Does Having an Empathetic Doctor Improve Clinical Outcomes?

Expressions of empathy by a doctor can improve patient satisfaction.[1] And there is evidence that expression of empathy can be taught.[2] But does consultation with an empathetic doctor result in better symptom control than consultation with a doctor who is not as empathetic? A recent review of the topic found seven RCTs addressing this point.[3] The outcomes were measured on a continuous scale so they could be combined through their standardised mean difference in a meta-analysis. All point estimates were favourable and there was a statistically significant effect across all seven included studies. The difference of 0.42 of a standard deviate for pain was described as ‘moderate’, but sounds impressive to me. The study also included trials of ‘positive communication’ designed to engender good expectations of treatment effect. Again, modest effects were noted, but I am cautious about such an approach as it may topple over into dishonesty. Empathy is a different matter and this paper provides yet more evidence on how important it is. Doctors need to display warmth, consideration and appropriate affect – they have to make an effort and do ‘emotional work’. Building a relationship with patients is the essence of practice. Making the diagnosis and performing procedures is the easy bit for experienced doctors. But there is no ceiling to excellence in patient communication.

— Richard Lilford, CLAHRC WM Director


  1. Kim SS, Kaplowitz S, Johnston MV. The Effects of Physician Empathy on Patient Satisfaction and Compliance. Eval Health Prof. 2004; 27(3): 237-51.
  2. Lilford RJ. Is it Possible to Teach Empathy? NIHR CLAHRC West Midlands News Blog. 10 November 2017.
  3. Howick J, Moscrop A, Mebius A, et al. Effects of empathic and positive communication in healthcare consultations: a systematic review and meta-analysis. J Roy Soc Med. 2018.

The Slow March of Epidemiology: From Disease Causation to Treatment to Service Delivery

Traditional epidemiology was concerned with the causes of disease – many of the great medical discoveries, from malaria to the effects of smoking, can be credited to classical epidemiology. The subject continues to make great strides thanks to modern developments, such as genome-wide association studies and Mendelian randomisation. Approximately 70 years ago Austin Bradford Hill ushered in the days of clinical epidemiology.[1] Epidemiological methods were used to study the diagnosis and treatment of disease, rather than simply the causes and prognosis. Randomised trials and systematic reviews became the ‘stock in trade’ of the clinical epidemiologist.

As more and more effective treatments were discovered, people started to worry about large variations in practice and in the quality of care. Service delivery health research and the ‘quality movement’ were born. Researchers naturally felt the need to measure quality. Progress was slow, however. First, quality improvement was initially dominated by management research; a subject that does not have a strong tradition of measurement, as I have reported elsewhere.[2] Second, the constructs that quality researchers were dealing with were much harder to measure than clinical outcomes. For example, an attempt was made to correlate the safety culture with standardised mortality rates across intensive care units. The result was null, but this might have resulted entirely from measurement error; mortality rates suffer from unavoidable signal to noise problems,[3] while the active ingredient in culture is hard to capture in a measurement.[4] As the subject of the quality of care seemed to become bogged down with measurement issues, the patient safety movement became dominant. Initially people focused on psychology and organisational science. However, no science can mature without, at some point, making its central concepts quantifiable. As Galileo (allegedly) said, “Measure what can be measured, and make measurable what cannot be measured.” So it became necessary to try to measure safety, and all the problems of quality measurement re-surfaced.

Most sensible people now realise that impatience does more harm than good; shortcuts lead nowhere and we simply have to work away, measuring and mitigating measurement error as bast we can. As stated, and as I have argued elsewhere,[5] clinical outcomes are insensitive to many service interventions. This is a lesson that those of us with a background in classical or clinical epidemiology have been slow to learn. Trying to copy clinical epidemiology, and to rely entirely on clinical endpoints, has driven service delivery research into two camps – qualitative researchers who eschew quantification, and quantitative researchers who want to apply rules of evidence that served them well in clinical research. However, there really is a third way. This method is based on observations across the causal chain linking intervention to clinical outcome. I have long argued that it is the pattern of data (qualitative and quantitative) across a causal chain that should be analysed.[5] Since then, people have started to pay attention, not just to the outcome at the human level, but also to mediating variables. More recently still, I have argued for the use of Bayesian networks to synthesise information from the causal chain in a particular study, along with evidence from reviews in salient topics.[6] Note that while coming from the same, realist, epistemology as ‘mixed-methods’ research, mediator variable analysis and Bayesian networking developed mixed-methods to another level, since they enable data of different sorts to be captured in a clinical outcome of sufficient importance to populate a decision model. The use of proxy outcomes acquired a bad reputation in clinical epidemiology. However, carrying this idea over into service delivery research is extremely limiting. It is also unscientific, since science is dependent on induction, and induction can only be carried out if the causal mechanisms behind the results obtained are understood.

— Richard Lilford, CLAHRC WM Director


  1. Hill AB. The environment and disease: Association or causation? Proc R Soc Med. 1965; 58(5): 295-300.
  2. Lilford RJ, Dobbie F, Warren R, Braunholtz D, Boaden R. Top-rated British business research: Has the emperor got any clothes? Health Serv Manage Res. 2003; 16(3): 147-54.
  3. Girling AJ, Hofer TP, Wu J, Chilton PJ, Nicholl JP, Mohammed MA, Lilford RJ. Case-mix adjusted hospital mortality is a poor proxy for preventable mortality: a modelling study. BMJ Qual Saf. 2012; 21(12): 1052-6.
  4. Mannion R, Davies H, Konteh H, Jung T, Scott T, Bower P, Whalley D, McNally R, McMurray R. Measuring and Assessing Organisational Culture in the NHS (OC1). 2008.
  5. Lilford RJ, Chilton PJ, Hemming K, Girling AJ, Taylor CA, Barach P. Evaluating policy and service interventions: framework to guide selection and interpretation of study end points. BMJ 2010; 341: c4413.
  6. Watson SI & Lilford RJ. Essay 1: Integrating Multiple Sources of Evidence: a Bayesian Perspective. In: Challenges, solutions and future directions in the evaluation of service innovations in health care and public health. Southampton (UK): NIHR Journals Library, 2016.

‘A-PINCH’ medicines – Those Largely Responsible for Medication-Related Harm

Last year the WHO announced their third global safety challenge.[1] [2] Three main challenges are included:

  1. High-risk situations.
  2. Polypharmacy.
  3. Transitions of care – an increasing priority for CLAHRC WM.

And the medicines to be wary about are the small proportion that are responsible for a large proportion of medication-related harm, covered by the ‘A-PINCH’ acronym: Antibiotics, Potassium and other electrolytes, Insulin, Narcotics (and other sedatives), Chemotherapeutic agents, and Heparin (and other anti-coagulants). Medicines review is important when patients transition across care providers.

— Richard Lilford, CLAHRC WM Director


  1. Shiekh A, Dhingra-Kumar N, Kelley E, Kieny MP, Donaldson LJ. The Third Global Patient Safety Challenge: Tackling Medication-Related Harm. Bull World Health Organ. 2017; 95: 546.
  2. Donaldson LJ, Kelley ET, Dhingra-Kumar N, Kiney M-P, Shiekh A. Medication Without Harm: WHO’s Third Global Patient Safety Challenge. Lancet. 2017; 389: 1680-1.

Which Combinations of Chronic Conditions have the Greatest Burden of Disease?

CLAHRC WM investigators are carrying out work on multi-morbidity. Which combinations of diseases are most debilitating for patients? According to Jia, et al. conditions individually with the greatest burden of disease are heart failure and depression.[1] The worst combination among 13 conditions was any combination including heart failure and depression.  High blood pressure and arthritis have much lower burdens of disease as measured by QALYs, either singly or in combination.

— Richard Lilford, CLAHRC WM Director


  1. Jia H, Lubetkin EI, Barile JP, Horner-Johnson W, DeMichele K, Stark DS, Zack MM, Thompson WW. Quality-adjusted Life Years (QALY) for 15 Chronic Conditions and Combinations of Conditions Among US Adults Aged 65 and Older. Med Care. 2018; 56(8): 740-6.

Before and After Study Shows Large Reductions in Surgical Site Infections Across Four African Countries

This important study from the WHO Infection Prevention and Control global unit was based on a multi-component intervention.[1] The intervention consisted of a number of specific measures, including not shaving the skin prior to surgery, antibiotic prophylaxis and proper skin preparation. It also included some generic cultural components, including the promotion of operating theatre discipline.

The incidence of post-operative infections of the surgical site was halved from 8% to about 4% following implementation of the intervention. No contemporaneous national initiative took place in any of the countries concerned, so I think that a general temporal effect, or rising tide,[2] is unlikely. Moreover, process measures improved in line with the clinical outcomes. A cause-effect explanation is thus plausible.

Yet, people will be properly sceptical of this result. Determining a surgical site infection (SSI) is subjective, as has been shown in many empirical studies.[3] Such measurements are likely to be reactive, meaning that there is an interaction between the intervention and the outcomes observed. The way to get around this is to objectify the observations in some way, such as by blinding the observers. Reactivity is a limitation on most studies of SSIs, whether randomised or observational. In my opinion, it is worth spending additional money to avoid this problem, which cannot simply be wished away. Possible methods to get around the problem include: use of observers who move from institution to institution and who do not know where or when interventions were implemented; or use of images scored blindly by independent observers. If this is too expensive, then independent sampling of wards taken at random by a truly blinded observer should be used. In the meantime, given the likely reactivity of the measurement, it is prudent to interpret RCTs of methods to reduce SSIs with caution.

— Richard Lilford, CLAHRC WM Director


  1. Allegranzi B, Aiken AM, Kubilay NZ, et al. A multimodal infection control and patient safety intervention to reduce surgical site infections in Africa: a multicentre, before-after, cohort study. Lancet Infect Dis. 2018; 18: 507-15.
  2. Chen YF, Hemming K, Stevens AJ, Lilford RJ. Secular trends and evaluation of complex interventions: the rising tide phenomenon. BMJ Qual Saf. 2016; 25: 303-10.
  3. Taylor JS, Marten CA, Potts KA, et al. What is the Real Rate of Surgical Site Infection? J Oncol Pract. 2016; 12(10): e878-83.

Retrospective Study of the Quality of Care Given to Patients Who Have Died – A Design That Should Be Laid to Rest According to Bach, et al.

Prof Tim Hofer (University of Michigan) drew this important article to my attention.[1] Bach and colleagues consider studies of the quality of care given to people who have recently died. Such studies sound laudable, but they are a poor reflection of the care given to dying patients.  Why is it that these apparently well-meaning studies provide biased results? There are two main problems. First, when patients are still alive it is not known who will die – many who die are not identified as dying, while many identified as dying do not do so. Thus, dead patients are a poor reflection of dying patients; they are a highly skewed group. Second, when the care given to dead people is examined, a time interval prior to death must be specified and the results obtained are highly sensitive to this arbitrary threshold. This study used real, population-based cohorts to show how differences in subject selection and time lead to massive bias. Retrospective case series, based on events that precede eligibility for inclusion in the cohort, simply should not be used to quantify the quality of care.

— Richard Lilford, CLAHRC WM Director


  1. Bach PB, Schrag D, Begg CB. Resurrecting Treatment Histories of Dead Patients. JAMA. 2004; 292: 2765-70.

Public versus Private Providers: Simple Solutions are Simplistic!

The CLAHRC WM Director has always been interested in public versus private provision of services. Very few people wish to privatise the judiciary or the army, and very few want to bring supermarkets or car factories into public ownership. However, there is plenty of dispute about many other services, such the rail tracks or care homes.

Previous news blogs summarised the result of studies of private versus government schools in Pakistan [1] and in the USA.[2] Tony Culyer compared private with public care homes and found no difference; they were equally bad.

So, the CLAHRC WM Director was fascinated by a recent article on government outsourcing in The Economist (30 June).[3] The conclusion here is nuanced. It turns out that over many industries public provision can be just as good as private provision, but usually, when the public provision involves competition. Thus, we should expect a very different outcomes when a public provider competes with a private provider, versus public providers in a system with no private providers. So, it seems to come down to monopoly versus multiple providers and having some element of competition.

In healthcare there is a further factor at play, the need to integrate over many service providers. Introducing competition in such a system is more difficult than simple competition between providers. Of course, healthcare also suffers from endemic market failure due to information asymmetries.

Many areas and economics remain unresolved. It is often much easier to identify what not to do, than to identify precisely what to do.

— Richard Lilford, CLAHRC WM Director


  1. Lilford RJ. League Tables – Not Always Bad. NIHR CLAHRC West Midlands News Blog. 28 August 2015.
  2. Lilford RJ. Government vs. Private Schools. NIHR CLAHRC West Midlands News Blog. 23 June 2017.
  3. The Economist. Britain’s outsourcing model, copied around the world, is in trouble. The Economist. 28 June 2018.

Adrenaline vs. no Adrenaline for Out of Hospital Cardiac Arrest

An iconic trial of over 8,000 participants was published last week by CLAHRC WM affiliate Gavin Perkins and colleagues comparing administration versus non-administration of epinephrine (adrenaline) for out of hospital cardiac arrest.[1] Overall only about one person in fifty survived such an event. The trial showed a higher survival rate among patients in the adrenaline group, but a lower rate of neurologically intact survivals. This can be accounted for by a higher rate of severe neurological damage in the adrenaline group. This result is broadly consistent with trials of high versus low dose adrenaline, where the higher dose is associated with more survival, but a higher proportion of survivors with neurological damage.

Congratulations to the trial team who brought many patients and the public on side before conducting the trial. This ensured that the trial would be completed under a fair amount of press scrutiny and even criticism.

The trial shows a trade-off between survival and intact survival. Public consultation carried out by the investigators suggested that people are more concerned with long-term than short term outcome. It would be extremely interesting to elicit trade-off functions from representative members of the public, say by means of discrete choice experiments. Here is a situation where consent cannot be acquired at the point of treatment and so, absent a close relative at the scene, choice of option has to be guided by public norms.

— Richard Lilford, CLAHRC WM Director


  1. Perkins GD, Ji C, Deakin CD, et al. A Randomized Trial of Epinephrine in Out-of-Hospital Cardiac Arrest. New Engl J Med. 2018;.

Sustainability and Transformation Partnerships: Why they are so Very Interesting

There is a strong international, national and local initiative to develop services generically by integrating care across multiple providers and many diseases, rather than to focus exclusively on disease ‘silos’. However, integrating care across providers runs into immediate problems because the interests of these different providers are seldom aligned. For instance, providing care in the community may reduce earnings in a hospital where money follows patients.

Integrating care across multiple providers can take different forms, which might play out in different ways. The least radical solution would consist of informal alliances to help plan services. At the other end of the scale organisations merge into common legal entities with consolidated budgets (so-called Responsible Care organisations). Between these two extremes lie formal structures, but where the budgets and legal responsibility remain with local providers.

The Sustainability and Transformation Partnerships (STPs) in England are a good example of the intermediate arrangement. They are part of official government policy, have some funding, and have generated considerable local buy-in.

However, the interests of local providers cannot be overridden by the STP. It is tempting to say that they are unlikely to be very successful given that, inevitably, the interests of the different organisations are not the same. However, there is some evidence that this might not be the inevitable, dismal outcome. The evidence comes from Elinor Ostrom, Nobel Prize winner for economics. We have cited her work previously within this news blog.[1][2] She describes the conditions under which collaboration can take place, even when the interests of the collaborating organisations are imperfectly aligned:

  1. Clearly defined boundaries.
  2. Congruence between appropriation/provision rules and local conditions.
  3. Collective choice arrangements.
  4. Monitoring.
  5. Graduated sanctions.
  6. Conflict-resolution mechanisms.
  7. Local autonomy.
  8. Nested enterprises (polycentric governance).

Ostrom’s work was carried out in the context of protection of the environment; fisheries, farms, oceans, forests and the like. So, it would be extremely interesting to examine STP using Ostrom’s findings as an investigative lens. Working with CLAHRC London we plan to conduct numerous case studies of STPs that exhibit different features or philosophies. We expect that we will uncover differences in structure and culture that play out differently in different places. Among other things, we will see whether we can replicate Ostrom’s findings in a health care context. On this basis, we may be able to develop a tool that could help predict how well an organisation, such as an STP, is working. In the long-term we would examine (any) correlation in adherence with Ostrom’s criteria and the overall success of an STP.

Of course, this is not an easy topic for study. That is precisely why we think it is a good topic for a capacity development centre, such as a CLAHRC, to tackle. There is an inverse relationship between the importance of a topic, and its tractability. This is where various tools that we have developed, such as Bayesian networks, come into their own. These tools make intractable subjects, such as evaluating the success of STPs, a little more tractable.

— Richard Lilford, CLAHRC WM Director


  1. Lilford RJ. Polycentric Organisations. NIHR CLAHRC West Midlands News Blog. 25 July 2014.
  2. Ostrom E. Governing the Commons: The Evolution of Institutions for Collective Actions. Cambridge University Press: Cambridge, UK; 1990.

Calling All Service Delivery Researchers

Daw and Hatfield draw attention to an important source of bias in non-experimental matched before and after studies.[1] Matching can introduce a bias under these circumstances. The bias results when regression to the mean is a possibility. Consider an intervention targeted at institutions with a high mortality rate. Matching will introduce bias because the intervention cluster has a low mortality relative to its group and the control cluster has a high mortality relative to its group. If this were not so, then they would not have matched. So a difference between the two groups may emerge over time, and be ascribed to any intervention, even when there was no intervention effect.  Traditionally, we say that as confounder is associated with the intervention (treatment assignment) and the outcome. However, a confounder can be associated with the intervention and the propensity of the outcome to change over time. This applies in before and after studies (difference-in-difference studies; studies with control for baseline conditions). The article confirms this theoretical problem by means of Monte Carlo simulations. This seems a very important point for all health service researchers to be aware of.

— Richard Lilford, CLAHRC WM Director


  1. Daw JR & Hatfield LA. Matching and Regression to the Mean in Difference-in-Differences Analysis. Health Res Educ Trust. 2018.