Truncate all drug brand name searches

If you search a drug brand name, be sure to truncate the end. Sometimes the brand name will appear with the TM symbol at the end, and if you search without truncation, you will miss those.

e.g. suboxone*[tiab] will find slightly more results than suboxone[tiab]

Watch out for advil*, though - there is a thing called advillin.

Substance Related Disorders

This is pretty easy in Medline:

"Substance-Related Disorders"[Mesh]

Just remember the hyphen and you're all good.

In PsycInfo, a little trickier, just because it doesn't explode properly:
Drug addictions:
DE "Substance Use Disorder" OR DE "Drug Abuse" OR DE "Alcohol Abuse" OR DE "Alcoholism" OR DE "Binge Drinking" OR DE "Drug Dependency" OR DE "Drug Addiction" OR DE "Inhalant Abuse" OR DE "Glue Sniffing" OR DE "Polydrug Abuse" OR DE "Alcoholism" OR DE "Alcoholic Psychosis" OR DE "Alcoholic Hallucinosis" OR DE "Delirium Tremens" OR DE "Korsakoffs Psychosis" OR DE "Wernicke's Syndrome" OR DE "Heroin Addiction"
Other addictions:
DE "Internet Addiction" OR DE "Process Addiction" OR DE "Pathological Gambling" OR DE "Sexual Addiction" OR DE "Workaholism"OR DE "Addiction"


In EMBASE, really tricky!
exp addiction/ or exp alcohol abuse/ or exp drug abuse/ or exp substance abuse/

You might think that you've caught all of the relevant stuff when you find the EMTREE term "addiction", but there are more terms under "abuse". But doing exp abuse/ is no good because of all the other, non-substance related narrower terms.

Comorbidity

As with many of the non-disease, non-treatment MeSH, Comorbidity is used inconsistently by indexers. Here are some keywords to help you:

"Comorbidity"[Mesh] OR comorbid*[tiab] OR co-morbid*[tiab] OR co-occur*[tiab] OR simultaneous*[tiab] OR superimpos*[tiab] OR coexist*[tiab] OR co-exist*[tiab]

If there is a sequence, that is if the patient first had one disease and then acquired another, you can also use:
preexist*[tiab] OR pre-exist*[tiab] OR antecedent[tiab]

Of course, your best bet is to combine the first disorder with the second disorder, without using any of the above. This is for cases where you're looking broadly (e.g. all of the diseases which can be comorbid with a particular disorder), or when there is another combination of the two terms that is cluttering up the search, for instance if you're searching for people who are addicted to two substances, and come across a lot of articles about treatment of patients who each have one of a variety of substance use disorders.

For the specific instance of a mental disorder and a substance abuse disorder, there is:
"Diagnosis, Dual (Psychiatry)"[Mesh] OR "dual diagnosis"[tiab] OR "dual diagnoses"[tiab]

CINAHL default search doesn't search the abstract?!?

OK, so if you're searching in CINAHL, and using a combination of Headings and keywords, the default setting to just search a keyword is "Select a field". I recall someone telling me that this was not an optimal search, but didn't realise how bad it was until today.

So, I had this:
S7     abscess*     7,986
S6     (MH "Abscess+")     5,152

Combined it with the rest of my search, got 28 results.

Then, I looked at my other search terms without the Abscess terms, which meant that there were 93 articles. However, one of those articles had the term abscess in the abstract. In fact, it appeared several times in the abstract and was obviously a significant part of the study results!

So, I searched again, and this time I received fewer results!

S10     TI abscess* OR AB abscess*     6,000

But when I combined them with my other search, there were more results, including the one which I had found.

The really weird thing is that when I tried S10 NOT S7, there were zero results! How is that possible?

Weird, weird, weird.

Prognosis

So, this is a tricky one. It's usually medicolegal, and usually it doesn't specify how the patient was treated, if at all.

"Prognosis"[Mesh:NoExp] OR "Disease Progression"[Mesh] OR "Return to Work"[Mesh] OR "Recovery of Function"[Mesh] OR "Convalescence"[Mesh] OR "Mortality"[Mesh] OR "Disease Free Survival"[Mesh]

There's also rehabilitation as a subheading. Probably best to tie it to the disease/disorder in question, if possible, and only free-float it if there isn't a disease in the question.

That's starting to get into the "how long will it take to get better" stuff, but here are a few more for that aspect:
"Time Factors"[Mesh] OR "Length of Stay"[Mesh]

Depending on the search, you may also wish to try:
"Treatment Outcome"[Mesh]

You can also do something straightforward, like:
prognosis[title] OR prognostic[title] OR predict*[title]
Why not?

EMBASE:

exp prognosis/ or disease course/ or exp adverse outcome/ or exp convalescence/ or exp disease clearance/ or exp disease duration/ or exp disease exacerbation/ or exp general condition deterioration/ or exp general condition improvement/ or exp illness trajectory/ or exp relapse/ or exp remission/ or exp survival/ or exp return to work/ or exp work resumption/ or exp mortality/

Suboxone gets its own MeSH heading!

And huzzah say I!

But, to get all that you can,

"Buprenorphine, Naloxone Drug Combination"[Mesh] OR ("Buprenorphine"[Mesh] AND "Naloxone"[Mesh]) OR suboxone*[tiab] OR bunavail*[tiab] OR zubsolv*[tiab] OR "buprenorphine naloxone"[tiab] OR "bup nal"[tiab] OR "bup nlx"[tiab] OR "bup nx"[tiab] OR "bnx"[tiab]

Note the asterisk after the suboxone keyword. This is important because some articles put the TradeMark symbol at the end of Suboxone, and this way you'll catch those as well.

What's really annoying is all the variant abbreviations that they use for this. I mean, seriously, check out this list:

• BUP-NAL
• BUP-NLX
• BUP-NX
• BPN-NAL
• BPN-NLX
• BPN-NX
• BNX

And that's not even mentioning the whitespace character variants, which fortunately the databases will ignore for you. However, if you are highlighting in Word, you'll want to replace the - with a / and then a +. As in, BUP/NAL and BUP+NAL.

In fact, I'm wondering if, when highlighting in Word, it might be faster to do: BUP-, BUP/, BUP+, -NAL, /NAL, +NAL, BPN-, etc... - That's 5 abbreviations times 3 whitespace characters for a total of 15 Find/replaces. Whereas the other way it's 6 combinations times 3 whitespace characters for a total of 18 Find/replaces. Plus, if some daft nutter has come up with another variant, you stand a chance of finding it by one or the other half of the combo.

EMBASE:

exp buprenorphine plus naloxone/ became an EMTREE heading in 2006.

Subheadings and explosions

In the Ovid Medline interface, subheadings are not autoexploded. You have to select subheadings from a list of checkboxes. Importantly, it looks like checkboxes are only provided for the subheadings associated with the MeSH term you're looking at - NOT for its narrower terms.

So, let's say that you're searching for statistics about corneal diseases. There exists a MeSH term Corneal Diseases, with narrower terms of course. You explode and go through to the subheadings page, and select epidemiology, ethnology, and mortality. You notice that "statistics and numerical data" is not available in the list.

You are, however, skeptical, so you try typing in exp Corneal Diseases/sn. Lo and behold, some results! A quick NOT shows that some of these were not found with the ep,eh,mo search.

Complete Reference view shows that the articles in this list are all narrower terms of Corneal Diseases, with subheading sn, such as Keratoconjunctivitis/sn.

So, if your search term has narrower terms and doesn't have a checkbox for a particular subheading, it's worth typing in the combination to see if any results from narrower terms are found.

Another example:

Searched Smoking and found three relevant subheadings: dt,pc,th

Searched "Tobacco Use" which is a broader term of Smoking. Only 2 subheadings found: pc,th

This meant that Smoking/dt returned 130 articles not found by the Tobacco Use search.

I was able to search exp "Tobacco Use"/dt,pc,th and retrieve the additional articles.

It may be helpful to search all subheadings related to your topic even if they are not listed.