Thinking about medicine*
I swear by Apollo the physician and Asclepius and Hygieia and Panacea and all the gods and goddesses, making them my witnesses, that I will fulfil according to my ability and judgement this oath and this covenant.
To hold him who has taught me this art as equal to my parents and to live my life in partnership with him, and if he is in need of money to give him a share of mine, and to regard his offspring as equal to my own brethren and to teach them this art, if they desire to learn it, without fee and covenant. I will impart it by precept, by lecture and by all other manner of teaching, not only to my own sons but also to the sons of him who has taught me, and to disciples bound by covenant and oath according to the law of physicians, but to none other.
The regimen I shall adopt shall be to the benefit of the patients to the best of my power and judgement, not for their injury or any wrongful purpose.
I will not give a deadly drug to anyone though it be asked of me, nor will I lead the way in such counsel.1 And likewise I will not give a woman a pessary to procure abortion.2 But I will keep my life and my art in purity and holiness. I will not use the knife,3 not even, verily, on sufferers of stone but I will give place to such as are craftsmen therein.
Whatsoever house I enter, I will enter for the benefit of the sick, refraining from all voluntary wrongdoing and corruption, especially seduction of male or female, bond or free.
Whatsoever things I see or hear concerning the life of men, in my attendance on the sick, or even apart from my attendance, which ought not to be blabbed abroad, I will keep silence on them, counting such things to be as religious secrets.
If I fulfil this oath and do not violate it, may it be granted to me to enjoy life and art alike, with good repute for all time to come; but may the contrary befall me if I transgress and violate my oath.
Advice for doctors
• Do not blame the sick for being sick.
• Seek to discover your patient’s wishes and comply with them.
• Work for your patients, not your consultant.
• Respect opinions.
• Treat a patient, not a disease.
• Admit a person, not a diagnosis.
• Spend time with the bereaved; help them to shed tears.
• Give the patient (and yourself) time: for questions, to reflect, and to allow healing.
• Give patients the benefit of the doubt.
• Be optimistic.
• Be kind to yourself: you are not an inexhaustible resource.
• Question your conscience.
• Tell the truth.
• Recognize that the scientific approach may be finite, but experience and empathy are limitless.
The National Health Service
‘The resources of medical skill and the apparatus of healing shall be placed at the disposal of the patient, without charge, when he or she needs them; that medical treatment and care should be a communal responsibility, that they should be made available to rich and poor alike in accordance with medical need and by no other criteria…Society becomes more wholesome, more serene, and spiritually healthier, if it knows that its citizens have at the back of their consciousness the knowledge that not only themselves, but all their fellows, have access, when ill, to the best that medical skill can provide…You can always ‘pass by on the other side’. That may be sound economics. It could not be worse morals.’
Aneurin Bevan, In Place of Fear, 1952.
In 2014, the Commonwealth Fund presented an overview of international healthcare systems examining financing, governance, healthcare quality, efficiency, evidence-based practice, and innovation. In a scoring system of 11 nations across 11 categories, the nhs came first overall, at less than half the cost per head spent in the usa.1 The King’s Fund debunks the myth that the nhs is unaffordable in the modern era,2 although funding remains a political choice. Bevan prophesied, ‘The nhs will last as long as there are folk left with the faith to fight for it.’ Guard it well.
‘There is a good deal of hit and miss about general medicine. It is a profession where exact measurement is not easy and the absence of it opens the mind to endless conjecture as to the efficacy of this or that form of treatment.’
Aneurin Bevan, In Place of Fear, 1952.
A qaly is a quality-adjusted life year. One year of healthy life expectancy = 1 qaly, whereas 1 year of unhealthy life expectancy is worth <1 qaly, the precise value falling with progressively worsening quality of life. If an intervention means that you are likely to live for 8 years in perfect health then that intervention would have a qaly value of 8. If a new drug improves your quality of life from 0.5 to 0.7 for 25 years, then it has a qaly value of (0.7 − 0.5)×25=5. Based on the price of the intervention, the cost of 1 qaly can be calculated. Healthcare priorities can then be weighted towards low cost qalys. The National Institute for Health and Care Excellence (nice) considers that interventions for which 1 qaly=<£30 000 are cost-effective. However, as a practical application of utilitarian theory, qalys remain open to criticism (table 1.1). Remember that although for a clinician, time is unambiguous and quantifiable, time experienced by patients is more like literature than science: a minute might be a chapter, a year a single sentence.3
Table 1.1 The advantages and disadvantages of qalys
Transparent societal decision making
Focuses on slice (disease), not pie (health)
Common unit for different interventions
Based on a value judgement that living longer is a measure of success
Allows cost-effectiveness analysis
Quality of life assessment comes from general public, not those with disease
Allows international comparison
Potentially ageist—the elderly always have less ‘life expectancy’ to gain
Focus on outcomes, not process ie care, compassion
The inverse care law, equity, and distributive justice:
The inverse care law states that the availability of good medical care varies inversely with the need for it. This arises due to poorer quality services, barriers to service access, and external disadvantage. By focusing on the benefit gained from an intervention, the qaly system treats everyone as equal. But is this really equality? Distributive justice is the distribution of ‘goods’ so that those who are worst off become better off. In healthcare terms, this means allocation of resources to those in greatest need, regardless of qalys.
How to formulate a diagnosis
Diagnosing by recognition:
For students, this is the most irritating method. You spend an hour asking all the wrong questions, and in waltzes a doctor who names the disease before you have even finished taking the pulse. This doctor has simply recognized the illness like he recognizes an old friend (or enemy).
Diagnosing by probability:
Over our clinical lives we build up a personal database of diagnoses and associated pitfalls. We unconsciously run each new ‘case’ through this continuously developing probabilistic algorithm with increasing speed and effortlessness.
Diagnosing by reasoning:
Like Sherlock Holmes, we must exclude each differential, and the diagnosis is what remains. This is dependent on the quality of the differential and presupposes methods for absolutely excluding diseases. All tests are statistical rather than absolute (5% of the population lie outside the ‘normal’ range), which is why this method remains, like Sherlock Holmes, fictional at best.
Diagnosing by watching and waiting:
The dangers and expense of exhaustive tests may be obviated by the skilful use of time.
Diagnosing by selective doubting:
Diagnosis relies on clinical signs and investigative tests. Yet there are no hard signs or perfect tests. When diagnosis is difficult, try doubting the signs, then doubting the tests. But the game of medicine is unplayable if you doubt everything: so doubt selectively.
Diagnosis by iteration and reiteration:
A brief history suggests looking for a few signs, which leads to further questions and a few tests. As the process reiterates, various diagnostic possibilities crop up, leading to further questions and further tests. And so history taking and diagnosing never end.
Heuristics are the cognitive shortcuts which allow quick decision-making by focusing on relevant predictors. Be aware of them so you can be vigilant of their traps.7
The diseases that we remember, or treated most recently, carry more weight in our diagnostic hierarchy. Question whether this more readily available information is truly relevant.
The hunt for, and recall of, clinical information that fits with our expectations. Can you disprove your own diagnostic hypothesis?
A good bedside manner is dynamic. It develops in the light of a patient’s needs and is grounded in honesty, humour, and humility, in the presence of human weakness. But it is fragile: ‘It is unsettling to find how little it takes to defeat success in medicine… You do not imagine that a mere matter of etiquette could foil you. But the social dimension turns out to be as essential as the scientific… How each interaction is negotiated can determine whether a doctor is trusted, whether a patient is heard, whether the right diagnosis is made, the right treatment given. But in this realm there are no perfect formulas.’ (Atul Gawande, Better: A Surgeon’s Notes on Performance, 2008)
A patient may not care how much you know, until they know how much you care. Without care and trust, there can be little healing. Pre-set formulas offer, at best, a guide:
Introduce yourself every time you see a patient, giving your name and your role.
‘Introductions are about making a human connection between one human being who is suffering and vulnerable, and another human being who wishes to help. They begin therapeutic relationships and can instantly build trust’
Kate Granger, hellomynameis.org.uk, #hellomynameis
Be friendly. Smile. Sit down. Take an interest in the patient and ask an unscripted question. Use the patient’s name more than once.
Listen. Do not be the average physician who interrupts after 20–30 seconds.
‘Look wise, say nothing, and grunt. Speech was given to conceal thought.’
William Osler (1849–1919).
Increase the wait-time between listening and speaking. The patient may say more.
Pay attention to the non-verbal. Observe gestures, body language, and eye contact. Be aware of your own.
Explain. Consider written or drawn explanations. When appropriate, include relatives in discussions to assist in understanding and recall.
Adapt your language. An explanation in fluent medicalese may mean nothing to your patient.
Clarify understanding. ‘Acute’, ‘chronic’, ‘dizzy’, ‘jaundice’, ‘shock’, ‘malignant’, ‘remission’: do these words have the same meaning for both you and your patient?
Be polite. It requires no talent.
‘Politeness is prudence and consequently rudeness is folly. To make enemies by being…unnecessarily rude is as crazy as setting one’s house on fire.’
Arthur Schopenhauer (1788–1860).
Address silent fears. Give patients a chance to raise their concerns: ‘What are you worried this might be?’, ‘Some people worry about. …, does that worry you?’
Consider the patient’s disease model. Patients may have their own explanations for their symptoms. Acknowledge their theories and, if appropriate, make an effort to explain why you think them unlikely.
‘A physician is obligated to consider more than a diseased organ, more even than the whole man - he must view the man in his world.’
Harvey Cushing (1869–1939).
Keep the patient informed. Explain your working diagnosis and relate this to their understanding, beliefs, and concerns. Let them know what will happen next, and the likely timing. ‘Soon’ may mean a month to a doctor, but a day to a patient. Apologize for any delay.
Summarize. Is there anything you have missed?
Communication, partnership, and health promotion are improved when doctors are trained to KEPe Warm:10
• Knowing—the patient’s history, social talk.
• Encouraging—back-channelling (hmmm, aahh).
• Physically engaging—hand gestures, appropriate contact, lean in to the patient.
• Warm up—cooler, professional but supportive at the start of the consultation, making sure to avoid dominance, patronizing, and non-verbal cut-offs (ie turning away from the patient) at the end.
Shared decision-making: no decision about me, without me
Shared decision-making aims to place patients’ needs, wishes, and preferences at the centre of clinical decision-making.
• Support patients to articulate their understanding of their condition.
• Inform patients about their condition, treatment options, benefits, and risk.
• Make decisions based on mutual understanding.
Consider asking not, ‘What is the matter?’ but, ‘What matters to you?’.
Consider also your tendency towards libertarian paternalism or ‘nudge’. This is when information is given in such a way as to encourage individuals to make a particular choice that is felt to be in their best interests, and to correct apparent ‘reasoning failure’ in the patient. This is done by framing the information in either a positive or negative light depending on your view and how you might wish to sway your audience. Consider the following statements made about a new drug which offers 96% survival compared to 94% with an older drug:
• More people survive if they take this drug.
• This new drug reduces mortality by a third.
• This new drug benefits only 2% of patients.
• There may be unknown side-effects to the new drug.
How do you choose?
Consult the bnf or bnf for Children or similar before giving any drug with which you are not thoroughly familiar.
Check the patient’s allergy status and make all reasonable attempts to qualify the reaction (table 1.2). The burden of iatrogenic hospital admission and avoidable drug-related deaths is real. Equally, do not deny life-saving treatment based on a mild and predictable reaction.
Table 1.2 Drug reactions
Type of reaction
All medications have side-effects. The most common are rash, itch, nausea, diarrhoea, lethargy, and headache
Due to inter-individual variance. Dosage regimen normally corrects for this but beware states of altered drug clearance such as liver and renal (p[link]) impairment
Reaction due to drugs used in combination, eg azathioprine and allopurinol, erythromycin and warfarin
Check drug interactions meticulously.
Remember primum non nocere: first do no harm. The more minor the illness, the more weight this carries. Overall, doctors have a tendency to prescribe too much rather than too little.
Consider the following when prescribing any medication:
1 The underlying pathology. Do not let the amelioration of symptoms lead to failure of investigation and diagnosis.
2 Is this prescription according to best evidence?
3 Drug reactions. All medications come with risks, potential side-effects, inconvenience to the patient, and expense.
4 Is the patient taking other medications?
5 Alternatives to medication. Does the patient really need or want medication? Are you giving medication out of a sense of needing to do something, or because you genuinely feel it will help the patient? Is it more appropriate to offer information, reassurance, or lifestyle modification?
6 Is there a risk of overdose or addiction?
7 Can you assist the patient? Once per day is better than four times. How easy is it to open the bottle? Is there an intervention that can help with medicine management, eg a multi-compartment compliance aid, patient counselling, an it solution such as a smartphone app?
8 Future planning. How are you going to decide whether the medication has worked? What are the indications to continue, stop, or change the prescribed regimen?
The placebo effect
The placebo effect is a well-recognized phenomenon whereby patients improve after undergoing therapy that is believed by clinicians to have no direct effect on the pathophysiology of their disease. The nature of the therapy (pills, rituals, massages) matters less than whether the patient believes the therapy will help.
Examples of the placebo effect in modern medicine include participants in the placebo arm of a clinical trial who see dramatic improvements in their refractory illness, and patients in severe pain who assume the saline flush prior to their iv morphine is opioid and reporting relief of pain before the morphine has been administered. It is likely that much of the symptomatic relief experienced from ‘active’ medicines in fact results from a placebo effect.
The complementary therapy industry has many ingenious ways of utilizing the placebo effect. These can give great benefits to patients, often with minimal risk; but there remains the potential for significant harm, both financially and by dissuading patients from seeking necessary medical help.
Why evolution has given us bodies with a degree of self-healing ability in response to a belief that healing will happen, and not in response to a desire for healing, is unclear. Perhaps the belief that a solution is underway ‘snoozes’ the internal alarm systems that are designed to tell us there is a problem, and so improve the symptoms that result from the body’s perception of harm.
Many patients who receive therapies are unaware of their intended effects, thus missing out on the narrative that may give them an expectation of improvement. Try to find time to discuss with your patients the story of how you hope treatment will address their problems.
The ward round
• All entries on the patient record must have: date, time, the name of the clinician leading the interaction, the clinical findings and plan, your signature, printed name, and contact details. Make sure the patient details are at the top of every side of paper. Write legibly—this may save more than the patient.
• A problem list will help you structure your thoughts and guide others.
• BODEX: Blood results, Observations, Drug chart, Ecg, X-rays. Look at these. If you think there is something of concern, make sure someone else looks at them too.
• Document what information has been given to the patient and relatives.
• Make sure you know when and where to attend.
• Make sure you understand what you need to do and why. ‘Check blood results’ or ‘Review warning score’ is not enough. Better to: ‘Check potassium in 4 hours and discuss with a senior if it remains >6.0mmol/L’.
• Write it down.
• The abcde approach (p[link]) to a sick patient is never wrong.
• Try and establish the clinical context of tasks you are asked to do. Prioritize and let staff know when you are likely to get to them.
• Smile, even when talking by phone. Be polite.
• Eat and drink, preferably with your team.
Making a referral
• Have the clinical notes, observation chart, drug chart, and investigation results to hand. Read them before you call.
• Use SBAR: Situtation (who you are, who the patient is, the reason for the call), Background, Assessment of the patient now, Request.
• Anticipate: urine dip for the nephrologist, pr exam for the gastroenterologist.
Resilience and coping
‘Burnout’ is common in clinical medicine. It is a syndrome of lost enthusiasm, reduced empathy, increased cynicism, and a decrease in the meaningfulness of work. Coping styles and resilience can protect doctors and better equip them to meet, and learn from, the challenges of clinical practice:12
• Self-directedness correlates strongly with resilience. A personal sense of responsibility allows learning from mistakes and moving on.
• Cooperativeness is the ability to work with opinions and behaviours different to your own, preventing them becoming a source of stress.
• Clinicians who are low in harm avoidance are better able to accept uncertainty and a degree of risk. This facilitates decision-making as it is unclouded by anxiety and pessimism about potential problems. Supervised experience outside your comfort zone may help you deal better with uncertainty.
• Be persistent but set realistic goals. Perfectionism can be detrimental.
• Task-orientated coping occurs when a situation is seen as changeable. This is associated with less burnout than emotion-orientated coping when situations are considered unchangeable: don’t just do something, stand there.
• Be self-aware. Development or modification of your personality traits may reduce your vulnerability.
Would you be surprised if your patient were to die in the next few days, weeks, or months? If the answer is ‘no’ then end-of-life choices, decisions, and care should be addressed.
Consider: decline in functional performance, eg in bed or chair >50% of day, increasing dependence, weight of co-morbidity, unstable or deteriorating symptom burden, decreased treatment response, weight loss >10% in 6 months, crisis admissions, serum albumin <25g/L, sentinel event, eg fall, transfer to nursing home.
Over 6000 people are waiting for an organ transplant in the uk and approximately 1000 people in need of a transplant will die each year (see p[link]).
Any patient who is a potential donor can be referred to a specialist organ donation service. That service will provide advice as to suitability for transplantation and will coordinate the approach to families. They are contactable 24 hours a day and their details will be held in your a&e and/or itu departments.
Organs can be retrieved from:
• donor after brainstem death or heart-beating donor.
• donor after cardiac death or non-heart-beating donor. Includes death following unsuccessful cpr and patients for whom death is inevitable but do not meet the criteria for brainstem death.
There are two legislative frameworks for organ donation:
• Opt-in. Donors give their explicit consent
• Opt-out. Anyone who has not actively refused consent is a donor.
The association between an opt-out system and higher organ donation rates is complicated by the presence of multiple cofounding factors. Non-legislative change including national coordination, support and training of clinicians, routine discussion as part of end-of-life care, and efficient organ retrieval also increase donation rates. The ethics of presumed consent should also be considered: the absence of an objection would not be an acceptable substitute for informed consent in other areas of clinical practice.
In the uk, although consent for transplantation rests with the deceased, if the patient’s family or representative cannot support donation then it will not go ahead. Register your decision on the nhs Organ Donor Register (https://www.organdonation.nhs.uk/register-to-donate/register-your-details/) and more importantly, let your family know your wishes.
‘Our clinical practice is steered by ethical principles. They guide the decisions we make in our clinics and ward rounds, what we tell our patients, and what we omit to tell them.’
Tony Lopez, Journal of Royal Society of Medicine 2001;94:603–4.
In the silences of our consultations it is we who are under the microscope, and we cannot escape our destiny in the sphere of ethics. To give us courage in this enterprise, we can recall the law of the aviator and seagull: it is only by facing the prevailing wind that we can become airborne, and achieve a new vantage point from which to survey our world. We hope for moral perception: to be able to visualize the morally salient features of a situation. For without this, ethical issues may float past never to be resolved. Be alert to words which may carry hidden assumptions: ‘futility’, ‘consent’, ‘best interests’.16 Consider WIGWAM in your routine patient review:
• Wishes of the patient: are they known or unknown?
• Issues of confidentiality/disclosure.
• Goals of care: are they clear? Whose are they: yours or the patient’s?
• Wants: to decline treatment or discharge against advice.
• Arguments between family/friends/doctors.
• Money: concerns of the patient, concerns of the healthcare provider.
Self governance: the ability of a patient to make a choice based on their own values and beliefs.
The obligation to benefit patients. Links with autonomy as benefit is dependent upon the view of the patient.
Do not harm. Or more appropriately, do no overall harm: you should stick a needle into someone when they need dialysis.
Four quadrants method
Identify the clinical problem, treatment options, goals of treatment, and likelihood of success.
What is the patient’s autonomous decision? (And is the patient capable of making one? If not, look for previously expressed wishes from advanced directives, family, friends, gp.)
Quality of life:
How will the proposed treatment affect quality of life? This is subjective: recognize your own biases and accommodate those of the patient.
The wider context: legal, cultural, religious, familial, and anything else that may impact.
These frameworks describe individual voices within the ethics choir. Sometimes there is a beautiful harmony, but how should you act when there is discordance? There is no hierarchy within the frameworks. Each component is binding unless it is trumped by a stronger principle. How you weigh up and balance the ethical components of a situation is not easy, but it should be clear and justified. Know the patient. Consult others, especially those who hold different opinions to yourself. Can you adequately defend your decision to the patient? Their family? Your consultant? Another consultant? A lawyer? If an investigative journalist were to sit on a sulcus of yours, having full knowledge of all thoughts and actions, would he be composing vitriol for tomorrow’s newspapers? If so, can you answer him, point for point?
‘Body and soul cannot be separated for purposes of treatment, for they are one and indivisible. Sick minds must be healed as well as sick bodies.’
C Jeff Miller, 1931.
Mental state examination:
• Appearance and behaviour: dress, hygiene, eye contact, rapport.
• Speech: volume, rate, tone.
• Emotion: mood (subjective and objective), affect (how mood is expressed with behaviour—appropriate or incongruent?).
• Perception: hallucinations—auditory (in the second or third person)?, visual?
• Form: block, insertion, broadcast, flight of ideas, knight’s move.
• Content: delusions, obsessions, phobias, preoccupations, self-harm, suicide.
• Insight: ask the patient why they have presented today.
• Cognition: orientation, registration, recall, concentration, knowledge.
Do not be afraid to ask about suicidal thoughts and plans. Remove yourself from the situation if you feel threatened.
Two questions can be used to identify depression:19
1 During the last month, have you been bothered by feeling down, depressed, or hopeless?
2 During the last month, have you often been bothered by having little interest or pleasure in doing things?
If a person answers ‘yes’ to either question they should undergo mental health assessment including a risk assessment of self-harm and suicide. Appropriate treatments include psychosocial intervention (guided self-help, cognitive behavioural therapy, structured physical activity) and medication. Treatment choice depends on disease severity, previous psychiatric history, response to treatment, and patient preference. If medication is indicated, a generic ssri should be considered first line after consideration of gi bleeding risk, drug interactions, toxicity, overdose, and discontinuation symptoms. The full effect of medication is gradual, over 4–6 weeks.
The Mental Capacity Act (mca) 2005 has a two-stage test for lack of capacity:
1 There is an impairment or disturbed functioning of the mind.
2 The patient is unable to make a decision.
Decision-making is impaired if the patient is unable to: understand the relevant information, retain it for long enough to make a decision, weigh up the information, communicate their decision. Capacity is decision-specific not patient-specific. When treatment is proposed to those who lack capacity, a capacity advocate should be provided. Even patients without capacity should be as involved as possible in decision-making.
Mental Health Act and common law
A patient can be detained under common law (subject to a test of reasonableness) or under the mha, only if they lack capacity to remain informally and are a danger to themselves or others. You will have more experience in verbal and non-verbal communication, than in detention under the mha, so use these skills first to try and de-escalate the situation. If rapid tranquillization is needed, be familiar with dosage, side-effects, and the need for ongoing observation. If there is no history to guide choice of medication, intramuscular lorazepam can be used.20
Doctors and mental health
Suicide rates are three times higher in doctors compared to the general population. Up to 7% of doctors will have a substance abuse problem within their lifetime. Do not ignore feeling low, poor concentration, and reduced energy levels. Do not self-diagnose and manage. Avoid ‘corridor consultations’. Trust your gp. Seek support:
‘To know how to grow old is the master-work of wisdom, and one of the most difficult chapters in the great art of living.’
Henri Amiel, Journal Intime, 21 Sept 1874.
Ageing is an inevitable and irreversible decline in organ function that occurs with time, in the absence of injury or illness, and despite the existence of complex pathways of maintenance and repair.
Healthy ageing is the maintenance of physical and mental abilities that enable wellbeing and independence in older age.
Do not presume ageing. Look for preventable and reversible pathology. Old age does not cause disease (although it can increase vulnerability and recovery time).
Look for ways to reduce disability and support older people in their own homes.
Differences in the evaluation of the older person
1 Multiple pathologies: Elderly patients have, on average, six diagnosable disorders. Effects may be multiplicative. Treatment must be integrated.
2 Multiple aetiologies: One problem may have several causes, eg falls. Treating each alone may do little good, treating all may be of great benefit.
3 Non-specific/atypical presentation: Delirium, dizziness, falls, mobility problems, weight loss, and incontinence can be due to disorders in more than one organ system. Typical signs and symptoms may be absent. Ask about functional decline in activities of daily living—this may be the only symptom.
4 Missed or delayed diagnosis: The older person may decline quickly if treatment is delayed. Complications are common. Use a collateral history: what is the patient usually like?
5 Pharmacy and polypharmacy: nsaids, anticoagulants, anti-parkinson drugs, hypoglycaemic drugs, and psychoactive drugs can pose a particular risk in the older patient. Double check for interactions. Consider body weight, liver and renal function—drug doses may need to be modified. The stopp/start criteria detail >100 potentially inappropriate prescriptions and prescribing omissions relevant to the older patient.21
6 Prolonged recovery time: Anticipate and plan for this. Don’t forget nutrition.
7 Rehabilitation and social factors: Essential for healthy ageing.
50% aged >80 will fall at least once per year. Falls23 lead to injury, pain, distress, loss of confidence, loss of independence, and mortality. Cost to the nhs is £2.3bn/year.
• History: frequency, context and circumstances, severity, injuries.
• Multifactorial risk assessment: gait, balance, muscle strength, osteoporosis risk, perceived functional ability, fear of falls, vision, cognition, neurological examination, continence, home and hazards, cardiovascular examination, medication review.
• Interventions: strength and balance training, home hazard intervention, correct vision, modification/withdrawal of medication (cardiovascular, psychotropic), integrated management of contributing morbidities. Consider barriers to change, eg fear, patient preference.
Pre-existing conditions and non-obstetric disease cause more maternal deaths in the uk than obstetric complications.24
Pregnant women should receive the same investigations and treatment as non-pregnant patients, with avoidance of harm/potential harm to the fetus whenever possible.
Most mistakes made in the medical management of pregnant women are due to acts of omission caused by inappropriate weighting of risk and benefit.
Physiological changes in pregnancy
Clinical assessment in pregnancy requires knowledge of the physiological changes associated with the gravid state. Expected changes and guidance on when to investigate for possible underlying pathology is given in table 1.3.
Table 1.3 Physiology and pathology in pregnancy
↓bp before 20 weeks’ gestation
Diastolic bp >80mmHg in 1st trimester
↑ Heart rate
Sustained tachycardia >100/min
Compensated respiratory alkalosis
Serum bicarbonate <18mmol/L
No change in pefr
Decrease in pefr
↑Respiratory rate by 10%
Respiratory rate >20/min
↑gfr and creatinine clearance
Creatinine >85μmol/L (egfr not valid in pregnancy)
Protein:creatinine ratio >30mg/mmol
Altered glucose handling
Fasting glucose >5.0mmol/L
Hb <10.5g/dL, platelets <100×109/L
If the uterus is positioned outside the imaging field of view, the radiation dose to the conceptus is minimal. Exposure from the following investigations is well below the threshold of risk to the fetus:
• Plain radiograph: chest, extremities, spine.
• ct: head, chest (but consider radiation to maternal breast in pregnancy/lactation).
Ultrasound and mri are preferentially used when imaging the abdomen.
Reassure your pregnant patient that a chest x-ray is safe. It is the equivalent of 3 days of background radiation. Do not presume it is not required—how else will you pick up the widened mediastinum as a cause for her chest pain?
Table 1.4 Drugs in pregnancy
Trimethoprim (1st trimester)
nsaids (3rd trimester)
Treatment for asthma: salbutamol, ipratropium, aminophylline, leukotriene antagonists
Live vaccines (mmr, bcg, Varicella)
‘The work of epidemiology is related to unanswered questions, but also to unquestioned answers.’
Patricia Buffler, North American Congress of Epidemiology, 2011.
Who, what, when, where, why, and how?
Epidemiology is the study of the distribution of clinical phenomena in populations.It analyses disease in terms of host, agent, and environment (the ‘epidemiologist’s triad’). It elucidates risks and mechanisms for the development of disease, and reveals potential targets for disease prevention and treatment. Epidemiology does not look at the individual patient, but examines a defined population. How applicable its findings are depend upon how well the sample population mirrors the study population, which must, in turn, mirror the target population. Does your patient fit in this ‘target’? If ‘yes’, then the epidemiological findings may be applicable.
Measures of disease frequency
Incidence proportion is the number of new cases of disease as a proportion of the population. Synonyms include probability of disease, cumulative incidence, risk.
Incidence rate is the number of new cases per unit of person-time, ie one person observed for 5 years contributes 5 person-years of follow-up.
Prevalence is the number of cases that exist at a given time (point prevalence) or time-frame (period prevalence), divided by the total population being studied. For example, the lifetime prevalence of hiccups is ~100% and incidence is millions/year. However, the point prevalence at 3am may be 0 if no one is actually having hiccups.
Comparisons of outcome frequency
Differences in outcome rates between populations point to an association between the outcome and factors distinguishing the populations (eg a smoking population compared to a non-smoking population). Challenges arise as populations tend to differ from each other in many ways, so it may not be clear which factor(s) affect outcome frequency. This leads to confounding. For example, we might find that heart disease is more common in those who use walking sticks. But we cannot conclude that walking sticks cause heart disease as age is a confounding factor: age is causal, not walking sticks.
Ways of accounting for associations:
a may cause b (antacids cause cancer), b may cause a (cancer causes antacid use), a 3rd unknown agent x (eg age) may cause a and b, or the association may be a chance finding. When considering the options, it is useful to bear in mind the Bradford Hill ‘criteria’ for causation (nb he did not claim any were essential):
1 Consistency of findings: among different populations, studies, time periods.
2 Temporality: the effect must occur after the cause.
3 Biological gradient: a dose response whereby more exposure = more effect.
4 Specificity: exposure causes a single outcome (smoking does not conform!).
5 Strength of association: strong associations are more likely to be causal.
6 Biological plausibility: there is a mechanism linking cause and effect.
7 Coherence: the relationship is supported by current disease knowledge.
8 Experiment: does removal of exposure reduce outcome frequency?
Studies should be designed to give an adequate answer to a specific research question. Samples need to be representative and of sufficient size to answer the question.
Outcome rates are examined in different populations, eg trend over time, geographically distinct groups, social class. Populations rather than individuals are the unit of study.
Longitudinal (cohort) studies:
Subjects are followed over time with measurement of exposure and outcome.
Patients with the outcome of interest are identified and past exposure is assessed in comparison to ‘controls’ who did not develop the outcome. Cases and controls should be adequately matched for other factors that may affect outcome, or these differences should be corrected for (mathematical assumption).
In a randomized controlled trial (rct), participants are allocated to an intervention/exposure (eg new drug treatment) or no intervention (eg placebo, standard care) by a process which equates to the flip of a coin, ie all participants have an equal chance of being in either arm of the study. The aim is to minimize bias and attempt to get at the truth as to whether the intervention is any good or not. Both groups are followed up and analysed against predefined end-points.
Done with the aim of eliminating the effects of non-studied factors. With randomization (and sufficient study size) the two arms of the study will be identical (on average), with the exception of the intervention of interest.
There is a risk that factors during the trial may affect the outcome, eg participant or clinician optimism if they know the patient is on active treatment, or an unwillingness to expose more severe disease to placebo. If the subject does not know which intervention they are having, the trial is single-blind. Ideally, the experimenter should not know either, and the study should be double-blind.
In a good trial, the blind lead the blind.
When a randomized controlled trial might not be the best method
• Generating new ideas beyond current paradigms (case reports).
• Researching causes of illnesses and prognoses (cohort studies).
• Evaluating diagnostic tests (cohort study and decision model).
• Where the researcher has no idea of the effective dose of a drug (dose-ranging adaptive design).
• When recruiting of patients would be impossible or unethical.
• When personalized medicine is the aim, eg treatments matched to patients’ biomarker profiles (adaptive design, cohort study).
In the end, all randomized trials have to submit to the ultimate test when the statistical collides with the personal: ‘Will this treatment help me?’, ‘Will this procedure help you?’ No randomized trial is complete until real-life decisions taken in the light of its findings are scrutinized. Remember Osler: ‘no two individuals react alike and behave alike under the abnormal conditions which we know as disease. This is the fundamental difficulty of the physician’. Do not ask for definitive trials: everything is provisional.
‘When you can measure what you are speaking about, and express it in numbers, you know something about it; but when you cannot measure it, when you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory kind.’
Lord Kelvin, 1883.
Comparisons between ‘exposed’ and ‘unexposed’ populations are made in terms of the risk or likelihood of an outcome This can be appreciated by plotting a 2×2 table (table 1.5).
Table 1.5 2×2 table analysis
• Absolute risk difference (attributable risk) = disease frequency in exposed minus the disease frequency in unexposed. Example (table 1.5): (70/90) − (120/570) = 0.57 ∴ exposure increases risk by 57%.
• Relative risk = ratio of outcome in exposed population compared to unexposed. Relative risk of 1 means risk is same in both populations. Relative risk >1 means exposure increases risk. Relative risk <1 means exposure lessens risk, eg vaccination. Example (table 1.5): (70/90)÷(120/570) = 3.69 ∴ risk is 3.69 × higher with exposure.
• Odds ratio = ratio of the probability of an outcome occurring compared to the probability of an outcome not occurring. Example (table 1.5): (70/20)÷(120/450) = 13.13 ∴ odds of outcome are 13.13 × higher with exposure.
Relative risk is easier to interpret than odds ratio but relies on a meaningful prevalence/incidence. For the individual, absolute risk difference may be most relevant.
The validity of a test which dichotomizes study participants can be assessed by examining the results from the test against a standard reference (or outcome: did the participant actually have the disease?) (see table 1.6).
Table 1.6 Table of possible test results
Patient has condition
Patient does not have condition
True positive (tp)
False positive (fp)
False negative (fn)
True negative (tn)
Sensitivity tp/(tp + fn) = of those with the condition, how many test positive? A sensitive test is able to correctly identify those with the disease.
Specificity tn/(tn + fp) = of those who do not have the condition, how many test negative? A specific test is able to correctly identify those without a disease.
‘Do they have abdominal pain?’ as a test for appendicitis will have ↑sensitivity (most cases have pain), but ↓specificity (many patients with pain do not have appendicitis).
Positive predictive value (tp/(tp + fp) indicates how likely it is that someone with a positive test result has the condition.
Negative predictive value (tn/(tn + fn) indicates how likely it is that someone with a negative test result does not have the condition. When you receive a test result, you need to know how likely it is to be correct.
Number needed to treat (nnt) is a useful way of reporting the results of randomized clinical trials. It is the reciprocal of the absolute risk difference: 1 ÷ arr.
A large treatment effect means that fewer patients need to receive treatment in order for one to benefit. It is specific to the chosen comparator (eg placebo or usual care), the measured outcome (eg death, blood pressure fall), and the duration of treatment follow-up used in the study. Look carefully at the details of the question that the nnt is attempting to quantify.
• Advantages: easily calculated, single numerical value for efficacy, can be used to examine harm (becomes the number needed to harm).
• Disadvantages: confidence intervals are difficult when the differences between treatments are not significant.
ebm is the conscientious and judicious use of current, best research evidence to optimize management plans and integrate them with patients’ values by:
1 Asking answerable questions.
2 Finding the best information.
3 Appraising the information for quality, validity, and relevance.
4 Dialogue to find out what the patient wants.
5 Applying data to patient care.
The amount of evidence
More than 2 million new biomedical papers are published each year including >20 000 new randomized trials. Patients benefit directly from a tiny fraction of these papers. How do we find them?
• A hierarchy of evidence (fig 1.3) is used to identify the best research available to answer our question.
• Specialist ebm journals, eg Evidence-based Medicine, appraise published information for quality, relevance, and interest on our behalf.
• The Cochrane Collaboration gathers and summarizes best evidence, free from commercial sponsorship and conflicts of interest. >37 000 researchers from 130 countries contribute.
• The concept of scientific rigour is opaque. What do we want? The science, the rigour, the truth, or what will be most useful to patients? These may overlap, but they are not the same. Can average cohort results inform clinical decisions on an individual level (especially in the context of comorbidity)?
• Can we really appraise all the evidence?27 We are hindered by publication bias. Around half of all clinical trials remain unpublished. See www.alltrials.net for the campaign to register all trials, and ensure methods and summary results are available.
• Evidence can be expensive. Who paid the bill and what is their vested interest?
• Is the result clinically significant? What is the level of benefit to the individual, as opposed to the population? Is the ebm tail wagging the clinical dog?
• How is our innate hierarchy of evidence constructed? Do we maintain the same standard of the evidence for all changes to our practice?
• Have you checked the correspondence columns in journals from which winning papers are extracted? It may take years for unforeseen flaws to surface.
• There is a danger that by always asking, ‘What is the evidence?’, we will divert resources from hard-to-prove areas (eg psychosocial interventions).
• ebm is never 100% up to date and reworking meta-analyses takes time and money. Specialists may ostensibly reject a new trial due to a tiny flaw, when the real dread is that it might flip their once-perfect formulation.
• ebm lies uncomfortably in a world of clinical intuition and instinctual premonition. Yet these instincts may be vital.
• If ebm is prescriptive, patient choice declines. Does our zeal for ebm make us arrogant, mechanical, and defensive? Where is the shared decision-making (p[link])?
• By focusing on answerable questions, ebm can distract us from our patients’ unanswerable questions; questions that still require time and acknowledgement.
The practice of ebm must be informed by clinical judgement and compassion.
Using Illness as a Metaphor,8 Susan Sontag describes two kingdoms: that of the well, and that of the sick. She describes our dual citizenship, and the use of a passport to travel from one kingdom to the other. But medicalization blurs this distinction. The boundary between the ‘Kingdom of the Sick’ and the ‘Kingdom of Well’ is lost and there is an anschluss of healthy people annexed into the potentially predatory and frightening kingdom of the sick from which there may well be no escape.
‘Too much medicine’ occurs as a result of:
• Overdiagnosis: Labelling an (asymptomatic) person as ‘sick’ despite the fact that subsequent treatment, lifestyle advice, or monitoring provides no benefit to their outcome (and potentially causes harm), eg non-progressive breast cancer.
• Overdetection: Increasingly sensitive tests identify pathology that is indolent or non-progressive, eg subsegmental pulmonary emboli diagnosed on ct angiography.
• Overdefiniton: Expansion of disease definitions or lowering of disease thresholds, eg an egfr diagnosis of chronic kidney disease now means that 1 in 8 adults are labelled with the disease, many of whom will never progress to symptomatic kidney failure; 15% of pregnant women now have subclinical hypothyroidism without evidence that thyroxine replacement is beneficial (2016).
• Disease mongering: The creation of pseudodiseases which pose no threat to health, eg restless legs, sexual health dysfunction, multiple chemical sensitivity.
• Overutilization: Healthcare practice that provides no net benefit, eg routine mri for lower back pain.
• Overtreatment: Treatment that is of no benefit (and may cause harm), eg antibiotics for viral infections, polypill for the population.
Too much medicine arises from the fear of missing a diagnosis, and concern about avoidable morbidity or mortality. A punitive society means there is a perceived need for more tests, to seek more certainty. But certainty is the holy grail of myth and legend. The individual patient is a unique set of symptoms, stoicism, experience, and need. And by the nature of life, all cure can only ever be temporary.
choosing wisely is an initiative to change doctors’ practice away from interventions that are not:
• supported by evidence
• free from harm
• truly necessary (including duplicative tests).
The top 5–10 interventions that should not be used routinely are given for each specialty. Search for those relevant to your current post at: www.choosingwisely.org/doctor-patient-lists/.
We thank Dr Kate Mansfield, our Specialist Reader, for her contribution to this chapter.
4 Sebastian Faulkes, Human Traces, 2005.
5 Milan Kundera, The Unbearable Lightness of Being, 1984.
6 Philoctetes by Sophocles 409 bc (translation Phillips and Clay, 2003).
1 This is unlikely to be a commentary on euthanasia (easeful death) as the oath predates the word. Rather, it is believed to allude to the common practice of using doctors as political assassins.
2 Abortion by oral methods was legal in ancient Greece. The oath cautions only against the use of pessaries as a potential source of lethal infection.
3 The oath does not disavow surgery, merely asks the physician to cede to others with expertise.
7 The Man Who Shot Liberty Valance, 1962: Ransom Stoddard (Jimmy Stewart) becomes a legend after killing Liberty Valance in a duel. It does not matter that the real shooter was Tom Doniphon (John Wayne) all along.
8 Susan Sontag, Illness as a Metaphor, 1978