Trip Database Blog

What has TRIP got to do with Somalia? 

The truth is very little.  However, at the start of this month we secured the domain name evidently.so (really http://www.evidently.so – but don’t click there as there’s nothing to see just yet). The .so domain is the country code extension for Somalia.  I purchased it on a whim, I had no idea we’d get it or what we’d use it for.  I just liked the phrase – it fits in with the ‘evidence’ concept and is an easy to remember name.

But we’ve got it and the more I reflect on it the more pleased I am with it.  I think it’s got all the potential to be a strong brand.

As for what to do with it, I’m currently thinking it might be a good home for our ‘new evidence’ service – which is currently just a vague concept. One to ponder, but if you have any suggestions?

I’ve been thinking how useful all the new content added – each month – to TRIP is.  We typically add between 400-700 articles each month manually (these are typically secondary evidence type articles), of these about 60% are recent publications and the rest from previous years.  In addition to this manual update we typically add around 3-5,000 new articles automatically (via various mechanisms). 

I typically view the manually added content as being the ‘best stuff’!  So, can we use this ‘best stuff’ in a meaningful way?  I believe so, but have no idea how!  Identifying it’s easy – displaying it is more problematic.  So, in an effort to help me understand the issues I’ve decided to experiment!  Therefore, I’ve taken March’s new content and extract all recent articles that relate to cardiology – 52 in total.  And below is my first effort at presenting them in a structured form (it’s a pragmatic structure – if that’s not obvious!) – please let me know what you think!

Lipids

• Mipomersen – Familial hypercholesterolemia – third or fourth line (National Horizan Scanning Centre)
• Rosuvastatin and the JUPITER trial. A critical appraisal (DTB, Navarre)
• Continue to follow NICE guidance when prescribing statins (NPC)
• Ezetimibe/simvastatin may reduce CV events in advanced CKD: but is it better than simvastatin alone? (NPC)
• Limited observational evidence for CV risk reduction with exenatide (NPC)
• Managing lipids (National Prescribing Service)
• Protection Against Nephropathy in Diabetes with Atorvastatin (PANDA): a randomized double-blind placebo-controlled trial of high- vs. low-dose atorvastatin (Evidence Updates)
• The cost-effectiveness of C-reactive protein testing and rosuvastatin treatment for patients with normal cholesterol levels (Evidence Updates)

Stroke

• Prognostic Risk Stratification Instruments for TIA (Washington University Emergency Medicine Journal Club)
• Elevation of serum magnesium may improve clinical outcome after aneursymal subarachnoid haemorrhage (BestBETs)
• Current evidence does not support the use of a negative D-dimer to rule out suspected pulmonary embolism in pregnancy (BestBETs)
• Stroke Prevention in the Setting of a Patent Foramen Ovale: A “Hole” in the Evidence (Clinical Correlations)
• Neurothrombectomy devices for the treatment of acute ischemic stroke: state of the evidence (DARE)
• An assessment of the incremental value of the ABCD2 score in the emergency department evaluation of transient ischemic attack (Evidence Updates)
• Computed tomographic angiography and venography for young or nonhypertensive patients with acute spontaneous intracerebral hemorrhage (Evidence Updates)
• Association between obesity and mortality after acute first-ever stroke: the obesity-stroke paradox (Evidence Updates)
• Very early mobilization after stroke fast-tracks return to walking: further results from the phase II AVERT randomized controlled trial (Evidence Updates)
• The safety of proton pump inhibitors and clopidogrel in patients after stroke (Evidence Updates)
• iScore: A Risk Score to Predict Death Early After Hospitalization for an Acute Ischemic Stroke (Evidence Updates)

Myocardial Infarcation

• Long-term effects of an expanded cardiac rehabilitation programme after myocardial infarction or coronary artery bypass surgery: a five-year follow-up of a randomized controlled study (Evidence Updates)
• High-dose tirofiban pretreatment reduces the need for bail-out study medication in patients with ST-segment elevation myocardial infarction: results of a subgroup analysis of the On-TIME 2 trial (Evidence Updates)
• Comparison of thrombolysis in myocardial infarction, Global Registry of Acute Coronary Events, and Acute Physiology and Chronic Health Evaluation II risk scores in patients with acute myocardial infarction who require mechanical ventilation for more than 24 hours (Evidence Updates)
• Clinical Events as a Function of Proton Pump Inhibitor Use, Clopidogrel Use, and Cytochrome P450 2C19 Genotype in a Large Nationwide Cohort of Acute Myocardial Infarction (Evidence Updates)
• With the “universal definition,“ measurement of creatine kinase-myocardial band rather than troponin allows more accurate diagnosis of periprocedural necrosis and infarction after coronary intervention (Evidence Updates)

CV Risk

• Renin-angiotensin system blockade and cognitive function in patients at high risk of cardiovascular disease: analysis of data from the ONTARGET and TRANSCEND studies (Evidence Updates)
• Antidepressant medication use and future risk of cardiovascular disease: the Scottish Health Survey (Evidence Updates)
• Cost-Utility of Aspirin and Proton Pump Inhibitors for Primary Prevention (Evidence Updates)

Clinical Rules

• Prognostic Risk Stratification Instruments for TIA (Washington University Emergency Medicine Journal Club)
• An assessment of the incremental value of the ABCD2 score in the emergency department evaluation of transient ischemic attack (Evidence Updates)
• The Randomised Assessment of Treatment using Panel Assay of Cardiac Markers (RATPAC) trial: a randomised controlled trial of point-of-care cardiac markers in the emergency department (Evidence Updates)
• Comparing the diagnostic performance of 2 clinical decision rules to rule out deep vein thrombosis in primary care patients (Evidence Updates)
• Genetic warfarin dosing tables versus algorithms (Evidence Updates)
• Predictive accuracy of SYNTAX score for predicting long-term outcomes of unprotected left main coronary artery revascularization (Evidence Updates)
• The cost-effectiveness of C-reactive protein testing and rosuvastatin treatment for patients with

Surgery

• Impact of clopidogrel loading dose on clinical outcome in patients undergoing percutaneous coronary intervention: a systematic review and meta-analysis (Evidence Updates)
• Comparison of impact of mortality risk on the survival benefit of primary percutaneous coronary intervention versus facilitated percutaneous coronary intervention (Evidence Updates)
• Predictive accuracy of SYNTAX score for predicting long-term outcomes of unprotected left main coronary artery revascularization (Evidence Updates)
• Randomized comparison of percutaneous coronary intervention with sirolimus-eluting stents versus coronary artery bypass grafting in unprotected left main stem stenosis (Evidence Updates)
• Effect of Timing of Chronic Preoperative Aspirin Discontinuation on Morbidity and Mortality in Coronary Artery Bypass Surgery (Evidence Updates)
• MADIT-CRT (Multicenter Automatic Defibrillator Implantation Trial With Cardiac Resynchronization Therapy) Trial (Evidence Updates)
• Thoracic epidural anesthesia for cardiac surgery: a randomized trial (Evidence Updates)

Other

• Sildenafil (Revatio) IV – for the treatment of patients with pulmonary arterial hypertension who are currently prescribed oral sildenafil and who are temporarily unable to take oral medicine (SMC)
• Estimating lifetime cardiovascular risk – we can, but should we? (NPC)
• Using low molecular weight heparin (LMWH) (WeMeReC)
• Haematological interventions for treating disseminated intravascular coagulation during pregnancy and postpartum (Cochrane)
• Evaluation of the diagnostic and prognostic value of plasma D-dimer for abdominal aortic aneurysm (Evidence Updates)
• Preoperative high-dose atorvastatin for prevention of atrial fibrillation after cardiac surgery: a randomized controlled trial (Evidence Updates)
• Home-based cardiac rehabilitation is as effective as centre-based cardiac rehabilitation among elderly with coronary heart disease: results from a randomised clinical trial (Evidence Updates)
• Rapid fluid removal during dialysis is associated with cardiovascular morbidity and mortality (Evidence Updates)
• Effect of adding pharmacists to primary care teams on blood pressure control in patients with type 2 diabetes: a randomized controlled trial (Evidence Updates)
• Six-month exercise training program to treat post-thrombotic syndrome: a randomized controlled two-centre trial (Evidence Updates)
• Trimetazidine: a meta-analysis of randomised controlled trials in heart failure (Evidence Updates)
• Early activation of intracranial collateral vessels influences the outcome of spontaneous internal carotid artery dissection (Evidence Updates)
• Minimized cardiopulmonary bypass reduces retinal microembolization: a randomized clinical study using fluorescein angiography (Evidence Updates)
• Effects of n-3 Polyunsaturated Fatty Acids on Left Ventricular Function and Functional Capacity in Patients With Dilated Cardiomyopathy (Evidence Updates)
• Screening for atrial fibrillation: sensitivity and specificity of a new methodology (Evidence Updates)

This is such an crucial video NICE guidance. Is it inaccessible? It’s been produced by a general practitioner from South Wales (Anne Marie Cunningham).  Although Anne Marie focuses on NICE guidance the issues affect other guidelines and many other sources of evidence.

The challenges Anne Marie raise are important and people interested in EBM/EBHC/EBP need to understand the problems front-line clinicians face. When Anne Marie can get the information she needs quickly and with little effort we’ll be in a better place.

However, we’re miles away from there and the challenges are daunting!

Each month we add add content via two main ways:

• Automatically – this uses various mechanisms (e.g. RSS)
• Manually – I go and look at various sites, typically those dealing with secondary evidence and grab any new content.

For this month (March 2011) I’ve just uploaded the monthly manual content.  This consisted of approximately four hundred articles dated 2011 and four hundred from 2010 (or earlier).  As I’m looking at ways of allowing users to more easily view new content I’ve experimented with a wordcloud using Wordle.  As many people have trouble viewing these I’ve embedded a screen grab below. To view it as large size you’ll need to click on it.

NOTE: I’ve edited the words (taken from the document titles only) to exclude certain non-clinical terms (e.g. trial, versus, randomized).

We’ve just had our 16,000 registered user of TRIP, that’s 1,000 new registrants in 3 weeks.

Our 16,000 user was a student from Fiji!

Here are a few recent Facebook posts (to see our full Facebook page click here).  Remember, the Facebook posts are typically small posts that are perhaps not ‘worthy’ of a blog post (but feel free to contradict me!)

• The BNF is now live on the TRIP Database (February 12th)
• Just about to add 300+ clinical guidelines from the Australian guidelines portal (February 15th)
• Just added records from two sites aimed at undergraduate medical students: Almostadoctor & MediVids (February 17th)
• Highlighted how good TILT was at highlighting significant new research (February 25th)
• Work on updating the systematic review filter that auto-searches PubMed (February 28th)
• Reported on the positive response to a training session I ran for a NHS organisation and offering to run more courses – just ask (March 2nd)
• I’m looking to introduce OpenURL into TRIP and gradually understanding the issues/complexity (March 3rd)
• Exploring hooking TRIP up to Google Translate (e.g. Spanish & Welsh) (March 7th)

We’ve been busy!

When users register on TRIP they can record keywords of interest and/or clinical areas of interest.  By doing this we send them a monthly email with new content that matches their interests. This information is displayed in the following format (click on image to enlarge):

I’m not a huge fan of this look as I think it’s too complicated. I raised this with Phil (the TRIP techie) and he wasn’t convinced by some of my ideas – so he suggested I ask the users what they thought!  The results are below…

When the monthly email arrives from TRIP, do you

• Ignore it – 2%
• Skim read it – 35%
• Follow some of the links – 62%
• I never receive them – 1%

The TRIP emails currently contain links to all new content. We are considering including the “top” 2 or 3 articles in each subject/interest area – highlighting them (title, publication and URL) in the email. What do you think?

• Sounds good, yes please – 77%
• Sounds good, but I’m worried it would make the email too long – 15%
• No thanks, there’s already too much information – 8%

Overall, how would you rate the emails

• Great – 22%
• Good – 56%
• Ok – 17%
• Poor – 5%
• Awful – 0%

What do you most like about the current emails?

• short, infrequent
• They are useful for concise
• They are well-organized and easy to use
• All wanted information is very clear arranged and I can grasp on more details with only one klick.
• Helps me with current awareness derived from a credible source! I often suggest my clients sign up for these emails directly

What do you most dislike?

• tough to see the significance without ANY detail
• Some features didn’t appear to work, specifically, I’ve clicked and chosen some of the links and have tried both the email and the send to rtf file but neither of those features have worked for me.
• Not enough information to help me decide whether to click on links
• Too wordy. Bullet points with links if the item looks interesting
• There is a lot of unrelated information in the links. Articles come up that have nothing to do with the topic.
• No information about the top articles in the emails compels readers to click on the multiple links. If article titles are listed, like Table of Content alerts from Journals, would be more user friendly.

We also asked for additional comments and got a number – including some significant love and fondness for TRIP.  A few raised issues of accuracy of the results (an issue we’re aware of an looking into).  The issue of primary care content raises it head yet again. The content for the emails is automated and I can think of no way of automatically deciding if an article is suitable for primary care or not.

Overall, it appears that most people are happy with the emails (I’m surprised) but a number would like additional information displayed to give a flavour of the content.  In the above example, with 12 interests and over 3,000 links – how do you select 4-5 that give a flavour?

These are all challenges that we need to deal with.  Fortunately, as people are broadly happy with the emails there is no huge pressure – so we’ll include it in our next upgrade (in 2-3 months). 

But if anyone cares to make a suggest we’d love to hear from you…

A while ago I read this article on TechCrunch Marissa Mayer’s Next Big Thing: “Contextual Discovery” — Google Results Without Search and it inspired me.  It helped me discover a recurring theme in my thoughts (and on this blog) that search is useful for known unknowns. The flipside of that is unknown unknowns, something you didn’t know you didn’t know.  In the above article Mayer talks about pushing information to people.  She talks alot about location but it doesn’t need to be restricted to that.

Currently on TRIP we have the ability for people to record their clinical area of interest and/or specific keywords of interest.  The specific keywords works reasonably well.  You say you’re interested in stable angina, and we email you every article that is about stable angina – simple and the feedback is positive.  Move to the wider clinical areas of interest (e.g. cardiology) and things get more problematic as it’s so broad.

But what if you’re like one of the people who doesn’t register or if they do they don’t record any interests?  Fine you use TRIP, it works well and that’s arguably enough.

But I want TRIP to do more.

If we encourage people to login we can start to note the articles they click on and start to create clinical hot spots of interest.  When they next login to TRIP (or we email them) we show them new articles that match there previous interests.  They don’t need to search.  There are a few issues (e.g. does previous searching predict future interests?) but I feel this is an avenue worth pursing.  I’m actually confident we can create these clinical hotspots and use that to the benefit of our users.

But what do you think?  Please comment to let me know.

In the middle of last year I attempted to draw conclusions about the difference in costs between TRIP and NHS Evidence (click here for the full blog post).  Since then I’ve been trying to understand the costs of NHS Evidence and have managed to arrive at the following figures (NOTE: I’ve had to use Freedom of Information requests which are tedious see here & here.  I welcome NHS Evidence correcting any figures below):

• Total budget – £24,438,000
• Content costs (BNF, Cochrane) – £10,675,971
• Remaining costs (This includes the budget for the specialist collections which I understand is approximately £2,000,000 for all 34) – £13,762,029

Of those remaining costs, here’s a flavour of the spending:

• Consultants and other temporary staff – £2,813,258
• Staff costs (the 41 employees) – £2,863,742
• Marketing – £730,000

Using the same search figures as before = 15,811,716 searches at a cost of £13,762,029 (excluding content costs from the total budget).  BTW since I quoted that figure a few people have suggested that actual figure for number of searches is too high as it includes search figures for other databases managed by NHS Evidence.  However, a specific question, via Freedom of Information, refutes this – so I will use this higher figure.

Each search on NHS Evidence costs 87.04 pence

TRIP on the other hand has now reduced its costs (on the TRIP Database) and will this year run on a maximum of £35,000.  Using the same search figures as before (8,058,648) this equates to a cost per search of 0.43 pence per search.

Therefore, each search on NHS Evidence costs 202 times more than on TRIP.

Why does this bother me?  A number of reasons:

• The NHS is facing massive challenges to the budget yet NHS Evidence appears immune.
• NHS Evidence distorts the market and TRIP suffers. 
• NHS Evidence does not appears to be engaging with librarians (that’s certainly the feel I get from conversations and emails). The recent decision to give the specialist collections to non-librarian consortia reinforces this feeling.

But the main one is that I don’t feel they offer a superior service to TRIP.  I would go as far as to say that I feel TRIP is significantly better at supporting real, frontline clinical staff.  However, I still stand by my main conclusion of last year that search isn’t the answer to properly supporting clinicians to practice evidence-based healthcare.

With regard to these latter points I’d love to have the funds to test this.  NHS Evidence got any spare change?

UPDATE: We’ve had an email from someone who used to work on the predecessor of the NHS Evidence – the National Library for Health (NLH).  They point out that the old budget for the NLH was approximately £9,000,000.  The only significant difference between the two (in relation to funding content (is the £5,000,000 paid to the BNF.  So, comparing the old NLH and the new NHS Evidence there is a difference in funding level of around £10,000,000. They would like to know what does that extra money get you?

I’ve no idea, perhaps NHS Evidence can tell us?

UPDATE 2: I tweeted about this post and subsequently saw this tweet from Ben Goldacre (@bengoldacre) he of Bad Science fame:

A search on NHS Ev costs 200X one on TRIP: I find TRIP better