Important Ovarian Cancer Trial Sets Scene for Multiple Indication Review

Sean Kehoe and colleagues compared chemotherapy then surgery versus surgery then chemotherapy in a landmark trial of 552 women with advanced cancer of the ovary. Though not reaching significance on the frequentist test, there was a 13% advantage for the chemo-first protocol. In addition less surgical deaths were recorded in the chemo-first group, this time at a high level of statistical significance.[1] We have compared adjuvant chemotherapy versus no adjuvant chemotherapy across all cancers, showing that the hypothesis of a similar effect across all cancers is better supported than the hypothesis of different effects by cancer type.[2] It would be fascinating to compare chemotherapy versus surgery first trials in a multiple-indication review [3] across all cancers. Any takers?

— Richard Lilford, CLAHRC WM Director


  1. Kehoe S, Hook J, Nankivell M, et al. Primary chemotherapy versus primary surgery for newly diagnosed advanced ovarian cancer (CHORUS): an open-label, randomised, controlled, non-inferiority trial. Lancet. 2015; 386: 249-57.
  2. Bowater RJ, Abdelmalik SM, Lilford RJ. Efficacy of adjuvant chemotherapy after surgery when considered over all cancer types: a synthesis of meta-analyses. Ann Surg Oncol. 2012;19(11):3343-50.
  3. Chen YF, Hemming K, Chilton PJ, Gupta KK, Altman DG, Lilford RJ. Scientific hypotheses can be tested by comparing the effects of one treatment over many diseases in a systematic review. J Clin Epidemiol. 2014;67(12):1309-19.

How Can Research in Low- and Middle-Income Countries (LMICs) Help People in High-Income Countries?

International research is undergoing a renaissance. Universities all over North America, Europe and Australia are establishing ‘Centres for Global Research’. Such centres draw funds from local donors and it is reasonable to ask whether research conducted abroad is completely altruistic, or whether it might also benefit the high-income countries that sponsor the research. Knowledge exchange is a two-way street and we should expect increasing traffic on the North-bound carriageway. The CLAHRC WM Director proposes the following classification for the potential local benefits of overseas research:

  1. The most obvious category relates to infectious disease. Research carried out abroad may provide early intelligence on impending risk so that countries may take steps to prevent spread, as in the recent Ebola epidemic (discussed in a previous post), or make preparations to contain the infection, as in the case of influenza, MARS and SARS.
  2. Research carried out in countries where a disease is common may provide information on how to treat it in countries where the diseases is rare. This would obviously apply to all tropical diseases that may affect returning travellers, visitors or immigrants, such as malaria and schistosomiasis. It would also apply to infections, such as leishmaniasis or West Nile fever that can be contracted in Europe, but with less frequency than in tropical countries. This type of knowledge transfer would also benefit people at risk of non-infectious diseases arising from habits imported from abroad, such as use of the areca nut, as discussed in a previous post.
  3. Providing larger populations to evaluate treatments where it would be very hard to accrue sufficient patients locally. Lelia Duley’s trial of magnesium for the treatment of eclampsia [1] and Ian Roberts’ ‘CRASH-2’ trial of tranexamic acid for massive haemorrhage [2] were both carried out over three continents. Yet the results drive practice across the world, including the UK.
  4. Providing a means to explore heterogeneity and thus glean deeper understanding of the role of context. For instance, the Cochrane review of trials of the effectiveness of providing additional support in labour through a layperson (so called doula) show that the service is effective in lowering the Caesarean rate where women are not accompanied by their partner, but not in countries like the UK where they usually are. The expeience of two people dying of terminal cancer, one in Kenya, the other in Scotland provides a further vivid example of the role of context.[3]
  5. The success of an intervention in LMIC may encourage people to try it locally. For instance, the success of ‘women’s groups’ in improving perinatal outcomes in India, Nepal and Bangladesh [4] have encouraged CLAHRC North Thames to replicate the method among Bangladeshi communities in East London, as mentioned in a previous post. But we should be alert to the danger of leaping too rapidly to the conclusion that what works in one place will necessarily work in another. Studies carried out in our CLAHRC shows that clinical research produces essentially the same results when carried out in North America or Europe,[5] but very different results across Europe and Asia.[6]
  6. Research methodologies of generic utility may be developed to deal with issues in LMIC. The fabled stepped wedge design widely used in CLAHRC WM [7] was first used in West Africa.[8] The lesson that much more can be learned by juxtaposing quantitative and qualitative research in systematic reviews than by either method alone, was ably demonstrated by twin Cochrane reviews on the subject,[9] [10] (mentioned in a previous post).

More speculatively, the CLAHRC WM Director posits a category where there is no specific nugget of information that is returned, but rather tacit knowledge about universal features of the human condition. Certain general principles may be derived by examining health improvement projects across many countries, rich and poor, as recently pointed out by previous NHS Chief Executive, Lord Nigel Crisp.[11] More indirect still are the cultural and political benefits of human interaction, and putative benefits of seeing the world in less parochial ways.

I invite readers of this News Blog to share other types of benefits and help populate this framework with lots of examples where LMIC research has benefited UK patients.

— Richard Lilford, CLAHRC WM Director


  1. Duley L. Magnesium and eclampsia. Lancet. 1995; 346:1365.
  2. CRASH-2 trial collaborators. Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial. Lancet. 2010; 376(9734): 23-32.
  3. Murray SA, Grant E, Grant A, Kendall M. Dying from cancer in developed and developing countries: lessons from two qualitative interview studies of patients and their carers. BMJ. 2003; 326: 368.
  4. Prost A, Colbourn T, Seward N, et al. Women’s groups practising participatory learning and action to improve maternal and newborn health in resource-limited settings: systematic review and meta-analysis. Lancet. 2013; 381(9879): 1736-46.
  5. Bowater RJ, Hartley LC, Lilford RJ. Are cardiovascular trial results systematically different between North America and Europe? A study based on intra-meta-analysis comparisons. Arch Cardiovasc Dis. 2015;108(1):23-38.
  6. Hartley LC, Girling AJ, Bowater RJ, Lilford RJ. A multistudy analysis investigating systematic differences in cardiovascular trial results between Europe and Asia. J Epidemiol Community Health. 2015;69(4):397-404.
  7. Hemming K, Haines TP, Chilton PJ, Girling AJ, Lilford RJ. The stepped wedge cluster randomised trial: rationale, design, analysis, and reporting. BMJ. 2015; 350: h391.
  8. The Gambia Hepatitis Study Group. The Gambia hepatitis intervention study. Cancer Res. 1987;47:5782-7.
  9. Lewin S, Munabi-Babigumira S, Glenton C, et al. Lay health workers in primary and community health care for maternal and child health and the management of infectious diseases. Cochrane Database Syst Rev. 2010; 3: CD004015.
  10. Glenton C, Colvin CJ, Carlsen B, Swartz A, Lewin S, Noyes J, Rashidian A. Barriers and facilitators to the implementation of lay health worker programmes to improve access to maternal and child health: qualitative evidence synthesis. Cochrane Database Syst Rev. 2013; 10: CD010414.
  11. Crisp N. Chapter 5: Turning the World Upside Down. In: Commonwealth Health Minister’s Update 2010. Geneva: World Health Organization. 2010. pp.89-933.


A Fabulous Paper on P Values, Confidence Limit, and, yes, Bayes

I thank my friend and colleague Alan Girling for drawing my attention to a recent issue of the statistical journal ‘Significance’.

This issue of the journal followed close on the heels of the UK general election and so, not surprisingly, the failure of polls to predict the outcome provided the topic for the feature article. But it was another article on representing significance in the psychological and medical literature that Alan suggested I should read.[1] The author, Andrew Gelman, discusses small studies with statistically significant results. Such studies tend to exaggerate effect sizes when the signal is associated with a lot of noise, and in the usual situation where true positive effects are of modest magnitude. Throw in a little publication bias and the literature becomes yet more severely distorted. I entirely agreed with the statement that use of confidence intervals “will not solve any problems: checking whether a 95% interval excludes zero is mathematically equivalent to checking whether p<0.05.” Gelman goes on to say that the problem of exaggerated claims is worse in psychology than medicine because there are fewer obstacles to carrying out small studies and, arguably, because the signals, relating as they do to latent mental constructs, tend to be more ‘noisy’ than those in medicine. Gelman comes down strongly in favour of the Bayesian approach, which brings sobriety to bear through the ‘prior’ probability density (see previous post).

— Richard Lilford, CLAHRC WM Director


  1. Gelman A. Working through some issues. Significance. 2015; 12(3): 33-5.

Every Now and Then, an Earth Shattering Trial with Immediate Implications for Policy

The CLAHRC WM Director has long thought that it made sense to treat all HIV patients with antiretrovirals and not wait until the CD4+ count had fallen to a specified threshold – a threshold that has been gradually increasing over the years. On this particular point, the Director was right. An RCT of 4,685 HIV-positive patients across 35 countries showed that immediate therapy, rather than therapy deferred until the CD4+ count had fallen below the current recommended threshold, reduced AIDS complications by over 50% within tight confidence intervals (P<0.001).[1] In addition, there is evidence that antiretroviral therapy can reduce the risk of HIV sexual transmission.[2]

— Richard Lilford, CLAHRC WM Director


  1. The INSIGHT START Study Group. Initiation of antiretroviral therapy in early asymptomatic HIV infection. N Engl J Med. 2015. [ePub].
  2. Cohen MS, Chen YQ, McCauley M, et al. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011; 365: 493-505

Yet Again, Low Proportion of Hospital Deaths Judged Preventable

Hogan and colleagues have reported another study on preventable mortality based on case-note review among 34 hospitals.[1] Only 3.6% of deaths were thought to have been preventable on the balance of probability. Preventability rates did not vary widely between hospitals.

Of course, this might be something of an underestimate because deaths where the probability of preventability was less than 50% are not included. The CLAHRC WM Director calculates preventability as the sum of all cases that may have been preventable, weighted by the probability that they were preventable. He also likes to adjust for the reviewer effect to minimise the influence of unusually ‘hawkish’ reviewers.

Despite these precautions, preventability is “in the eye of the reviewer,”[2] and may be over-estimated because of hindsight bias, or under-estimated because some practices that may increase the risk of death cannot be discerned from case-notes.

— Richard Lilford, CLAHRC WM Director


  1. Hogan H, Zipfel R, Neuburger J, Hutchings A, Darzi A, Black N. Avoidability of hospital deaths and association with hospital-wide mortality ratios: retrospective case record review and regression analysis. BMJ. 2015; 351: h3239.
  2. Hayward RA, Hofer TP. Estimating hospital deaths due to medical errors: preventability is in the eye of the reviewer. JAMA. 2001; 286(4): 415-20.

More, Yet More, on Pure, White and Deadly

Yes, the more you eat and the fatter you become, and the more likely you are to get diabetes. Subsidiary questions:

  1. Are some diets more likely to make you fat than others? Yes, sugar is worse than other diets in making you fat (see previous posts). I think this is because they are less satiating, but they also have worse metabolic effects (again, see previous posts).
  2. At a given level of weight gain, is a high intake of sugar more likely to cause diabetes?

A recent meta-analysis of prospective studies of sugar-sweetened beverages in seventeen cohorts including no less than 38,253 people were analysed. Each extra ‘serving’ of sugar-sweetened beverage was associated with an 18% increase in the risk of diabetes even after correcting for adiposity.[1]

The CLAHRC WM Director says “think carefully before you correct for a variable on the causal chain between exposure and outcome because you will likely mask a true association.” These results are therefore impressive because they show an increased risk of diabetes with sugary drinks, net of the increased risk of obesity itself. Sugary drinks cause a very high spike of glucose, which may predispose to diabetes over a more gradual increase following, for example, a rice meal. Sugar is made up of a glucose and fructose molecule, and fructose increases insulin resistance. It is therefore likely to be more diabetogenic than the carbohydrate found in rice, wheat and potatoes, which is pure glucose. The astute reader will have noted that fruit juice will also cause a severe spike in glucose levels and also contains sugar, and hence fructose. Unlike whole fruit, which protects against diabetes, fruit juice also seems to be diabetogenic net of weight gain. Incidentally, this is a beautifully conducted and analysed study from Japan, the USA and the UK, which is worth reading for this reason alone.

— Richard Lilford, CLAHRC WM Director


  1. Imamura F, O’Connor L, Ye Z, et al. Consumption of sugar sweetened beverages, artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes: systematic review, meta-analysis, and estimation of population attributable fraction. BMJ. 2015; 351: h3576.

Nice Theory, but Ridding Children of Those Revolting Worms Does Not Seem to Affect Health

Wouldn’t it be nice if we could find a simple physical cause for stunted-growth and impaired school performance? Worm infestations provide such a target; round worms (which may compete for the child’s nutrition), and hook worms (which suck blood through the intestinal wall). Alas the attractive notion that we can do a great deal of good by the simple distribution of de-worming pills does not seem to be true. The recent meta-analysis from Paul Garner’s group [1] provides little support for de-worming policies. Even among children known to be infected, de-worming does not do anything dramatic, and surprisingly (and disappointingly) the trials examining this issue are of poor quality.

It could be argued that de-worming does not do any harm, and that worms are revolting, and therefore that de-worming programmes should continue. But in a world of scarce resources, this argument might not hold. And then of course some people say that worms do some good by providing an inflammatory response that reduces atopic conditions, such as hay fever of asthma.[2]

Science cannot prove a negative, merely exclude effects of measurable size, and we may never have a conclusive answer to this question. Comments are invited on this controversial issue.

— Richard Lilford, CLAHRC WM Director


  1. Taylor-Robinson DC, Maayan N, Soares-Weiser K, Donegan S, Garner P. Deworming drugs for soil-transmitted intestinal worms in children: effects on nutritional indicators, haemoglobin, and school performance. Cochrane Database Sys Rev. 2015; 7: CD000371.
  2. Quinnell RJ, Bethony J, Prichard DI. The immunoepidemiology of human hookworm infection. Parasite Immunol. 2004; 26(11-12): 443-54.


An Article on Check-lists in Nature (Yes Nature!)

Further to our recent post on check-lists, the topic has piqued the interest of the famous scientific journal Nature.[1] They discuss the variable success of check-lists and effectively conclude that “it ain’t what you do, it’s the way that you do it.” This makes the CLAHRC WM Director consider what a check-list really is – is it an intervention or just one of many tools that can be used to change behaviour? When the CLAHRC WM Director was a pilot, check-lists were there to remind you of things that might otherwise have been forgotten. However, it is possible to hypothesise that check-lists in most health contexts function more as a cultural tool. If so, they are but one of many routes to behaviour change. The CLAHRC WM Director posits that once the care has improved, check-lists can be removed with impunity – their effect is indelible. This is a testable hypothesis.

— Richard Lilford, CLAHRC WM Director


  1. Anthes E. The Trouble with Checklists. Nature. 2015; 523: 516-8.

DevoManc: a ‘courageous’ decision?

The devolution of health and social care budget in Greater Manchester represents a bold, radical, and courageous approach for the NHS; albeit this increased localism is not unusual in an international context. At an event held recently in Manchester it was good to see the optimism and enthusiasm from many in the room both at the opportunity to try something different and at the possibilities offered. However, there is no escaping the bottom line, which is that significant savings estimated at between 13% and 17% need to be found against a backdrop of several years of static budgets and efficiency savings. This, combined with the risk of double running costs (for which there is no additional funding) as new models of care are introduced and the risk of additional need being uncovered as part of service redesign pose a huge financial risk to the proposal. The overwhelming likelihood is that, in response to this, thresholds for services and the types of care offered are likely to be raised, which could bring conflict with the NHS Constitution that gives patients rights around access, timeliness and choice of provider. It also risks losing public support so patient involvement and consultation throughout the formulation and delivery of the project will be key.

Geographically too there is a danger that the process is overly-focused on Manchester, when it is in fact a regional re-organisation, and there is a risk that those living on the periphery of the area will choose to register across the boundary if they perceive the service there to be preferential as has been seen on the England/Wales border. This geographic inequality could be worsened because the easiest way to generate savings will be to close acute beds, and the risk is that this will be targeted at the less densely populated areas and less specialist Trusts.

The CLAHRC West Midlands will be watching closely; both the process of devolution and the evaluation of it. There seems to be considerable political traction associated with the DevoManc and there is a strong possibility that other regions – willingly or unwillingly – may follow. The benefits for any Secretary of State for Health are clear; they can distance themselves from poor performance or access to services by directing grievances towards local decision-making and prioritisation by the elected mayor and his or her delegates, rather than an overall lack of central funding, which in a time of austerity must be attractive for a politician of any denomination. Those in Manchester responsible for delivery clearly understand this, which makes it a truly courageous undertaking, but as fans of ‘Yes Minister’ will recall: “a controversial policy will lose votes, whilst a courageous one will lose the election.”

— Paul Bird, CLAHRC Head of Programme Delivery (Engagement)

High Side-Effect Rates with Statins in Ordinary Practice but not in RCTs – the Enigma is Explained

Why do such a high proportion of patients who take statins complain of muscle pain in ordinary practice, while RCTs find little increase in this unintended effect? Could it be because the RCTs recruited a low-risk population? No – over 170,000 patients have been entered into high-quality RCTs, many had multi-morbidities and this was not a ‘sanitised’ population. Could it be because unintended effects were not assiduously recorded in RCTs? Again, no.

A recent paper in JAMA [1] examines re-introduction studies and find that the great majority of patients are able to tolerate treatment when it is restarted (admittedly, often in a different form). A particularly interesting study was carried out among people who had been ‘intolerant’ of statins. They were randomised to statin and non-statin treatments following a preparation phase where they were given placebo and could withdraw if intolerable symptoms recurred. Muscle pains were no more prevalent among those randomised to receive statins than among those in the non-statin group. It would appear that the very high incidence of muscle pain in standard practice is the result of psychological expectation, not pharmacological action. What do readers think?

— Richard Lilford, CLAHRC WM Director


  1. Newman CB, & Tobert JA. Statin Intolerance: Reconciling Clinical Trials and Clinical Experience. JAMA. 2015. 313(10); 1011-2.