Statistic of the year: 90.5%

What is 90.5%? It’s the the proportion of plastic waste that has never been recycled. This startling number has been named the Statistic of the Year by the Royal Statistical Society. You can read the full story at https://www.statslife.org.uk/news/4026-statistics-of-the-year-2018-winners-announced.

Advertisements

Larry Hedges awarded 2018 Yidan Prize for Educational Research

The newsletter of the American Statistical Association carried this good ews stry in its November 2018 issue. Larr Hedges is (co)-author of some of the seminal work on meta-analysis, including the books “Statistical Methods for Meta-Analysis” and “The Handbook of Research Synthesis and Meta-Analysis”.

The ASA article states that Larry’s current work is at “the smallest scale of educational research, so called ‘single case’ designs involving a single individual. Such studies are crucially important in areas such as the study of rare disabilities, where it is unrealistic to assemble a large sample.” It’s interesting that a health example was the one given here, becasue studies where a vast array of data is collected on just one person (personal omics) hit the research stage in 2012, and in general I feel that studies where data sets are wide but not deep are becoming more and more common in biomedical research.

The ASA goes on to say that “The Yidan Prize is named for its founder, Charles Chen Yidan who established the prize in 2016 with a mission to make the world a better place through education.” Can’t fault that as a goal! Well done Larry Hedges!

Our cultures count: examining the measurement of cultural participation for Aboriginal people’s health and well-being in Central Australia

Alyson’s Thesis Proposal Review was presented in the NCEPH seminar series on Thursday 13 December.

She emphasised the notion of strength-based evidence rather then a deficit model, as well as the important principle of “No survey without service” to a local community.

The Mayi Kuwayu (MK) study is Alyson’s context for data collection and policy implication. Central Australia Land Council is her area of interest, which encompasses a large slide of the Northern territory including the towns of Alice Springs and Yulara, and 15 language groups. Alyson spoke of her interest in the tension (or is it a mutually informative relationship?) between epidemiology and ethnography and how they might both inform her research. One of her key research questions will be to see the extent to which national data collections can support local community / place-based decisions.

Her quantitative analyses will focus on responder profiles for MK, reliability and baseline analysis. Her qualitative analysis will include in depth interviews with 15 people regarding how they answered MK questions. Field notes kept on an ongoing analysis will form the basis of the reflective part of the thesis.

In conclusion, Alyson declared that she’s looking forward to coddiwompling – travelling purposefully towards an as-yet-unknown destination.

At the end of the seminar, an intersting line of questions developed around the meaning of reliability and validity in the MK context.

Collaboration across boundaries in Population Health

My entry in the poster competition part of “Collaboation across boundaries” on Tuesday 4 December was a renovated version of the globe that won People’s Choice last year. No such award this year, but mine was the only 3D poster apart from a piece of furniture from Design researchers. The fabric cover makes it so much easier to place text, and there have been other enhahcements since last year too. So nice to get more than one use out of a poster!

NECTAR 2018 globe Alice

 

Collaboration across boundaries

This day-long meeting of early and mid-career researchers took place at University House, ANU, on Tuesday 4 December. The keynote speaker, Professor Eleanor Huntington, Dean of CECS at ANU, began the day’s proceedings with an amazing inspirational speech. Her topic was :Why do we need to blur the boundaries in Australian research for it to benefit society and for us to lead a better future? She spoke passionately, personally, and without a single note on paper or screen. Two of the stories she told particularly resnated with me. The first was on how research letting arts studets loose on a pile of craft items revealed that some were problem finders (sorting through the materials to idntify patterns first) and some were problem solvers (jusy got on and made somethign as instructed). Actually it’s the problem finders who are more creative than the problem solvers. The second was on four characteritic of great collaborators: deep expertise, divergent thinking, motivation, confidence. I’m off to get me some more of all of those!

The second session of the day was occupied with a dozen pitches for collaborative research projects. These were refreshingly diverse, much more so than the stage show that the 3MT has become. ot every project was fully formed, which was also refreshing.

Ekavi Georgousopoulou, wants to use electronic patient data management input to multivariable models, neural networks and decision trees to predict outcome of an admission. One of my favourites (from a biased point of view!) she ended up with prizes for both poster and pitch.

Matt Thompson, is looking for collaborators to take the research on nanostructured surfaces on silicon and germanium to applications e.g. solar panels and batteries.

Xinyuan Xu is researching machine help for mental health in order to Iidentify at-risk individuals to enable push notifications.

Jin Huang is interested in material chemistry of 3D printing, aiming for a polymer that might be self-healing, stimulus-responsive, and so on.

Negar Ardakani, Arabic & Islamic Studies, reads Avicenna’s thousand-year old text for evidence of Ancient Persian medicine and CFS.

Jaban Moses, health informatics at Deakin University. is aiming to address the lac of cloud-based solutions for data analytics in sleep studies.

Claudia Munera is interested in collaborative networks tackling climate change.

Youssef Hafiz, is keen to improve ceramics used in knee replacements.

Caitlyn Byrt, Adelaide Uni, studies aquaporins, cells that do water transport across cell membranes and can be used for astronaut hydration too. One do my favourites, I was also on the money there as she came first in the pitch competition.

Joshua Chu-Tan and Nilisha Fernando study macular degeneration in a mouse model.

Trang Tan, is passionate about rubbish and recycling – like in Taiwan you meet the truck as yo would an ice cream van. We need solutions to deal with our own waste here, especially food waste. This was also one of my favourite presentations.

Andrew Ross of the Fenner School spoke about his experiences with the Ground water Solutions Initiative for Policy and Practice (GRIPP).

Ekavi and her prize-winning poster appear below.

NECTAR 2018 Ekavi poster

Eat cheese, live forever. Eat cheese, never die.

I’m delighted to present a guest post from Keeley Allen. She wrote this piece for a Biostatistics assignment and really nailed the learning outcomes I was aiming to address. The whole class had fun reading the press release, NBC report and Washington Post comments on the health benefits of cheese.

“There was a lot of excitement in the media at the end of August concerning cheese. Yes, cheese. While I get pretty excited by the prospect of some brie or gouda, this excitement wasn’t so much from the cheese itself as from the misinterpretation of a study’s results and mistaken causation. This excitement reminds us to carefully read through findings before making conclusions and to follow the golden rule of epidemiology: just because an exposure is associated with a particular health outcome does not automatically imply causation[1].

The source of all this excitement was a study presented at ESC Congress by Professor Banach and his co-researchers. This research included a longitudinal cohort study and meta-analysis of other studies considering the association between dairy consumption and different forms of mortality. Through their research, Banach and his co-researchers found statistically significant associations between dairy consumption and lower rates of certain kinds of mortality. The study found cheese consumption acts as a potential protective factor for total mortality risk and dairy consumption overall was associated with a lower total mortality risk and lower risk of cerebrovascular mortality[2]. The press release notes that causality is difficult to ascertain, especially in light of the finding that milk consumption is associated with higher risk of coronary heart disease mortality[3].

Seems reasonable, right? We have an association, no one is jumping to causation, and we have some suggested changes to dietary intake guidelines. Well, that is until NBC ran this headline and by-line:

“It’s time to reconsider low-dairy diets, new study suggests.

Cheese and yogurt were found to protect against death from any cause, and also against death from cerebrovascular causes, like stroke.”[4] (Emphasis added)

And the jump from association to causation is made. While it is pretty clear that cheese can’t protect you from every cause of death like bear attack or fire as one author has amusingly highlighted[5], the headline suggests a definitive, causal relationship that we cannot be sure of from the research. Translating ‘total mortality’[6] to ‘any cause of death’[7] is problematic and misleading for the reader.

By the nature of the study design, an observational study such as a cohort study, can identified association but cannot definitively determine causation. An experimental design, such as an intervention study where some people follow a dairy diet and a control does not, could potentially identifying a causal relationship, if such a relationship exists.

What the NBC reporter could have done is stress that this is not a definitive finding upfront. Making clear that the finding is an association and not a causation upfront can help the readers make sense of what the study has found and make a reasonable assessment of the findings and the impact on their lives. Less definitive language would also go a long way to prevent the jump to causation. The press release could have helped too by emphasising that their findings can only be an association by the nature of the study and further research would be needed to investigate a potential causal relationship.

As an aside, the study summary also doesn’t tell us how much dairy consumption provides these significant results. It’s unclear if ‘consumption’ of cheese or milk once a day, a week, a month or some other frequency has returned these results, so we may have to wait until the study is published before we take an extra big slice of camembert.

[1] Webb, P., Bain, C., Page, A. (2017) Essential Epidemiology: An introduction for students and health professionals. Cambridge University Press. Cambridge. Page 275.

[2] European Society of Cardiology (2018) Current advice to limit dairy intake should be reconsidered. Press Release. 28 August. Accessed 7 October 2018: https://www.escardio.org/The-ESC/Press-Office/Press-releases/Current-advice-to-limit-dairy-intake-should-be-reconsidered

[3] European Society of Cardiology (2018) Current advice to limit dairy intake should be reconsidered. Press Release. 28 August. Accessed 7 October 2018: https://www.escardio.org/The-ESC/Press-Office/Press-releases/Current-advice-to-limit-dairy-intake-should-be-reconsidered

[4] Charles, S. (2018) ‘It’s time to reconsider low-dairy diets, new study suggests’, NBC News, 29 August, Accessed 7 October 2018: https://www.nbcnews.com/health/heart-health/it-s-time-reconsider-low-dairy-diets-new-study-suggests-n904491

[5] Petri, A (2018) ‘EAT CHEESE LIVE FOREVER EAT CHEESE NEVER DIE’, Washington Post, 31 August, Accessed 7 October 2018: https://www.washingtonpost.com/news/opinions/wp/2018/08/31/eat-cheese-live-forever-eat-cheese-never-die/?noredirect=on&utm_term=.5f610146ea48

[6] European Society of Cardiology (2018) Current advice to limit dairy intake should be reconsidered. Press Release. 28 August. Accessed 7 October 2018: https://www.escardio.org/The-ESC/Press-Office/Press-releases/Current-advice-to-limit-dairy-intake-should-be-reconsidered

[7] Charles, S. (2018) ‘It’s time to reconsider low-dairy diets, new study suggests’, NBC News, 29 August, Accessed 7 October 2018: https://www.nbcnews.com/health/heart-health/it-s-time-reconsider-low-dairy-diets-new-study-suggests-n904491

From global to local: invasive and antimicrobial-resistant Salmonella

I knew I should not have taken my lunch to this seminar. With a slightly runny egg yolk in the tuna salad that had been on the bench for an hour or so, I was staring a plate of potential salmonella in the face before the speaker even began.

Andrea Parisi gave this PhD exit seminar on Thursday 29 November. Her supervisors are Professors Martyn Kirk, John Crump (Uni Otago), Benjamin Howden (Uni Melbourne) and Associate Professor Katie Glass.

She began with a detailed history and clinical manifestation of Salmonella, then it was on to the five key research questions of her PhD; burden, health outcomes, sources, epidemiology and risk factors of (invasive) non-typhoidal salmonella (NTS).

The burden component of the project was done through a systematic review and meta-analysis (4566 papers boiling down to 67) organised through the Global Burden of Disease mechanisms. They have data entry sheets and software, DisMod-MR to help wrangle summary statistics from papes into a usable meta-analysis.

Although sub-Saharan Africa had far and away the highest prevalence of NTS, it was Vietnam that was the focus of her primary data collection to investigate sources of salmonella inection. Andrea used some elegant violin plots in R to display the varying sources of NTS in humans.

With one published paper and at least three in draft form, Andrea has clearly had a successful experience as a PhD student in NCEPH. Best wishes for the home stretch!