How We Manage the Weekly List of MS Papers
Every Friday, MSDF editors use a curated PubMed search to link to every newly published MS-related paper. Here’s how we do it.
Each week, somewhere between 30 and 110 papers related to multiple sclerosis are published in the scientific literature. At MSDF we endeavor to list them all, publishing links to a curated set of each week’s new papers every Friday. Here’s how we do it.
We’re currently using the following PubMed query, which has evolved over the 3 years of MSDF’s existence:
"multiple sclerosis" OR "demyelinating" OR "optic neuritis" OR "acute disseminated encephalomyelitis" OR "neuromyelitis optica" OR "transverse myelitis" OR "NMO" OR "ADEM" OR "experimental autoimmune encephalomyelitis” OR “cuprizone” OR “neurodegeneration” OR “myelin” OR “microglia” OR “demyelination” NOT adem [Author]
This query returns many false positives. I read all the titles and most of the abstracts and make judgments about which of those are directly relevant to MS or related disorders. Today, for example, the query returned 139 papers and in my judgment only 58 of them (42%) were truly MS-related. Some weeks the proportion is even lower than that.
I believe the query terms “neurodegeneration,” “myelin,” “microglia,” and “demyelination” are responsible for most of the false positives. “Neurodegeneration,” in particular, returns many references related to other neurodegenerative disorders, such as Alzheimer’s disease, Parkinson’s disease, amyotrophic lateral sclerosis, stroke, and hypoxia, to name a few. Some papers mentioning those disorders mention MS as well, so we can’t add terms such as “NOT ‘Alzheimer’” to the query.
My judgments on which articles are relevant are often subjective, and I frequently struggle with those decisions. It’s easy to decide relevance when an article actually mentions multiple sclerosis. It’s harder when it mentions only myelin or only Th17 cells. When I’m on vacation, and one of my colleagues is curating the list in my absence, his or her judgments often differ from my own.
If you think we’ve missed an important MS-related article—or have included an irrelevant article—I hope you’ll let me know. And I’m open to suggestions on how to adjust our PubMed query to decrease false positives and false negatives.
Once I’ve chosen which of the articles to include in the week’s list, I then select between two and four of them as Editors’ Picks. Those are the week’s articles that seem to me to be the most important or interesting or intriguing. Once again, I invite readers to take issue with my choices. I’d love to hear about important articles that I have not designated Editors’ Picks or, on the contrary, Editors’ Picks that don’t deserve the honor.
Key open questions
- Which search terms should be added to or subtracted from the PubMed query?
- Is there any way to make the selection of papers less subjective?
- How useful do you find the weekly list of papers?
Comments
I think you are providing a great service to the MS community, a forum for ideas, discussions and more importantly, distilling information for our readers, Bob.
I do believe that you ought to also search for clinically isolated syndrome, radiologically isolated syndrome and perhaps, types of MS (RRMS, PPMS, SPMS), as well as Marburg variant, Balo's concentric sclerosis and ADEM to complete the circle. Some day you could also bring in pediatric MS.
Thank you for the suggestions. We already have ADEM in the search term, and I'm certain it's not necessary to specifically search for RRMS, PPMS, SPMS, and pediatric MS, since the "multiple sclerosis" search term will capture those. I'm going to investigate whether including "clinically isolated syndrome" and "radiologically isolated syndrome" will pick up additional papers. And I definitely like your suggestions for Marburg variant and Balo's concentric sclerosis, and I'll be adding those search terms.
So I conducted a few tests, searching PubMed for
"clinically isolated syndrome" NOT "multiple sclerosis" and
"radiologically isolated syndrome" NOT "multiple sclerosis"
To my surprise, both of those searches brought up numerous papers. To my greater surprise, many of those papers are in the journal Multiple Sclerosis! So I've learned that a PubMed search doesn't include the name of the journal.
I'll be adding both CIS and RIS to our search terms.
Very useful! Thank you.
a PS ....one could almost say getting exactly what you want out of PubMed is a science in itself!