Definitions of chronic disease need to be more patient centred

Abstract

Early, small abnormal changes should be defined as disease only if there are effective interventions that can make an important difference to patients, argue Zu-Yao Yang and Jin-Ling Tang

The proposal for new diagnostic criteria for hypertension in November 2022 caused heated debate in China. The recommendations, from a guideline group of leading Chinese clinicians,¹ followed changes to the American hypertension guideline in 2017 to reduce the cut-off blood pressure for diagnosing hypertension from 140/90 mm Hg to 130/80 mm Hg.² Adoption of this new diagnostic cut-off value for the Chinese population would have created around 350 million new patients with hypertension and increase the total number with the diagnosis by 130%.³ The new recommendation quickly caused intense nationwide debates, and the National Health Commission finally said no to it.⁴

The controversy over the diagnostic criteria of hypertension, which has also occurred for other non-communicable diseases, raises a fundamental question for medicine: what should count as a chronic disease? Historically, disease was known as sickness or illness, which is defined and diagnosed according to symptoms and signs and is in essence the patient’s current experience of their health. In some sense, it is patients who decide what a disease is. This has all been changed since medicine started coupling with modern science and technology.

Expanded definitions

Over the past few decades, physiological, biochemical, histological, and imaging technologies have greatly extended our ability to “see” within the body. This led to the discovery that disease begins with small changes at the cellular, molecular, biochemical, or physiological level and that symptoms and signs become apparent only later, when these changes are sufficiently severe. As a result, the definition of disease has been considerably expanded. Today, more and more diseases are defined at a cellular, molecular, biochemical, physiological, and structural level and can exist independent of patients’ current experience. This is particularly true for chronic diseases. For example, diabetes mellitus is defined biochemically according to the blood sugar concentration, cancer histologically according to the morphological structure of cells, and hypertension physiologically according to blood pressure.

Traditionally, these early abnormalities are considered risk factors for future diseases and real symptoms—for example, hypertension is a risk factor for cardiovascular events. A large, long term follow-up study in China estimated the lifetime risk of cardiovascular disease in patients with hypertension (blood pressure ≥140/90 mm Hg) was 26%, 12% higher than that in normotensive people,⁵ implying that 86% of patients with hypertension will not develop a cardiovascular event in their lifetime. Similarly, the small mortality:incidence ratio of thyroid cancer and prostate cancer detected in screening programmes implies that most patients who have early stage cancer detected will not eventually die from the cancer.^{6 7} Thus, early stage cancer could also be viewed as a risk factor for symptomatic disease and death.

What is an appropriate threshold for diagnosis?

The probabilistic relation between a risk factor and the risk of future disease raises a second important question: how far from normal values should a risk factor or a probability of disease be considered and treated as a real disease? For hypertension, the core issue is the cut-off blood pressure for diagnosis, but there are similar decisions for other chronic diseases. Diagnosis of conditions such as hyperlipidaemia, type 2 diabetes, renal failure, coronary artery disease, and aortic aneurysm, for example, all rely on an “arbitrary” cut-off value. Cancer is not exempt from this problem, although it is less conspicuous. The diagnostic cut-off value in the size of cancer is seldom debated but is being pushed down by the fast progress in imaging and histological technologies, and any “visible” cancer is clinically diagnosed as cancer. What cut-off value should be used?

Some may argue that any increased risk of future disease should be medically diagnosed and corrected. However, a small change in a diagnostic cut-off value will create massive numbers of new patients, and interventions in these patients are often of small benefit or ineffective. For example, changes to the cut-off values for diagnosing hypertension, hyperlipidaemia, and diabetes in China around the beginning of the century doubled the total number of patients and created 359 million new patients between 2002 and 2009.⁸ If all these new patients had been treated with drugs of average price in 2010, it would have consumed 56% of the government’s total health expenditure of ¥480bn in that year.⁸ It is simply not affordable. Furthermore, prescribing antihypertensive drugs, for example, in these new cases can prevent only two out of 100 patients from developing a major cardiovascular event within five years, whereas 98 patients have to pay the cost, bear the risk of harms, but have no benefits.⁹

To make the situation worse, the spectrum of chronic disease can be, and in fact has already been, further expanded by applying different (eg, diabetes) or more sensitive methods (eg, cancer) or by creating pre-disease conditions (eg, diabetes). However, randomised controlled trials show that cancer screening programmes, treatment of early hypertension, comprehensive cardiovascular screening, and general health checks often have small or no benefits,^{10 11 12 13 14 15} suggesting that many efforts targeted at early, small abnormalities or risk factors are futile. This raises questions about the value of unchecked expansion of technology based diagnoses of diseases in terms of patient care.

Effect on patients

Another important consequence of defining an abnormal test or examination result as disease is that patients will lose, and in fact have often already lost, their say in whether they are sick, whether they need a treatment, and whether the treatment is effective. This loss is of no small matter. As a result, their health and rights cannot be adequately protected partly because of the competing interests of other stakeholders in the business of healthcare. Furthermore, interventions cannot make patients truly satisfied if they cannot sense what is wrong with their body and the benefit of interventions on it. Lastly, making small abnormal changes a disease may also increase overdiagnoses, cause health inequity,^{16 17} and compromise care quality and safety. Even the best evidence based policies will fail if clinicians do not have sufficient time to follow them, not to say the unsubstantiated recommendations, as clinicians have already been overloaded.^{18 19}

It is now evident that disease is not a black and white entity and ever earlier or smaller forms always exist. Undoubtedly, health technology will continue to advance so that in the future we can see even smaller changes in the body. Early detection and treatment are certainly beneficial for many diseases. The question is whether we should make a small abnormality a disease for which early diagnosis has little or no benefit and how much further we should go in this direction. Given the widely existent overdiagnosis and ever increasing health expenditure, we argue, for the purpose of patient care that is medicine’s fundamental goal, we should rethink the way we define chronic diseases and make early, small abnormal changes a disease only if there are effective interventions that can make a clinical difference that is worthwhile to patients. To this end, the interventions’ absolute benefit and cost effectiveness in the newly labelled patients should be highly emphasised and carefully evaluated before changing a diagnostic cut-off point or modifying a disease definition.