The biopsychosocial model of illness is central to the understanding of rehabilitation, and central to all aspects including its practice and research. This part of the site could not cover everything, nor can it give references or evidence for everything. It does, however, acknowledge there are criticisms, and potential weaknesses. The weaknesses are not sufficient, in my view, to warrant abandoning it, but they should be used as a stimulus to refining and improving it.
This page will cover several matters. Press the red link to jump to a section:
- What is a model. here
- Why is a model needed? here
- The biomedical model outlined. here
- Why is a new model needed? here
- The biopsychosocial model introduced. here
- A holistic biopsychosocial model developed. here
- A systems model explained. here
- The perceived weaknesses discussed. here
What is a model?
Most people are probably unaware that they have any ‘model of illness’. And few could explain what model they have. Yet, from an early age, we all develop and use a model of illness. So first, what is a model?
When your car starts to malfunction, your immediate thoughts include why has this happened, why now, what is wrong, what will happen if I ignore it, and what can I do about it. Your answers will depend upon you understanding of how your car works. You may understand all about the engine (diesel, petrol, or electric), and other mechanics of the car, or you may know very little. Most people know that fuel is needed, and will check that. Some people will automatically check the oil level, or radiator fluid, or something else because that was the problem last time.
Eventually most people will go to a garage. This is because the engineers have a model of how cars work, and know the details of each different type of car. The reality is that, now few people have a sufficiently detailed model in their mind to analyse the problem and find a solution. Engineers, on the other hand, have excellent models.
In other words, a model refers to the representation in the mind of some object or process or activity, that is used to analyse how changing one factor might alter the situation, This helps identify the best action to improve a situation or solve a problem. For example, when talking to someone you will have a ‘model’ that helps you predict how they will respond to a request or remark. If you want them to do something, you might think through their response to one way of asking, then another, until you find the approach most likely to succeed. Models help you choose actions, by allowing you to predict what is most likely to achieve your goal.
Why a model of illness?
You are a conscious being, which means that you have a personal awareness of what sensations you are feeling, of the quality of your performance of activities, and of your emotional state. Each of these can be compared to what you would expect. At some point you may notice something unexpected. If the unexpected phenomenon continues, or if the difference is severe or very significant, you will start to analyse it.
Research shows that, when faced with an unexpected bodily or personal experience, people ask much the same questions in all cultures:
- What. What is the immediate, direct cause for the unexpected experience?
- Why. What factors or events led to this experience occurring now?
- How. How does the cause lead to the change or experience?
- Future. What will happen over the next minutes, hours, days, weeks?
- Action. What can I do, or what can others do, to lessen any consequences?
Each person will have their own explanatory model of illness. Some people will attribute most changes to allergy, or to diet, or to electric fields. Others, now as well as in the past, will attribute experiences to others cursing them, or casting evil spells. Whoever we are, we will have a preferred and more likely way of analysing unexpected bodily function.
The biomedical model
Although everyone’s personal model differs, in most modern Western cultures, and increasingly elsewhere, a biomedical model is the socially dominant model. The essential, key features are that:
- all illness is due to disease, which is located within the body;
- and all abnormal experiences are symptomatic of disease
- disease occurs outside the control of the patient
- treatment, and cure, is achieved by external agents (drugs, surgery etc)
- and cure will reverse illness and return the person to health
This model is closely associated with several other cultural phenomena:
- a reductionist, scientific approach;
- for example one disease accounts for all abnormal experiences
- a social ‘contract’ whereby the ill person
- is expected to seek expert help as soon as possible
- is excused of some or all social responsibilities
- is expected to strive to return to health as quickly as possible
- a complete separation between mental and physical phenomena
- referred to as ‘Cartesian dualism’
This model, at present, is the basis for almost all political and management decisions made about healthcare, and is the basis for most funding decisions and policies. It is so ingrained that it is next discussed, and is not mentioned anywhere. Indeed, it is hard to find any definition or discussion, except in article discussing the biopsychosocial model of illness. The first two articles identified by Google on a search for “The biomedical model of illness” are by me, writing about the biopsychosocial model.
Why is a (new) model needed?
This biomedical model started to develop in the early 1500s, when the scientific revolution started, and over 400 years it led to great advances. However, early in the 1900s disquiet began, particularly around the separation of mental health. Then, over the twentieth century, other developments occurred, which drew attention to weaknesses. Among others were:
- the arrival of holism (1932), developing with general systems theory and the recognition of complexity;
- the development of sociology, and the recognition of the influence of culture on illness and behaviour;
- the increasing recognition of the social, environment and financial influences on illness and health;
- more recently the emergence of multi-morbidity, where patients often have multiple long-term diseases;
- and the recent recognition of functional disorders and the major influence of mental health on bodily function and personal behaviour.
The biomedical model of illness does not recognise any of these factors. It cannot use them when analysing health problems. The healthcare systems are still based on a biomedical model, and this is one important cause of the increasing failure of healthcare systems to manage population health effectively.
The original article about the biopsychosocial model of illness was published in Science in 1977, by Engel. However this was the culmination of work undertaken by many people from many disciplines over many years. The first reference I can find was in an article by Adolf Meyer, an influential psychiatrist, published in the Journal of the American Medical Association in 1917 which refers to a presentation in San Francisco in 1915. His interest probably predated 1915.
In his article Engel wrote, presciently, “‘But nothing will change unless and until those who control resources have the wisdom to venture off the beaten path of exclusive reliance on biomedicine as the only approach to health care.” Twenty-seven years later there was no change; forty years later little had changed.
The original article, which is well worth reading, was centred on psychiatry, and on understanding the influence of non-disease factors on health and illness. Examples used included insulin-dependent diabetes and grief. It did not explicitly set the model in the context of rehabilitation.
The framework, the conceptual ideas, was rapidly used by the World Health Organisation to classify the consequences of disease in the International Classification of Impairments, Disabilities and Handicaps (WHO ICIDH, 1980). In the light of much, quite justified, criticism, the framework was improved and published in 2000 as the International Classification of Functioning, Disability, and Health. (WHO ICF).
The two main improvements were (a) a change in terminology, introducing ‘activities’ in place of disability and ‘participation’ in place of handicap, and (b) the addition of contextual factors including the physical, social, and personal contexts.
The vital point to understand about the WHO ICF is that it is using the biopsychosocial model as a framework, purely for the purpose of classifying the non-disease aspects of illness, including some factors that have an impact on illness. The WHO ICF is not, in itself, a model. It makes no comment on, and it does not document or quantify, the nature or direction of any inter-relationships. It is summarised in this graphic.
A holistic model
Although Engel’s article acknowledged many factors including culture, the model was initially made operational in the WHO ICIDH, which was limited to person-related items. The move to the WHO ICF framework was a great advance, because it incorporated contextual factors. Some other items were still missing.
Over about 10-12 years, my friend and colleague Peter Halligan and I discussed patients for hours most evenings at the end of the working day. In these conversations I usually framed my contribution within the biospychosocial model. He would then ask pertinent questions, challenging my model. This was a productive dialogue that culminated in our paper in the British Medical Journal in 2004 (see here) where we added more features, and this development was continued in 2017.
The latest version of this framework, that I have termed the holistic biopsychsocial model added four features that make it much more person-centred.
The first two concern time. People working with adults easily overlook the importance of time, but paediatricians do not. Everything in paediatrics has, as its first consideration, the child’s age. Children change and mature rapidly. In adulthood, though biological change is slower, other contextual aspects change quite significantly. Therefore some consideration of a person’s age, or more accurately, the stage of their life is a very important matter.
The person’s stage within their particular illness is also an important factor, and this is the second temporal domain within the framework.
The third and fourth aspects of the model both make the model personal to the specific patient.
The first is choice. We all have the ability to choose, and in long-term illnesses patients make many choices. Some of those choices are against medical recommendations. Any model of an illness that does not factor in a person’s choice will fail more often than it succeeds.
The second is quality of life, alternatively described as wellbeing. The quantification of this rests entirely with the patient. Only the patient can (a) identify what factors are important in determining his or her quality of life and (b) decide what weight or importance to attach to each factor. These weights may well determine the person’s choices.
The addition of these four items returns the model from its original impersonal use, in the WHO ICIDH to a much more person-centred framework. It also now includes two items which fall outside the range of items that can, to some extent, be observed or deduced – quality of life, and individual choice. Both these items are determined entirely by the person. They will be influenced by the value that the person places upon items within one or more of the model’s domains. Each persons choice of items and domains is their own.
A systems model
Finally, we can move on to the model as a way of analysing a situation and predicting consequences of actions. We can consider the important characteristics of the model.
Engel stressed that his model fitted within the realm of General Systems Theory. This is theory underlies other, similar models such as Kielhofner’s Model of Human Occupation. The vital characteristic to note is that the model is highly complex. The key features of a complex system are:
- multiple factors can influence one factor
- one factor can influence many others
- the relationships can be in any direction, or bidirectional
- the relationships are not linear
- there are direct and indirect relationships
See here for further discussion of complexity.
A further vital point is that a system (or domain) is itself comprised of lower-order units that are also systems. The body is comprised or organs; organs are composed of cell; cells have many intracellular organelles and so on. At the same time, most systems are one system among other superordinate systems. The person is part of their family; the family is part of a neighbourhood etc. At the same time, the person many be part of a workgroup that employs him or her; and that group may be part of a department or service; and so on.
From a rehabilitation perspective, one cannot hope to collect information on and analyse the effects of all superordinate systems that a person is involved in, nor the interactions between all subordinate systems. The biopsychosocial model provides a framework collecting into manageable units the different contextual systems that may have an effect, and the different internal systems of relevance. The skill of a rehabilitation expert is to identify, quickly, the important factors in a particular case.
And the weaknesses ..
There are people and articles that criticise the biopsychosocial model. Almost all are in fact critical of the way it is used or interpreted, rather than the model itself. There are also other criticisms made that I have heard but not (yet) seen written.
Use in disability & employment allowance disputes.
There are many articles, some in scientific journals, others in more journalistic publications, which criticise the model for decisions that are made about disabled people and their right to employment or other disability allowances. These criticisms are inherently invalid as the model does not make the decision. It is not the framework which is to blame. The model or framework is intended to help in the analysis and understanding of a situation. It cannot itself make any decision.
It may well be that the assessor is undertaking a poor analysis, not recognising the existence of a factor, or its ability to have an influence. Alternatively, the assessor may feel that, although factors A, B, and C are present and might cause the observed disability, they (the assessor) believe that the person has chosen to act in a disabled way.
In other words, it is not the model that is faulty. It may be that the analysis is inadequate, perhaps because the assessor simply does not have an appropriate level of knowledge and/or appropriate skills. It maybe that the interpretation is wrong or, perhaps more accurately, at variance with the interpretation of others. It could be that the assessor is misinterpreting the legal or policy guidance.
Use in analysing some conditions.
A second area of public debate and concern relates to the analysis of functional disorders, with chronic fatigue syndrome being a particular case in point. I do not wish to take any specific view here, but will explain the issues.
About 25% of people attending outpatients do not have any disease that can plausibly account for their prevention problems – a follow-up confirms that few in fact go on to manifest any disease. This problem is often referred to as a functional disorders, and these are well recognised in most medical specialities such as neurology. It is unknown why any particular person may develop such a disorder, but some risk factors are known,
It is one of the defining characteristics of complex systems that they can malfunction even when all the individual components are functioning properly. This is a special case of a general property of any system, known as an emergent property. The most obvious example is consciousness. We undoubtedly have consciousness (though its nature is uncertain), but it cannot be explained on the basis of neurophysiological processes within the brain.
One hypothesis is that functional disorders represent an emergent property of the holistic biopsychosocial systems-model. Indeed, almost all complex systems experience failures, yet investigation does not find any specific abnormality. Many common examples occur within healthcare, when a patient is harmed not through anyone’s wilful behaviour, or through anyone’s culpable negligence, but simply because healthcare is an extraordinarily complex system.
It is people who use this model in an attempt to understand this class of illness. They do so to identify better treatments, and helpful interventions, such as cognitive behavioural therapy and exercise (being active) are two. These are not cures, but ways of improving the life of the patient.
Its not our responsibility …
A third criticism I have heard, specifically relating to its use in rehabilitation, is that it is not ‘our responsibility’ (e.g. as an inpatient rehabilitation team) to sort out housing, organise transport for someone to attend a day centre, help the person apply for financial befits they need, provide the wheelchair they need to be able to get around etc etc. Therefore, we should not ask or be concerned about these matters.
I have no fixed opinion on whether or not any of the examples are or are not the direct responsibility of a specific rehabilitation service, but ..
A) I would argue strongly that knowing about these problems is a vital responsibility, for without that knowledge much of your patient’s behaviour and many of there choices may simply appear irrational or cussed.
B) I would also argue that, although it may not even be in your power to do something, it is probably in someone’s power, and therefore you need to contact that person or organisation and ask for their input. This is not only morally correct, but it usually pays off in terms of a more effective period of rehabilitation and a shorter stay, because most of the problems will, if not addressed, delay discharge and/or increase the risk of re-admission to an acute hospital,
I conclude that the model is the best available at present. Doubtless it can be and, I hope, will be improved over the next 40 years. It is much more powerful as an analytic tool than the biomedical model or any other model that I am aware of. It includes almost all existing models within it.