screening. For many diachronic diagnostic situations there is no gold standard by which the test can be judged.
For example a screening test can lead to a biopsy reported as borderline: this might be called ‘indolent tumour’ appearances or DCIS (though both appear identical under the microscope). The pathologist may only be able say that this tissue has somewhere between a 1 in 4 and 1 in 20 chance of causing life threatening physical harm over the next 10 – 20 yrs, and it might be that the risk of coming to harm from other things is even greater, and that treatments themselves can cause harm (radiotherapy and heart/lung disease) , and there is lack of certainty about whether treatment actually reduces overall the death rate from all causes.
So even if the population level of risk is known, e.g. for 50 yr old women of a certain social class and nationality and with certain lifestyle risk factors there is no gold standard synchronic test that can be used to provide certain information about what will happen in the future. The best that can be done is to follow screened and unscreened populations over many years to see what is the excess level of diagnosis in the screened group above and beyond the number of deaths from the screened condition reduced in the screened group compared to the non screened group. This would represent over diagnosis. The problem with estimating sensitivity and specificity like this is that a) it does not reflect prevalence for other populations and does not reflect an individual’s unique potential for developing the target condition, b) the effect on treatment recommendations/decisions depends upon the cut off between low and high risk, which is in a circular fashion determined by a decision to make a recommendation for treatment even when the level of risk is highly uncertain and is even contested, and c) the screening test itself varies with time.
Another example might be the diagnostic criteria for the diagnosis of Munchausen Syndrome By Proxy (MSBP) also now known as Fabricated and Induced Illness (FII), these criteria were ‘described’ by Roy Meadow to characterise the mothers of women deemed to have deliberately murdered their babies. Several convictions partly based on Meadow’s ‘misleading’ evidence have been overturned, but many Family Court custody decision cannot be challenged because they are held in secret and the participants have to sign secrecy non-disclosure clauses. The diagnosis was in response to the problem of ill children without a diagnosis, and included the conundrum of cot death, or Sudden Infant Death Syndrome (SIDS). Many critics have claimed the ‘test’ for MSBP was over sensitive and not specific enough. It is possible its ‘cut off’ point was determined by an over zealous need to ‘know’ the cause of these children’s illnesses and deaths. Too many innocent families ended up being caught up in the process. I believe there are parallels with over zealous screening programmes in Modern Medicine today.
The contestations about the ‘right ‘cut-off for a diachronic test arise, I believe, when the test is incapable of providing reliable information about the future of a biological system. We know that so much of the future of biological systems is likely to be due to chance, literally due to random mutations as cells’ DNA replicates billions of times in the lifetime of one organism, (always remembering that such random mutations can also reverse harmful changes which is why some cancers may regress). The contestations are also influenced by three factors: a) the intrinsic level of variability confounding attempts to provide ‘scientific’ reliable information, and b) an intrinsic accultured scientific abhorrence of a vacuum of knowledge that demands that tissues yield up their futures to us now – the result of a scientific Cartesian enlightenment in the 17th Century when Man’s thought achieved a rational primacy underwritten by a faith in God’s benevolence, and c) a politically and economically motivated medical doctrine that places a primacy on preventing death over and above maximising quality of life today. b) above reflects man’s passion for ignorance, where here ignorance is the disavowal of simply ‘not-knowing’, the ignorance of our real ignorance. Modern Medicine in the West so far refuses to challenge the ideological strategies enforcing diachronic test cut offs that coerce individuals into harmful treatment programmes, and refuses to acknowledge that some (borderline) tissue appearances are uninterpretable, much like language is uninterpretable in the sense that we never can see inside to the meaning and intention of the speaker. The injunction “Be Normal!” of Modern Diachronic Medicine – objectifies the individual, and enforces a closure, gating of the universal within a normality-pathology opposition that cannot face up to the ontological difference of Heidegger, wherein the representation is different from the object by precisely the representation itself (see ref) . What the tissue gives up to the test, its sensed representation, is precisely what the tissue is not. Diachronic Modern Medicine dehumanises in the name of a Cartesian humanism that is deluded and grandiose.