1. Dolphins and cows Neal Maskrey / Jonathan Underhill
National Prescribing Centre
Liverpool UK

2. what should we be doing differently?
why evidence-based practice doesn’t happen as often as it sometimes should
what should we be doing differently?

5. BTS/SIGN Asthma guidelines 2004, 2005, 2007
BTS Asthma guidelines 1997
BNF says “but inhalation preferred”
BNF for children “not recommended”
BTS/SIGN Asthma guidelines 2004, 2005, 2007

7. The doxazosin arm of ALLHAT
The doxazosin arm of ALLHAT? Hypertension 2003; 42: , Ann Intern Med 2002; 137: , JAMA 2000; 283:
Increased risk of stroke (RR 1.26), combined CVD (RR 1.20) and heart failure (RR 1.80) with doxazosin
The ALLHAT Study also included an alpha blocker arm. However, because of an apparent excess risk of CVD in this groups, this arm was stopped early. Final analysis confimred an increased risk of CVD and heart failure, without any difference for the primary outcome of coronary death/ MI.

8. Prescribing of alpha-blockers in the US following ALLHAT Stafford RS, et al. JAMA 2004; 291: 54-62
In the US, active marketing of doxazosin was stopped and the amount prescribed reduced appropriately.

9. ALLHAT (doxazosin arm)
PUBLISHED
Bu this was not the case in the UK, where doxazosin prescribing continued to increase…

10. Trends in the prescribing of alpha blocking drugs in England
……..and is still doing so. Remember, that guidelines now place alpha-blockers as at least a 4th line measure for treatment of hypertension. There are of course caveats with this prescribing data – prescribing for benign prostatic hypertrophy being the major confounder – but they have BPH in the States too.

11. Cox IIs MeReC Briefing 2002; 20
The GI safety of rofecoxib and celecoxib has been assessed in large clinical outcome trials which, on first analysis, show benefits over non-selective NSAIDs in the incidence of serious upper GI complications. However, longer-term GI data from the celecoxib study (CLASS) and cardiovascular adverse event data from the rofecoxib study (VIGOR) have questioned the risk/benefit profile of these new drugs and, until they are better understood, it seems sensible not to use them routinely in large numbers of people.

12. Trends in Prescribing of NSAIDs in General Practice in England
Well, some people may have taken notice, but this data seems to indicate that lots didn’t. The lilac coloured bar is rofecoxib which was of course withdrawn voluntarily following increased CV event rates, and the peach bar is celecoxib which increased following the rofecoxib withdrawal and which then fell away again as the MHRA issued warnings about cardiovascular risks of all coxibs.
All this is still subject to litigation but calculations of the numbers of avoidable or premature heart attacks associated with the prescribing of these agents is more than 100,000 in the USA alone. 12

13. MeReC Extra 30 November 2007 NSAIDs - GI and CV risk
Long-term, randomised controlled trials (RCTs) have demonstrated that coxibs cause a small increased risk of thrombotic events in comparison with placebo. The excess risk was estimated to be about three cases per 1000 users treated for one year. This risk appears to increase with dose and persists throughout treatment.

14. ….following a review of the evidence in October 2006, the Commission on Human Medicines (CHM) advised that there was sufficient evidence to suggest that traditional NSAIDs may also be associated with a small increased risk of thrombotic events when used at high doses and for long-term treatment. Furthermore, they identified that not all traditional NSAIDS carried the same CV risk:
Naproxen 1000mg/day may be associated with a lower risk of thrombotic events than coxibs.
Ibuprofen …..at low doses (e.g. 1200mg/day) …. does not suggest an increase [in] thrombotic risk
Diclofenac 150mg/day has a thrombotic risk profile similar to that of etoricoxib▼ 60/90mg, and possibly other coxibs.

15. Diclofenac accounted for 46% of all NSAID prescribing …in primary care in England ……. If the excess risk for CV events is the same as COX-2 inhibitors (3 per 1000 patients per year) then approximately 2000 additional or premature CV events per year could be caused by diclofenac prescribing, compared with no treatment.

17. #1 However good the information is, on its own it is
usually insufficient to change practice

18. Work harder
Be more effective
Be imaginative and try some new things

19. 1. Work harder I’m not working hard enough My workload is just right
I’m working too hard

20. Without innovation, public services costs tend to rise faster than the rest of the economy. Without innovation, the inevitable pressure to contain costs can only be met by forcing already stretched staff to work harder.
Mulgan G & Albury D (2003).
Innovation in the Public Sector.
Strategy Unit, London.

21. #2
It’s hard to find NHS workers who have
substantial spare capacity

22. 2. Be more effective

23. 2003
1994
2007

24. Summary of recommendations
Incentive scheme for all practices
Better co-ordination of resources promoting rational prescribing
Better information
Education
Audit
Formularies and guidelines
Use skills of community pharmacists
Practice visits should be better targeted, more focussed, effectively followed up
Summarise information for prescribers
Improve local consensus
Involve and educate patients

25. The median effect size overall was approximately 10% improvement in absolute terms. Grimshaw J, et al. Implementing clinical guidelines: current evidence and future implications: Journal of Continuing Education in the health professions 2004; 24: S31-S37.
The only factors that had a significant relationship with implementation of these important prescribing changes were an innovative approach among the doctors (most practitioners were cautious of change), and fundholding status. Salisbury C, et al. The implementation of evidence-based medicine in general practice prescribing. BJGP 1998; 48;
102 trials of educational interventions in the health care professions published between 1970 and 1993 inclusive.
Dissemination-only strategies, such as conferences or the mailing of unsolicited materials, demonstrated little or no changes in health professional behaviour or health outcome when used alone.
More complex interventions, such as the use of outreach visits or local opinion leaders, ranged from ineffective to highly effective but were most often moderately effective (resulting in reductions of 20% to 50% in the incidence of inappropriate performance). Oxman AD, et al. CMAJ 1995; 153:
Grimshaw J, et al. Implementing clinical guidelines: current evidence and future implications: Journal of Continuing Education in the health professions 2004; 24: S31-S37. One of the most common findings from health services research is a failure to routinely translate research findings into daily practice. Previous systematic reviews of strategies to promote the uptake of research findings suffered from a range of methodologic problems that have been addressed in a more recent systematic review of guideline dissemination and implementation strategies. Changes in practitioner behavior; in the desired direction, were reported in 86% of the comparisons made. The median effect size overall was approximately 10% improvement in absolute terms. The review suggests that interventions that were previously thought to be ineffective (e.g., dissemination of educational materials) may have modest but worthwhile benefits. Also, multifaceted interventions, previously thought to be more effective than single interventions, were found to be no more effective than single interventions. Overall, there is an imperfect evidence base for decision makers to work from. Many studies had methodologic weaknesses, and reporting of this kind of research is generally poor, making the generalizability of study findings frequently uncertain. A better theoretical underpinning of studies would make this body of research more useful.

26. #3 If it was possible to deliver implementation
strategies with 100% effectiveness, the best results
we could expect are a shift in the desired direction of
10-50%.

27. Let it happen Help it happen Make it happen
Greenhalgh T, et al. NCCSDO. April 2004
Let it happen
Help it happen
Make it happen
Marketing: based on health promotion and social marketing theories. Attractive product, adapted to local needs (after assessment). Stepwise approach, multiple channels for communication
Educational: based on adult learning theories. Focus on internal motivation of professionals. Bottom up, PBL, small interactive groups
Epidemiological: based on rational, cognitive theories; information. Guideline development; dissemination via courses, journals, mailing

28. Let it happen Help it happen Make it happen
Social interaction: based on social learning and innovation theories, social influence and power theories. Peer review in local networks, outreach visits, opinion leaders, patient mediated interventions
Organisational: based on management & system theories. Create structural and organisational conditions to improve care. Re-engineering care processes, TQM, team building, leadership.
Behavioural: based on learning theory. Control performance by audit, feedback, reminder systems, £, sanctions.
Coercive: based on economic, power and learning theories. Regulations, budgeting, contracting.

29. #4 The NHS mostly does things to clinicians,
rather than helping them acquire for
themselves the knowledge, skills and attitudes to
do a better job

30. The NHS professional bureaucracy
Front line staff have a large measure of control over decisions
Hierarchical directives issued by those nominally in control often have limited impact – even resisted
Negotiation, not imposition
Collegial influences – credibility of professionals at their core, not simply those in formal positions of “authority”
Mintzberg H. The Structuring of Organisations: a synthesis of research. 1979

31. 3. Be imaginative

32. information management
cognitive psychology
explaining risks and benefits to patients

33. information management
So let’s start, as we did, with Information Management.
information management

35. information management
Volume of published material is unmanageable
Reading is haphazard
Formal CPD is haphazard

36. "We surveyed one acute medical take in our hospital
"We surveyed one acute medical take in our hospital. In a relatively quiet take, we saw 18 patients with a total of 44 diagnoses. The guidelines that the on call physician should have read remembered and applied correctly for those conditions came to 3679 pages. This number included only NICE, the Royal Colleges and major societies from the last 3 years. If it takes 2 min to read each page, the physician on call will have to spend 122h reading to keep abreast of the guidelines" (for one 24h on-call period).
Allen D, Harkins KJ. Lancet 2005; 365: 1768

37. More reading? Potential journals 10,000
Potential new articles per week 40,000
Even if 97% are not relevant (no POOs) 1,200
Time to read each article 15minutes
10h a day, 6 days a week = articles.
So at the end of the first week you are about 4 weeks behind in your reading.
At the end of the first month, you are 4 months behind in your reading.
And at the end of the first year you are almost 5 years behind in your reading.

39. Effect of exercise on pain in knee OA Roddy E, et al
Effect of exercise on pain in knee OA Roddy E, et al. Ann Rheum Dis 2005; 64: 544-8
There is reasonable evidence to support the use of exercise in patients with knee osteoarthritis.
This systematic review considered RCTs of trials of exercise against non-exercise
Both aerobic walking and home based quadriceps strengthening exercise reduce pain and disability (data not shown) from knee osteoarthritis.

40. Clinician reading journals
Evidence-based treatment
for the patient
Clinician reading journals

41. Information Mastery requires two different approaches to managing information:
Foraging - a method of being alerted to new relevant, valid information when it is published
Hunting - a method of finding information when it is needed
NPC would add a third:-
Hot-synching – clinicians rapidly checking once or, at the most, twice a year that their key approaches for the management of conditions they see commonly still match the best evidence
DN Need a link on NPCi to Dave and Allen’s paper describing how to select a foraging service.
Both clinicians and patients rate “keeping up to date” very highly. We discussed in the second workshop in this overview the information explosion and the resulting overload. The question is what tactics can be employed to manage the information.
Two American professors – Dave Slawson and Allen Shaughnessy – came up with the Information Mastery concept and first published on it in They’ve described foraging as the first key approach. Everyone needs a service that alerts them to the key – and we’ll define the characterisitics that comprise “key” in a moment – the key new research in their specialty. Let’s use our NPC approach as an example. Five years ago we developed our own in house electronic currrent awareness bulletin. Our information scientist selected key websites with us – sites that would publish on a daily weekly, monthly or irregularly new developments relating to prescribing therapeutics and medicines management. So for example we’d capture policy developments as well as new clinical trials. We developed some selection criteria for the new information published on those sites…….and the service became so useful to us in-house that we rapidly made it available to anyone who wished to subscribe. We also devised some criteria for selecting the greatest hits of the daily digest and provide a weekly digest for those wanting a once a week update, rather than a once a day. More than 4,000 now subscribe.
There are lots of services available that do this now for different audiences– and some are much better than others.
As well as simple lists like our eCAB, there’s a need for a service that takes just the key new developments and both critically appraises the new data and sets those findings in the context of the rest of the evidence. We do that with our NPCi blog service – 2 or 3 key papers a week.
The key feature of relying on foraging is that some one else – preferably a trusted public sector organisation – does the selection and critical appraisal for the subscribers. We don’t expect clinicians to read journals, select the papers to appraise and then incorporate the new data into their practice. We do the selection and appraisal, leaving clinicians to concentrate on what they need to do and are skilled at doing.
This is a young science and developments are taking place all the time. We’ve recently recognised we need to put levels on evidence into the blogs, for example.
As well as foraging, clinicians need an approach to finding information when its needed, when they are “stuck”. The appraoch need to be to finding the best answer , not just an answer. We’ll come to that in a minute.
Over time, we’ve added a third approach – which links to a new and more logical approach to continuous professional development. It’s crazy to expect clinicians to be up to date with conditions they see rarely – the volume of material is too large to handle and by the time they see a condition its likely they’ll have forgotten what they learnt – or worse remembered what they learned incorrectly. So its unreasonable to expect a GP to know without checking how to manage hypertrophic cardiomyopathy, but entirely reasonable to expect a tertiary centre cardiologist to. Similarly, its unreasonable to expect a cardiologist to be up to date in managing depression in the postnatal period, but you’d expect a GP to be on the button. And yet the cardiologist ought to be spot on when managing depression in people after a heart attack.
Multiple surveys have shown clinicians can spend a maximum of an hour a week keeping up to date. That’s enough time for a little foraging and then each year as a minimum hot synching. Hot synching is a term used in computing when one device connects to another and their memories update each other. I was describing the process of covering the key evidence in the 30 or so conditions a clinician sees commonly swiftly on a regular cycle when talking to some family medicine residents in the States and one of them said “You mean like hot synching” – and that’s exactly it. You’ll note we’ve designed NPCi specifically to enable hot synching for the common conditions in primary care and those topics as appropriate for some of the requirements of some specialist too. But with our resources we’re never going to be able to cover, for example lots of oncology because there are a hundred different cancers all with their own evidence base, or cover the requirements of orthopaedic hand surgeons, liver transplant teams and so on. But the same principles apply to people working in those fields.

42. Usefulness = Relevance x Validity Work
What are the criteria used when looking for the best answer or important evidence?
Slawson DC and Shaughnessy AF. J Am Board Fam Pract 1999; 12:
Usefulness = Relevance x Validity Work
Dave and Allen recognised the three issues in using evidence – work, relevance and validity - when they came up with the usefulness equation.

43. How can we quickly spot what is NOT important to us?
Not RELEVANT
Upstream to clinical decisions being made, e.g. animal or in vitro studies
Study populations and / or settings do not reflect question type, practice population and settings
Not VALID
Poor study design
Bias and confounding
Measurement validity
Insufficient power

44. So, filtering for relevance
Feasible (intervention)
Outcomes (patient-orientated)
Common (condition)
Change in practice required

45. Disease / condition DOO POO
DOOs can mislead and don’t always relate to POOs Ebell M, et al. Am Fam Physician 2004; 69: 548–56
Disease / condition
DOO
POO
Doxazosin for BP
In Blacks, reduces BP
Increases mortality
Lidocaine for arrhythmia after MI
Suppresses arrhythmias
Finasteride for BPH
Improves urinary flow rate
No clinically important change in patient symptom scores
Arthroscopic surgery for OA knee
Improves appearances after debridement
No change in function or symptoms at 1 year
Sleeping infants on their stomach or side
Based on anatomy and physiology, will decrease risk of aspiration
Increase in SIDS

46. Disease / condition DOO POO
Vitamin E for heart disease
Reduces levels of free radicals
No change in mortality
HRT for CHD prevention
Reduces LDL, increases HDL
No decrease in CV or all cause mortality, increase in CV events over 60 years
Insulin for PWT2D
Keeps blood glucose below 6.7mmol/L
No reduction in mortality
Sodium fluoride for fracture prevention
Increases bone density
Does not reduce fracture rate
Lidocaine prophylaxis after MI
Suppresses arrhythmias
Increases mortality
Beta blockers for heart failure
Reduce cardiac output
Reduce mortality

47. If the answer to any of those is “no”
I don’t know and I don’t care

48. After checking it is relevant, is the answer likely to be valid?
How to quickly spot the fatal flaws:
Is it a high level of evidence?
Is it statistically significant?
Is it clinically significant?:
Do you understand what the the numbers tell you?
Absolute vs. relative risk vs. NNT
Was there enough people in the study for long enough?
Was the allocation concealed?

49. Reading and critical appraisal MUST (largely) be replaced by reading pre-digested sources of evidence from trusted sources

50. Be an Information Master! Slawson DC and Shaughnessy AF
InfoRetriever, DrCompanion, self-assembly
Cochrane Library
NICE, (NSFs)
Usefulness
Clinical Evidence
InfoPOEMs, Prodigy
BestTreatments
NPC ref sheets
EBM DTB MeReC
Bandolier
“Ivy League” journals
Textbooks
Medline

51. The five S approach Haynes RB. Evidence-Based Medicine 2006;11:162-4
NPCi therapeutics floors, Clinical Evidence, CKS (PRODIGY), etc
NPCi blogs and podcasts, EBM, ACP Journal Club, etc
Cochrane reviews, HTAs, NICE full guidelines, etc

52. information management
Stop reading journals (to stay up to date)
Read trusted summaries of evidence
Concentrate CPD on the conditions you see commonly (c30 for a GP = two thirds of problems presenting)

53. #5 Information management is nowhere near being a
key component of undergraduate
or postgraduate curricula – yet without it how
can we expect clinicians to be up to date?

55. Comments / Questions