What is a diagnosis? Part 1
Last updated: November 23, 2025
In healthcare, a diagnosis is a label assigned to a health condition. The Oxford English Dictionary defines diagnosis as: “the identification of the nature of an illness or other problem by examination of the symptoms,” and the Cambridge Dictionary as: “a judgment about what a particular illness or problem is, made after examining it.” Its Ancient Greek root was “to distinguish,” with two parts: to know (gnosis) and to be apart (dia). This meaning closely relates to a formulation.
Labels, on the other hand, are “a classifying phrase or name applied to a person or thing, especially one that is inaccurate or restrictive.” [OED] or “a word or a phrase that is used to describe the characteristics or qualities of people, activities, or things, often in a way that is unfair”.
The definitions expose the dual nature of a diagnosis; labels can be inaccurate and unfairly applied. Therefore, diagnoses may be vital because of their consequences, but they can also cause lasting harm through other effects. Someone with autism needs the diagnosis because “Being undiagnosed prevents autistic people from accessing vital statutory support.” Presenting to healthcare with a problem without a diagnosis can be distressing for the patient, who then worries about the cause of their symptoms and their future. The patient’s diagnosis often influences the hospital’s income. An insurance company’s payment for healthcare depends on the diagnosis; some diagnoses may render you uninsurable.
I will explore issues surrounding diagnosis. For example, are all the uses based on a single unifying phenomenon or construct, and are all the uses valid and appropriate? I start from the position that no single construct could be valid for all the uses made of a diagnosis. I have continued in What is a diagnosis? Part 2.
Table of Contents
Introduction
Surprisingly, the literature on diagnosis and its meaning is sparse. As the philosophers Julian Reiss and Rachel Ankeny wrote about diagnosis in their text on Philosophy of Medicine, “Although a critical part of the practice of medicine, it has been relatively neglected in the literature of philosophy of medicine …”. Yet the patient’s diagnosis is central to much of healthcare; treatment decisions, funding, access to services, research programmes and studies, and many other things all depend on a person’s diagnosis.
Understanding the meaning and critically evaluating the use of a diagnosis is essential in rehabilitation. Although rehabilitation professionals may say, “I treat the patient, not the disease,” they must know the diagnosis because it will usually influence rehabilitation: access to rehabilitation services is often dependent on diagnosis, and diagnosis is essential when estimating prognosis.
The continuing impact of diagnosis rather than need influences many rehabilitation services. For example, some spinal cord injury services exclude patients with non-traumatic spinal lesions despite their identical clinical needs, and traumatic brain injury units may exclude patients having similar problems after subarachnoid haemorrhage.
In their philosophical treatise, Julian Reiss and Rachel Ankeny also wrote about diagnosis: “Diagnosis is the process through which a clinician determines what is wrong with a patient who is ill or ailing in some way.” and “Diagnosis, the process of determining the nature of a disease or disorder and distinguishing it from other possible conditions.” In other words, diagnosis may refer to both the process and its outcome.
The Oxford English Dictionary [OED] is ambiguous on this issue. “The identification of …” implies a process, yet the quotation used as an example says, “A diagnosis of Crohn’s disease was made”.
In healthcare, the term diagnosis is almost exclusively used to refer to the outcome. A diagnosis identifies a person as belonging to a group, such as people with lung tumours, fibromyalgia, or paraplegia. I will not discuss its use in describing the process.
History frequently explains and helps analyse complex matters. Thus, to clarify the meanings and the difficulties they entail, I will initially consider the word’s origins and the context in which it was developed.
What is a diagnosis? A historical perspective.
In 2006, Geza Balint and colleagues published a brief history of medical taxonomy and diagnosis, which I have used to outline the history.
Hippocrates studied illnesses in Ancient Greece around 400 BCE. He classified identified diseases using a framework based on symptoms and their anatomical location. Balint et al. suggest that his approach was holistic, in contrast to Chrysippos (300 BCE), who also classified conditions using the affected organ as the framework. Individuals developed other systems for classifying healthcare conditions over the following centuries.
Thomas Sydenham (1624-1689) promoted classifying diseases by description. His method resembled that of Linnaeus in categorising species according to their traits. In fact, Linnaeus published a classification of diseases in 1763, dividing them into 11 major groups.
As science gradually gained more knowledge about diseases, pathological changes observed during autopsies were increasingly used to classify diseases. This system was followed by classifications based on organ systems, chemical changes, or nearly all biological domains. Medical diagnoses became increasingly precise, with diagnoses such as apoplexy being refined; stroke was initially identified as one cause of apoplexy, and later, different types and locations of stroke were recognised.
Formulation
The dictionary’s meaning is much closer to a formulation, a term developed and used widely in psychiatry. A formulation is holistic, usually based on the biopsychosocial model of illness, and will identify relevant factors:
- Predisposing factors that render the person at a higher risk of the condition
- Precipitating factors that lead to the condition occurring
- Perpetuating factors that prolong the condition in the person
- Protecting factors that reduce the risk or the consequences of the condition
Wikipedia (March 21, 2024) summarised the concept of clinical formulation thus: “A clinical formulation ……. is a theoretically-based explanation or conceptualisation of the information obtained from a clinical assessment. It offers a hypothesis about the cause and nature of the presenting problems and is considered an adjunct or alternative approach to the more categorical approach of psychiatric diagnosis. In clinical practice, formulations are used to communicate a hypothesis and provide a framework for developing the most suitable treatment approach. It is most commonly used by clinical psychologists and is deemed to be a core component of that profession. Mental health nurses, social workers, and some psychiatrists may also use formulations.”
Depersonalisation of illness
The process of categorising illness, generally based on changes in the structure or function of bodily organs, has led to a depersonalisation of illness, “the action of divesting someone or something of human characteristics or individuality” [OED].
This stems from an inherent human tendency to classify and label illnesses, shifting the focus from the individual with a condition to the category or group of patients sharing a particular cause, leading to similar diseases. The label assigned is the diagnosis. The taxonomy employed is inconsistent, with different criteria applied to other groups.
Consequently, the patient’s diagnosis does not carry the same meaning for all illnesses. Though the meaning of the word diagnosis may be explicit (i.e. it is a categorical label used within a taxonomy), the meaning of the given diagnosis will vary from patient to patient. Much of this ambiguity is best understood by considering the models of illness and their influence on taxonomy.
Biomedical or biopsychosocial diagnosis?
Successful taxonomies depend on a comprehensive theory of the concerned construct. Medical science had no such all-encompassing theory when classification began in Ancient Greece. The initial theories appeared as science developed from 1592 onwards, and soon the biomedical model started to take shape, evolving as science advanced. However, the biomedical framework remains a patchwork, lacking a unifying system. This leads to the present inconsistent system.
International Classification of Diseases
The first International Classification of Diseases (ICD) was published in 1893 when the only systematic model of illness was biomedical. This was implicit, not explicit, and it considered all illnesses attributable to an observable dysfunction or structural change within the body. The taxonomy used was based on Bertillon’s classification of the causes of death. This was biomedical but mixed an organ-based system with ‘general diseases’, primarily infections.
The International Classification of Diseases and Related Health Problems continues to use a mixed taxonomy; the list of chapter headings for the ICD-10 Version: 2019 illustrates this mixture of constructs.
The biomedical approach to illness has several problems, which I have discussed in detail. For example, the following assumptions are untrue:
- An illness has only one cause;
- All primary causes of illness are within the body;
- All symptoms signify a bodily dysfunction;
- Reversal or removal of the cause will cure the illness.
Nevertheless, the ICD-10 classifies illnesses according to the biomedical framework or theory, which causes considerable difficulty. Many illnesses lack an obvious cause, and their diagnosis relies on the history and pattern of symptoms. They are syndromes—”a group of symptoms which consistently occur together, or a condition characterised by a set of associated symptoms” [OED]. Examples include irritable bowel syndrome, dyslexia, depression, and chronic low back pain.
Syndromic illnesses.
Syndromic illnesses are common. Without any other classification system, the ICD-10 usually places them in a chapter dedicated to an organ system. For instance, headaches and migraines are classified as nervous system disorders, while fibromyalgia is categorised under diseases of the musculoskeletal system and connective tissue.
This reveals a significant flaw; some diagnoses are based on observable pathology confirmed as the cause, but others within the same group (e.g., nervous system diseases) are based solely on symptoms and may lack organ-based pathology.
Furthermore, a diagnosis is defined as “identifying the nature of an illness”, and a single word or phrase label based on actual or presumed organ pathology often provides no indication of the illness’s true nature.
Psychiatric diagnoses
It is now worth a detour into psychiatry. Although it uses medical diagnoses as labels, psychiatric practice has retained a more holistic approach, referring to its approach as formulation. One reason for psychiatry’s dependence on a medical model is that funding requires single diagnostic labels.
The Diagnostic and Statistical Manual of Mental Disorders (DSM-5-TR) is the equivalent of the International Classification of Diseases (ICD-10). Virtually all the diagnoses are syndromes, with lists of symptoms and algorithms to deduce whether the person has the diagnosis.
The original DSM-I in 1918 had 22 diagnoses; DSM-II had 182, DSM-II had 265, and DSM-IV had 297. Shadia Kawa and James Giordano outline a more detailed history. It is also essential to recognise the strong influence culture and social factors have on psychiatric diagnoses; for example, homosexuality was once considered a psychopathology. Thomas Szasz questioned the validity of all psychiatric diagnoses in 1960, and his view is still relevant today.
The use of symptoms for making a diagnosis is questioned in psychiatry. As Martin Scheffer and colleagues say, in a review of the theory of a dynamical systems view of psychiatric disorders, “symptoms people present with do not fit neatly into single diagnostic buckets. Thus, disorders exist on a spectrum of possible states (chronic anxiety, dysfunctional cognitions, depressed mood, etc) that are not clearly separated and vary widely between individuals.”
They further stress that symptoms change over time, such that patients “… may follow a roller coaster of disordered states over the course of their lifetime, reflected in a largely unpredictable succession of different diagnoses.” The theory is based on well-established mathematical principles of chaotic systems; essentially, people move into one of a set of steady-state positions (“attractors”), which are then named as diagnoses.
So, what is a diagnosis?
This analysis suggests current diagnoses fall into two classes.
The first group encompasses people diagnosed with an unmistakable and observable disorder of cellular or organ structure or function, either intrinsic (e.g., cancer or genetic) or extrinsic (e.g., infection or trauma), that is highly likely to be the proximate cause of the person’s symptoms. These are the biomedical diagnoses. The classification system is not logically coherent, but most diagnoses depend on externally observable features.
The second group concerns patients with a history and pattern of symptoms similar to those reported by other patients, but without any definite or proven single bodily changes causing the syndrome. Even if a patient in this group has observable changes in organ structure and function, these changes are unlikely to be associated with the symptoms.
The latter group are best referred to as syndromic diagnoses. They include most psychiatric diagnoses and a substantial proportion of current medical diagnoses, such as migraine, irritable bowel syndrome, and hypermobility syndrome.
Biomedical diagnoses typically:
- Are, in principle, provable with observable changes in bodily structure or function
- Allow targeted intervention or remove, reverse, or control the causative bodily changes
- Determine a patient’s prognostic field and the pattern of change, and may suggest items that can give more predictive information.
Biomedical diagnoses do not:
- Explain all the symptoms a person has
- Always give rise to symptoms associated with them
- Exclude a syndromic diagnosis additionally being present
- Determine any of the consequences of the disease, such as disability and dependence
- Have a significant relationship to an individual’s healthcare needs and the resources required to satisfy those needs
Syndromic, biopsychosocial diagnoses typically:
- Are not provable or defined by any specific observable feature,
- Have no clear boundaries separating them from healthy people or other similar syndromic diagnoses,
- Rarely, if ever, have a single cause
- Are associated with many contributing factors
Syndromic diagnoses do not:
- Explain any of the symptoms a person has or indicate a specific cause,
- Exclude a biomedical diagnosis being present,
- Determine any of the consequences of the given diagnosis, such as disability and dependence
- Have a significant relationship to an individual’s healthcare needs and the resources required to satisfy those needs
The ICD-10 diagnoses used now:
- Are based on a single aspect of a person’s illness and the choice of aspect
- is inconsistent between diagnostic groups
- is not based on a uniform system
- often means a person can be given several diagnoses
- May be based on an ill-defined or imprecise collection of symptoms with
- No clear separation between healthy people and people with the diagnosis
- No clear separation between people with other syndromic diagnoses
- No substantial evidence to support the diagnosis as a unique entity
- Are weakly related or unrelated to the person’s experience of the illness
Conclusion
The meaning and use of diagnoses vary greatly, yet the word is often treated as if it signifies a single concept within a unified framework of illness. As a result, a person’s diagnosis is frequently used to justify decisions that cannot legitimately be based solely on the diagnosis. This typically happens when syndromic diagnoses such as autism, attention deficit hyperactivity disorder, or chronic fatigue syndrome are mistaken for a biomedical disease with a single, proven cause and, usually, a specific treatment. The improper use of many diagnoses is causing increasing problems within healthcare. This article explains why or how these issues develop.
A second post explores more effective approaches. A diagnosis must be contextualised for the person, using a formulation or description that acknowledges uncertainty and the influence of other factors. This is best achieved using the biopsychosocial model of illness. Within this framework, a diagnosis is not a single-word descriptor usually based on an organ system. Instead, a diagnosis is a multi-axial account of interrelated factors that explain the person’s illness.