Finding Relevant Evidence

Part A was where evidence lives. This half is the mechanics: how you get it out, from the question you built in unit 2 to a yield small enough to read. One technique carries most of this lecture, and it is unglamorous. Enter each concept alone and look at what it returns, before you combine anything. Combining first hides which of your terms is doing the damage, and that is why searches fail in ways people cannot diagnose. By the end you can take terms out of a PICOT question and say why keywords are not enough; use subject headings alongside them; combine concepts with AND and OR and say which way each moves the yield; rebuild a search that came back thin, and name the limit most likely to hide what you needed; and say what makes a search reproducible and how you know to stop.

Five videos taking the question apart into search terms, and reaching the limit that keywords alone run into.

Where the search begins.

EBP 0337 Start From the PICOT Question

The search starts from the question you already built. Not from a phrase typed whole into a box.

How terms come out of it.

EBP 0338 Take Keywords From the Question

The elements of the question supply the search terms directly. That is what the format was for.

What a keyword search really is.

EBP 0339 What a Keyword Search Actually Does

A keyword search matches the text of a record. It finds the word, not the concept.

Why start there anyway.

EBP 0340 Why Start With Keywords

Keywords are where you start because you can always type them, and they show you the vocabulary the literature uses.

And the limit you hit.

EBP 0341 Keywords Are Ambiguous

The same idea has several names and the same word has several meanings. Keywords alone find some of what exists.

Three videos on the instrument that fixes the keyword problem, and how the two are used together rather than instead of each other.

The instrument that fixes it.

EBP 0342 Subject Headings

A subject heading is a controlled term assigned by an indexer, so every record on a concept carries the same tag whatever words the authors used.

The best-known one.

EBP 0343 MeSH

Medical Subject Headings, the controlled vocabulary used in MEDLINE and PubMed.

How they are used together.

EBP 0344 Headings and Keywords Together

Headings catch what keywords miss and keywords catch what has not been indexed yet. Use both, combined with OR.

Seven videos on the mechanics. The technique that matters most is entering each concept alone before combining anything.

A decision made before you type.

EBP 0345 Choose the Databases Before Searching

Choose your databases before you start, based on the question. Deciding as you go is how coverage gets uneven.

The technique the rest depends on.

EBP 0346 Search Terms One at a Time

Enter each concept alone and see its yield. Combining first hides which term is doing the damage.

The word for what comes back.

EBP 0347 Hits

The number of records a term or combination returns.

How the concepts join.

EBP 0348 Boolean Connectors

AND, OR and NOT. Three connectors, and two of them do almost all the work.

The first connector.

EBP 0349 AND Narrows

AND requires both terms, so it always returns fewer. It is how you combine different concepts.

The second.

EBP 0350 OR Broadens

OR accepts either, so it always returns more. It is how you gather the several names for one concept.

How the pieces assemble.

EBP 0351 Build Search Statements in Steps

Build each concept as its own statement, then combine the statements. The history shows you exactly where the yield changed.

Five videos on fixing and tuning a search. Watch the full-text limit, which is the one most likely to hide the study you needed.

What to do when almost nothing comes back.

EBP 0352 Rebuilding a Search That Returns Too Little

Broaden the population, add synonyms with OR, check headings, and strip limits. Too little is more often a search fault than a literature fault.

A small technique with large effect.

EBP 0353 Truncation

A symbol on the stem of a word catches its endings, so reposition catches repositioned and repositioning.

Filters, and when to apply them.

EBP 0354 Limits

Limits narrow by language, age, date or type. Apply them last, after you know what the unlimited yield looks like.

The limit that quietly loses evidence.

EBP 0355 The Full-Text Limit

The full-text limit discards everything your library could obtain another way. It is the limit most likely to hide the study you needed.

What makes that limit unnecessary.

EBP 0356 Interlibrary Loan

A library can obtain an article it does not hold. That is why searching only what is immediately free is a false economy.

Five videos on what closes a search. Saving the history is the step that makes the whole thing systematic rather than merely thorough.

The step that makes it systematic.

EBP 0357 Save the Search History

Saving the history is what makes the search reproducible, and therefore what makes it systematic at all.

Why a search has a shelf life.

EBP 0358 A Search Is Dated

A search describes the literature on the day you ran it. Before anything is published, it is re-run.

Why the same search moves badly.

EBP 0359 Databases Search Differently

The same strategy behaves differently in different databases. Each is rebuilt rather than pasted across.

The word for what you end with.

EBP 0360 Yield

The set of records a completed search returns, which is what you then screen.

How you know when to stop.

EBP 0361 Knowing You Have Found It

The same studies begin recurring across databases and the reference lists stop offering new names. That is saturation, and it is the signal rather than a feeling.

Three videos on what people believe about searching that is untrue. The first is the one that ends searches early.

The first of three errors.

EBP 0362 Misconception: Finding Nothing Means There Is Nothing

Finding nothing usually means the search failed. Rebuild before concluding the literature is empty.

The second.

EBP 0363 Misconception: The Best Evidence Is the Newest

Newest is not best. A well-conducted older trial can outrank a recent weak one, and strength is level and quality together.

The third, which unit 8 takes up.

EBP 0364 Misconception: A Guideline Answers Your Question

A guideline is built for a population and may not address your patient. Unit 8 is about appraising one before adopting it.

One procedure taking a question from its elements to a readable yield.

The unit assembled into one search.

EBP 0365 Conducting a Systematic Search

One question taken from PICOT elements through headings, Boolean combination and limits to a yield small enough to read.

That is step two, complete. Against the outcomes. The question supplies the terms directly, which is what the PICOT format was for. But keywords match text rather than concepts, so the same idea under a different name is missed. Subject headings fix that, because an indexer assigns one controlled term to every record on a concept. Headings catch what keywords miss, keywords catch what has not been indexed yet, and you use both joined with OR. AND requires both terms and always narrows. OR accepts either and always broadens. Concepts are built as separate statements and then combined, so the history shows exactly where the yield changed. When a search comes back thin, rebuild before concluding the literature is empty: broaden the population, add synonyms, check headings, strip limits. And the full-text limit is the dangerous one, because it discards what your library could have obtained through interlibrary loan. Saving the search history is what makes a search reproducible, and reproducibility is what makes it systematic at all. A search is dated, so it is re-run before anything is published. You stop when the same studies recur across databases and the reference lists stop offering new names. Unit 4 is step three: deciding whether to believe what you have found.