Can Medicine Be Cured

by Seamus O'Mahony

  It is in everyone’s interest to solve this. I am not optimistic, however, that this can be fixed because it would require the confluence of several events. The clinicians would have to wrest power back from the managers. They would have to negotiate a new concordat with the nursing profession. Medical litigation would have to change to a non-adversarial ‘no-fault’ compensation model. Regulatory bodies such as the General Medical Council would have to be radically reformed. Accountability would have to be realigned with power. The profession would need to open a dialogue with their patients, with society, with politicians, with the media about what medicine can and cannot do. How might such reforms be achieved? I am not optimistic that the coalface clinicians will suddenly rise up and cast off their shackles. The professional bodies are sclerotic, their only concern being self-preservation. Managers will not willingly hand over power; it will have to be taken from them. The incentives for the present system to continue are just so much greater than the incentives to reform. Most of all, medicine needs leadership, not limp-wristed virtue signalling. It is difficult to know where this is going to come from. Perhaps it is a problem for the next generation to solve.

  I teach the next generation. At a recent tutorial with final-year medical students, it occurred to me that the ward where I was teaching was the same ward where I had worked as an intern in 1984. I told my tutorial group that, in that prelapsarian Eden, all my patients were on one ward – not scattered all over the hospital, or on trolleys in the emergency department. When my boss – the professor of medicine – did a ward round, he was accompanied throughout by the senior ward sister, and the round concluded in her office over tea. Perhaps the many intervening years have given an excessively rosy glow to my memories, but when I compared it to the kind of round I do now, it seems like a very heaven. Following my last weekend on-call for General Medicine, I had over fifty patients under my care, scattered across fifteen locations within the hospital. I start at 7 a.m. so the junior doctors on-call the night before can go home at 9 a.m. All of these rounds start in the emergency department, where I have ten patients: four in cubicles and six on trolleys in the corridors. Since it is so early, the corridors are dark and many of the patients are asleep. Trying to examine an elderly patient on a trolley in the corridor is quite a challenge, with the noise, the lack of privacy and the passing traffic. With each patient, the routine is the same. I get a quick summary from the junior doctor who admitted them, but very often this doctor was on an earlier shift, and has gone home. I look at the patient and decide if they are sick, or not. I write a brief note in the chart, as my conclusions are so often different from the admitting junior doctor’s. I check the drug chart to see what has been prescribed. The protocols for sepsis and venous thromboembolism ensure that several patients are on intravenous antibiotics who shouldn’t be, and nearly all are on an anticoagulant. I see a man in the emergency department who has had a major stomach haemorrhage. He needs an urgent endoscopy, but this cannot happen until he is allocated a bed. A twenty-two-year-old with liver cirrhosis (due to alcohol) has been admitted with jaundice and fluid retention. He will be dead within a year. A man with advanced dementia in long-term psychiatric care has been sent in with a chest infection. Because he coughs and splutters when swallowing (this is long-standing) the nurses have him assessed by a speech and language therapist who concludes that the man has an ‘unsafe’ swallow, and cannot be allowed to eat or drink. I reverse this order, to the great relief of the patient and the annoyance of the therapist. It takes several hours to see the patients, partly because they are so scattered – this is called a ‘safari’ ward round. On most wards, the nurses do not accompany us, so much vital information is lost or never relayed. Over the weekend, I am obliged to join morning conference calls with the clinical director and nurse managers. Much of the discussion involves potential breaches of targets for patients accommodated on trolleys in the emergency department.

  I eventually finish at 12.30 p.m., five-and-a-half hours after starting in the emergency department. Most of my patients are elderly people, many from nursing homes. Most are on several long-term medications – some are taking as many as twenty – all, no doubt, prescribed by their well-meaning GPs, who are simply following guidelines. A few of these elderly patients are under my care at least in part because of a side effect of one of these drugs. The day is one of intense pressure and concentration; clinical judgement is the principal value I bring to this round. My role is to identify the patients who have not been sorted out, whose admitting diagnosis is incorrect, who are not getting the right treatment. This judgement, in the world of protocolized, evidence-based medicine, is not highly regarded. And despite all the squalor, chaos, distraction and poor communication, the responsibility for all these patients is mine, and mine alone.

  The work of the junior doctors and nurses is ruled by protocols. The nurses have ‘early warning scores’ which trigger an urgent call to a doctor; most are false alarms. The junior doctors, meanwhile, are forced by protocols to prescribe antibiotics and anticoagulants to large numbers of patients who do not require them, and who may suffer harm. All of this induces a feeling of weariness and futility. The research laboratory seems a long way from this ward round, and evidence-based medicine, which led to the proliferation of these protocols, has become a sulking bully. Our patients are no longer people, but abnormal physiological and laboratory parameters, malfunctioning organs, and what Ivan Illich called ‘bundles of diagnoses’. The doctors, nurses and other health professionals have all done their job correctly and by protocol. Just as the medical–industrial complex sees sick people as malfunctioning machines, it regards doctors as standardized and protocolized functionaries. Patients are a problem to be processed by the hospital’s conveyor belt; it is hardly surprising that they often feel that nobody seems to be in charge, or cares about them as individuals. My own hospital is proud of its ‘patient flow’ initiative, which gives explicit institutional approval and branding to the factory model of health care, where sick people are a quantifiable input that must be processed into an output, with the turnover ever shorter. We are treating, but we are not healing.

  After the round, I retreat to my office for a sandwich, and catch up on my emails. There is an invitation to a workshop on ‘open disclosure’, an initiative aimed to diminish the fallout from medical error by a policy of admission of culpability from the beginning. I note that none of the four ‘facilitators’ of this workshop is a doctor; none has had to tell a patient and a family that they have made a mistake. Another email informs me that my attendance at a training session on hand-washing is compulsory. Yet another, from my professional indemnity organization, invites me to attend workshops on mastering difficult interactions with patients and shared decision-making. A message from one of the royal colleges invites me to attend a seminar on ‘leadership for clinicians’. I ponder the paradox of those shrewd doctors who escape the stresses of the clinical front line by teaching other doctors how to be leaders.

  12

  The McNamara Fallacy

  The main ideological difference between doctors and managers is that managers believe that health care can be run like a business, and that data are the key to success: they are ‘Dataists’. In a system such as the NHS, which is entirely state-funded, politicians have a duty to ensure that the service is accountable and answerable to that mythical figure, ‘the tax-payer’. This not unreasonable obligation led, over many decades, to an obsession with metrics – that which is measurable. This feverish monomania created the cancerous target culture that now pervades the NHS.

  Over-emphasis on metrics is often referred to as the McNamara fallacy. Robert Strange McNamara (1916–2009) was US Secretary of State from 1961 to 1968, during the presidencies of John F. Kennedy and Lyndon B. Johnson. His career was, by any standard, stellar: after graduating in economics at Berkeley, he took an MBA at the Harvard Business School, and became its youngest professor at the age of twenty-four. During the Second World Wa
r, McNamara served in the US military’s Department of Statistical Control. He applied the rigorous statistical methodology which he had learned at Harvard to the planning and execution of aerial bombing missions, achieving a dramatic improvement in efficiency. McNamara served in Europe and the Far East, where he assisted General Curtis LeMay in the planning of the fire-bombing of Japanese cities, when over 100,000 civilians perished in one night. After the war, the Ford Motor Corporation recruited several members of Statistical Control, including McNamara; these clever young men were nicknamed the ‘whiz kids’. The once-great company was then in disarray and losing money. McNamara and his fellow whiz kids applied their skills of rational statistical analysis to the problems of the ailing giant, achieving huge improvements, and returned Ford to profit. In 1960, at the age of forty-four, he was appointed president of the Ford Corporation, the first person outside the Ford family to achieve this distinction. After less than two months in this post, McNamara was offered a cabinet position by President-elect John F. Kennedy. He rejected the initial offer of treasury secretary, believing he was not qualified for this position, but accepted the post of defense secretary.

  McNamara used the same rigorous systemic analysis at the Pentagon that had worked so well at Ford, reducing costs and increasing efficiency. As the conflict in Vietnam escalated in the early 1960s, McNamara applied this quantitative approach to the prosecution of the war. He believed that as long as enemy casualties exceeded the numbers of US dead, the war would eventually be won: ‘Things you can count, you ought to count; loss of life is one.’ He set a figure of 250,000 enemy casualties as the ‘crossover point’, beyond which the North Vietnamese would be unable to replace their dead with new troops. By 1967, however, the US was no nearer to concluding the war, despite a massive increase in ground troops and aerial bombardment of North Vietnam. Public opposition to the war had grown: even McNamara’s son Craig, a student at Stanford, took part in an anti-war demonstration. McNamara concluded that the conflict was unwinnable, and wrote a memo to President Johnson and the joint chiefs of staff advising de-escalation and commencement of a peace settlement. Johnson didn’t reply to this memo, and in late 1967, eased McNamara out of the administration. ‘I wasn’t sure if I resigned or was sacked’, McNamara later wrote. In April 1968 he became president of the World Bank, a position he held until 1981. Clark Clifford, who succeeded McNamara as defense secretary, observed:

  Vietnam was not a management problem, it was a war, and war is about life and death, filled with intangibles that defy analysis. He [McNamara] had never been in a war, and perhaps he did not fully appreciate at first its stupid waste and its irrational emotions, and the elusiveness of facts and truth when men are dying. Nor did he fully understand the political roots of the conflict until it was too late. He had tried to master the war as he had everything else in his remarkable career, using pure intellect and his towering analytical skills – but Vietnam defied such analysis.

  Three years after McNamara’s resignation, the sociologist and political analyst Daniel Yankelovich (1924–2017) coined the phrase ‘The McNamara Fallacy’ (the Anglo-Irish author Charles Handy popularized the phrase in his 1994 book The Empty Raincoat, and is often erroneously credited as the originator of the concept):

  The first step is to measure whatever can easily be measured. This is OK as far as it goes. The second step is to disregard that which can’t be easily measured or to give it an arbitrary quantitative value. This is artificial and misleading. The third step is to presume that what can’t be measured easily really isn’t important. This is blindness. The fourth step is to say that what can’t be easily measured really doesn’t exist. This is suicide.

  Medicine is, and always has been, messy, imprecise and uncertain; the McNamara fallacy is the delusion that all of this complexity can yield itself to numerical analysis. This leads to over-reliance on crude metrics, such as hospital mortality rates, and the setting of arbitrary targets, many or most of which do not improve patient care, and some of which cause harm. Meanwhile, unquantifiable attributes, such as continuity of care and compassion, are neglected.

  Professor Sir Brian Jarman is British medicine’s Robert McNamara. In the 1990s, he developed a new metric, the Hospital Standardized Mortality Ratio (HSMR). This metric was the single biggest factor which precipitated the scandal at Stafford Hospital, still regarded as the greatest catastrophe in the history of the NHS. Even the name ‘Stafford’ is now used as a kind of short-hand to encapsulate everything that is wrong with hospital care in Britain. The scandal was created by a unique confluence of the McNamara fallacy, the target culture, managerialism, political and media opportunism, and common-or-garden unkindness. Jarman’s HSMR was widely and erroneously believed to measure ‘avoidable’ deaths in hospitals; that is, deaths directly caused by poor care. The Stafford scandal led to several official inquiries, including two by Sir Robert Francis QC. There was a further inquiry in 2013 by the NHS’s medical director, Professor Sir Bruce Keogh, into fourteen other hospitals with a high HSMR. The story of Sir Brian Jarman and his HSMR is a stark warning on how metrics mislead medicine.

  Jarman worked initially as a GP in London, but moved into academe and was appointed professor of primary health care at St Mary’s Hospital in 1984, and later head of the division of primary care at Imperial College. He became an eminent member of the British medical establishment: he was knighted in 1998, and elected president of the British Medical Association in 2003. Jarman was keen on statistics and health informatics and developed socio-economic indicators such as the Underprivileged Area Score, or Jarman Index. In the early 1990s, he turned his attention to hospital mortality rates, and developed the Hospital Standardized Mortality Ratio (HSMR). The formula for the Hospital Standardized Mortality Ratio is: HSMR = (Actual deaths/expected deaths) x 100. An HSMR of 100 therefore means the actual death rate was the same as the expected death rate. The cause of death was taken from the discharge ‘codes’: after the death of a patient, a hospital clerk – a ‘coding officer’ – goes through the notes and allocates a specific code for the main or primary diagnosis, and also codes for other diagnoses, or ‘co-morbidities’. These numerical codes are based on the international classification of disease (ICD). This coding is not always accurate, being subject to the clarity of the clinical notes and the vigilance of the coding officer. The health informatics expert Paul Taylor explained how the HSMR is determined:

  Data on actual deaths were taken from HES [Hospital Episode Statistics] and restricted to in-hospital deaths. A logistic regression model was used to calculate the risk of death for the 50 most common diagnoses, which account for over 80 per cent of admissions, based on a set of factors: sex, age, admission method (non-elective or elective), socio-economic deprivation quintile of the area of residence of the patient, diagnosis/procedure subgroup, co-morbidities, number of previous emergency admissions, year of discharge, month of admission and the source of admission, and the use of the ICD code for palliative care. Using data on the mix of these factors seen by each trust, they calculated the expected death rate.

  In 1999, Jarman joined the panel of the Kennedy Inquiry into the children’s heart surgery unit at Bristol Royal Infirmary. The scandal was a turning point in the history of the NHS: British medicine would never be quite the same. It started when Steve Bolsin, an anaesthetist at the hospital, expressed concern about the high mortality rate among babies and children undergoing cardiac surgery. The media reported a simplistic narrative of incompetent surgeons presiding over a blood-bath of dead babies. The truth was both more complicated and banal: the surgeons took on cases they should have sent elsewhere, and the cardiac surgery unit was under-staffed and underfunded. One of the key recommendations of the Kennedy Inquiry was that hospital mortality rates should be made more widely available. In the same year that Jarman joined the Bristol Inquiry, he and his colleagues at Imperial published a paper in the British Medical Journal entitled ‘Explaining differences in English hospital death rates us
ing routinely collected data’. Coming out in the middle of the Bristol scandal, this was hot stuff. This paper outlined the HSMR statistical methodology, and reported variations in mortality rates across hospitals in England. The paper had a political message, showing a strong link between hospital death rates and the ratios of doctors (both in hospitals and general practice) to number of population served. Jarman wrote to the secretary of state for health, Frank Dobson, asking if he could publish the HSMRs of various English hospitals; Dobson refused. In September 2000, Jarman met with two journalists, Tim Kelsey of the Sunday Times and Roger Taylor of the Financial Times. They had both covered the Bristol story, and saw an unexploited commercial potential in Jarman’s HSMR. Kelsey and Taylor co-founded Dr Foster Intelligence (DFI), whose first Good Hospital Guide appeared in 2001. DFI was established as a commercial enterprise, which would earn its income from the Good Hospital Guide (which published HSMRs and hospital league tables) and by selling its services to hospital trusts. It claims: ‘We are the leading provider of healthcare variation analysis and clinical benchmarking solutions worldwide.’ The former health secretary Alan Milburn was an enthusiastic supporter, and hospitals were told to collaborate with DFI whether they liked it or not. Milburn’s successor, Patricia Hewitt, bought the company in 2006 for £12m; this deal went through without a competitive tender. Roger Taylor is still director of research at Dr Foster, and describes himself as an ‘entrepreneur, journalist and author’ – ‘statistician’ is conspicuously absent from this list. Tim Kelsey is now CEO of the Australian Digital Health Agency. Taylor and Kelsey must be very grateful indeed to Jarman for giving them such a commercial opportunity. Confusingly, maddeningly and bizarrely, the academic unit that Jarman led at Imperial College was named the Doctor Foster Unit (DFU). Jarman has always claimed that the Doctor Foster Unit is completely independent of Dr Foster Intelligence: he admitted, however, to the second Francis Inquiry that 47 per cent of DFU’s funding came from DFI.