Return to site

Why you should search PubMed and other health literature databases

Oliver Zolman MD

· Health Evidence

Dr Google might not suffice if you are looking to find the most reliable health information.

Whilst you might be shocked to see your doctor using google, they are (hopefully!) not googling your symptoms to look for information from Wikipedia or WikiHow. Instead, they are often googling national clinical guidelines - such as NICE guidelines (National Institute of Health & Care Excellence), approved patient information leaflets to print out - such as from NHS patients, or another sort of clinical guideline database called UpToDate.

Compared to clinical guidelines, patient information leaflets/websites or UpToDate, which doctors generally base their care on, studies on PubMed can be years more up to date.

However, what would be very unusual to see your doctor doing on their computer is using a medical database with all the latest medical papers. The best of these is called MEDLINE, which is a database with tens of thousands of peer-reviewed medical journals are listed on, with over 50 million total medical papers, over 10 million completely free to access full text papers (via something called PubMed Central), and over 1 million new medical papers added every year. The majority of the other papers on MEDLINE which are not free to see the full text (rather than just the 'abstract' which is a free summary of the key parts of the paper), you can use Deep Dyve ( (no affiliation) to gain a personal or small group subscription to these for probably 10 to 100 times cheaper monthly price than traditional journal subscription packages.

What would be very unusual to see your doctor doing on their computer is browsing a medical database with all the latest medical papers

MEDLINE can either be searched via a paid database searching software called Ovid, which can cost many thousands of dollars/pounds/euros a year to access unless you are at a university, or completely for free via PubMed - a USA government website created by the US national centre for biotechnology information (NCBI).  

PubMed can search the latest clinical (when we say 'clinical', this is just a fancy way of saying 'human') studies that have just been published a few days ago, and even ones dating back to the early 1900s.

Compared to clinical guidelines, patient information leaflets/websites or UpToDate, which doctors generally base their care on, studies on PubMed can be years more up to date - often taking years for the latest studies to trickle down into these clinical guidelines, patient information leaflets/websites and UpToDate. For many cutting edge topics, they may not be covered by these aforementioned information sources at all. 

Even when patient information leaflets/websites do exist, as discussed in our other blog post on 'Fast and Slow Health Decisions' - many patient information resources that cover topics relevant to informing a patient to make a health decision don't meet International Patient Decision Aid Standards (IPDAS).

Other databases exist, such as Cochrane Foundation database and Google Scholar. However, the Cochrane Foundation database can be searched from within PubMed and Google Scholar provides very broad, inaccurate results in its searches as papers are not tagged and reviewed by humans like in PubMed, and generally the search features are less advanced compared to PubMed. (Google Scholar is best used as a search engine after doing a PubMed search).

So why isn't my doctor searching PubMed?

From my experience, comprehensive PubMed/MEDLINE searching skills to find and access the papers relevant to a given health problem for a patient are not taught routinely in medical school or residency training. Skills in 'critical appraisal' (i.e. critique of the accuracy and relevancy of studies, including biostatistics) are also not routinely covered in depth in medical training. Both of these skills are typically reserved to epidemiology or evidence based medicine masters degree programs, or those undertaking research masters, PhDs or academic clinical training (such as academic foundation programs or academic clinical fellowships in the UK) outside of medical training.

Another huge barrier is time. With 15 minutes for GP/family doctor appointments typical, with at least 5 of these minutes spent recording notes typically, fitting in a quick pubmed search could be tricky. The search would likely have to be under 5 minutes, and there is a learning curve to be quick enough to perform such searches. 5 minute searches only have limited potential too; to get maximum assurance you have covered all the latest studies may take an hour or even many hours in complex health questions. One also has to take into account the extra time the clinician has to spend communicating this new information to the patient, or the extra time to develop a decision aid to help the patient decide between existing and new options that may be found in the search. This also raises further questions such as what if a clinician identifies an expensive new private therapy that is not approved on public healthcare insurance, or a test or therapy that is only available internationally or in a certain clinical trial.

Of course, resistance to change - as the National Health Service (NHS) in the UK is often critiqued of, can also be an inevitable barrier. Recently I read a 2019 LinkedIn article by an NHS doctor stating it took them 18 months to implement a new process of 'sponging up blood and weighing the sponge before and after' to better estimate blood loss in obstetrics and gynaecology wards (a test with sufficient evidence base), rather than trying to guess the blood loss volume by looking at the amount of blood on bedding and pads etc. Other articles have found that it took around 10 years for the latest studies to be implemented into average clinical practice (although this was using studies from over a decade ago).

Lack of funding at the national level could also be argued to be an issue - with the main UK guideline organistaion, NICE, having funding of around £50M GBP per year. This means NICE can only make so many guidelines, and only update them so often (quite often every 5 years). With 5 yearly updates, and over 5 million new peer-reviewed medical papers published in this time, unless there is no innovation at all in a field, even the best guideline at the time will quickly be missing new papers that could change health decisions.

Medical disclaimer: Our services and guides do not provide medical advice. Do not make any changes to your health behaviours based on this article without consulting your licensed clinicians first. See our medical disclaimer.

All Posts

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!