Wikipedia:Peer review/Concussion/archive1

Article (Edit|History) • Article talk (Edit|History) • Watch articleWatch peer review

This peer review discussion has been closed.
I'm mainly looking for people familiar with the subject to check facts, because I'm not an expert, and it's possible that I've introduced some distortions or inaccuracies (though I've compulsively referenced!). Particularly I'd be interested to know if folks think I've done the structural damage debate justice. I'd like to know what it would take to get this to GA and beyond, so of course any other suggestions would be great too. Thanks much, delldot talk 14:25, 22 January 2008 (UTC)[reply]

  • Good work in general. Suggest applying WP:MEDMOS (manual of style for medical articles). Consider adding reference to English head injury guideline from NICE (CG56). JFW | T@lk 20:01, 23 January 2008 (UTC)[reply]
    • Thanks for the advice! I'll definitely work on better MEDMOS compliance. Do you recommend following the section header advice strictly (e.g. getting rid of the grades section or incorporating it into diagnosis)? Should the "Causes and risks" section be split into separate "causes" and "epidemiology" sections? Are there other specific aspects of MEDMOS that need to be worked on?
    • I couldn't find a concussion-specific page in NICE. Were you suggesting linking to the page you provided, or a more specific one? delldot talk 23:08, 23 January 2008 (UTC)[reply]
  • On the whole a good treatment of the topic. I had a hard time reading the Definition section; it needs some information not given until much later in the article, in the Diagnosis section. The Definition section states that there are multiple definitions but does not define concussion; the reader may expect to read next what those multiple definitions are. The lead uses the term "functional", but what this means is not clear. Much later in the article, it appears "functional" may refers to some manner of biochemical or metabolic imbalance. Consider reorganizing the sections. If you want some "color", it may help to mention the United States Polo Association "three concussions rule": USPA member clubs are required to report injuries sustained during USPA-sanctioned polo matches, and a player who receives 3 concussions becomes ineligible to play. --Una Smith (talk) 21:21, 23 January 2008 (UTC)[reply]
    • Thanks, this is very helpful. It sounds like I should explain functional better and maybe I should move the info from Diagnosis up, hopefully I'll have this done soon. I'll look into the Polo Association rule too. delldot talk 23:08, 23 January 2008 (UTC)[reply]
    • Well, I removed functional from the lead and defined it lower down. And I added some to the definition section, hopefully it's better defined now, but it may need more. delldot talk 00:43, 24 January 2008 (UTC)[reply]
Much better. I cannot find the USPA "three concussions rule" on the USPA public website, but they have members-only content. It may be there. I suggest ask their safety committee. --Una Smith (talk) 23:59, 26 January 2008 (UTC)[reply]
All right, thanks! delldot on a public computer talk 00:23, 27 January 2008 (UTC)[reply]
I've included a more general sentence: "More than one grade III concussion during a professional or collegiate athlete's carreer may lead health professionals to recommend that the player cease playing the sport." I think I'll use this instead of the polo reference, which might be too specific. delldot talk 10:15, 29 January 2008 (UTC)[reply]
    • Thanks so much, very helpful! Looking at your edit summaries, here's my to-do list:
      • templates belong at top of sections done
      • Redundant, weasley words I think I've taken care of these, unless there are problems with passive voice and other vague wording. I hope someone will reread it and let me know if it's ok now.
      • No access date for PMID (also for url's to online journals, if I'm reading MEDMOS right)
      • nbsp's for 5 minutes, grade I etc.
      • Consistency for numbers in lists done, unless a sentence with "grade I" requires a numeral.
      • format=PDF for pdf's in refs
      • Check proper date formatting
      • Hyphen consistency
      • Peer reviewed refs - replace news & web & emed - news and emedicine sources are gone, a few web sources such as the CDC that I deemed reliable are still there - let me know what you think.
      • Publishers for web sources - this one won't be a problem when the web sources are gone ;-)
    • Hope I haven't missed anything. When I find the original studies, should I get rid of the popular news sources, or just leave them in in addition to the peer reviewed ref? Thanks again! delldot on a public computer talk 05:37, 27 January 2008 (UTC)[reply]
    • Do you think sources for organizations like the Brain Injury Association, Neurosurgery Today, the Mayo Clinic, etc. are ok, or should I replace those with the original studies as well?
    • The weasel words are giving me no end of heartache. What do you recommend doing with a sentence like "Some professionals believe concussion is under-diagnosed..." where what I mean to say is actually that "a portion of scientists do, the rest don't," and there's really no percentage or camp other than the 'believers in this concept' that I can point to? Does citing a source help with the weasliness? How about "some sources such as so-and-so"?

Thanks again, sorry to be such a pain. delldot on a public computer talk 07:40, 27 January 2008 (UTC)[reply]

Simply say Concussion may be under-diagnosed and cite sources arguing both ways. --Una Smith (talk) 16:46, 27 January 2008 (UTC)[reply]
Thanks! I think I'm about half done with these, but I'm still finding it difficult to avoid wording like "controversy exists", or "but it has not been demonstrated conclusively". In some cases (erk!) it seems like you can't find a study stating the negative - they just don't say anything. delldot talk 07:51, 29 January 2008 (UTC)[reply]

I can't add much to the above, and unfortunately I'm not the expert that you are seeking, but on reading the article a couple of things bothered me: the word "patients" is used 20 times, making the reader feel that the article is written for health care professionals, (Polio uses the word 7 times). I know this is difficult to avoid entirely, but can you use some alternatives where possible, such as "victim", "sufferer" or simply "people"? Also, I saw the word "multiple", does this mean "many"? A good article. --GrahamColmTalk 10:50, 2 February 2008 (UTC)[reply]

Thanks! Yeah, I can replace some instances of "patients", no problem. In the cumulative effects section, multiple means more than one (maybe "two or more" would be better? But it's bulky). In the other sections it was just lazy writing ;) delldot talk 13:05, 2 February 2008 (UTC)[reply]
My, but that was a lot of patients! I have it down to 7 now, hopefully that's acceptable. delldot talk 14:57, 2 February 2008 (UTC)[reply]

Review in response to request by DellDot

edit

In general, the article is in very good shape. The references are mostly good, although they tend to focus more heavily than I would like to see on sports medicine journals as opposed to more basic medical journals. There are a number of images, which all seem to be correctly licensed, and the flow of the article is fairly coherent. I assume, based on above comments that it conforms to MEDMOS.

As it’s currently written, the article seems to be fairly sports-heavy in terms of examples, discussion and images. Perhaps some additional examples can be used to broaden the overall impact. For example, two of the examples of concussion involve boxing photos. How common are head injuries in different lines of work, for example? I'm sure that roofers are much more likley to suffer a concussion than a desk-jockey like me.

I have a number of comments, numbered below.

  1. There is one sentence in the lead that betrays a conceptual confusion: “concussion can cause a variety of physical, cognitive, and emotional symptoms. The most often affected mental faculties are memory, attention, speed of information processing, and reaction time to stimuli.” Information processing speed and reaction time to stimuli are confounded. Indeed, one way to measure information processing speed is to measure reaction times to simple stimuli. There are other tests to distinguish information processing slowness (sometimes referred to as bradyphrenia) from slowness of action (sometimes referred to as bradykinesia) which is an especially important issue in elderly populations. It might be worth checking the references to see if they’ve clearly distinguished these two possibilities.
  2. Definition: The changes in the definition, such as the NIHCE report which includes more about synaptic damage, probably should be better placed in their historical context. That is, as we have gained better and better methods of examining the brain, we find damage that would have gone undetected by previous methods. So, the NIHCE report, which is the most recent standard, will take into account the increasing precision of our ability to examine the living human (and non-human) brain after injury.
  3. Pathophysiology: This section probably should be expanded. As it stands right now, it would seem to be very dense for the lay reader. It’s very good, and well-referenced, but really seems like a hard read for someeone not familiar with all of the jargon. Perhaps much of the jargon can be converted to lay terminology, and then the technical terms put in parentheses, or the [[link|linktext]] code could be used to put more general terms in place of the technical terms, while still linking to the tehcnical terms.
  4. The article probably should have a section on social consequences and costs of concussion, especially given the high degree of returning veterans with TBI of some sort after the Gulf War [1] [2](check with MEDMOS for where it should be placed). I'm sure that this isn't only a problem for athletes and soldiers, so there is probably some estimate of the social costs of concussion each year.
  5. Also, recent studies of Iraq vets have suggested that concussion in Iraq increases the likelihood of developing PTSD. [3] [4]

Overall, I think that it's a good article, and it's clearly come a long way in the couple of months that DellDot has been working on it [5]. Edhubbard (talk) 17:38, 2 February 2008 (UTC) Edhubbard (talk) 17:38, 2 February 2008 (UTC)[reply]

ps: I just found this, which might be useful... [6] Edhubbard (talk) 17:42, 2 February 2008 (UTC)[reply]

Thank you so much! You clearly put a lot of work into this very thorough review. A couple responses:
  • Good point about the sportiness. I think it has to do with the fact that sports seems to be the only context that consistently uses "concussion" rather than "mild traumatic brain injury" or a mix of the two. I'm not sure what to do about the problem, since they're sometimes treated as synonymous and sometimes not. I've avoided using sources that only say "mTBI". Maybe a separate article should be written for mTBI to explain the difference, rather than having a redirect as we currently do.
  1. I think I've solved this by removing the reaction time bit, so it now reads, "The most often affected mental faculties include memory, attention, and speed of information processing."
  2. I think there's a significant group that defines concussion as having no structural damage, i.e. they wouldn't classify an injury that had such detectable damage as concussion. I hope I've accurately represented this, that's one reason I wanted an expert to look at it.
  3. Good advice on the pathophysiology section, I'll work on making it more accessible.
  4. What do you think of adding info on social consequences and costs into the epidemiology section?
  5. I interpreted that paper to mean that PTSD increased the likelihood that people would continue to have symptoms because of psychological factors, but they did mention the impaired cognitive processing would impede dealing with the traumatic events. I'm delighted to report that one of the authors of the study, Charlescengel, is a Wikipedian! Hopefully he can clear this up.
Thanks again for the terrific review. delldot talk 22:44, 2 February 2008 (UTC)[reply]

Review of epidemiology section

edit

Here's a quick review of Concussion#Epidemiology. I am reviewing this version.

  • "estimating an overall incidence of over 600 per 100,000 people" cannot possibly be right. An incidence rate is per person-year, not per person. Or perhaps you meant to write "annual incidence"?
    • Good catch on the incidence thing! The paper actually uses "population-based rate", does saying "estimating a rate of over 600 per 100,000 people" fix it?
      • No, because it doesn't state the period the rate is being determined over. Incidence rate is cases per person-year; see Incidence (epidemiology). Eubulides (talk) 17:35, 4 February 2008 (UTC)[reply]
        • Actually, looking closer at this, I think it's two different rates. The direct quote is "...the rate of hospital-treatment in adults with MTBI ranges from about 100 to 300/100,000. We also accept that that a large number of MTBI cases are not treated at hospitals and that the actual rate of all MTBI is probably in excess of 600/100,000." Now I think the first figure is per 100,000 people treated in hospitals and the second is per 100,000 people per year in the general population. Do you think that's right? Thanks again for taking all this time with this.
          • For something like that you have to look at the context; this will require reading the paper fairly carefully in some cases (you may even need to read some of its citations) to make sure you understand the terms it's using. Sorry, but papers are sometimes sloppy about epidemiology. All that being said, from the above quote it appears that the paper is trying to distinguish between the true annual incidence rate of MTBI (6 per 1,000) and the annual incidence rate of hospitalizations for MTBI (3 per 1,000). The first number is more important for studying concussions itself; the second number is more important for (say) hospital administrators who need to budget provisions for treating concussions. It is probably useful to mention both numbers, but the former is the more important (you might just say something like "it's estimated that half of concussions are treated" or something like that). Two other things. First I'm assuming these are annual incidence rates (i.e., MTBI cases per person-year); you should double-check that and then write it down explicitly in the article. Second, this is the sort of area where review papers are really valuable; you can probably find incident rate estimates all over the map and the review paper can put them into context. Eubulides (talk) 20:09, 4 February 2008 (UTC)[reply]
  • "The lack of the highly noticeable signs and symptoms, such as coma, frequently present in other forms of head injury, could lead clinicians to miss the injury; this possibility has led to the suspicion that concussion may be under-diagnosed." This belongs under "Diagnosis", not "Epidemiology".
    • Moved.
  • "A U.S. study found that children under five had the highest rate of mTBI." That's a bit vague; compared to who? The abstract of the cited study (PMID 15841752) makes it sound like it's compared to Alaska natives, which is pretty weird. What's going on here?
    • Added "of all age groups" to the sentence.
  • "Another study found that the incidence of mTBI in Canada was 200 per 100,000" Hold on a minute. The U.S. study reported average annual incidence of 500 per 100,000, which is more than twice as high. These figures don't match at all. The U.S. results do match the cited WHO results, though. What is going on here?
    • The discrepancies are due to widely varying definitions of concussion and mTBI, as well as problems with underreporting. For example, the WHO study made the 600 estimate based on their guess that the condition was being underreported. The actual reports led them to 100-300/100,000 and they extrapolated. I can add a couple sentences about this, hopefully it will at least make the problem with our lack of knowledge clearer.
  • I didn't have time to read the rest of the article, but I noticed that a good chunk of "Causes" actually belongs under "Epidemiology". This includes the figure, and the discussion of risk of concussion. Pretty much all the counts and percentages in "Causes" belong under "Epidemiology". "Causes" should just talk about the nature of the causes; it should let "Epidemiology" do the risk analysis.
    • After doing this, I found that I had one sentence in causes, so I moved it to the lead. Do you think this is OK?
  • It seems odd putting "Epidemiology" so early in the article. WP:MEDMOS hints at putting it later. See Tourette syndrome for an example of a better order.
    • Moved.
  • What is the association between mTBI and other diseases, e.g., Alzheimer's, Parkinson's? I'm surprised to not see this mentioned here.
    • There's a short mention under prognosis, I can add a bit to it.
  • There should be a brief summary of the epidemiology section (at least 1 sentence, say) in the lead. (This is true for all the sections, not just epidemiology.)
    • Thanks for catching that. I think there's now something from every section other than prevention, which I couldn't think of anything that's not super obvious to say about.
  • This point is not just about epidemiology. I'd like to see more review articles, if possible. There is too much emphasis on primary studies. See WP:MEDRS for details on why secondary sources are preferred. Here are some review-article possibilities I found in a brief search (I haven't read them): Kiraly & Kiraly 2007 (PMID 18040539), Petchprapai & Winkelman 2007 (PMID 17966292), Cook et al. 2006 (PMID 16884134), Comper et al. 2005 (PMID 16296570), Heegaard & Biros 2007 (PMID 17826211), Hootman et al. 2007 (PMID 17710181). I'm sure you can find more.
    • Great suggestions, I'll definitely use them. Do you think I should be removing info from primary sources as well, or just adding secondary ones?
      • If a secondary source reviews several primary sources, you typically don't need to cite the primary source directly (unless it's really notable). If the primary source is so new that it isn't reviewed yet, you may need to cite it directly, but here you must be quite careful since you are using your own judgment as to which primary sources are worth citing. Eubulides (talk) 17:35, 4 February 2008 (UTC)[reply]
        • What if I use info from a primary source's intro section, for example, that's citing other authors' work? For example, in a sentence that says "neuropsychological tests exist to test speed of information processing", citing an article that used such tests and discussed them in the intro?
          • That's certainly better than nothing, as the intro to a primary source is often a mini-review. However, review papers are usually better sources. First, primary-source intros are typically not checked by editors as carefully as review papers are (the editors are far more interested in the primary source's method, results and discussion). Second, primary-source intros are typically written from a particular viewpoint and are often not as balanced as reviews are. Eubulides (talk) 20:09, 4 February 2008 (UTC)[reply]
            • Ok, I've been working on getting rid of these. Do you think I should get rid of the two figures in the epidemiology section based on original studies? That is, assuming I can't find a review citing the info. delldot talk 16:50, 5 February 2008 (UTC)[reply]
              • I'd keep the figures. Figures are nice. Figures from secondary sources would be better but primary sources are better than nothing. However, the first figure's labels aren't right. The top line should say "MTBI annual incidence by age", not "MTBI incidence by age". The vertical axis's label should read "MTBI per 100,000 person-years", not "MTBI per 100,000 people". The horizontal axis's label somehow omits people aged 35; surely that's not intended. And the caption should say "Annual incidence of mTBI by age group in Canada", not "Incidence of mTBI by age group in Canada". See Incidence (epidemiology) for details about why one cannot say an incidence rate is "200 per 100,000 people". Eubulides (talk) 07:27, 9 February 2008 (UTC)[reply]

Hope this helps. Sorry, I don't have time to read the whole article carefully. Looks like a good article. Eubulides (talk) 07:30, 4 February 2008 (UTC)[reply]

Thanks very much! This is very helpful. I'll work on your suggestions over the next few days. delldot talk 08:27, 4 February 2008 (UTC)[reply]

Colin

edit

Just a quick look for now, I'm afraid.

  • I see the earlier comments about sports-related sourcing. I appreciate the issue but be aware that that population group doesn't reflect the wider population. It would typically involve young healthy men and there's probably first-aid or better medical support to hand. So, if you are making a general comment, try to find a more general source if possible.
    • Good points from both of you. I'll find more general sources wherever possible.
  • Eubulides points out the benefit of reviews. I'd like to add another: you should usually be able to write an entire sentence or two based on one source. At the moment, some sections have a distinct paper cited per phrase or fact. Not only is this visually distracting, the use of separate sources to build a sentence may lead to accusations of synthesis. For example, surely one source can be found to list the common symptoms. Another example is "The symptoms of most concussions are resolved within days or weeks,[63] but problems may persist.[48]" Why can't that all come from one source?
  • Oddly, some sections are citation free: "Post-concussion syndrome", "Dementia pugilistica", "Second-impact syndrome" (only two). I hope you've not been reading any nonsense about summary-style meaning you don't have to provide references.
    • Heh, that's exactly the nonsense I was reading. Sourced.
  • Try to avoid contiguous sets of references if you can, if you don't need all the refs. For example "The CDC estimates that 300,000 sports-related concussions occur yearly in the U.S.,[72][77]" - that's just one fact, so surely one reference will do.
    • I think I'll take these out last, so I can choose the best one to keep at the end.
  • Don't provide a URL to the full text unless it is free. I found some after checking a random small selection, but you should really go through them all. If you've got journal access that is "remembered" using cookies then it may help to use a different computer or browser from normal. If the paper is free but requires registration, my preference is to indicate this after the citation, in parenthesis: "(requires free registration)".
    • Done, I'm pretty sure.
  • Per WP:MEDMOS, I don't feel that an access-date is required for URLs to online versions of print medical journals. These aren't dynamic documents in the way that a web site is.
    • Done.
  • Overall, I found it readable and generally lay-friendly. Colin°Talk 20:48, 4 February 2008 (UTC)[reply]