Another fortnight, so many new topics! On April 1st a friend sent me a text stating that April fools day is the one day of the year where we read content online or in the newspapers and question its veracity.
This resonated with me and so today’s blog focuses on content, content sharing, and content checking. I have checked my facts, so hopefully you will not find any errors.
Topics for today’s blog are:
– Think before you like: Why context matters in health news
– Abstract submissions to the Medical Information and Communications meeting
– How I use ChatGPT and why it is not a magic bullet
– What ChatGPT thought of my abstract on digital islands
Think Before you like: Why Context Matters in Health News
A LinkedIn post from a week ago reported:
- That the UK government will introduce a legal requirement to make all clinical trial results public within 12 months of trial completion
- That companies in breach of the law will be refused permission to start new trials
The writer went on to say “The UK law sets a new global benchmark for transparency in medical research. Comparable disclosure laws in the European Union and the United States only cover some types of trials, and so far remain unenforced.” And concluded “This is a major victory for patients in the UK and sets a new global gold standard for transparency”.
Comments ranged from those commending the UK government, those saying that it’s about time too, to those questioning how enforcement might look. I added a comment that European Medicines Agency (EMA) has substantially comparable guidance in place (as does the US), and an observation from the UK Medicines and Healthcare Regulatory Agency (MHRA) website “Since January 2021 the MHRA no longer has access to the EudraCT database and therefore there was a danger that UK clinical trials would not appear in a public register, as trial sponsors themselves cannot directly publish their trials in the EU Register”
While the post is not incorrect, the context I have added in would help readers understand that the UK law is in fact substantially in line with global clinical trial disclosure standards from the EMA and the US Food and Drug Administration (FDA). The key difference between the UK law and EMA and FDA regulations is, as best I can tell, that companies in breach of the law, won’t be allowed to initiate new trials. This is laudable but it remains to be seen how the UK will enforce the law. However, if that is the only difference, then the conclusion “This …. sets a new global gold standard for transparency” doesn’t seem defensible.
The reason I am sharing these thoughts today is that one commentator wrote, I paraphrase “Does it matter that what the UK is doing is not truly novel and that this post is perhaps mainly a PR activity? After all the patients that we serve are the biggest losers in all of this.”
After some reflection, my answer is yes it does matter. The post I’m mentioning above was reposted 164 times and liked by 1616 people. Estimating a conservative 300 connections per person the content reached between 50 000 (reposts) to 500 000 (likes) people.
I found the post because my connections liked it. My first thought was “clinical trials are not well regulated, which puts patients at risk” but I knew that this was incorrect, and the global gold standard comment made me think, I looked up the context, and I realised that the post by itself is, while not inaccurate, beyond the last statement, is potentially misleading. For readers who are not involved in the pharmaceutical industry, their conclusion might just remain what my initial conclusion was: “clinical trials are not well regulated, which puts patients at risk”
This can ultimately impact people’s willingness to participate in research, and erode trust in trial sites, sponsors and the pharmaceutical industry and so ultimately harm patients, instead of helping, So, I think, yes context and fair balanced writing matters.
Key takeaways: Think before you like, or repost. Is the information comprehensive? Does the article sound balanced? Is the source a reputable journal like Nature, or HBR, or is it someone’s opinion. If the latter, is the content polarising or does it seem fair and balanced. If in doubt consider fact checking or looking for references.
N.B. for interested readers I am happy to share the link to the original post.
Abstract submissions to the Medical Information and Communications meeting
We recently had the call for abstracts for the Medical Information and Communications meeting, which will be held in Brussels 27th to 29th September. Register here.
Following on from my previous article on the importance of fair balanced content and context I thought it might be useful to share some thoughts on abstract writing.
You can pick any topic. Basically, anything you find interesting in the Medical Information or associated space could be of interest to the audience. You can write your abstract in whatever form you would like. However, the classic abstract format is as follows and I generally find it very useful. 1) Background – share why the project is important 2) Materials and Methods – what did you do, why, how, who was engaged 3) Results – what were the outcomes 4) Conclusions.
When submitting an abstract to the programme committee the more information you provide the easier it will be for us to assess. Timelines are as follows: submit poster and presentation abstracts by the 19th of April. I look forward to reading your submission.
Key takeaway: if you think it is interesting, we will likely think so too! If you are in doubt contact any member of the programme committee, including myself, we are happy to give guidance. And last but not least don’t miss the deadline.
How I use ChatGPT and why it is not a Magic Bullet
One of the key skills in research, interviewing, querying datasets and using tools like ChatGPT or TripAdvisor is the ability to ask the right question and know enough about the subject at hand to be able to put the answers into context.
Here are some ways I have used ChatGPT:
- Quality control: When I have reservations on something I have found online I may put it into ChatGPT to see if I can validate the content/substantiate my qualms with sources, generally after I have done my own research already. In this situation, it has become very apparent that while the tool is useful, it has limits, and you need to know the subject area very well yourself, so that you can put the data retrieved into context and know where to dig further.
- As a title generator: I provide the key words, specify the type of content I need a title for, abstracts, panel discussions, books etc., and the maximum length for the title and see what ChatGPT produces.
- As a test reviewer for an abstract I was submitting, which I will share.
- To generate title suggestions for a 15-minute slide presentation outlining key data on one of the biggest Swiss newspapers as a test on how a student of my acquaintance might use it.
- When looking for regulations and guideline texts and computer error fixes.
I don’t use the text as is but as an inspiration or sparring partner as I develop ideas and when I use it always ask it to provide links to the sources used.
Key takeaway: ChatGPT is not the great equaliser I initially imagined. It’s extremely useful, but only if you know your subject well, and you know how to research, do quality control and check the data, the output you get will reflect any deficiencies you have. If you are creative, I suspect you can use it to generate innovative content, in much the same way that a loom can help you make a beautiful tapestry but only if you are able to provide the image yourself. It is truly a useful tool, but it’s not the magic bullet it’s been hailed as, I don’t believe.
What ChatGPT Thought of my Abstract on Digital Islands
As shared in my last blog, I will be hosting a panel at Connect in Pharma in June in Geneva. We will discuss among other things, digital islands and how to overcome them. After much deliberation, the title I decided on was “Marooned on a digital island in a sea of data?”
Having written the abstract, I sent it over to a friend for comment. He sent me ChatGPT’s feedback back instead.
Here it is:
As an AI language model, I think the abstract sounds like an interesting and timely topic for a panel discussion. The title “Marooned on a digital island in a sea of data?” is catchy and thought-provoking, and the description raises important issues about the challenges of managing data in today’s rapidly evolving digital landscape.
The panellists, Dr med. Isabelle C. Widmer and Peter Shone, bring different perspectives and expertise to the discussion with their backgrounds in elytra GmbH and iEthico, respectively.
The abstract also outlines some of the common problems that companies face with data management and highlights the steps they are taking to address them. The focus on implementing a company-wide data management strategy is particularly relevant as it emphasizes the need for a holistic approach to managing data.
The inclusion of AI as a possible solution to data management challenges is also noteworthy, and the panellists’ exploration of the promise and limitations of AI should provide valuable insights for the audience.
Overall, the abstract sets a clear agenda for the panel discussion and promises to provide actionable insights for attendees.
My friend added: As a humble human being I think your abstract is good as well.
Key takeaway: ChatGPT wrote a pleasing text based on my abstract. However, the resulting text feels as though the words I used in my abstract have merely been jumbled and structured into new sentences that read differently but mean the same thing. As said above, ChatGPT is great at filler text.
I hope my blog provides you with some useful insights and, as ever, I look forward to hearing your thoughts.
If you are struggling with your job, not sure where to go from here, want to develop further in your career, or want to look at repeat patterns in your daily work and private life, that you suspect may be blocking you from reaching your full potential and you would like to discuss how executive coaching could help, please contact me for a confidential and informal chat.
Isabelle C. Widmer
University Hospital Basel
Photo by Isabelle Widmer