Does global health need a ‘red team’?

Looking at population health, time-series data it is easy to imagine that everything is getting better and better. What is more, as your eye tracks the line into some imaginary future, it is easy to believe that things will continue to get better and better.  It is a soothing balm to the more insidious thought, that doom awaits us around every corner.  In the world of stock pickers and equities experts, the balm is the Ying of the bull to the Yang of the bear. Hope versus despair.

The late Hans Rosling has done more to ground people in that hopeful view of the future than any other person.  The gapminder website, his creation, provides clear, firm evidence of global improvements in health and well-being across a wide range of outcomes.  As you follow the motion picture trends, countries improve. Some occasionally collapse, horribly. Then they recover. And on average, all improve.  Poverty, life expectancy, education, the infant mortality rate — it does not matter what you focus on, the world has been getting better and better

Figure 1 is a quick snapshot of this improvement in life expectancy from 1915 and 2015. In both years, higher national wealth was associated with better life expectancy.  In 1915, a country with a GDP/capita (adjusted for inflation and price) of $1,000 had a life expectancy of 30 years. In 2015 a country with a GDP/capita (adjusted for inflation and price) of $1,000 had a life expectancy of 60 years.

 

Figure 1. The left and right panels show the countries’ life expectancy in relationship to the GDP/capita (adjusted for inflation and price) 100 years apart. In 1915 a country with a GDP/capita of $1000 had a life expectancy around 30 years. In 2015 it was around 60 years — a difference of about 30 years. Source: Gapminder

In contrast, in the middle of the 18th Century, life expectancy was similar across all countries, without regard to national wealth. Little had changed by the middle of the 19th Century. Sixty years later (1915), there was a strong association between national wealth and life expectancy; and over the next 100 years, things became much better for everyone.

Will this continue?

Let’s hope that it will.  There are however significant threats visible on the horizon — and I would argue that Global Health needs a strong Red Team to make plain that dreadful prospect, often and forcefully. And as the Red Team argues their side we should hope fervently that they are utterly and comprehensively wrong! We should nonetheless listen to the arguments and not glaze over or dismiss them as we would Cassandra.

Red Teams arose in the US military and intelligence communities. They were there to argue against self-satisfied complacency. If the majority view was purple, they argued orange, if Winter, then Summer. Their purpose was to find the weaknesses in the status quo. One of the most extraordinary examples of the power of a contrarian view was the Millenium Challenge 2002, in which Paul van Riper showed that a demonstrably weaker force (the Red Team) could be devastatingly effective against the powerful (Blue Team) when they were prepared to play outside the constrained paradigm of accepted norms.

In Global Health the situation is, of course, entirely different — we do not battle each other, but we do struggle with (and against)  nature and the environment.  What is not different between Intelligence agencies and Global Health agencies is that views become entrenched. The Philosopher of Science, Thomas Kuhn, described the entrenchment of scientific ideas in terms of normal science: “the regular work of scientists theorizing, observing, and experimenting within a settled paradigm or explanatory framework”. These “settled paradigms” can permit significant new developments, but they brook no serious opposition (only tinkering at the margins). They are the VHS manufacturer to the plucky Betamax.

“Beta what?”, I hear you ask, and the point is made.

Global Health has large, powerful groups that are in danger of playing a form of technocratic hegemony — Global Health, normal science.  It’s incremental, unabrasive, and potentially wrong or ineffectual. Some of the possible threats to global health are well known, and if we focus only on those related to climate change and population growth the following is a reasonable starting list:

The global expansion of humans over the past 10,000 years was made possible by the growth of agriculture, which in turn was made possible by a stabilisation in the climate about … 10,000 years ago.  Our current success is again a product of agricultural developments. Paul Ehrlich, in his 1968 book The Population Bomb wrote a Malthusian tale of global starvation.  His prediction failed to take account of Norman Borlaug’s green revolution, and the development of semidwarf wheat, which saw grain yields triple in the 1960s and 1970s. The predicted cycle of devastating starvation was averted.

Success in the past, unfortunately, does not tell us anything about the future. Timely science then does not predict timely science now. Although Borlaug’s work saw Ehrlich’s predicted threats displaced in time, towards the end of his Nobel Prize acceptance speech, Borlaug said:

Malthus signaled the danger a century and a half ago. But he emphasized principally the danger that population would increase faster than food supplies. In his time he could not foresee the tremendous increase in man’s food production potential. Nor could he have foreseen the disturbing and destructive physical and mental consequences of the grotesque concentration of human beings into the poisoned and clangorous environment of pathologically hypertrophied megalopoles. Can human beings endure the strain? Abnormal stresses and strains tend to accentuate man’s animal instincts and provoke irrational and socially disruptive behavior among the less stable individuals in the maddening crowd.

We must recognize the fact that adequate food is only the first requisite for life. For a decent and humane life we must also provide an opportunity for good education, remunerative employment, comfortable housing, good clothing, and effective and compassionate medical care. Unless we can do this, man may degenerate sooner from environmental diseases than from hunger.

So far, the international, multilateral approach to a possibly gloomy future is to seek hope — it does, after all, spring eternal.  We will reduce greenhouse gas emissions, tackle global poverty through economic growth, and increase food production. We will not need to tackle population growth, nor will we have to make do with less. We write about planetary health, but we do not develop strategies for a planet that is less human-friendly tomorrow than it is today.

I hope that global health and well-being will improve well into the future, well past my life and I hope well past that of my children, (and their children, …). In case it does not, I would like to think that there is a Global Health Red Team that does not just echo gloomy news in the halls of power, but argues for and develops strategies suitable for the world in which we are all worse off.  What should our goal be in that worse off world?  Is it a global goal, an equitable goal of mutual pain, or is it a “My Country First”, Shakespearean tragedy of the commons?

There is an ironic twist to the use of Red Teams in the US military that may have some bearing on their use in Global Health.  In the Millenium Challenge 2002 when the Red Team devastated the Blue Team in the first few days of a fortnight-long exercise, the judges reset the clock. They hamstrung the Red Team, and then let everything play out in a way that would ensure that normal (military) science came out unscathed.

Global Health needs to be intellectually braver.

 

Guidelines for the reporting of COde Developed to Analyse daTA (CODATA)

I was reviewing an article recently for a journal in which the authors referenced a GitHub repository for the Stata code they had developed to support their analysis. I had a look at the repository. The code was there in a complex hierarchy of nested folders.  Each individual do-file was well commented, but there was no file that described the overall structure, the interlinking of the files, or how to use the code to actually run an analysis.

I have previously published code associated with some of my own analyses.  The code for a recent paper on gender bias in clinical case reports was published here, and the code for the Bayesian classification of ethnicity based on names was published here. None of my code had anything like the complexity of the code referenced in the paper I was reviewing.  It did get me thinking however about how the code for statistical analyses should be written. The EQUATOR (Enhancing the QUAlity and Transparency Of health Research) Network has 360 separate guidelines for reporting research.  This includes guidelines for everything from randomised trials and observational studies through to diagnostic studies, economic evaluations and case reports. Nothing on the reporting of code for the analysis of data.

On the back of the move towards making data available for re-analysis, and the reproducible research movement, it struck me that guidelines for the structuring of code for simultaneous publication with articles would be enormously beneficial.  I started to sketch it out on paper, and write the idea up as an article.  Ideally, I would be able to enrol some others as contributors.  In my head, the code should have good meta-data at the start describing the structure and interrelationship of the files.  I now tend to break my code up into separate files with one file describing the workflow: data importation, data cleaning, setting up factors, analysis.  And then I have separate files for each element of the workflow. My analysis is further divided into specific references to parts of papers. “This code refers to Table 1”.  I write the code this way for two reasons.  It makes it easier for collaborators to pick it up and use it, and I often have a secondary, teaching goal in mind.  If I can write the code nicely, it may persuade others to emulate the idea.  Having said that, I often use fairly unattractive ways to do things, because I don’t know any better; and I sometimes deliberately break an analytic process down into multiple inefficient steps simply to clarify the process — this is the anti-Perl strategy.

I then started to review the literature and stumbled across a commentary written by Nick Barnes in 2010 in the journal Nature. He has completely persuaded me that my idea is silly.

It is not silly to hope that people will write intelligible, well structured. well commented code for statistical analysis of data.  It is not silly to hope that people will include this beautiful code in their papers.  The problem with guidelines published by the EQUATOR Network is in the way that journals require authors to comply with them. They become exactly the opposite of guidelines, they are rules — the ironic twist on the observation by Geoffrey Rush’s character, Hector Barbossa in Pirates of the Caribbean.

Barnes wrote, “I want to share a trade secret with scientists: most professional computer software isn’t very good.”  Most academics/researchers feel embarrassed by their code.  I have collaborated with a very good Software Engineer in some of my work and spent large amounts of time apologising for my code.  We want to be judged for our science, not for our code.  The problem with that sense of embarrassment is that the perfect becomes the enemy of the good.

The Methods sections of most research articles make fairly vague allusions to how the data were actually managed and analysed.  One may make references to statistical tests and theoretical distributions.  For a reader to move from that to a re-analysis of the data is often not straight forward.  The actual code, however, explains exactly what was done.  “Ah! You dropped two cases, collapsed two factors, and used a particular version of an algorithm to perform a logistic regression analysis.  And now I know why my results don’t quite match yours”.

It would be nice to have an agreed set of guidelines reporting COde Developed to Analyse daTA (CODATA).  It would be great if some authors followed the CODATA guidelines when they published.  But it would be even better if everyone published their code, no matter how bad or inefficient it was.

 

Who will guard the journals? Gender bias in the “Big Five” medical journals.

Journals, by which I mean Editors, have shaped modern science, particularly in medicine. The publication policies of journals now direct the kinds of ideas that are acceptable, how to present the ideas and the ethical frameworks that should govern data collection, authorship, treatment of participants, and data sharing. Journals will refuse to publish a paper if they are not satisfied that the authors have fulfilled those requirements. The journals have become both arbiters and gatekeepers of sound scientific practice. A recent journal issue on conflicts of interest appearing in the Journal of the American Medical Association (JAMA) is a case in point (2 May 2017).

Editors will also self-publish encyclicals of good conduct, laying down the rules of engagement for the future. The JAMA editorial supporting the recent special issue is such an example. When the journals involved are at the top of their fields, these views reverberate. In medicine, the Big Five journals in general and internal medicine are the New England Journal of Medicine (NEJM), Lancet, JAMA, British Medical Journal, and Annals of Internal Medicine. When the Editors speak, the field listens. Their role is revelatory. It is an imperfect conduit of nature’s voice whispered to researchers in their labs and clinics.

The rules do not, unfortunately, prevent the publication of bad science. The Autism-MMR paper in the Lancet is an excellent example of bad science slipping into the field. In general, however, failures of science lie at the feet of the scientists. The journals rise above it. A retraction here, a commentary there, and the stocks or pillory of peer humiliation are kept for the authors.

It is easy when criticism can be deflected, and laid at the feet of authors. What, however, should the response be when researchers identify a bias in the Big Five journals? Bias in medicine is a serious issue. It indicates a skew in the published science – a tendency to emphasise one kind of science over another or the promotion of one interest over another. It carries risks into the future of skewing practice and funding.

In 2016, Giovanni Filardo and colleagues identified a gender bias in first authors of research articles published in the Big Five. The journals were more likely to publish articles with a man as a first author than a woman. The most biased journal was NEJM. You will not have read about the research in that journal, however, because they rejected the paper when it was submitted. Unfortunately, the bias in the gender of published first authors is not a local, journal issue. The bias has a larger and more insidious career effect. Women are less likely to be in the prestigious position of the first author in prestigious Big Five journals, and ceteris paribus they are disadvantaged in funding applications, job applications, receipt of awards, and recognition.

My co-authors and I recently published an investigation of gender bias in clinical case reports. You may be unsurprised to learn that clinical case reports are more likely to be about men. Apparently, clinical cases about a man are just more interesting than a clinical case about a woman. All but one of the investigated journals showed a gender bias, and the most biased journal was NEJM.

Of course, journals can and should reject research papers that are not relevant or deficient in quality. And our paper may have been both. The fact that a journal like the NEJM should have rejected two recent papers that identified the journal as being the most gender-biased among the Big Five begins to look like an avoidance of criticism.

If there is a tendency to avoid self-reflection, particularly in an area as important as bias in science, then the editorial decisions begin to have much greater significance, and at least a whiff of hypocrisy. The origins of a bias may be authorial. A greater proportion of articles written by men than women are submitted to the journal; a greater proportion of clinical case reports about men rather than women are submitted to the journal. The Editors are in a position to correct the submission bias, just as they vigorously correct other biases. The Big Five would have acceptance rates below 10%; they presumably have a bias towards higher rather than lower quality science. We are suggesting that in exercising their Editorial judgment they could include factors they have (presumably) hitherto not noticed in their own behavior. They might find it easier to explain these editorial shifts if they based it on scientific research published in their own journals. At the very least it indicates that the issue is taken seriously.

This article was co-written by Daniel D Reidpath and Pascale Allotey

Sharing data while not sharing data

There has been a major shift among journals towards making data available at the time of publication.  The PLoS stable of journals which includes PLoS Medicine, PLoS Biology, and PLoS One, for example, have a uniform publication policy that is quite forthright about the need to share data.

I have mixed feelings about this.  I have certainly advocated for data sharing and (with Pascale Allotey)  conducted one of the earliest empirical investigations of data sharing in Medicine.  I can understand, however, why researchers are reluctant to provide open access to data. The data can represent hundreds, thousands, or tens of thousands of person-hours of collection and curation. The data also represent a form of Intellectual Property in the development of the ideas and methods that lead to the data collection. For many researchers, there may be a sense that others are going to swoop in and collect the glory with none of the work. There have certainly been strong advocates for data sharing where the motivation looked to be potentially exploitative (see our commentary).

I recently stumbled across a slightly different issue in data sharing.  It arose in an article in PLoS One by Buttelmann and colleagues. Their study looked at whether great apes (Orangutans, Chimpanzees and Bonobos) could distinguish in a helping task between another’s true and false beliefs.  The data set comprised 378 observations from 34 apes in two different studies, and they made their data available … as a jpg file.  A small portion of it appears below, and you can download the whole image from PLoS One.

Partial data from Buttleman et al. (2017)

It seems strange to me to share the data as an image file.  If you wanted people to use the data, surely you would share it as a text file, CSV, xlsx, etc.  If the intention was to satisfy the journal requirements but discourage use, then an image file looks (at first glance) to be a perfect medium.  Fortunately, there are some excellent online tools for optical character recognition (OCR), and the one I used made quick work of the image file.  I downloaded it as in xlsx format, read it into R, and cleaned up a few typographical errors that were introduced by the OCR. You can download their data in a machine-readable form here. I have included in the download an R script for reading the data in and running a simple mixed effects model to re-analyse their study data. My approach was a little better than theirs, but the results look pretty similar. I am not sure why they did not account for the repeated measurement within ape, but ignoring that seems to be the typical approach taken within the discipline.