"A medical degree is not required for this job...but it wouldn't hurt!"

I had a very interesting clinical query last week which took me completely out of my comfort zone and made me think afterwards "I must blog about this!" We had a call from a registrar on the wards on Tuesday asking if we could track down the lowest ever recorded sodium level (in a human) in the literature. He had a patient with a sodium level of 100 and he wanted to see if there had been any recorded lower. "Well that shouldn't be too difficult" thought I, but having sat myself down in front of Medline it took me ages to think how to approach the question in a standard database query sort of way. I started out very basically entering "low sodium", "lowest sodium" or "sodium requirements" as a title search but just got loads of articles about dietary sodium levels and things to do with animals. I then branched out to a thesaurus search for SODIUM and HUMANS and "low*" but again, too broad. I had to take a break then to do a one-to-one session with an OT but I asked the library assistant and senior library assistant to have a think as well.

An hour or so later I was back on the case. The library assistant had found a wonderful site on medical world records but unfortunately sodium levels were not exactly exciting enough to merit an entry. The senior library assistant suggested I consult some books on fluids and electrolytes which proved to be an excellent move because from these books ("Fluids and Electrolytes Made Incredibly Easy" and "Fluids and Electrolytes: A 2-in-1 Reference for Nurses") I discovered that a low sodium (or serum sodium) level was known medically as hyponatremia and it is measured in terms of MeQ/L. A serum sodium level of 100 is dangerously low (low is considered to be between 120 and 135) which made me hope our registrar was treating this patient and not sitting around waiting for my answer!

I returned to my Medline search and search hyponatremia in the title and the thesaurus, combined these searches with OR then searched "100 MEQ/L" OR "less than 100 MEQ/L" and added these searches to the hyponatremia search. I got 6 results, which were not bad but still not quite what I was after. I started working down from 100 MEQ/L to 90, then 80, 70 and 60 but that didn't work very well so I then tried combining the hyponatremia searches with "severe" and "serum sodium level" which, while not giving me the answer, helped me to gain a better knowledge about what I was looking for. In the end I found the combination of hyponatremia (title and thesaurus) AND "serum sodium.ti,ab" AND "MEQ/L" AND "severe.ti.ab" gave me a good range of low sodium levels, both case studies and research on groups of patients and I could be reasonably sure that the lowest serum sodium level ever recorded in the literature is 99 MEQ/L.

It was a very interesting search, not least of which the way it illustrated the different ways humans and computers "think". Funnily enough, the library assistant told her friend about the search a few days later and he promptly put in "lowest sodium level recorded" into Google and came up with a relevant article mentioning 99 MEQ/L as the lowest ever but without having done the search myself on a database I couldn't have been certain that this was the right answer. I certainly now know more about sodium levels than I ever wanted to know!