Preventing Re-admissions

CLAHRC WM is working with Sandwell and Birmingham Hospitals group to improve care of patients on discharge from hospital. This is a worthwhile exercise since handover from hospital to community is a ‘fault-line’ for safe care. Some think that improving care over this transition may reduce re-admission rates, and indeed differences in re-admission rates across hospital sites within the group prompted the above initiative in the first place.

However, the CLAHRC WM Director is circumspect regarding the prospects for reduced re-admissions. His argument is simple: most re-admissions result from inter-current or progressive disease, while the proportion of re-admissions that are preventable is small, especially beyond the first four weeks after discharge. It follows that re-admissions are a small signal easily buried in noise.[1] This does not, of course, mean that improving care at discharge is not a worthwhile objective.

A recent RCT of an expensive intervention based on one-to-one self-management education from a discharge nurse, backed up by telephone calls after discharge, did not lead to reduced re-admissions and may have actually increased hospital contact overall.[2] Is this yet a further example of an intervention motivated by the need to reduce healthcare utilisation that results instead in improved care but no reduction, or even an increase, in healthcare costs?

— Richard Lilford, CLAHRC WM Director


  1. 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 Quality & Safety. 2012; 21: 1052-6.
  2. Goldman LE, Sarkar U, Kessell E, et al. Support From Hospital to Home for Elders: A Randomized Trial. Ann Intern Med. 2014; 161(7): 472-81.

Failure to replicate earlier findings

A recent paper by Horton [1] attracted 1,576 views and 123 tweets within ten days of its publication.[2] A major scientific breakthrough? Hardly, the paper reported failure to replicate an earlier finding. Whose earlier finding? Horton’s! Horton suggests two reasons for his failure to replicate earlier results: lack of generalisability or type 1 error (false positive result arising through the play of chance). Both papers dealt with memory clues through associations – the idea that a clue, such as a place or a person that was present when the memory was stored in the brain, could prompt its recall.

The CLAHRC WM director thinks Type 1 error is the likely explanation. A positive value is much less impressive evidence against the null hypothesis than many suppose.[3] A Bayesian approach [4] would have forestalled such a volte-face.

— Richard Lilford, CLAHRC WM Director


  1. Brown-Schmidt S, Horton WS. The Influence of Partner-Specific Memory Associations on Picture Naming: A Failure to Replicate Horton (2007). PLoS ONE. 2014; 9(10): e109035.
  2. Reas E. This Week’s Most Discussed PLOS Neuroscience Article: The Influence of Partner-Specific Memory Associations on Picture Naming: A Failure to Replicate. [Online]. 2014.
  3. Goodman S. A Dirty Dozen: Twelve P-Value Misconceptions. Semin Hematol. 2008; 45(3): 135-40.
  4. Goodman SN. Toward Evidence-Based Medical Statistics. 2: The Bayes Factor. Ann Intern Med. 1999; 130: 1005-13.

Behaviour Change – Special Issue of Psychology and Health

Readers of this News Blog know that CLAHRCs are interested in behaviour change – CLAHRCs not interested in this subject should send the money back! So a recent special issue of Psychology and Health on the risk of bias in RCTs of behaviour change interventions should pique our interest. Unsurprisingly, much of the material is old hat to clinical and service delivery researchers, and the issues discussed are not specific for behaviour change interventions. Drug trials are the exception in not having to cope with difficulties such as in blinding therapists (leading to co-intervention or contamination), blinding patients and observers (leading to detection bias for subjective outcomes), and isolating or standardising the active ingredient of the intervention. The above problems are shared with trials of most types of intervention; surgery, physiotherapy, targeted service change, generic service change, and so on. One author conflates randomisation (a procedure to guard against selection bias) with other procedures, such as double blinding (which guards against performance and detection bias).[1] In fact, they are separate causes of bias and it is possible to have one without the other.

If you have time for only one article, I recommend the paper by Jim McCambridge [2] on the social psychology of research participation. This includes question-behaviour effects where consent procedures or outcome questionnaires (applied to control and intervention groups) interact with the intervention to attenuate or amplify its effects. To deal with this, they recommend the Solomon-4 design where randomisation is to both intervention and (enhanced) questionnaires in a 2×2 factorial design. A real example where filling in a lengthy questionnaire interacted synergistically with an intervention is given. McCambridge makes the excellent point that the problems don’t go away just because a study is not randomised. The article, however, also deals with randomisation itself. Being assigned to a control group might be associated with ‘resentful demoralisation’. Here Zelen randomisation (no consent from control group) is one possibility. Another, oft recommended by the CLAHRC WM Director, is ensuring that only patients in equipoise [3] enter a trial, as originally recommended by Brewin and Bradley.[4]

— Richard Lilford, CLAHRC WM Director


  1. Tarquinio C, Kivits J, Minary L, Coste J, Alla F. Evaluating complex interventions: Perspectives and issues for health behaviour change interventions. Psychol Health. 2015; 30(1): 35-51.
  2. McCambridge J. From question-behaviour effects in trials to the social psychology of research participation. Psychol Health. 2015; 30(1): 72-84.
  3. Lilford RJ, Jackson J. Equipoise and the ethics of randomization. J R Soc Med. 1995; 88(10): 552-9.
  4. Brewin CR, Bradley C. Patient preferences and randomised clinical trials. BMJ. 1989; 299(6694): 313-5.

Measuring Quality of Care

McGlynn and Adams [1] repeat a point frequently made by the CLAHRC WM Director – before using outcomes to judge the quality of care, first model plausible effects.[2] [3] Only a small fraction of an outcome may be amenable to improved care.

The rate of hospital deaths in the UK is about 3%. Allowing a generous 20% of those to be preventable sets an upper headroom for improvement of 0.6%. So don’t expect quality of care to show up in mortality statistics. Or, to take another example, about 1% of hospital patients suffer a preventable medication related adverse event.[4] So don’t expect improved medicine management to show up in quality of life scores among the hospital population.

— Richard Lilford, CLAHRC WM Director


  1. McGlynn EA, Adams JL. What makes a good quality measure? JAMA. 2014; 312(15): 1517-8.
  2. Yao GL, Novielli N, Manaseki-Holland S,Chen YF, van der Klink M, Barach P, Chilton PJ, Lilford RJ. Evaluation of a predevelopment service delivery intervention: an application to improve clinical handovers. BMJ Qual Saf. 2012; 21(s1): i29-38.
  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 Quality & Safety. 2012; 21: 1052-6.
  4. de Vries EN, Ramrattan MA, Smorenburg SM, Gouma DJ, Boermeester MA. The incidence and nature of in-hospital adverse events: a systematic review. Qual Saf Health Care. 2008; 17(3): 216-23.

Two Provocative Papers on Diet and Health

People are always extremely interested in research on diet and health. Who would have thought that milk was so bad for you? Not only does milk increase risk of heart disease, but it aggravates the one condition that one might have supposed it would protect against, namely osteoporosis. Why doesn’t this high calcium drink prevent calcium loss from bone? The sugar in milk is lactose, a disaccharide derived from glucose and galactose. Galactose may be great for babies, but it is a powerful oxidant, and oxidants are harmful for adults. So it turns out milk aggravates osteoporosis leading to more fractures.[1] However, if you ferment the sugar to lactic acid, for example in producing yoghurt or soured milk, then these negative associations disappear. So the loss of bone in elderly people is not a calcium deficiency disease in most cases, and milk is positively harmful for bone maintenance. Most animals become lactase deficient, losing the ability to digest lactose after weaning.[2] This is not just an accident; it protects them from the harmful effect of galactose. For some reason humans are not so lucky and retain the enzyme. We therefore need to pre-ferment our lactose before consumption. Not surprisingly, a lively correspondence ensured after publication of this provocative paper, but the authors mount a convincing defence.

The question of what sort of diet to take to lose weight is a long-standing controversy – low carbohydrate or low fat? Well a recent randomised trial shows that a low carbohydrate option is much superior in terms of both weight loss and lipid profile.[3] I don’t think we should be surprised: very high carbohydrate diets are an anomaly that would not have occurred in human evolution before the relatively recent discovery of agriculture. It is a pity that the good Dr Atkins didn’t live to see his theory vindicated.

— Richard Lilford, CLAHRC WM Director


  1. Michaëlsson K, Wolk A, Langenskiöld S, Basu S, Warensjö Lemming E, Melhus H, Byberg L. Milk intake and risk of mortality and fractures in women and men: cohort studies. BMJ. 2014; 349: g6015.
  2. Desai BB. Handbook of Nutrition and Diet. New York, NY: Marcel Dekker, Inc. 2000.
  3. Bazzano LA, Hu T, Reynolds K, Yao L, Bunol C, Liu Y, Chen C-S, Klag MJ, Whelton PK, He J. Effects of Low-Carbohydrate and Low-Fat Diets: A Randomized Trial. Ann Intern Med. 2014; 161(5): 309-18.

Wealth and Happiness

The CLAHRC WM Director has written before about happiness. Not his own mood, but that of people living under different conditions! His previous reading of the literature is that increasing wealth has a rather small effect on happiness, both at the individual and population levels. However, he may have underestimated the hedonic effect of wealth, at least at the national level, according to researchers from the Pew Research Center.[1] It transpires that self-reported well-being in high-income countries is considerably higher than in middle-income countries. This difference is diminishing as their economies converge. Of course there is an obvious assumption here, namely that the wealth is causing the happiness, not the other way around.

— Richard Lilford, CLAHRC WM Director


  1. Pew Research Center. People in Emerging Markets Catch Up to Advanced Economies in Life Satisfaction. 2014.

Is Low Fertility a Problem for High-Income Countries, but a Boon For Low-Income Countries?

The perceived wisdom is that low fertility is bad for national wealth in high-income countries, but good news in low-income countries. A UN report found that 54 high- and middle-income nations are following pro-natal policies, at least in part, because of their putative economic advantages.[1]

So let’s start with the basics. The middle-aged (working) population supports the childhood and elderly population through public (e.g. education) and private (e.g. direct payment) transfers. A large elderly population supported by a relatively small working population is bad news for public finances.

But that’s not the end of the story according to a recent paper by Lee and Mason.[2] Public finances are only part of a country’s economy and it is important to consider also private inter-generational transfers. It is also important to factor in the costs of educating and bringing up children. As the proportion of older people rises, so private transfers from old to young increase and the costs of bringing up the next generation decrease.

The above study is based on detailed analysis in forty countries using standardised methods to estimate production and consumption of goods and services, along with public and private inter-generational transfers. The authors use the data to calculate the fertility rate that maximises material living standards overall. The results obtained from their model confirm the above point regarding the narrow issue of public finances in high-income countries. They are maximised by fertility rates of about 3 births per woman – well above the replacement rate. Similar effects are seen in middle-income countries, but in low-income countries low fertility rates (down to 1%) maximise public finances. This is because such a low replacement rate provides a big proportional reduction in the costs of rearing children.

So much for public finances, but what about the economy overall – is it true that living standards fall in high-income countries when fertility falls below the replacement rate of ~2.1%? In fact, the optimal fertility level is about 1.8 in high-income countries, falling to about 1.5 in low-income countries. To put this another way, the combined effects of inter-generational transfers and having a lower proportion of children to rear, exceed losses due to relatively smaller working-age populations, irrespective of whether the country has high or low per capita GDP.

What about immigration in high-income countries? To cut a long story short, us immigrants are chameleons, taking on the behaviour of our adoptive country. So we provide a short-term boost but fairly neutral effects in the long-term.

Of course there are many assumptions in these calculations notwithstanding the empirical source of data to populate the model. Nevertheless, the accepted wisdom that high fertility rates are bad news in low-income countries, is supported. However, in contrast to the prevailing view, modest reductions in population growth might actually benefit high-income countries. The paper quoted here is not an easy read but I strongly recommend it for your next long haul flight.

— Richard Lilford, CLAHRC WM Director


  1. Department of Economic and Social Affairs: Population Division. World Population Policies 2013. New York: United Nations. 2013.
  2. Lee R, Mason A, members of the NTA Network. Is low fertility really a problem? Population aging, dependency, and consumption. Science. 2014; 346(6206): 229-234.

Anti-Obesity Interventions

The CLAHRC WM Director spotted an article on obesity prevention in a recent issue of the Economist.[1] It was based on a systematic review and quantitative analysis of the literature covering 74 anti-obesity interventions, classified into four groups according to the mechanism of action.[2] The main findings are very much in line with current opinion:

  • Highest impact interventions rely on restricting choices (through regulation or structuring the environment differently), rather than individual will-power.
  • Structural solutions, such as provision of healthy food at schools, apply to wider populations and tend to be more enduring than those targeting behaviour on individual / small group basis.
  • However, there is no magic bullet and investing in the lower impact measures is still worthwhile; we cannot rely only on regulation and structural solutions, and a number of CLAHRCs are investigating methods to change individual behaviour.
  • Strategies relying on conscious effort have ephemeral effects, but some more so than others. Exercise alone is least effective in reducing weight in the short-term and these minimal effects are not enduring. Diet and exercise is more effective than diet alone in the short-term, but they end up about the same (mean weight loss of 5kg) at 50 months. Of course, exercise has benefits apart from weight-loss.
  • Advertising campaigns that address social norms and self-image are particularly effective in primary prevention – for example, stigmatising drunk drivers. However, the CLAHRC WM Director thinks that such messages would have to be carefully framed to avoid “victim blaming” in the context of obesity.

The authors have a provocative message for researchers that is relevant to the prevention themes within CLAHRCs. They make two points:

  1. It is difficult to measure the effect of some interventions, such as making cycle lanes available.
  2. Some worthwhile effects are very small and hence hard to measure.

These are important points to which the Director makes the following responses:

  1. It is crucially important not to conflate “no evidence of effect” with “evidence of no effect” – a lack of precise and accurate evidence is, by itself, a prescription for neither action nor inaction.
  2. Evidence of take-up of healthy behaviour can be used to model downstream effects and hence can help in deciding whether, on balance, a certain intervention is worthwhile. It is possible to model, for example, the potential effects on health of cycle lanes on the basis of changes in cycling behaviour.
  3. Because such models must be populated with a wide range of parameters, many of which are very uncertain, Bayesian methods should be used to calibrate effects and their credible ranges.[3] [4] [5]
  4. It is possible that the effect of introducing a wide range of interventions in parallel is more than the sum of each individual intervention effect. To put this another way, multiple interventions across society may generate a change in attitude – a culture change.

— Richard Lilford, CLAHRC WM Director


  1. The Economist. The War on Obesity: Heavy Weapons. The Economist. 2014.
  2. McKinsey Global Institute. Overcoming Obesity: An Initial Economic Analysis. New York, NY: McKinsey & Company. 2014.
  3. Yao GL, Novielli N, Manaseki-Holland S, Chen YF, van der Klink M, Barach P, Chilton PJ, Lilford RJ; European HANDOVER Research Collaborative. Evaluation of a predevelopment service delivery intervention: an application to improve clinical handovers. BMJ Qual Saf. 2012; 21 (s1): i29-38.
  4. 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.
  5. Lilford RJ, Girling AJ, Sheikh A, Coleman JJ, Chilton PJ, Burn SL, Jenkinson DJ, Blake L, Hemming K. Protocol for evaluation of the cost-effectiveness of ePrescribing systems and candidate prototype for other related health information technologies. BMC Health Serv Res. 2014. 14: 314.

Conflict of Interest in NICE, CLAHRCs and Other Independent Organisations

In a previous blog the CLAHRC WM Director hailed the creation of NICE as one of the crowning achievements of the previous Labour administration – up there with granting independence to the reserve bank. NICE epitomised enlightenment values by bringing a highly rationalist approach to bear on NHS procurement decisions. Interventions had first to be effective and, if effective, they had also to be a better buy than the (nominal) activities displaced within a fixed budget. However, that was before the 2008 crash. The government, desperate to kick-start the economy, became susceptible to arguments to make NICE more responsive to the needs of industry. Companies would produce their own models, ‘single technology reviews’, to be critiqued by the NICE ecosystem, rather than the other way around. NICE would support industrial innovation for devices through a separate system of External Assessment Centres.

The economy is generally conceptualised in terms of the demand for, and supply of, products and services. The economy can be strengthened on both sides – providing better information strengthens the demand side, while innovations to meet demand strengthens the supply side. Can one organisation really do both simultaneously? Not according to a recent BMJ article reporting ‘insiders’ concerns that the new minister with responsibility for NICE will have dual appointments across the Department of Business, Innovation and Skills, and the Department of Health.[1] According to the BMJ article the new minister, George Freeman, is aware of the potential risk and will take steps to mitigate it.

So how may NICE manage this putative conflict of interest, thereby preserving its currently colossal international reputation? Well, it so happens that CLAHRCs also have a responsibility to work with industry and applicants had to say how they would do so on the application form. Three separate “ways of working” can be distinguished in which an independent organisation (such as NICE or CLAHRCs) may contribute to the national wealth agenda:

  1. Strengthening the demand side of the health economy by evaluating cost-effectiveness of interventions – in the case of NICE there are particular clinical treatments, while in the case of CLAHRCs they are the services that support individual treatments.[2]
  2. At the supply side, by strengthening industry generally – the industry, as opposed to any particular industry. One way of doing this is by supplying knowledge and tools that might be helpful to commercial enterprises in a certain sector. For example, CLAHRC WM has developed methods for health economic evaluations at the design and development stages of a new technology.[3] [4] [5]
  3. At the supply side, by collaborating with a particular commercial organisation.

It is only in the third of these “ways of working” that the potential conflict arises. To manage the risk we propose that:

  1. The potential risk should be acknowledged, not ‘pushed under the carpet’, since it is based on extensive empirical evidence.[6] [7] [8] [9] [10]
  2. The “way of working” should be crystal clear for any project.
  3. There should be no overlap between personnel involved in supply or demand side evaluations of a particular product at any time in its life cycle.
  4. Ideally an organisation should not be involved in supply and demand side evaluation of a particular product at the same time (but this criterion may be difficult to meet in a large organisation, such as a university or NICE).

These ideas have been “road-tested” in a presentation to the NIHR Office for Clinical Research Infrastructure (NOCRI) and to the NIHR Biomedical Centres and new CLAHRC Directors, but we would welcome comments and feedback.

— Richard Lilford, CLAHRC WM Director


  1. Cohen D. Insiders say NICE is being encouraged to be more favourable to industry. BMJ. 2014; 349: g6387.
  2. 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.
  3. Girling AJ, Lilford RJ, Young TP. Pricing of medical devices under coverage uncertainty – a modelling approach. Health Econ. 2012; 21(12): 1502-7.
  4. Girling A, Young T, Brown C, Lilford R. Early-Stage Valuation of Medical Devices: The Role of Developmental Uncertainty. Value Health. 2010. 13(5): 585-91.
  5. Vallejo-Torres L, Steuten L, Buxton M, Girling AJ, Lilford RJ, Young T. Integrating health economics modelling in the product development cycle of medical devices: a Bayesian approach. Int J Technol Assess Health Care. 2008; 24(4): 459-64.
  6. Ethical Standards in Health & Life Sciences Group. Guidance on collaboration between healthcare professionals and the pharmaceutical industry. [Online]. 2012.
  7. Fletcher SW. Continuing Education in the Health Professions: Improving Healthcare Through Lifelong Learning. Chairman’s Summary of the Conference. New York: Josiah Macy, Jr Foundation. 2008. pp. 13-23.
  8. Spurling GK, Mansfield PR, Montgomery BD, Lexchin J, Doust J, Othman N, Vitry AI. Information from Pharmaceutical Companies and the Quality, Quantity, and Cost of Physicians’ Prescribing: A Systematic Review. PLoS Med. 2010; 7(10): e1000352.
  9. Steuten L, Buxton M. Economic evaluation of healthcare safety: which attributes of safety do healthcare professionals consider most important in resource allocation decisions? Qual Saf Health Care. 2010; 19: 1-6.
  10. Wang AT, McCoy CP, Murad MH, Montori VM. Association between industry affiliation and position on cardiovascular risk with rosiglitazone: cross sectional systematic review. BMJ. 2010; 340: c1344.

In-Vehicle Technology to Reduce Road Traffic Incidents

A 2013 report from the World Health Organization (WHO) highlighted that 1.24 million people die every year from road traffic incidents (RTI) and 50 million are injured.[1] Traffic accidents are the second biggest killer of children (5-14 years) and young people (15-29 years).[2] Projections suggest that they will be the seventh largest cause of all death by 2030,[3] and third in the league table for burden of disease.[4] In response, the General Assembly of the United Nations called for “effective international collaboration on road safety issues”. The Global Status Report on Road Safety (2013) identifies an inverted U-shaped curve for road traffic fatalities by stage of economic development: 20.1 per 100,000 population in middle-income countries, 8.3 per 100,000 in high-income countries, and 18.3 per 100,000 in low-income countries.[1]

A number of interventions have been shown to reduce the frequency and severity of crashes. Cars can be made safer by standard engineering solutions, such as seat belts and airbags.[5] Civil engineering projects include ‘traffic calming’ and street lighting.[6] Enforcement diminishes driving while intoxicated and reduces speed. Various behavioural techniques, such as roadside speedometers and encouraging passenger activism are also effective.[7] [8] However, many of these interventions, such as road improvement, are expensive and cannot be implemented quickly.

In-vehicle technologies offer considerable promise in many ways. First, they can be used to report crash data autonomously to a remote computer and hence identify crash ‘hot spots’ and contribute to the evaluation of interventions. Second, they can alert the police and emergency services that a crash has occurred, where it occurred, and expectations of serious injuries.[9] Third, they can provide a record of driver behaviour for driver feedback or for rewards or sanctions, such as decreased/increased insurance premiums. Fourth, they may incorporate driver warning functions to alert the driver when speed limits are breached or when potential hazards, such as a cyclist in the driver’s blind spot, are present. Fifth, they can take control of a vehicle in an emergency, for example to rectify drift towards opposing traffic when the driver is fatigued or distracted. The first three uses are telematic and the fourth and fifth are generally referred to as driver assistance functions.

We hypothesise that the dismal extrapolation in the Global Status Report cited above can be ameliorated if telematic and driver assistance technology is implemented in a way that is psychologically, socially, and technically appropriate. We have therefore formed an international collaboration with the University of Michigan Research Institute in Ann Arbor, Michigan, The Indian Institute of Technology in New Delhi, and the DY Patil University in Mumbai to adapt and evaluate this technology in middle-income countries.

— Richard Lilford, CLAHRC WM Director
— Chetan Trivedy, Academic Clinical Lecturer in Emergency Medicine, University of Warwick


  1. World Health Organization. Global Status Report on Road Safety 2013: Supporting a Decade of Action. Geneva: World Health Organization. 2013
  2. World Health Organization. World Report on road traffic in jury prevention. Ed. Peden M, Scurfield R, Sleet D, Mohan D, Hyder AA, Jarawan E, Mathers C. Geneva: World Health Organization. 2004
  3. World Health Organization. Projections of mortality and causes of death, 2015 and 2030. [Online]. 2011.
  4. World Health Organization. The Global Burden of Disease. 2004 Update. Geneva: World Health Organization. 2008.
  5. Cummings P, McKnight B, Rivara FP, Grossman DC. Association of driver airbags with driver fatality: a matched cohort study. BMJ. 2002; 324: 1119-22.
  6. Kjemtrup K, Herrdtedt L. Speed management and traffic calming in urban areas in Europe: a historical view. Accident Anal Prev. 1992; 24(1): 57-65.
  7. Pilkington P, Kinra S. Effectiveness of speed cameras in preventing road traffic collisions and related road casualties. BMJ. 2005; 330: 331-4.
  8. Habyarimana J, Jack W. Heckle and Chide: Results of a randomized road safety intervention in Kenya. J Public Econ. 2011: 95; 1438-46.
  9. Road Safety Observatory Review. Synthesis title: Telematics. 2013.