Statistics is often seen as a science (e.g., statistical science), but it can also be viewed as an art. Professor David Spiegelhalter’s latest book explains how the art of statistics can enlighten our life with wisdoms and insights.

David Spiegelhalter is an accomplished scientist – statistician. He is Chair of the Winton Centre for Risk Evidence Communication in the Statistical Laboratory at the University of Cambridge. He is Professor of Public Understanding of Risk at Cambridge University. He was elected as President of the Royal Statistical Society (2017-2018), Fellow of the Royal Society in 2005, awarded an OBE in 2006, and knighted in 2014. He is one of the world’ most-cited and influential scientists in the field of statistics.

However, don’t let his academic background put you off. He is a great teacher of statistics, and a wonderful communicator of risk. His style is fairly conversational, often warm, and sometimes witty. The Art of Statistics is a testament to those traits. Throughout the book, readers will not find any formulae. Introducing a tough topic such as statistics without a single formula is to me a sign of a masterful work.

To many people — in as well as out of academia — statistics is a hard topic. Lay people often understand statistics to be only about counts and frequencies. Most people can hardly make sense of proportions, odds, or average, let alone the coefficient of correlation. But they are not alone; generation after generation of university students also struggle to understand abstract statistical concepts operationalized by mostly esoteric formulae. I myself used to find statistical texts rather dull, because they are populated with coin flipping and gambling exercises which seem irrelevant, even thoughtless, to the real world of science and society. Spiegelhalter’s Art of Statistics is not like those texts; it is — or promises to be — a joyful reading.

The book can be seen as a thin veiled introduction to statistics, but it is not necessarily written for university students; it is for lay people and the public at large. Lay people love stories and questions, because they serve as catalysts for productive conversations. And, the book is structured with plenty of interesting stories that surely stimulate thinking in many creative ways. Spiegelhalter starts off with a tragic story about an English doctor named Harold Shipman who had killed at least 215 patients between 1975 and 1998. The Shipman’s story includes actual data on age, gender, time and year of death of all victims. A simple listing of those data does not tell us anything. Instead, Spiegelhalter applied some simple statistical graphics to visualize the data, and the result is an insight into the crime: most of Shipman’s victims were in their 70s and 80s, and they tended to die in the early afternoon. Of course, the analysis could not tell us why they died at that time, but an official enquiry into the crime revealed that Shipman often visited his patients at home after lunch, when he would inject them with a massive dose of diamorphine he said was to make them more comfortable. The simple analysis enable the data to ‘speak’.

Through the analysis, Spiegelhalter makes the point that “numbers do not speak for themselves, we are responsible for giving them meaning” (page 27). Statistics is a very powerful tool to extract lively information from cold data.

Fig 1: a simple scatter plot of age at death and year of death of 215 victims. The plot reveals that most victims were in their 70s and 80s. There is a ‘gap’ between 1990 and 1992 in the graph, because during that period Harold Shipman worked for a private practice.

From the tragic Shipman’s story, Spiegelhalter launches into a series of discussions on various statistical subjects. These include the analyses of categorical data and continuous data; using sample data to make inferences concerning population parameters; hypothesis testing; modelling relationships using linear regression and logistic regression; survival analysis; and prediction and analytics. The book also touches on topical issues such as reproducibility and risk communication. Spiegelhalter is a renowned Bayesian statistician, and readers can surely bet that he has a chapter on Bayesian inference which he termed ‘Learning from Experience‘.

Each of those subjects is commenced with a story or a question. Starting with the question of “How many trees are there on the planet“, Spiegelhalter discusses what constitutes a “tree” (which is incidentally defined as at least 12.7 cm in diameter breadth height), and how to answer the question by using sampling techniques. Other questions in the books are also well discussed: What is the cancer risk from bacon sandwiches? Can we trust the wisdom of crowds? Do busier hospitals have higher survival rates? Does going to university increase the risk of getting a brain tumour? Do statins reduce heart attacks and strokes? Is prayer effective? Why do old men have big ears? etc. These questions seem deceptively simple, but the answer — or the process of reaching an answer — is quite a scientific journey.

That scientific journey is called PPDAC which was advanced by New Zealand, a world leader in statistics education in school. In this approach, a research cycle is started off with a problem or research question (P). The second stage is to plan (P) a data collection procedure (i.e, an experiment), and to decide the variables to be measured in relation to the research question. The third stage is to actually perform the data collection (D), which means measurement, management, and cleaning. The fourth stage involves the analysis of data (A) which includes visualization, numerical description, and statistical modelling. The final stage is to make conclusion (C) or interpretation of the results of analysis. A new hypothesis or research question can be formed from the conclusion, and a new cycle of PPDAC starts again. I think that is a brilliant way to introduce statistics to students and the public at large.

Apart from those topics, Spiegelhalter briefly touches on ‘data science’, which is increasingly becoming a parlance in modern science these days. However, the definition of data science has not been consistent. Some think that data science, defined as ‘the extraction of knowledge from data‘, is simply a new name for statistics in the computer age; others argue that it is a combination of computer science and statistics. My view is that data science is an integration of statistics and informatics. Informatics here encompasses information science, computation, and software engineering. While some people view data science as a threat to statistics, I would consider that data science is an opportunity for statistics. The opportunity to learn new concepts and ideas, and to engage in shaping the research agenda of cutting edge science such as genomics.

The book is aptly subtitled ‘Learning from Data‘, suggesting an air of ‘machine learning’. Actually, it is about statistical learning, not machine learning. Using the data collected from the Titanic disaster, Spiegelhalter introduced the idea and method of classification and metrics for evaluating the accuracy of a classification model. Issues of over-fitting and training data are also discussed. Readers are reminded that a statistical model — whether it is about association, classification or prediction — is a simple representation of the complexity of the natural world. A statistical model has two key components, a deterministic one and a random error, with the latter being expressed in terms of a probability distribution. As such, all statistical models are wrong, but as Tukey said, some models are still useful. And, to evaluate the usefulness of a statistical model is quite an art itself.

No discussion of statistics is complete without mentioning the P-value and the concepts of hypothesis testing and significance testing. Spiegelhalter does not discuss the history of the P-value idea in the book, but he does give a very standard definition. He also puts the P-value in the context of the current research ecosystem by asking the following question: what would happen if 1000 studies aimed to test 1000 hypotheses, each designed with type I error of 5% and type II error of 20% (ie 80% power), and if 10% of the null hypothesis were true? The answer is that we would expect that 36% of these discoveries are false discoveries. This is indeed a depressing picture of our present research literature. It is not surprising that the current crisis of reproducibility has been blamed on the misue of P-values.

We are living in a world awash with data which are daily propagated by the popular media, and this book can help us make sense of data (and even distinguish between good data from bad). A very important thing to remember is that data are products of social activities. And, to make sense of data, we must consider data in social context. Data without context mean nothing. For instance, statistical correlation does not necessarily mean causation — a point raised by a commentator in Nature in 1900, when the coefficient was proposed by Karl Pearson.

A beautiful example of data-in-context is the story about bacon and cancer risk. In September 2015, the World Health Organization (WHO) and International Agency for Research in Cancer (IARC) announced that processed meat (eg bacon, sausages) was classified as ‘Group I carcinogen’, in the same category as cigarettes and asbestos. The reclassification was based on research data showing that each 50g of processed meat was associated with an 18% increase in the risk of bowel cancer. However, when the increased risk is considered within the background data (~6 out of 100 people are expected to get bowel cancer in their lifetime), the 18% number means one extra case, or 7 out of 100 lifetime bacon eaters, a very modest risk. The point of this simple exercise is that health risk should be contextualized and presented in absolute terms.

Sometimes, there is a total mismatch between what the research data show and what is conveyed to the public. The story about the ovarian cancer screening is a good case in point. On May 5, 2015, The Independent ran an article proclaiming that ‘Ovarian blood tests breakthrough: huge success of new testing method could lead to National Screening Program.‘ However, the claim could not be further from the truth. The fact is that in the UK ovarian cancer clinical trial (involved more than 200,000 women), researchers found no statistically significant benefit of screening. Thus, there was an obvious breakdown of communication between researchers who produced the statistics and the popular media that conveyed the statistics to the public. Spiegelhalter suggests a list of 10 questions to ask when confronted with a claim based on statistical evidence. The 10 questions are grouped into three broad groups: the trustworthy of the statistics; the trustworthy of the source that produced the statistics; and the trustworthy of the interpretation.

The Art of Statistics is an excellent and timely account of statistics in the age of competing interests between emerging fields such as statistical learning, machine learning, deep learning, and the likes. Irrespective of the perceived competing interests, statistics still remains a science of extracting knowledge from data. It is probably not an overstatement to say that statistical science has the potential to shape research agendas of cutting edge sciences, and deepen our understanding of the natural world. The Art of Statistics further enlightens our daily life with surprising wisdoms. It is a book that should be in the bookshelf of every research laboratory, and would be a terrific addition to all family libraries. I highly recommend it.

====

PS: All figures and data in the book are available from the following webpage:

https://github.com/dspiegel29/ArtofStatistics

The first IMRaD part of a scientific paper is the Introduction. The main purpose of the Introduction is to answer the following very simple question: ‘Why did you do this study?’ To answer that question, you need to briefly survey the literature, identify a gap in knowledge, and state your specific aim(s). In this note, I would like to offer you a strategy for writing the Introduction, and it is called “CaRS.”

Introduction: from general to specific

The Introduction of a scientific paper can be considered an educational material for readers. By reading the introduction, readers will have key concepts and ideas of the research project that the authors (you) are going to tell in the paper. Apart from learning about societal challenges, scientific issues, readers will learn scientific terminologies in the field. Thus, you can think of the Introduction as a way to educate your readers about your scientific concern and its significance.

The Introduction of a paper should be structured in the deductive format. When I said ‘deductive’, I mean the argument in the Introduction has to proceed from a general issue to a specific research question. In order words, the Introduction should begin with a big problem in the field, and end up with a specific research question or research aim.

A strategy for moving from general issue to specific issue is CaRS. CaRS stands for “Creating a Research Space.” The CaRS strategy is very popular among natural and social sciences writers. There are three steps in the strategy: first, you create a situation; second, you identify a niche for your study; and third, you occupy the niche. The ‘occupation’ of the nice means that you must state a research hypothesis (if possible) and then followed by a primary aim of the research project.

In the following, I am going to describe the CaRS strategy in more detail. I shall use two published papers on obesity (1) and rosiglitazone drug (2) to illustrate the writing strategy.

Tell readers: what are known

The first part of an introduction should be designed to create a situation. Here, you tell readers your research area, and the big issue that the field is currently facing. You also need to tell readers that the issue is of importance. And, there are several ways to demonstrate the importance, including the magnitude of the problem, the consequence of the problem in terms of human cost and societal impact. In short, the first part of an introduction is where you tell readers what is known.

For example, in a study of the relationship between obesity and body composition in Vietnam, the authors argue that obesity is an important global public health problem. They then characterize the importance in terms of high prevalence, morbidity consequences, and mortality. Here is their introduction (1):

“Obesity is recognized as a global health problem because it affects a large proportion of individuals in developed and developing countries. In the United States, 32% of adult men and 35% of adult women are obese (ie BMI ≥ 30 kg/m2). In Asia, approximately 17% of population is considered obese by World Health Organization (WHO). Obesity is associated with an increased risk of mortality, type 2 diabetes, cardiovascular diseases and cancer. Moreover, obese individuals have 7 times higher the risk of developing diabetes than individuals of a normal BMI. Since obesity is increased with advancing age, the on-going rapid aging of the population will further impose a greater burden on the society. Indeed, it has been estimated that by 2030 nearly one-third of the world population is overweight or obese.”

In a study of the adverse effects of rosiglitazone (a common anti-diabetic drug), the authors provide a brief but essential information concerning the use of thiazolidinedione. They write a nice very succinct introduction (2) as follows:

“Thiazolidinedione drugs are widely used to lower blood glucose levels in patients with type 2 diabetes mellitus. In the United States, three such agents have been introduced: troglitazone, which was removed from the market because of hepatotoxicity, and two currently available agents, rosiglitazone (Avandia, GlaxoSmithKline) and pioglitazone (Actos, Takeda). The thiazolidinediones are agonists for peroxisome-proliferator–activated receptor γ (PPAR-γ). PPAR-γ receptors are ligand-activated nuclear transcription factors that modulate gene expression, lowering blood glucose primarily by increasing insulin sensitivity in peripheral tissues. Rosiglitazone was introduced in 1999 and is widely used as monotherapy or in fixed-dose combinations with either metformin (Avandamet, GlaxoSmithKline) or glimepiride (Avandaryl, GlaxoSmithKline).”

Tell readers: what are unknown

The second part of an introduction should identify a nice, a gap in knowledge. The knowledge gap can be stated in terms of lack of data (or up-to-date data), a controversial viewpoint, or a novel method has not been applied. In short, here you tell readers what are unknown.

For the obesity study (1), the authors talk about the lack of consensus on the definition of obesity in Asian populations (1), and that can be considered a niche:

“Although it is believed that obesity is increasing in Asian populations, there is actually no consensus on the definition of obesity for Asians. In 2004, a WHO expert consultation concluded that Asian individuals are at greater risk of type 2 diabetes and cardiovascular disease with a lower BMI than their Caucasian counterparts, but the consultation did not come up with a consensus cut-off BMI for defining obesity in Asians. The consultation also proposed that the WHO BMI cut-off points should be retained as international classifications. In reality, some groups use the BMI ≥25 or BMI ≥27.5 as a cut-off value for the diagnosis of obesity in Asian men and women.”

For the rosiglitazone study, the authors identify the niche by pointing out that the drug was initially approved for its function of reducing blood glucose level, but its safety profile was not known. In addition, they point out that initial studies on potential adverse effects were based on small sample sizes, and therefore, could not detect a true association should it exist (2):

“The original approval of rosiglitazone was based on the ability of the drug to reduce blood glucose and glycated hemoglobin levels. Initial studies were not adequately powered to determine the effects of this agent on microvascular or macrovascular complications of diabetes, including cardiovascular morbidity and mortality. However, the effect of any antidiabetic therapy on cardiovascular outcomes is particularly important, because more than 65% of deaths in patients with diabetes are from cardiovascular causes.”

Tell readers: hypothesis and aim

Finally, the last paragraph of an introduction is where you occupy the niche that you have identified in the second part. “Occupation” here means that you propose a solution to the identified problem. The best way to state your solution is by stating a testable hypothesis and a specific research aim. Some people state more than 1 research aim, and that is fine too. I should mention that a hypothesis is a predictive statement, not simply a neutral statement.

In the obesity study (1), after pointing out that there has been lack of studies in the Vietnamese population, the authors state their research aim:

“Vietnam is a developing country with a population of approximately 92 million, representing 1.3% of the world population. Approximately 70% of the population lives in rural areas. During the past 20 years, the country has continued to be one of the world’s fastest economic growth, with annual growth rate of ~5%. Parallel with the economic development, Vietnam has also undergone remarkable changes in dietary patterns which led to a change in BMI. Therefore, the population is an ideal setting for studying the burden of obesity in transitional economies. However, no studies in the past have examined the burden of obesity using PBF as an indicator in Vietnam. Thus, in this study, we sought to analyze the association between PBF and BMI, and to define the prevalence of overweight and obesity using both BMI and PBF criteria.”

In the rosiglitazone study (2), having identified the lack of data concerning the potential adverse effects of the drug, the authors state their aim very clearly. They also make a preamble statement pertaining to the data that were used in the study:

“Therefore, we performed a meta-analysis of trials comparing rosiglitazone with placebo or active comparators to assess the effect of this agent on cardiovascular outcomes. The source material for this analysis consisted of publicly available data from the original registration package submitted to the Food and Drug Administration (FDA), another series of trials performed by the sponsor after approval, and two large, prospective, randomized trials designed to study additional indications for the drug.”

Common pitfalls

Although its structure is simple, it is not always straightforward to write a good introduction. In practice, I have seen many peer-reviewed papers that do not state any research aim or any hypothesis! Some common pitfalls I have seen are as follows:

· Exhaustive review of literature review. Some writers write a long introduction with so many historical information. That is probably not necessary because historical data can distract readers from the issue at hand. You don’t have to review the idea/issue from the time of Hippocrate; you simply get straight into the problem that you have identified.

· Failure to state the global problem behind your research. You should always start off with a big / global problem so that readers can have some idea which angle you are from.

· Failure to state the importance of the problem. Readers like to read papers that address an important problem, and it is your challenge to ‘hook’ them into your research world.

· Failure to state research aim. This is the worst pitfall in many many papers. It is annoying to read a 2-page introduction only to find out there is no specific aim! The lack of study aim or hypothesis is one of the common reasons for rejecting manuscripts.

In summary, the Introduction is an important component of a scientific paper, because it acts as an educational materials for readers. Based on the Introduction, readers will decide to continue reading the paper or not. The Introduction should ‘hook’ readers with 3-part structure: knowns, unknowns, and hypothesis/aim. The strategy for writing the Introduction is CaRS — creating a research space.

The introduction is sometimes likened to the gate of a scientific work. You, of course, want the gate to be attractive so that readers can enjoy your work. That means your introductory arguments should be clear, concise, analytic, and evidence-based. The clarity, precision and evidence will attract readers to explore other parts of the paper. To borrow the Greek motto, “a bad beginning makes a bad ending“, a bad introduction will likely make a bad conclusion.

Here is a check-list of the most important information that you should cover in the introduction:

• Have you answered the question: why did you undertake this study?

• Is there a specific aim or hypothesis?

• Is the research question important and impactful?

• Medical research paper: no more than 2 pages ?

Good luck!

_______

(1) https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0127198

(2) https://www.nejm.org/doi/full/10.1056/nejmoa072761

If you ask an experienced researcher ‘which section should I start composing a scientific paper’, you are likely to be advised that you should start with the Results section. Others may say that you should start with the Methods section, because it is probably the easiest section to write. I have a different opinion: I think you should start with neither Results nor Methods. Before writing a paper, you should sit down, get a cup of coffee (or tea), and sketch out key idea and information to be included in the paper. I called this step “set the stage“, similar to artists set the stage for a performance.

In the previous note, I have already provided some key principles for writing an impact title and an abstract. Title and abstract are two important components of a scientific papers, because they are the first contacts that most journal editors and readers have. Now comes the harder task: composing the paper. In this note I am going to share with you some experience in sketching out key ideas for composing a paper.

A scientific paper is usually structured according to the IMRaD format: Introduction, Methods, Results and Discussion. Although that is the standard format, the sequence of writing is not necessarily in that order. Many experienced researchers think that the Results is where you should start writing, but others suggest that it is better to start with the Methods section because most information of that section has already been written in a proposal. I have a different idea. I suggest that you start off with a sketch of key ideas for the paper; some people call this sketch a ‘micro paper’ (1).

Your micro paper should answer the following questions:

· What is the global / societal / general problem?

· What is the specific and scientific problem?

· What is the hypothesis?

· Key information pertaining to the study design / experiment

· What are new findings?

· Key figure and table?

· What is the importance of the result?

· What is the scientific / local implication of the finding?

· What is the global implication of the finding?

Micropaper: set the stage for writing

In the global problem/issue, you identify a single problem of ‘big picture’ nature. Just one or two sentences will do. For example, in a study comparing the percent body fat (PBF) between Australians and Koreans, I approach the problem from a big picture of obesity as a global public health problem. Obesity is also increasingly recognized as a major issue in Asian populations.

I then write down one of two local/scientific problems. The local/scientific problem is that BMI is not a good measure for diagnosing obesity, and that there is no validated criterion for the diagnosis of obesity in Asian people.

From the scientific problem, I state my hypothesis. There is a call for lowering the BMI threshold for the diagnosis of obesity in Asians, and this call was based on the assumption that for a given BMI, Asians have higher fat mass than Caucasians. However, based on some recent data, I hypothesize that for a given BMI, Asians actually have lower PBF than Caucasians. In other words, my hypothesis is a falsification of the common assumption.

Study design and methods. This is where I sketch out key details concerning study design and method. Again, one or two points are enough for further writing.

In the “key findings”, I write down two most important findings from the study. First, for the same BMI, Asians indeed had higher PBF than Caucasians. Second, the ‘effect’ was mainly seen in women, not in men.

The micro paper should have one or two key figures. Here, I sketch out the kind of graph that I wanted to present in the paper. I call these graphs “money data”, meaning that they are the most expensive pieces of data of the study.

Then, in the discussion, I sketch out one or two points concerning the scientific importance, new findings, local and global implications of the findings. That should conclude the paper.

So, before writing anything at all, you should find some time to sit down and write down key idea and data for the paper. I normally sketch out my idea and data in a paper (see photo). I don’t like to think in front of a computer screen, because it is highly distracting. With a blank paper and a pen, I can express my thinking on the paper clearer and more emotive.

====

(1) Eric Lichtfouse. Scientific Writing for Impact Factor Journals presentation. Web:

https://www.researchgate.net/publication/278769971_Scientific_writing_for_impact_factor_journals

I am pleased that our paper on the fracture attribution has recently been published by the Journal of Clinical Endocrinology and Metablism. We found that in men and women aged 60 yr and older only 20% of all fractures were attributable to osteoporosis. I think that this finding has important implication for the prevention of fracture in the general population. I have also had a chance to discuss our work with Endocrinology Today here.

The underlying idea of this work was actually derived from the axiom articulated by Goeffrey Rose more than 30 years ago. In an influential essay, “Sick individuals and sick population” (Int J Epidemiol 1985), Rose postulated that when the relationship between a risk factor and disease is continuous, a truncation of the distribution of the risk factor into “low risk” versus “high risk” group will result in the majority of disease cases in the low risk group rather in the high risk group.

In this study, my student (Ha Mai), my colleagues and I sought to test the Rose’s axiom by examining the relationship between bone mineral density (BMD) and fracture. We all know that people with lower BMD have higher risk of fracture, and that the relationship between BMD and fracture is continuous. However, in 1994, a WHO expert panel recommended that those with BMD T-score ≤ -2.5 be diagnosed as having “osteoporosis”.

I just would like to make a note here that at that time, some prominent experts were not impressed with the idea of T-score dichotomization. In a commentary in Clinical Rheumatology (1), Dr. Richard Wasnich writes that “What are the issues surrounding the use of T-scores, as recommended by the WHO panel? On the one side, they seemingly offer simplicity, which is sorely needed. However they are not readily translated into interventional guidelines. The opposing viewpoint is that T-scores are a major step backwards into the realm of fracture thresholds.” He continues: “Fundamental to this debate is the fact that bone density is a risk factor, and not a diagnostic test. So making a ‘diagnosis’ of osteoporosis based on the presence of a single risk factor, at a single point in time, is already a tenuous concept.” He concludes that “what we need is an estimate of absolute fracture rate […] There is no need to obscure this useful information by inventing a new statistic, e.g. the “T-score.“

At the time, I was a relatively young fellow who had just entered the field for only 5 years. However, those words of Dr. Wasnich have haunted me for a long time. I thought he was on something, but I could not have time to work it out. Then, I came across a Goeffrey Rose’s book, The Strategy of Preventive Medicine, which was an eye-opening reading for me. So, I was determined to pursue the idea proposed by Wasnich and Rose.

Back to to the T-score and osteoporosis, in subsequent years, epidemiologic studies showed that among post-menopausal women, the prevalence of osteoporosis was around 20%. More importantly, many RCTs showed that treating women with osteoporosis reduced their risk of fracture by about 50%. However, according to Rose’s reasoning, we can predict that the majority of fractures would be occurred in those without osteoporosis, and treating those with osteoporosis will do little to reduce the burden of fracture in the general population.

In this study, we used data from the famous Dubbo Osteoporosis Epidemiology Study that involved 3700 men and women, all aged 60 yr and above at baseline. Fracture was prospectively ascertained using X-ray reports. We used a statistical method called “heuristic population attributable risk” (Hanley, 2001) to work out the proportion of fractures that is attributable to osteoporosis and advancing age. We defined “advancing age” as those aged 70 yr and older.

So, what did we find? Well, we found many things, but some key findings can be summarized as follows:

• the prevalence of osteoporosis (using the WHO criteria) was 21% in women and 11% in men;
• the risk of fracture among people with osteoporosis was increased by 80% (compared with those without osteoporosis), but the magnitude of association progressively declined with time;
• approximately 20% of total fractures were attributable to osteoporosis within the first year of follow-up, decreasing to 17% by year 5, further decreasing to 14% by year 10 and rising to 19% by year 20.
• by accounting for advancing age, we found that ~35% of total fractures in women and men were attributable to advancing age and/or osteoporosis.

Those findings support the Rose’s axiom: a large majority (two-thirds) of fractures was not attributable to either osteoporosis or advancing age. From a population prevention point of view, this finding suggests that if we focus on treating those with osteoporosis, we would reduce a very small proportion of fractures in the general population.

I thought the finding therefore has important public health implication. I submitted the manuscript to BMJ, but they politely rejected our work, because — they said — it would be more suitable to a specialist journal. The manuscript was then submitted to JCEM where it received a more receptive review. One reviewer complained about the word “attributable“, because the reviewer thought that the word implies causation whereas all we can say is an association. Of course, the reviewer is correct, but the problem is the “population attributable risk” expression has been well established in the lexicon of epidemiology.

Another reviewer suggested to change the title to something more interesting. Our original title was “Fractures and fracture-associated mortality attributable to low bone mineral density and advancing age: a time-variant analysis“. However, the reviewer thought that the title did not capture the key finding of the study. We then came up with a declarative title: “Two-thirds of all fractures are not attributable to osteoporosis and advancing age: implication for fracture prevention“. The reviewer seemed happy with the change. I would like to take this opportunity to thank both reviewers for their helpful and really constructive comments.

So, what is next? Well, in the Endocrinology Today interview, I said that the clinical implication is that we can’t predict who will fracture from a measurement of bone mineral density alone. Therefore, the assessment of fracture risk for an individual should move beyond bone mineral density to take into account — among others — important factors such as a history of fracture, prior falls, smoking habit, and co-morbidities. Tools such as the Garvan Fracture Risk Calculator and FRAX can help doctors to evaluate a patient’s risk of fracture based on multiple risk factors. I also touches on the use of genomics in fracture prediction. Fracture is partly due to hereditary factors, and over the past ten years or so, we have identified many genes that were associated with fracture risk. So, the next step is to develop a test that combines the information of these genes to identify high risk individuals much earlier, even at birth.

I should mention that in our study, we also wanted to answer the following question: how many deaths are attributable to advancing age, osteoporosis and fracture? The answer is: almost 80% of mortality were attributable to advancing age, osteoporosis and fracture; however, most of the attributable proportion was accounted for advancing age. The important implication of this finding is that patients with a fracture have reduced survival. At present, there are effective anti-osteoporosis therapies that can reduce the risk of mortality among patients with a fracture. However, the reality is that most (approximately 70%) patients with fracture do not receive any treatment, and that is a crisis. Our finding implies that many lives can be saved if doctors initiate treatment early.

_____

(1) https://www.healio.com/endocrinology/bone-mineral-metabolism/news/online/%7B98f1be52-5d01-47e1-9a08-d4af61a9ef53%7D/two-thirds-of-all-fractures-not-attributable-to-osteoporosis

(2) Wasnich RD, Clin Rheumatol 1997;16:337-339.