Plenty of heat on whether mediated searches are of any value, Could I add
some evidence based light:

One-nil to the pharmacists! - MEDLINE searches performed by a drug
information pharmacist with online training, in response to drug information
requests, were judged by drug information pharmacists to be as useful as
searches performed by librarians.

Am J Hosp Pharm. 1988 Dec;45(12):2507-10.
Comparative usefulness of MEDLINE searches performed by a drug information
pharmacist and by medical librarians.
Wanke LA, Hewison NS.
The usefulness of MEDLINE searches conducted by a drug information
pharmacist who had received online training was compared with the usefulness
of searches conducted by medical librarians. Searches for literature
pertinent to select drug information requests received at a university drug
consultation service during a three-month period were conducted
independently by the pharmacist and one of three reference librarians. The
three librarians had received extensive online search training and were
experienced in conducting MEDLINE searches; the pharmacist was trained by a
fourth experienced librarian-searcher using National Library of Medicine
(NLM) training materials. For each drug information request, the pharmacist
and the librarian searched two MEDLINE files on the NLM service. The
printouts generated by the two searchers were evaluated by two drug
information pharmacists in terms of how useful each list of citations would
be in enabling the drug information pharmacists to answer the related
question. The majority of printouts in 48 sets of searches were judged to be
useful. Of the 96 possible "best search" votes (two votes for each of the 48
sets of printouts), the pharmacist's searches received 34 votes, the
librarians' searches received 28 votes, and there were 34 tie votes. The
numbers of useful primary and alternate citations were found to be the best
predictors of whether a given printout would receive the "best search"
rating. MEDLINE searches performed by a drug information pharmacist with
online training, in response to drug information requests, were judged by
drug information pharmacists to be as useful as searches performed by
librarians.


One-nil to librarians! Librarians had equivalent recall to, and better
precision than, experienced end-users.
Comput Biomed Res. 1990 Dec;23(6):583-93.
How good are clinical MEDLINE searches? A comparative study of clinical
end-user and librarian searches.
McKibbon KA, Haynes RB, Dilks CJ, Ramsden MF, Ryan NC, Baker L, Flemming T,
Fitzgerald D.
The objective of this study was to determine the quality of MEDLINE searches
done by physicians, physician trainees, and expert searchers (clinicians and
librarians). Its design was an analytic survey with independent replication
in a setting of self-service online searching from medical wards, an
intensive care unit, a coronary care unit, an emergency room, and an
ambulatory clinic in a 300-bed teaching hospital. Participating were all
M.D. clinical clerks, house, and attending staff responsible for patients in
the above settings. Intervention for all participants consisted of a 2-h
small group class and 1-h practice session on MEDLINE searching (GRATEFUL
MED) before free access to MEDLINE. Search questions from 104 randomly
selected novice searches were given to 1 of 13 clinicians with prior search
experience and 1 of 3 librarians to run independent searches (triplicated
searches). Measurements and main results from these unique citations of the
triplicated searches were sent to expert clinicians to rate for relevance
(7-point scale). Recall (number of relevant citations retrieved from an
individual search divided by the total number of relevant citations from all
searches on the same topic) and precision (proportion of relevant citations
retrieved in each search) were calculated. Librarians were significantly
better than novices for both. Librarians had equivalent recall to, and
better precision than, experienced end-users. Unexpectedly, only 20% of
relevant citations were retrieved by more than one search of the set of
three, with the conclusion that novice searchers on MEDLINE via GRATEFUL MED
after brief training have relatively low recall and precision. Recall
improves with experience but precision remains suboptimal. Further research
is needed to determine the "learning curve," evaluate training
interventions, and explore the non-overlapping retrieval of relevant
citations by different searchers.

Professions united! Increase in precision, recall, and user satisfaction for
the user-present search.
Bull Med Libr Assoc. 1982 Jul;70(3):298-304.
Being there: the effect of the user's presence on MEDLINE search results.
Morris RT, Holtum EA, Curry DS.
An evaluation was conducted at the University of Iowa Health Sciences
Library to determine what effect the user's presence had on MEDLINE search
results. One hundred users participated over a four-month period. Three main
criteria were used: search precision, search recall, and user satisfaction.
Each MEDLINE search was processed twice, once prior to the user's arrival
and a second time during the scheduled appointment with the user. The two
searches for each user were processed by different searchers, and four
searchers alternated processing the user-absent or the user-present search.
Users were asked to compare the citations on the two searchers by checking
each printout for relevant citations. A short questionnaire was administered
to determine general information about users and satisfaction measures for
each search. The extensiveness of the interview for the user-absent search
was varied in order to determine if the depth of the initial interview would
affect the search results. Evaluation of the findings indicated an increase
in precision, recall, and user satisfaction for the user-present search.
Some difficulties in designing the experiment are discussed and implications
of the evaluation results are considered.

And now for some non-evidence-based opinion:
I don't believe that health librarians working in the NHS can be viewed,
with individual exceptions, as expert searchers. They are the information
equivalent of the GP in that they are expected to handle any information
enquiry and do a pretty good job of it. (They also have to do so many other
things as well - managing staff and stock, end-user training, finance
etcetera). Expert searchers at centres for systematic reviews and guidelines
are the equivalent of consultants - they see a more limited "case" but
handle it more intensively and usually get much more credit!

I suspect that there is a volume and outcome issue here - just as surgeons
and endoscopists need to see a certain number of cases in order to keep
their hand in (which I suppose is anatomically true of vets too!) -
librarians presumably need to do a certain number of hours searching a year
to be considered experts. I think it would be wonderful if we could define
what that number is and then let anyone doing less than that acknowledge
that they are wasting their time and should let the end users do it
themselves!

Of course there are other issues such as how can librarians get better at an
activity that they usually do in isolation (mentoring or peer-review of
searches perhaps?) and whether certain cognitive styles are better than
others so we can spot who are the naturally-skilled searchers!

Without getting overly personal it is ironic that Reinhard who is one of the
best searchers I have ever known downplays librarian's skills while Mike who
has a breadth of other achievements builds them up. I suspect that the real
issue is whether we are talking in terms of "relative" or "absolute
values" - i.e. librarians are better than the alternative (but that alone
doesn't make them brilliant! cp. "Old age isn't so bad when you consider the
alternative") but are probably not as good as we encourage each other to
think we are!