Owen Yang
When it comes to research into the cause of a disease, clinicians tend to fall into a mind trap as to how to classify a disease.
Most of the time, the classification system of a disease is tied to how it should be treated. Classification for cancer and immune disease tend to be very complicated and with molecular signatures, making an illusion that there is something to do with the cause of the disease. But the main purpose of these classification is to assess how the disease is treated.
Lung cancer with certain pathological feature is more likely to respond to certain type of chemotherapy. Blood cancer with certain mutation can be targeted by medication specifically designed for them.
Occasionally there is coincidence between the cause of a disease and the classification of a disease. This is a natural coincidence, because some treatment of the disease is related to treating the cause itself. Type 1 diabetes is caused by immune destruction of islet beta cells, whilst type 2 diabetes starts with body resistance to insulin. But the classification itself was decided probably mainly from the perspective that type 2 diabetes can be treated with medications such as metformin, whilst type 1 diabetes has to be treated with insulin.
By contrast, many treatments are not designed to treat the root cause, but more to treat the consequence of an immediate cause. In heart attack we classify them by the extent of coronary artery occlusion, which does not really reflect the root causes such as smoking or a broader involvement of atherosclerosis.
Do not become the roadblock of prevention studies
It is very common that a clinician is very sensitive to the latest classification of a disease, and decide all studies without using the most up-to-date classification is useless. This narrow-mindedness can be a roadblock to many aetiology and preventative studies.
It is relatively easy for a treatment study to catch up with the latest classification of diseases in short-term scenarios, which is the main focus of most clinicians. We start following patients based on the latest classification which dictates treatment, and follow for a period of time to see whether the treatment has made a difference.
Even for a treatment study, it would be difficult to catch up with the latest classification if what we care is a long-term outcome, for example life-long cardiovascular risk among those who have hypertension in their 40s. By the time most cardiovascular events occur, at least in their 60s or 70s, the original classification is already 20 to 30 years late. It is very easy to forget that this is in fact the most up-to-date study possible, considering the follow up period needed. Sadly, I see so many clinicians just dismiss them as poor studies because the original classification seems out of date.
A good, unbiased aetiology study tend to be prospective, which means the knowledge of aetiology factor comes before the disease. This is either a clinical trial to prevent the disease or a cohort study to explore the risk factor of a disease. These studies are also likely to require a longer follow up, and therefore will have ‘out-of-date’ disease classification by the time when the outcome events are observed. If the observation or the treatment started 20 years ago, the events accumulates over the period of the past 20 years, and it is not usually possible to track back to the earlier cases to obtain information using the new disease classification.
In all of these scenarios, again, we need to be clear that the new disease classification does not necessarily reflect the aetiology of the disease. We need to be more open-minded with long-term studies, and do not be stubborn with the classification systems from the latest treatment guidelines.
