158 VOLUME 4 • NUMBER 2 • JUNE 2012 JOURNAL OF PRIMARY HEALTH CARE
Mangin D. Adherence to
evidence-based guidelines is
the key to improved health
outcomes for general
practice patients—the ‘no’
case. J Prim Health Care.
Dee Mangin MBChB,
Department of Public
Health and General
Practice, University of
BACK TO BACK
7. New Zealand Guidelines Group. New Zealand cardiovascular
guidelines handbook: a summary resource for primary care
practitioners. 2nd ed. Wellington: New Zealand Guidelines
Group; 2009. Available from: http://www.nzgg.org.nz/
8. National Clinical Guideline Centre. Chronic heart failure: the
management of chronic heart failure in adults in primary and
secondary care. London: National Clinical Guideline Centre;
2010. Available from: http://guidance.nice.org.uk/CG108/
9. General practitioner, Dr A. A Report by the Health
and Disability Commissioner (Case 10HDC00253).
2010. Available from: http://www.hdc.org.nz/
10. Late diagnosis of abdominal tumour and DvT (04HDC14223).
A Report by the Health and Disability Commissioner. 2005.
Available from: http://www.hdc.org.nz/decisions--case-notes/
11. Worrall G, Chaulk P, Freake D. The effects of clinical practice
guidelines on patient outcomes in primary care: a systematic
review. CMAJ. 1997;156:1705–12.
12. Doran T, Fullwood C, Gravelle H, Reeves D, Kontopantelis E,
Hiroeh U, Roland M. Pay-for-performance programs in family
practices in the United Kingdom. N Engl J Med. 2006;355:375–84.
13. Øvretveit J, Gustafson D. Improving the quality of health
care. Using research to inform quality programmes. BMJ.
Adherence to evidence-based guidelines is
the key to improved health outcomes for
general practice patients
‘Clinical Practice Guidelines’—a Google search
using this term netted 26 200 000 results in
0.43 seconds. Guidelines are as unmanageable as
the research they were designed to summarise.
Guidelines were intended to bring the best scien-
tific evidence to bear on primary care practice—
an upgrade from the Blue Book that we used to
carry in case of knowledge emergencies as a house
surgeon. Guidelines have now moved beyond
this—the quality of family practitioners’ care is
increasingly measured by guideline adherence.
Is adherence to guidelines the best way to im-
prove health outcomes? No—it may result in care
that seems measurably better, but is meaning-
fully worse for health outcomes. There are three
broad reasons for this—the quality of guidelines,
the quality of the available research data that
underpin them and their unfitness for purpose in
a primary care setting.
The quality of guidelines
If guidelines stuck to the data and critical as-
sessment of its gaps and uncertainties this might
be useful—but back-filling the gaps in data
with ‘consensus’ appears to be irresistible. In a
study of 2700 recommendations in the Ameri-
can Heart Association / American Cardiology
Association guidelines, only 10% were based on
high-quality RCT evidence.1 Half were simply
consensus. The widespread levels of conflict of
interest of group members with the manufactur-
ers amplifies the concern.
The label ‘level C evidence’ does not undo the air
of certainty of the written word on the page of a
guideline. One example is HbA1c target levels for
Type 2 diabetes, which are standards that increas-
ingly doctors are exhorted to adhere to, and in
some countries carry an income bonus. There is no
good evidence for treating to any particular target
HbA1c. Large well-designed studies have shown
the harm and increased mortality associated with
tight glucose control and the lack of meaningful
benefit of tight control on outcomes that matter
to patients. Yet guidelines continue to include
these targets, and do so inconsistently: targets in
recent Type 2 diabetes guidelines internationally
vary between <6.5% (<47.5 mmol/mol) and 8%
(<64 mmol/mol). Adhering to the targets speci-
fied in many guidelines for diabetes would kill
more patients than were helped. Forcing HbA1c
low also increases the risk of the patient suffering
hypoglycaemia, which does have an association
VOLUME 4 • NUMBER 2 • JUNE 2012 JOURNAL OF PRIMARY HEALTH CARE 159
BACK TO BACK
with a long-term outcome that matters to patients:
dementia. Similarly, large numbers of patients
take aspirin and statins for primary prevention
of cardiovascular disease, but recent research in-
dicates that the risk–benefit ratio is not generally
favourable for aspirin use in primary prevention,
in particular in older patients, as well as there
being increasing concerns about adverse effects.
However, reversing guidelines is like reversing an
ocean liner. Many guidelines are years out of date.
The quality of the evidence
The scientific base for guidelines has multiple
weaknesses. Research is increasingly com-
mercially constructed in a way that is likely to
obscure the real effects of treatments. Half of
efficacy and two-thirds of harm outcomes are
incompletely reported.2 Outcomes are biased in
favour of the funding company’s drug. Papers are
often ghost-written, trial data are not available
for public scrutiny, and publication decisions
are commercial ones. Sixty percent of clinical
trials remain unpublished and less than a third
of published meta-analyses obtain the individual
unpublished data.3–5 The ineffective efforts of
a Cochrane group to obtain Tamiflu data is just
One study redid 42 meta-analyses, adding in
unpublished trial data submitted by companies to
the FDA. In the re-analysis, 46% showed lower
efficacy of the drug, 7% showed identical efficacy,
and 46% showed greater efficacy. The re-analysis
showed more harm from the drug after inclusion
of the unpublished trial data.
Virtually all ‘evidence’ is generated in populations
highly selected to show maximum efficacy and
minimum harms. Populations with co morbidities
using multiple medications are usually excluded
from the clinical trials, yet these are the popu-
lations in which we use them—increasing the
potential for harm and reducing the potential for
benefit. Adherence to guidelines based on this sea
of uncertain evidence cannot be justified.
The misfit with primary care
The most compelling argument against adher-
ence is the mismatch between the partialist-
A 70-year-old woman with three chronic
diseases and two risk factors, if guidelines were
followed, would be prescribed 19 different
doses of 12 different medicines at five differ-
ent times of day. More importantly, there are
10 possibilities for significant drug interac-
tions, either with other medicines or with other
diseases. This prescriber would be rated as a
good physician using single-disease measures,
whereas the physician using wisdom and judg-
ment in avoiding polypharmacy would be rated
low on adherence. (Boyd JAMA 2006)
driven framework of guidelines and the general-
ist approach of primary care. One focuses on a
single disease in many people, while the other
focuses on a single person with many prob-
lems, including their particular set of values
and priorities. The idea of guideline adherence
cuts across the demonstrated benefit of patient-
centred primary care. The attention paid in
guidelines to the values of patient centredness
espoused in evidence-based medicine is sparse
and they provide inadequate quantitative data on
risks and benefits to support informed treatment
decisions.6 The consensus on risk thresholds
justifying treatment reflects physician values not
patients’. Most patients would not think taking
a statin justified at the kind of risk–benefit level
The idea of adherence to guidelines disempowers
doctors and patients in the use of their observa-
tion of individual response and needs. Even at
a simple pharmacological level, a single disease,
non-patient-centred approach is perilous (Box 1).
Guideline adherence is based on therapeutic posi-
tivism, undermining the skill involved and value
that should be placed on decisions not to give
treatments, as well as the improved health out-
comes. This is a specialist skill of primary care.
Adverse drug events are the fourth leading cause
of death in US hospitals. Polypharmacy is one of
the biggest threats to healthy old age. The quality
of primary care in coming decades is likely to be
160 VOLUME 4 • NUMBER 2 • JUNE 2012 JOURNAL OF PRIMARY HEALTH CARE
defined not by what we do give, but by how well
we make decisions not to give treatments.
Opportunities lost and
Procrustes was a figure in Greek mythology who
had an iron bed in which he invited travellers to
spend the night. If the guest was too short he
would stretch them to fit. If they were too tall,
Procrustes would amputate the excess length.
Adherence to guidelines is a procrustean ap-
proach to good quality care. Variation in care does
not necessarily mean poor care. It may represent
good care in a complex context.
Single-disease guidelines have had their day.
They are not fit for purpose for primary care and
adherence can be outright harmful, as well as
in surrogate process measures and intermediate
indicators, but in one analysis of five of the seven
studies that looked at actual patient outcomes
that matter, adherence to guidelines made
absolutely no difference.10 There is much better
evidence about the benefits of strong primary
care on health outcomes. We need to trust the
evidence of our own eyes as our patients beta
test treatments for the first time in the real
world of multi-morbidity, unknown adverse ef-
fects and individual preferences. The combined
knowledge of, and research by, family doctors
and their patients in understanding the effects
of these combinations of treatments, or of not
taking treatments, has a lot more to offer patient
outcomes than guideline adherence.
BACK TO BACK
It’s not possible to know when to follow and
not follow guidelines without a fundamental
understanding of the research data that does
(and does not) underpin them. The worry is
that critical reading and understanding of
research will be replaced by reading guidelines.
raising concerns about increasing inequity. They
are also not the best we have. There are sophis-
ticated and effective resources and programmes
available that offer the kind of critical appraisal
of the source research data that is of much more
value to practitioners than a flowchart—and
that make critical understanding of the evidence
fun.8,9 It’s not possible to know when to follow
and not follow guidelines without a fundamental
understanding of the research data that does (and
does not) underpin them. The worry is that criti-
cal reading and understanding of research will be
replaced by reading guidelines.
Theseus finally killed Procustes by making him
fit his own bed: studies of the effect of adher-
ence to guidelines usually only look at change
1. Tricoci P, Allen M, Kramer J, Califf R, smith sC, Jr. scientific
evidence underlying the ACC/AHA clinical practice guide-
lines. JAMA. 2009 2009;301(8):831–41.
2. Chan A, Hróbjartsson A, Haahr M, Gøtzsche P, Altman D.
Empirical evidence for selective reporting of outcomes in ran-
domized trials: comparison of protocols to published articles.
JAMA. 2004 May 26;291(20):2457–65.
3. Rising K, Bacchetti P, Bero L. Reporting bias in drug trials
submitted to the Food and Drug Administration: Review of
publication and presentation. PLos Med. 2008;5(11):e217.
4. Hart B, Lundh A, Bero L. Effect of reporting bias on meta-
analyses of drug trials: reanalysis of meta-analyses. BMJ. 2012
5. Ahmed I, sutton AJ, Riley RD. Assessment of publication bias,
selection bias, and unavailable data in meta-analyses using
individual participant data: a database survey. BMJ. 2012
6. McCormack JP, Loewen P. Adding value to clinical prac-
tice guidelines. Canadian Family Physician. 2007 August
7. Trewby PN, Reddy Av, Trewby Cs, Ashton vJ, Brennan G,
Inglis J. Are preventive drugs preventive enough? A study of
patients’ expectation of benefit from preventive drugs. Clin
8. McCormack J, Allen M. Therapeutics Education Collaboration
9. Richards D, Toop L, Graham P. Do clinical practice education
groups result in sustained change in GP prescribing? Fam
Pract. 2003 April 1;20(2):199–206.
10. Lawton RJ, Burton A. Clinical guidelines: A means to too many
ends? clinmed. 2000 August 21, 2000:2000070008.
Full reference list available on request