Measuring disease in a society
Assessing the effectiveness of treatment of a disease
The sociocultural context of medicine
Medicine cannot be practised without considering the societal contexts within which it exists: a new ‘wonder drug’ for hypertension will be ineffective if a patient does not take it; an expensive, complicated treatment for an advanced stage of a disease would be unnecessary if that disease could be prevented by more simple means; a patient who is efficiently diagnosed with gluten-sensitive enteropathy (coeliac disease) but who does not understand a doctor’s description of sources of gluten in the diet will continue to suffer from the effects of that disease.
There also needs to be an overall view or ‘tally’ of disease in society to look out for the rise of new diseases (e.g. Helicobacter, HIV, severe acute respiratory syndrome (SARS) and swine flu in the past three decades) and to balance the need and availability of resources for particular diseases. These areas are covered by specialisms such as psychology, sociology, epidemiology, and health economics, but are considered here under the generic title of medicine and society.
An individual doctor working in a hospital is unlikely to spot significant changes in the rate of a specific disease in the locality because the rate at which that doctor sees patients is determined by the availability of resources such as clinic appointments. General practitioners are better placed to detect short-term rises in diseases such as influenza, but they will not be able to detect changes in less common diseases because they are likely to see only one or two cases per year. The specialism of epidemiology (and the related specialism of public health medicine) provide techniques for measuring the overall burden of disease in a society. There are a few definitions which are important:
• Prevalence—the number of people in a defined population who have a disease at any one time point, e.g. all those with rheumatoid arthritis
• Incidence—the number of people in a defined population who are diagnosed with a disease in a defined time period, e.g. the number of women diagnosed with breast cancer in a year.
It can be seen that the prevalence and incidence of a disease can vary considerably depending on the biological course of that disease. If a disease is chronic, e.g. persisting for years—such as rheumatoid arthritis, then there may be many people in a population with that disease, i.e. a high prevalence, but the incidence could be low. If the time course of a disease is rapid then there can be a high incidence but a low prevalence, e.g. an influenza epidemic will have a high incidence in 1 year but a very low prevalence because people recover from flu within 10 days.
Epidemiological methods can detect changes in the incidence/prevalence of a disease in a particular population and this may lead to further investigations into the pathogenetic mechanisms of that disease. It is important here to recognize the differences between causes of disease and risk factors for a disease:
• Cause—is a factor which has been scientifically proven to be an integral part of the mechanism of causing a disease, e.g. cystic fibrosis is caused by a mutation in the cystic fibrosis transmembrane conductance regulator gene.
• Risk factor—is a factor which is identified as being statistically significantly associated with a disease but has not yet been proven to be an integral part of the disease mechanism. Risk factors are often identified in epidemiological studies, and widely reported in the popular media (e.g. ‘daily glass of red wine protects against prostate cancer’), but their association may only occur because they themselves are closely associated with another factor that is a real cause of the disease. Thus risk factors identified by epidemiological studies then need laboratory studies, using the classical scientific method of proving or disproving the null hypothesis, to show whether they are real causes of the disease in question.
In order to calculate the prevalence or incidence of a disease there have to be methods of detecting that disease in a population. Methods include:
• Specific surveys—e.g. healthcare workers going out into a population and performing tests that detect a disease, such as sputum testing for tuberculosis
• Notifiable diseases—in many countries there are specific disease which must be notified to authorities by healthcare workers whenever the disease is diagnosed. In the UK such diseases include poliomyelitis, anthrax, cholera, tuberculosis, whooping cough (all infectious diseases)
• Cancer registries—in many countries there is a flow of information from hospitals to registries whose aim is to record all cases of cancer and follow-up of those cases. There are obvious problems in establishing reliable flows of information so the figures are not wholly accurate
• Death certificates—death certificates include a doctor’s opinion of the cause of death of the person and this might be regarded as a useful source of information on diseases that cause death. However, many studies have shown that unless an autopsy has been performed, this information can be inaccurate, often misidentifying the pathology within a body system or even incorrectly implicating failure of a particular system as the cause of death
• Industrial diseases—there are some diseases which are specifically caused by work in industry which are recorded in government statistics, e.g. asbestos-related lung disease, including mesothelioma and asbestosis.
Another important aspect of public health medicine/epidemiology is measuring and assessing the effectiveness of a treatment of a disease. This can only be done by taking a society-wide view of the benefits and costs of a treatment. Many aspects of this are controversial and institutions that have been set up to advise on or regulate new treatments, e.g. the National Institute of Health and Clinical Excellence in the UK, are often subject to much pressure from patient groups and specialist doctors. In assessing treatments, a number of factors have to be taken into account:
• The effectiveness of the treatment—if a treatment is prescribed to the correct type of patient for the correct type of disease how effective is the treatment? Some treatments are almost self-evidently highly effective, e.g. administration of thyroxine for hypothyroidism and other hormone-replacement therapies, an antibiotic to which a bacterial organism is sensitive in a patient with septicaemia (large randomized controlled trials were not required to demonstrate the effectiveness of penicillin). Other treatments are much more difficult to assess and do require large randomized controlled trials with long-term follow-up—e.g. antihypertensive drugs, cancer chemotherapy
• The cost of a therapy—is important because extremely expensive treatments that only produce a small benefit may be too costly to be accepted as routine treatment in a state-funded health service. The whole cost of the treatment needs to be calculated and not just the simple cost of the drug, e.g. an anti-cancer drug (such as the new monoclonal antibody therapies) may cost £10 000 per patient per year but this cost may increase to £30 000 a year if the additional costs of administering the drug (e.g. inpatient admission for intravenous infusion) and monitoring the patient (e.g. blood tests to detect neutropenia) are included
• The resources required to administer the treatment—even if the financial cost of a treatment is thought to be good value when the benefit of the treatment is assessed it may be that there are insufficient resources available at that point in time to allow the treatment to implemented across a whole healthcare system, e.g. insufficient nurses trained to give chemotherapy.
Some of the measures used to measure the benefit of a therapy include:
• Life expectancy—is the average number of years of life predicted to be left for a person of a particular age. Thus treatments can be evaluated by the number of years that they increase life expectancy in patients. However, this does not take into account the quality of those ‘extra’ years
• Quality-adjusted life years (QALYs)—a widely used measure that takes account of the amount of extra years life a treatment will on average give a patient and the quality of life of those extra years. A treatment that gave 10 years extra life with no reduction in the quality of life in those years could be described as giving 10 QALYs. A treatment that gave 10 years extra life but with reduced quality of life (e.g. amputation of a leg for a malignant tumour) would be described as having <10 QALYs but >0 QALYs and the point which it lies between the two points would depend on a fairly subjective assessment of the reduction is quality of life, e.g. if it was thought to reduce the quality of life by 20% it would produce 10 × 0.8 = 8 QALYs. QALYs are a more useful measure of benefit than raw life years but the quality adjustment can be difficult to assess
• Disability-adjusted life years (DALYs)—if a disease does not cause premature death, then QALY will not be a useful measure of benefit (it will always be zero), so the concept has been extended to DALYs, where the number of years with/without a disability is included. This is very useful for chronic diseases that cause a significant reduction in the quality of life of a person but do not usually cause premature death; such diseases include osteoarthritis, rheumatoid arthritis, and dental caries.
• Healthy life years (HLY)—is a variant of QALYs and DALYs, describing the number of disease-free years of life. It is often used in calculating composite ‘health’ scores for comparisons between countries, especially in the European Union.
A hypothetically ‘technically perfect’ doctor who could diagnose any disease with which a patient presented and prescribe optimal treatment would be very ineffective if he or she did not establish a satisfactory working relationship with his or her patients. It is no use being able to diagnose diseases if you cannot explain that diagnosis in a way that the patient can understand and act upon. Treatment will be ineffective if a doctor does not empathize with a patient sufficiently well to understand the context of the illness within the patient’s overall life situation. In fact, a doctor is unlikely to make the correct diagnosis if he/she does not establish a satisfactory working relationship because he/she will not elicit all the symptoms and signs that will point to the correct diagnosis. Many aspects of the doctor–patient relationship are best learnt in simulated and real-life interviews with careful guidance, but there are some ground rules that can be learnt before this happens. A very useful formulation is that given by the General Medical Council in the Good Medical Practice guidelines:
‘To fulfil your role in the doctor–patient partnership you must:
a. Be polite, considerate and honest
b. Treat patients with dignity
c. Treat each patient as an individual
d. Respect patients’ privacy and right to confidentiality
e. Support patients in caring for themselves to improve and maintain their health
f. Encourage patients who have knowledge about their condition to use this when they are making decisions about their care.’1
One interesting area of the doctor–patient relationship is determining the reason that the patient has come to consult the doctor, especially in general practice, which is usually the first port of call for any patient. Whilst a doctor could assume that the patient has come to consult them about the illness which the patient is describing to them this would often be incorrect as demonstrated by sociological research, such as that of Irving Zola, who identified five triggers to consultation:2
1. The occurrence of an interpersonal crisis
2. The perceived interference with social or personal relations
3. ‘Sanctioning’ (pressure) from others
4. The perceived interference with vocational or physical activity
5. The ‘temporalizing of symptomatology’—setting a deadline for action (often manifest as a rush of consultations just before Christmas in the UK).
Whilst these are at least tangentially related to the ‘illness’ that the patient is describing, the doctor needs to find out which are these are the most relevant to this particular patient, because this will influence the advice and treatment that is recommended. For example, a man may present with the symptoms of urinary outflow obstruction due to benign prostatic hypertrophy which is causing little current interference with his life but he has come at this point in time because his wife’s brother has just been diagnosed with prostate cancer. In such circumstances, this man may not need immediate therapy for his prostatic hypertrophy but may need the ‘reassurance’ of a prostate-specific antigen blood test to reduce the probability of a diagnosis of prostate cancer.
The ‘clinical iceberg’ is a related concept which points out that there is often little difference between the symptoms experienced by people in the general population who do not go to consult their doctor and those who do. For example, levels of pain from osteoarthritis are similar, so there must be some other precipitating factor for the consultation.
It is easy to observe different patterns of behaviour that people exhibit when they have an illness, even from friends and associates when they contract a viral respiratory infection (‘man flu’ etc.). These patterns of behaviour will be influenced by a person’s upbringing and current life situation, including employment and religious beliefs. However, there are some generic patterns of behaviour that can be usefully identified because they provide some information which may be useful when treating a patient with that pattern of behaviour:
• Denial—a pattern of behaviour where the patient does not consciously acknowledge that they have the disease from which they are clearly suffering or by their behaviour, do not accept the presence of the disease. This may be seen as a beneficial adaptive behaviour when faced with an imminent terminal illness, although acceptance of such a diagnosis may be more beneficial for the person and their family. May occur with other illnesses at different times in a person’s life. For example, ‘denial’ of type 1 diabetes mellitus in teenage life with refusal to take insulin and repeated hospital admissions in diabetic ketoacidosis
• Medicalization—a behaviour pattern where all symptoms and behaviours are seen as having a strong relationship to defined medical illness rather than more ‘rational’ explanations such a significant life events etc. Doctors and other healthcare professionals may well induce this behaviour in patients by overemphasis on the medical aspects of a patient’s overall life situation. It differs from hypochondriasis in that it does relate to a real, rather than imagined, illness or state. There is long-standing debate over whether the medical profession has turned normal life events, such as childbirth, into ‘diseases’ by the process of medicalization.
• Risk taking—this is behaviour which threatens a person’s health because they are not making rational decisions about the risks attached to a certain behaviour. Of course this varies from the very obvious, e.g. unprotected sexual intercourse with multiple partners in a population with a high prevalence of HIV infection, to more widespread and subtle behaviour such as being overweight/obese over a long period of time despite knowledge of the health risks attached to this state (which many doctors themselves exhibit).
• Hypochondriasis—a strongly held belief by a person that they have a particular disease in the face of all evidence to the contrary. Can be mild and as such be adaptive behaviour to an unsatisfactory life situation, but can be so strongly held as to constitute a monoideistic delusion.
Farmer R, Lawrenson RE, Miller D (2004). Epidemiology and Public Health Medicine. Oxford: Blackwell Publishing.
Morrison V, Bennett P (2006). An Introduction to Health Psychology. Edinburgh: Pearson Education.
Scrambler G (2008). Sociology as Applied to Medicine, 6th edn. London: Elsevier.