Tuesday, April 29, 2008

Search challenge 8

With the ever-increasing growth of the medical literature, as well as seemingly increasing public scrutiny of the happenings at our medical institutions, it seems like finding good ways to "track" the publications of an institution and also how it's being mentioned in the popular media are challenges that health sciences librarians are being called to help address.

So, for this week, a multi-part search challenge:
- What strategies do you use to track what people from your institution (hospital, medical school, etc) are publishing (e.g. what databases, terms, other resources)?
- How do you track mentions of your institution in the popular media?
- How often do you update such searches?
- Who "needs" this information at your institution?


Search challenge 7: strategies

Last week's search challenge explored whether there is evidence in the literature to support the use of nutritional therapy in patients with pressure ulcers.

A PubMed search strategy that seems to work well:
(pressure ulcer[mh] OR pressure ulcer*[tiab] OR decubitus ulcer*[tiab]) AND (nutrition[tiab] OR nutritional[tiab] OR nutrition therapy[mh] OR diet therapy[sh] OR vitamins[mh] OR vitamin[tiab] OR vitamins[tiab]) AND eng[la] AND humans[mh]

I also tried the MeSH term "nutritional support," since I was seeing some articles that looked at feeding regimens which are often indexed in this broad MeSH category, but didn't find that it added anything to the retrieval, probably because of the text words I'd included.

In reviewing the results, there's a meta-analysis that includes the literature through 2004: Stratton RJ, Ek AC, Engfer M, Moore Z, Rigby P, Wolfe R, Elia M. Enteral nutritional support in prevention and treatment of pressure ulcers: a systematic review and meta-analysis. Ageing Res Rev. 2005 Aug;4(3):422-50 (PubMed abstract) -- this meta-analysis included studies that focused on either enteral nutrition formulas or other oral supplements (e.g. vitamin C). The meta-analysis seemed well-executed and had clear inclusion/exclusion criteria; it also provided a good overview of the previous literature on this topic.

For this kind of question, I think I'd be comfortable providing a good meta-analysis accompanied by any high-quality references (e.g. trials or cohorts) published since the time it was executed; I'd also want to look at the kinds of things that they excluded and consider whether any of those need to be mentioned. For example, there might be therapies represented in small experimental studies that wouldn't make it into a meta-analysis but still would be useful to mention to the requester as a potential option, so that I'd feel like I was thoroughly representing all of the possible "options" that are in the literature.

Labels: ,

Monday, April 21, 2008

Search challenge 7

Since this month's case study focuses on a geriatric issue, this week we'll tackle a geriatrics-focused search question: Does improved nutrition (defined broadly to include controlled diet, supplementation, PEG feeding, etc.) help older patients heal a pressure ulcer more quickly and/or more effectively?


Search challenge 6: strategies

As Martin points out in his comments on search challenge 6, some of the review-level literature on this topic is older than we've seen for some of the other search challenges.

If you enter "ischemic stroke" into the Entrez MeSH Browser, it links you to "Brain Stem Infarctions" - combining a location (the brain stem) with a particular kind of damage (infarction, i.e. tissue death due to lack of oxygen, often due to reduced blood supply). Infarction is one of the potential downstream effects of ischemia, either due to the length and/or the severity of the ischemic event.

Typing "ischemia" into the MeSH Browser lists "Brain Ischemia" as the 6th hit in the terms retrieved, defined in its scope note as
"Localized reduction of blood flow to brain tissue due to arterial obstruction or systemic hypoperfusion. This frequently occurs in conjunction with brain hypoxia (HYPOXIA, BRAIN). Prolonged ischemia is associated with BRAIN INFARCTION."
"Stroke" is also a MeSH term, defined as
"A group of pathological conditions characterized by sudden, non-convulsive loss of neurological function due to BRAIN ISCHEMIA or INTRACRANIAL HEMORRHAGES. Stroke is classified by the type of tissue NECROSIS, such as the anatomic location, vasculature involved, etiology, age of the affected individual, and hemorrhagic vs. non-hemorrhagic nature."
So I tried this search: brain ischemia[majr] AND stroke[majr] AND (etiology[tiab] OR etiology[sh] OR etiologic[tiab] OR cause[tiab] OR causes[tiab])
limited to All Child, and retrieved ~1000 citations.

And narrowing to exclude some of the (arguably, of course) lesser publciation types like this:
brain ischemia[majr] AND stroke[majr] AND (etiology[tiab] OR etiology[sh] OR etiologic[tiab] OR cause[tiab] OR causes[tiab]) NOT (case reports[pt] OR letter[pt] OR comment[pt] OR editorial[pt] OR news[pt])

reduced the retrieval to ~500 citations. I noticed that a number of the first several pages of hits discussed neonatal and infant stroke; for a "real" request, I imagine that I would either know what pediatric age group the requester is interested in, from our reference interview, or would need to get back in touch to see if we should define "pediatric" as all children, infants only, etc. Narrowing to English language narrows the set even further.

A quick Google search also revealed a report from the National Institute of Neurological Disorders and Stroke from 2005, "Recognition and treatment of stroke in children," which includes a section on etiology and risk factors.

UpToDate, for those who have access, also has a nice overview titled "Ischemic stroke in children and young adults: Etiology and clinical features."

Though I excluded case reports from one of the strategies above, this would likely be a question where I would really want to take a close look at the cases reported in the literature as a complement to the more "synthesized" material from the NINDS report, the UpToDate entry, and the review articles -- there are likely instances of ischemic stroke in children that have made their way into the case literature that have not yet entered into the larger case series or other retrospective works that still may be very useful in gaining a broad perspective on why a given child might have developed an ischemic event.

Adding "risk factors" to the etiology portion of the search string might also be useful too, though it does increase the volume of search retrieval.


Friday, April 18, 2008


This week's New England Journal of Medicine includes several articles discussing electronic health records.

- Two perspective pieces: Personally Controlled Online Health Data — The Next Big Thing in Medical Care? by R. Steinbrook and Off the Record — Avoiding the Pitfalls of Going Electronic by P. Hartzband and J. Groopman

- A Sounding Board piece: Electronic Health Records, Medical Research, and the Tower of Babel by R. D. Kush, E. Helton, F. W. Rockhold, and C. D. Hardison

Each item gives a little bit different perspective on the potential benefits and pitfalls of electronic health records and strategies for implementing and using them. Steinbrook focuses on consumer involvement in health care and issues of data ownership and portability, while Hartzband and Groopman focus on the clinician perspective. Kush et al. consider data standards for EHRs in relationship to clinical research.

Another Sounding Board item in this issue, Tectonic Shifts in the Health Information Economy by K. D. Mandl and I. S. Kohane, talks about HIPAA in relationship to portals like Google Health and Microsoft Health Vault, among many other interesting things.


Thursday, April 17, 2008

Narratives in clinical practice

I just found this brief article in my stack of "to be read" papers as I moved to a new office space -- I had printed it because it provides a really interesting take on the utility of the case report or case narratives in clinical practice.

Citation: Campo R (2006) “Anecdotal Evidence”: Why Narratives Matter to Medical Practice. PLoS Med 3(10): e423

Rafael Campos focuses on the potential usefulness of cases in generating research hypotheses, prompting connections between topics, visualizing new possibilities...

He notes,
The inscrutably enduring power of the anecdote itself is what incites all our most fearsome defenses. So furious are we in our rejection of the merely anecdotal one cannot help but begin to wonder at it. What is it in the ostensibly harmless tale my great-grandfather told about the secret of his longevity being the small glass of bitters mixed with a raw egg he downed before bedtime each night since the age of ten that rallies us to spend billions of dollars in grants from the National Institutes of Health , disbursed every year to scientists seeking their own more explicitly pharmaceutical recipes for living longer? Why does our clinging to superstition and our willingness to be intrigued by mystery provoke such an angry, unrelenting diligence? An anecdote, after all, is just a story.

The irony in our growing intolerance of the anecdote is that storytelling is full of lessons in imagination and invention so beneficial to the creative investigator.


Monday, April 14, 2008

Search challenge 6

This week's search challenge is another question that takes a look at the pediatric literature: What are the causes of pediatric ischemic stroke?

A few related issues to consider:
- how important are case reports for this question?
- is there a reference text or other resource that is useful, in addition to Medline?
- how broad does a Medline search need to be to retrieve most of the "good" items for this question?


Search challenge 5: strategies

In the comments on search challenge 5, Heather proposed an Ovid Medline strategy combining the concepts in a straightforward and effective way, retrieving just under 10 citations:
1 exp Osteomyelitis/di [Diagnosis]

2 exp Arthritis, Infectious/di [Diagnosis]

3 1 or 2

4 exp C-Reactive Protein/

5 erythrocyte sedimentation rate.mp. [mp=title, original title, abstract, name of substance word, subject heading word]

6 4 and 5

7 3 and 6
Another strategy, this one in PubMed format, has a little bit broader retrieval at about 30 citations:
"Blood Sedimentation"[Mesh] AND "C-Reactive Protein"[Mesh] AND ("Osteomyelitis"[Mesh] OR "Arthritis, Infectious"[Mesh]) AND (English[lang] AND ("infant"[MeSH Terms] OR "child"[MeSH Terms] OR "adolescent"[MeSH Terms]))

When I see these sizes of retrieval, I see it as a challenge to think even more broadly in terms of synonyms, abbreviations, etc. This search retrieves ~ 70 citations, adding some "false drops" but also including a few items that might be useful in addition to the hits from the above searches:
(blood sedimentation[mh] OR esr[tiab] OR erythrocyte sedimentation[tiab]) AND (c-reactive protein[mh] OR crp[tiab]) AND (osteomyelitis[mh] OR osteomyelitis[tiab] OR arthritis, infectious[mh]) AND eng[la]


Search challenge 4: strategies

In the comments on search challenge 4, Martin noted:
"Syncope, Vasovagal"[Mesh] AND "Blood Donors"[Mesh] gives 21 hits, the first one (17655598) is an intervention study, indicating that syncope incidence is quite different among subgroups. This make giving a general answer more difficult.

Some data on incidence comes from (16202053), which suggests a 0.9% incidence.
Other studies go for vasovagal reactions (not limiting to syncope) and, of course, give much higher numbers.

Data on the consequence of syncope comes from (11778059).
I tried a search that was a little more broad, which did draw in a few tangential things (like how to prevent/reduce complications such as fainting during or after blood donation) but still didn't yield an overwhelming volume, about 90 citations:
(vasovagal OR vagal OR fainting OR syncope) AND (blood donors OR blood donation)
As I was looking through the articles and a broad Google retrieval, I noted that there was a significant body of older literature that looked very relevant and potentially useful. IndexCat, the online system which allows users to search volumes from the Index-Catalogue of the Library of the Surgeon-General's Office. The database currently contains Series 1-5 of the Index (1880-1961), which includes 61 volumes in the 5 series, covering reports, journal articles and other materials from 1880 to 1961. Some of the WWII literature, with huge blood drives to support the war effort, had great surveillance data on side effects.

A quick overall summary (with a few sample articles thrown in for good measure): The majority of the original studies on this question were done between the 1940s and 1960s. Much of the data was generated from the mass blood drives conducted for WWII. Later articles refer to the earlier ones for incidence/frequency rates and are more concerned with predicting which donors are likely to exhibit negative reactions. Reported incidence rates range from less-than 1% to 10%; it is important to recognize that the definitions of syncope (fainting) varied somewhat from study to study.

Reference #1: Callahan R, Edelman EB, Smith MS, Smith JJ. Study of the incidence and characteristics of blood donor "reactors." Transfusion 1963; 3: 76-82.

This article reports on a retrospective review of the records of the Milwaukee Blood Center for blood donors from 1958 to 1962 with a more detailed analysis of the reactor records for the year 1960. They report an overall donor reaction rate of 4.5 percent with a higher rate seen in females. Reactions include mild, moderate and severe, where actual syncope is considered a 'severe' reaction. Of the reactions for all years, such severe reactions had a frequency of 5.7%. For the year 1960, 6.2% of the reactions were considered to be severe. Note that these results are retrospective and thus are subject to errors in recording and interpretation.

Reference #2: Moloney WC, Lonnergan LR, McClintock JK. Syncope in Blood Donors. N Engl J Med 1946; 234: 114-118.

This article reports a pair of prospective studies on the frequency and characteristics of syncope in blood donors. The first series was composed of 6,882 donors bled between 1942 and 1943; 162 total cases of syncope were identified with 17 'severe' cases. A severe case is defined as one in which loss of consciousness or convulsions were seen. The second series was composed of 9, 251 donors bled during 1943; of these 527 syncope cases were identified with 158 'severe' cases. The authors examine multiple factors including donor characteristics and environment. They report that younger donors, particularly women, were more prone to fainting.

Reference #3: Trouern-Trend JJ, Cable RG, Badon SJ, Newman BH, Popovsky MA. A case-controlled multicenter study of vasovagal reactions in blood donors: influence of sex, age, donation status, weight, blood pressure and pulse. Transfusion 1999; 39: 316-320.

This retrospective case-controlled study examined 1,890 blood donors with reported syncope from three U.S. blood centers, 1994-1995. These were compared to case-controls and random population controls in a logistic regression analysis to determine which of several factors correlated with the incidence of syncope. The authors report that age, weight and donation status (first-time or repeat donor) were significant but that sex, blood pressure and pulse were not. For the frequency of reactions in blood donors, they quote a 2-5% rate of vasovagal reactions and a 0.08-0.34% of syncope, based on nine previous reports in the literature including the Callahan article.

Reference #4: Greenbury CL. An analysis of the incidence of "fainting" in 5, 897 unselected blood donors. Br Med J 1942; 1: 253-255.

This study examined the incidence of syncope in blood donors and the influence of age, sex, occupation, lack of food, fatigue, and room temperature on it. A donor was considered to have fainted if any of the fainting-associated symptoms required an interruption of routine. Donors who fainted after leaving the building were not included. 5, 897 donors were randomly chosen (4127 females, 1760 males) from four London sectors. The percent fainting varies among age-grouped males and females from 1.63% to 7.62%. Incidence declined non-uniformly with age in both males and females. In most age groups, females had higher incidence rates. Incidence was higher in clerical workers compared to all other occupation groups. Hours of work (fatigue) and room temperature had no effect. Among female clerical workers, aged 18-25, time from last meal was sigificant, otherwise, it was not.

Reference #5: Boynton MH, Taylor ES. Complications arising in donor in a mass blood procurement project. Am J Med Sci 1945; 209: 421-436.

This article review the various types complications of blood donation seen among the millions of donors collected by the Blood Donor Service of the American Red Cross from 1941 to 1944. They classify complications into four categories, with group 2 being the one that includes 'faints' (syncope). They list four frequency figures from the literature: 2.8%, 5.5%, 4.2%, and 8.9%. The latter is noted to be simply a subjective report by donors that they suffered an ill-effect. The authors here report data on 2294 'reactors' - the frequency of symptoms within the group, ranging from pallor to loss-of-consciousness-to incontinence, and state that their results are in general agreement with the finding of the British Medical Council report.

Reference #6: Subcommittee of the Blood Transfusion Research Committee. Fainting in blood donors. Br Med J 1944; 1: 279-286.

In this study, several blood centers in England and Scotland were required to complete a record card for any donor who fainted during a blood collection procedure. A donor was considered to have 'fainted' if they showed signs and symptoms related to fainting which required a delay or change in normal procedure. Multiple symptoms were recorded. The next donor after the symptomatic one, bled by the same person, was recorded also as a control. They report the frequency of recorded faints in two Centers as 5.4% and 5.7% respectively. The effects of various factors on fainting and the incidence of specific symptoms are shown in Tables I - XI.

Reference #7: Poles FC, Boycott M. Syncope in blood donors. Lancet 1942; 2: 531-535.

This article reports on the experience of the British Army Blood Transfusion Service. Of their first 10,000 donors bled, they record a 2.8% fainting incidence. They define this as any case in which the donor lost consciousness or could not sit or stand without doing so. In one series of 4306 donors, they report 116 faints with a varied time of onset.


Tuesday, April 01, 2008

April JMLA posted in PubMed Central

And with it, the next installment in our case study series:

Cahall M, Jerome RN, Powers J. The impact of a literature consult service on geriatric clinical care and training in falls prevention. J Med Libr Assoc. 2008 April; 96(2): 88–100.

The case:
You frequently collaborate with a geriatrician in the adult primary care clinic of your large academic medical center. On a routine visit to his office to discuss his current needs for clinical evidence, he requests that you analyze the literature on effective interventions to reduce accidental falls in older persons. As he also provides geriatric care at the local Veterans Administration hospital and is the medical director at several local nursing homes and home health agencies, he notes that he is interested in literature describing interventions for use in private homes as well as those implemented in institutional settings.

More in the full article here.

Labels: ,

Search challenge 5

I'm still working up the post on the strategies for search challenge 4 (and may have to surrender my secret librarian decoder ring if I can't figure out what I'm doing wrong with my IndexCat search for fainting after blood donation...if you have tips, please contact me!), but in the meantime, a search for this week -- we had a request from a reader for comparison-focused searches - comparing different treatments, diagnostic tests, etc.

So a comparative search for this week: How do the erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) diagnostic tests compare and is it necessary to do both tests in pediatric patients with suspected osteomyelitis and/or septic arthritis?