Anda di halaman 1dari 6


Medical education school experiences. Trainees in graduate

................................................................................... medical education progress to acquire
competence, and expertise resides with

Finding safety in medical education the senior teaching faculty.9

The emergence of new knowledge,
however, drives constant and dynamic
D P Stevens reorientation of the teacher-learner rela-
................................................................................... tionship along the continuum. In the
traditional teacher-learner relationship,
The doctor-patient and teacher-learner relationships remain at senior (expert) doctors impart knowl-
edge to (novice and advanced beginner)
the core of the rapidly changing practice of medicine. Medical students and (increasingly competent)
education must embrace a safety culture if these relationships graduate trainees. That having been said,
are to serve patients well. most would concede that expertise in
information technology currently resides
“What gives value to travel is fear. It is the THE EVOLVING DOCTOR-PATIENT with the student and trainee, while the
fact that, at a certain moment, when we are AND TEACHER-LEARNER senior doctor is the novice. When it
so far from our own country . . . we are RELATIONSHIPS comes to patient safety, all learners along
seized by a vague fear, and an instinctive Consider the relentless transformation the continuum from medical student to
desire to go back to the protection of old that occurs in two essential bonds in teaching faculty are novices. In medi-
habits.” Albert Camus1 health care and medical education—the cine, rapidly accumulating new knowl-
doctor-patient and the teacher-learner edge increasingly merges the traditional
The doctor-patient relationship has been roles of teacher and learner.
at the core of medicine for centuries.
The traditional doctor-patient relation-
However, the last decade of the 20th IMPORTANCE OF A “SAFETY
ship is frequently amended by the im-
century has seen radical accelerating CULTURE” IN MEDICAL
perative for doctors to collaborate more
change in the context in which that rela- EDUCATION
effectively with each other and with other
tionship is embedded. It is increasingly By addressing the importance of estab-
health professionals. What has tradition-
complex and hurried2 and, disturbingly, lishing a “safety culture” in medical
ally been a “one to one” bond for the
it is fraught with substantial risk to the education, Aron and Headrick7 have
patient now may require effective integra-
patient.3 focused on a fundamental and strategic
tion with a “one to many” relationship.
Similarly, the teacher-learner relation- issue. In this regard there is much to
Another element that has dramatically
ship exists at the core of medical educa-
altered the interaction between doctors learn from the study of cultural realign-
tion, which must wrestle with the same
and patients is the Internet. There was a ment in other complex high risk organi-
accelerating change. While the focus on
time when medical knowledge was a zations. For example, in their efforts to
patient safety gains increasing attention
principal source of authority for the discern organizational models for safe
in the clinical setting, it is slow to gain
doctor.8 Now both the patient and doctor systems Weick and Sutcliffe11 have inves-
strategic awareness among medical edu-
have access to the same information. The tigated the culture of so called “high
cators. Were it of biological origin, the
discovery of an epidemic that results in doctor is now responsible for integrating reliability organizations” (HROs) such as
44 000–98 000 deaths annually in the US and customizing information for the ben- US Navy aircraft carriers and nuclear
alone3 would quickly find its way to the efit of the patient. While professional power plants, and have extended their
formal medical curriculum, for changes authority still plays an important role in observations to describe lessons for
in biology and technology (particularly how the doctor provides counsel, the healthcare systems.
of such magnitude) are readily incorpo- ubiquity of information brings about a It may be informative to extrapolate
rated by medical faculties. Patient safety substantial realignment. There exists little such lessons one step further to examples
pushes medical education into unfamil- formal education for dealing with these in systems for medical education. Weick
iar territory. Complex systems,4 culture,5 inevitable modifications of the doctor- and Sutcliffe11 point out that HROs adopt
and teamwork6 are not mainstream top- patient relationship. a culture that centers on mindfulness and
ics in the traditional curriculum. So, The teacher-learner relationship is constant attention to failures. An example for
where do we start? also evolving. Leach9 suggests that the medical education might be morbidity
In this issue of QSHC Aron and studies of Hubert and Stuart Dreyfus and mortality conferences that relent-
Headrick7 set out an excellent proposal provide a valuable insight into how doc- lessly explore the root causes of failure in
by offering a systems metaphor for tors learn. Working in the 1970s, the recent care events. HROs readily adapt
discerning safety in medical education. Dreyfus brothers studied how pilots organizational structure temporarily to meet
They argue persuasively that the “or- acquire knowledge and skills.10 They unusual situations—for example, the stu-
ganizational defences” of the medical described five progressive stages in the dent who teaches the teacher how to use
education system fail, and the result is continuum of learning: novice, advanced the most recent computer software for
inadequate education for doctors to pro- beginner, competent, proficient, and ex- electronic order entry. Finally, HROs are
vide safe care. Students and, impor- pert. The novice learns by careful appli- constantly mindful of the unexpected and
tantly, their future patients are at risk. cation of a defined set of rules. The exploit such events for organizational
The authors focus strategically on impor- advanced beginner demonstrates greater learning—for example, timely bedside
tant elements of medical education that skill by applying those rules to new teaching and learning that focus in depth
include entrance requirements, curricu- unforeseen situations. As learners ac- (perhaps even celebrate as a learning
lum, organizational culture, assessment, quire expertise, learning sheds rule opportunity) a trainee’s near miss in the
and accreditation. The list is daunting bound behavior and becomes more in- care of a patient.
but on target and highly strategic. All the tuitive. The Dreyfus model can be readily Weick and Sutcliffe11 refer to reliability
same, the strategy might benefit from applied to the continuum of medical as a “dynamic non-event”. Expressed
further reflection on deep seated trends education. The process begins when another way, when adverse events are
that envelop medical practice and, inevi- medical students move from novice to prevented, nothing happens. An edu-
tably, contemporary medical graduates. advanced beginner during their medical cational tradition that places greater

emphasis on disease treatment than on Aron and Headrick serve patients well REFERENCES
illness prevention is slow to reward doc- with their proposal that medical educa- 1 Camus A. Notebooks, 1935–1942. In:
tors and students when the “dynamic tors should radically rethink systems for Palmer P. The courage to teach. San
Francisco: Jossey-Bass, 1998: 39.
non-event” of patient safety happens. preparing future doctors.7 Improving 2 Ludmerer KM. Time to heal: American
Davidoff has suggested that one addi- health and health care begins with the medical education from the turn of the Century
tional cultural barrier to improvement in to the era of managed care. New York:
focus on improving medical education. Oxford University Press, 1999.
the healthcare system is shame because Strategic improvement—based on adop- 3 Institute of Medicine. To err is human:
“ . . . improvement means that, however tion of a systems approach, reflection on building a safer health care system.
good your performance has been, it is not Washington, DC: National Academy Press,
the realigned doctor-patient and teacher- 1999.
as good as it could be”.5 By extension, learner relationships, transformed cul- 4 Mohr JJ, Batalden PB. Improving safety on
educators who have devoted their ca- the front lines: the role of clinical
ture, and strong leadership—provides the
reers to educational systems that were microsystems. Qual Saf Health Care
appropriate start. The pace of change in 2002;11:45–50.
historically successful, but now are in-
medicine and health care insists on a 5 Davidoff F. Shame: the elephant in the room.
sufficient, must embrace the need for Qual Saf Health Care 2002;11:2–3.
valid improvement if knowledge for measure of urgency. Patients rightfully 6 Risser DT, Simon R, Rice MM. A structured
safety is to find its way to their students. trust the profession to educate doctors to teamwork system to reduce clinical errors.
incorporate such change into their care. Error reduction in health care. San Francisco:
Jossey-Bass, 1999: 235.
Finding safety in medical education can
“Improving health and provide reassuring confirmation of that
7 Aron DC, Headrick LA. Educating physicians
prepared to improve care and safety is no
health care begins with the trust. accident: it requires a systematic approach.
Qual Saf Health Care 2002;11:168–73.
focus on improving medical Qual Saf Health Care 2002;11:109–110 8 Starr P. The social transformation of American
medicine. New York: Basic Books, 1982.
education” 9 Leach DC. Competence is a habit. JAMA
Finally, effective and safe clinical sys- ..................... 10 Dreyfus SE, Dreyfus HL. A five stage model of
the mental activities involved in directed skills
tems require leaders who are relentlessly Correspondence to: Dr D P Stevens, Vice acquisition. Unpublished manuscript
committed to safety and reliability.4 11 This President, Medical School Standards and supported by US Air Force Office of Scientific
requirement may necessitate new criteria Assessment, Association of American Research under contract F49620-79-0063
for educational institutions in the recruit- with the University of California, Berkeley.
Medical Colleges, 2450 N Street NW, 11 Weick KE, Sutcliffe KM. Managing the
ment and promotion of their organiza- Washington, DC 20037, USA; unexpected. San Francisco: Jossey-Bass,
tional leadership. 2001.

Quality improvement research professionals, others on changing or-

................................................................................... ganisations or interactions between
parts of the system; some emphasise

Quality improvement research: self-regulation, others external control

and incentives; some advocate “bottom

understanding the science of change up” and others “top down” methods.
Despite their differences, however, each

in health care aims to contribute to better patient

care—and they might, but the evidence
for understanding their likely impact is
R Grol, R Baker, F Moss not robust and many seem based more
................................................................................... on belief than rigorous research of value,
efficacy, or feasibility.2 From what we
Essential for all who want to improve health care. know, no quality improvement pro-
gramme is superior and real sustainable
improvement might require implemen-
tation of some aspects of several

xpectations of healthcare services change so that care can be made safer
are ever increasing and those deliv- and better. approaches—perhaps together, perhaps
ering care no longer hold the mono- Everyone—authorities, policy makers, consecutively. We just do not know
poly of opinion on what constitutes good and professionals—seems to accept the which to use, when to use them, or what
or best care. To earn the label “good need for change. New initiatives aiming to expect.
enough”, care must meet standards to cure our ailing systems come in More evidence and understanding is
expected by consumers as well those of droves. This is an international phenom- required. At least 40 good systematic
enon. Many initiatives are linked to pro- reviews and numerous controlled trials
expert providers. Headlines in news-
grammes that capture a particular are available,3 4 but many of the trials can
papers, statements in policy documents,
approach—for example, evidence based be criticised because, for example,
and many analyses, surveys and reports medicine; accreditation and (external)
repeatedly highlight serious problems in randomisation or analysis was con-
accountability; total quality manage-
healthcare delivery related to underuse, ducted at the patient level while the
ment; professional development and
overuse, or misuse of care.1 Health revalidation; risk management and error intervention focused on professionals or
systems are sometimes unsafe and fre- prevention; organisational development teams, and outcome parameters are
quently we harm patients who have and leadership enhancement; disease often poorly chosen or are difficult to
trusted us with their care. There is an management and managed care; com- compare. Most studies were conducted
endemic failure to engage patients with plex adaptive systems; and patient em- in the USA, limiting generalisations to
decisions about their care. We know powerment. They may differ in perspec- other systems. Some strategies are better
there are problems; we just need to tive. Some focus on changing studied than others. We know more

about CME, audit and feedback, remind- There is a recognised process for the
Box 1 Some research
ers and computerised decision support development of new drugs, their intro-
approaches for quality duction into routine practice, and their
than about organisational, economic,
administrative and patient mediated improvement research establishment in the treatment of de-
interventions. New methods including fined conditions. As knowledge about a
• Observational studies of existing drug is accrued, new and better patterns
the effects of problem based education or
change processes of treatment gradually become estab-
portfolio learning, TQM, breakthrough
• In-depth qualitative studies on critical lished. Similar measured approaches are
projects, risk management methods,
success factors and barriers to needed to help develop and establish
business process redesign, leadership
change improvement programmes better, safer systems of care. “Change
enhancement, or sharing decisions with
• Systematic reviews of both the impact management” is a discipline central to
patients are not well studied. Studying
of different strategies and the influ- health care. The academic base that sup-
the effects of specific strategies in con-
ence of specific factors on change ports change management and quality
trolled trials will provide some answers
• Well designed cluster randomised improvement in health care should
to some questions about effective
trials underpin all clinical and managerial
change, but will not address some of the
• Systematic sampling and interpret- learning programmes. The science of
basic questions about the critical success
ation of experiences of change change management is not new, but
factors in change processes. They need to
• Methods for developing valid and there is a long way to go before we will
be complemented by observational and
sensitive indicators for measuring understand enough to be able to trans-
qualitative studies.
change form care so that it is “good enough” to
Health care is becoming increasingly
• Meta-analyses of large samples of meet everyone’s expectations of quality
complex and the problems are large. It is
improvement projects and safety.
unrealistic to expect that one specific
• Methods for evaluation of large scale
approach can solve everything. A quali- Qual Saf Health Care 2002;11:110–111
implementation and change pro-
tative study by Solberg et al5 of critical
grammes .....................
factors supporting implementation of
• Economic analyses of resources Authors’ affiliations
change showed that a mixture of profes-
needed for effective change and R Baker, R Grol, Guest Editors, Quality
sional and organisational factors is cru-
improvement of care Improvement Research Series
cial. “Give attention to many different F Moss, Editor in Chief, Quality and Safety in
• Statistical process control
factors and use multiple strategies” is Health Care
the message.6 Although we may know
Correspondence to: Professor R Grol, Centre for
that multifaceted strategies combining Research on Quality in Health Care (WOK),
different actions and measures linked to the intervention, small scale explanatory PO Box 9101, 6500 HB Nijmegen, The
specific obstacles to change are usually trial, followed by larger trials and re- Netherlands;
more successful than single search into long term implementation).
interventions,7 we know little about Clearly, different research methods are REFERENCES
1 Bodenheimer T. The American health care
which components of such complex required for different phases,9 but it is system. The movement for improved quality in
interventions are effective in different essential that, despite the eclectic base of health care. N Engl J Med 1999;340:488–
target groups. So, while there is some the research, researchers from different 92.
2 Grol R. Beliefs and evidence in changing
general knowledge, there is little detailed faculties and disciplines come together clinical practice. BMJ 1997;315:418–21.
understanding of the “black box” of to collaborate in this complex field and 3 Grol R. Improving the quality of medical care.
change. that the vogue for “quick fixes” is Building bridges among professional price,
We need to learn about change in the replaced with sustained research. payer profit, and patient satisfaction. JAMA
real world of health care and the crucial To stimulate and support debate about 4 Grimshaw JM, Shirran L, Thomas R, et al.
determinants of successful improve- research on quality improvement and Changing provider behavior: an overview of
ment. New thinking about healthcare change management in health care we systematic reviews of interventions. Med Care
2001;39(8 Suppl 2): II2–45.
settings as complex adaptive systems have commissioned a series of papers to 5 Solberg L, Brekke M, Fasio J, et al. Lessons
emphasises the importance of experi- provide an overview of some relevant from experienced guideline implementers:
menting with multiple approaches and methodologies. The first two papers are attend to many factors and use multiple
strategies. Jt Comm J Qual Improv
discovering what works best.8 Small published in this issue and more will fol- 2000;26:171–88.
changes can sometimes have large low. Pope et al10 explore some of the 6 Solberg L. Guideline implementations: what
effects—but we have little understand- qualitative methods that can be used to the literature doesn’t tell us. Jt Comm J Qual
Improv 2000;26:525–37.
ing about which small changes to use in gather information about the delivery of 7 Wensing M, van der Weijden T, Grol R.
which settings and their likely impact. good quality care, and Wensing and Implementing guidelines and innovations in
For real change and sustained im- Elwyn11 consider some of the key issues general practice: which interventions are
effective? Br J Gen Pract 1998;48:991–7.
provement a tailored research method- related to measurement of patients’ 8 Plsek PE, Greenhalgh T. The challenge of
ology is essential. The full range of views. Forthcoming issues of QSHC will compexity in health care. BMJ
methodology has yet to be established, include papers that describe research 2001;323:625–8.
9 Campbell M, Fitzpatrick R, Haines A, et al.
but will include contributions from methods for indicator development in Framework for design and evaluation of
epidemiology, behavioural sciences, edu- primary care; a methodology for evaluat- complex interventions to improve health. BMJ
cational research, organisational and ing small scale improvement projects; 2000;321:694–6.
10 Pope C, van Royen P, Baker R. Qualitative
management studies, economics, and methods for evaluating quality improve- methods in research on healthcare quality.
statistics (box 1). Theoretical models of ment programmes; research designs for Qual Saf Health Care 2002;11:148–52.
evaluations of complex interventions randomised controlled trials in quality 11 Wensing M, Elwyn G. Research on patients’
views on the evaluation and improvement of
propose a phased approach (theoretical improvement; and economic evaluations quality of care. Qual Saf Health Care
phase, definition of the components of of change management. 2002;11:153–7.

Patient safety medicinal products and medical de-

................................................................................... vices, and creation of a culture of
safety within healthcare organisa-

Championing patient safety: going tions.

• Development of mechanisms, through
global accreditation and other means, to rec-
ognise the characteristics of health-
care providers that offer a benchmark
Sir Liam Donaldson for excellence in patient safety inter-
................................................................................... nationally.
• Encouragement of research into pa-
A resolution on patient safety being considered by the World tient safety.
Health Assembly will move patient safety on to the world
health stage. In May 2002 the World Health Assem-
bly will consider the resolution and, if
accepted, the drive for safer health care

he early days of any programme to on to the world health stage. In January
will be a worldwide endeavour, seeking
improve the quality of health care 2002 the Executive Board of the World
to bring benefits to patients in countries
are always the most challenging, but Health Organisation (WHO)8 passed a
rich and poor, developed and developing,
potentially the most exciting. Raising resolution put forward by the govern-
in all corners of the globe.
awareness of the issues, inspiring front ments of the UK, Japan, Belgium, the
line staff to rise to the challenges, Islamic Republic of Iran, and Italy calling Qual Saf Health Care 2002;11:112
influencing policy makers to make a on the WHO to establish a programme
commitment and invest resources, per- on patient safety. The Executive Board is
suading a health service to accord it a the executive committee of the WHO,
priority, reassuring the doubters, picking consisting of one representative from Correspondence to: Sir Liam Donaldson, Chief
each of 32 of its 191 member states. It Medical Officer, Department of Health, Room
out the leaders—these are the chal- 111, Richmond House, 79 Whitehall, London
lenges. And so it has been with patient sits twice a year with the Director
safety. The problem of medical error had General of the WHO and her staff in
attendance to discuss policy and
been present since organised clinical
progress on major global health matters.
practice began, yet no one recognised it
Selectively, and with careful deliberation, REFERENCES
as a fundamental concept. Airline safety 1 Leape L. A systems analysis approach to
it identifies important new areas of work
had been systematically improved over medical error. J Eval Clin Pract
and makes recommendations to the 1997;3:213–22.
three decades by understanding and
World Health Assembly. 2 Reason J. Human error. Cambridge:
strengthening systems, yet few people
The resolution on patient safety set Cambridge University Press, 1990.
were struck by the parallels with health 3 Berwick DM, Leape LL. Reducing errors in
out four proposed areas for action:
care. Lives were being lost and people medicine. Qual Health Care 1999;8:145–6.
being made ill daily in hospitals around • Determination of global norms, stand- 4 Kohn LT, Corrigan JM, Donaldson MS, eds.
ards and guidelines for the definition, To err is human: building a safer health
the world, yet no pattern was recognised. system. Washington, DC: National Academy
Over a period of 5 years spanning the end measurement and reporting of ad- Press, 2000.
of the 20th century and the beginning of verse events and near misses in health 5 Department of Health. An organisation with
care and the provision of support to a memory. Report of an Expert Group on
the 21st, this position has been trans- Learning from Adverse Events in the NHS.
formed. countries in developing reporting sys- London: The Stationery Office, 2000.
Influential individual thinkers1–3 and tems, taking preventive action, and 6 Department of Health. Building a safer
implementing measures to reduce NHS for patients: implementing ‘An
major governmental reports4–6 have Organisation with a Memory’. London:
quantified the problem of medical error risks. Department of Health, 2001.
and its impact on patients, have de- • Promotion of framing of evidence- 7 US Department of Health and Human
Services Press Release. US and UK sign
scribed the context of risk and unsafe based policies including global stand- agreements to collaborate on health care
systems, and have scoped the action nec- ards that will improve patient care, quality and fighting bioterrorism, 10 October
essary to produce change. with particular emphasis on 2001 (
This year patient safety has begun to such aspects as product safety, safe 8 World Health Organisation. World Health
Organisation Executive Board Resolution
move from in-country programmes and clinical practice in compliance with EB109.R16, 18 January 2002
some between-country collaborations7 appropriate guidelines and safe use of (

Patient safety research studying the effects of action to improve

................................................................................... safety. Aggregated statistics are notori-
ously unreliable.1 2 Review of case notes

Patient safety research: does it have is a widely used method for measuring
error rates. However, it is subject to a
number of identified biases3 4 and the
legs? problem that sicker patients have more
opportunity for error. Enhanced tech-
R J Lilford nologies are being developed to measure
error, such as digitally imaging endo-
scopic surgery and installation of cam-
Patient safety research: where does it fit in? eras in operating theatre lights. The defi-
nition and unbiased measurement of
error will be discussed at a forthcoming

esearch into patient safety is highly Patient safety research has a role in: Anglo-American conference on method-
topical. The Agency for Health Care • Identifying the nature, extent, and ological issues in patient safety research.
Research and Quality spends about context of iatrogenic injury (including
£40M per year under this heading and errors of omission) DISCOVERING UNDERLYING
the UK has established a Patient Safety CAUSES
• Uncovering the factors antecedent to
Research Programme which I direct. Deeper understanding of the causes of
injury, especially the underlying be-
Patient safety research is somewhat error builds on extensive work in other
havioural causes
unusual in that it works back from effect industries such as air traffic control, the
to cause; while most research asks about • Developing and evaluating interven- nuclear industry, and others. Evolution-
the effects of structures and process on tions designed to reduce error ary selection did not equip the human
outcomes, patient safety research starts mind for the complex technologies
with the outcome—iatrogenic injury— All of these involve a wide range of which it went on to create, so we are now
and asks how it might be avoided. A research methodologies. prey to a disturbing range of psychologi-
research programme with an emphasis cal inadequacies.5 High risk industries
on safety is needed to determine how IDENTIFYING THE NATURE, reduce this problem by automation and
and why safety is undermined and hence EFFECT, AND CONTEXT OF close coupled systems. However, the con-
to develop and evaluate practices target- IATROGENIC INJURIES tinued presence of the human operator is
ing safety as their main objective. How- Enumerating and categorising error can required for those functions not easily
ever, many other research programmes be done by counting reports (reporting automated and to intervene when events
concerned with improving quality gener- systems, litigation records) or by investi- move outside system parameters. This
ally will impact on safety and a dialogue gating all cases of opportunity for error latter, particularly, is a task for which
with these programmes is essential. in an attempt to ascertain both numera- human cognition is supremely ill suited.
Similarly, managerial organisations with tor and denominator information and In these highly automated environ-
special responsibility for safety have hence measure incidence. Yesterday’s ments, where error is rare but cata-
come into being in many countries. Such research project can be today’s routine strophic, the human operator is the
organisations, which focus specifically data system, and many countries have “intelligent knowledge base” in the
on safety (such as the English National established standing mechanisms to so- system, yet it is precisely this knowledge
Patient Safety Authority (NPSA)), need licit, record, and act on reports of based problem solving which fails under
to mesh with other organisations (such untoward incidents. Such systems go stressful conditions leading, for exam-
as the Commission for Health Improve- beyond traditional reporting procedures ple, to an incident at a nuclear plant in
ment) responsible for quality generally. for drug reactions, device failures, trans- Ohio.6 On the other hand, well practised
Patient safety can be seen as a kind of fusion reactions, falls from bed, and nee- routine procedures which have become
knowledge management, continually dle stick injuries. The English pro- intuitive can also fail, for example, if the
learning, educating and motivating. Pa- gramme has commissioned research into operator’s attention is distracted—a fac-
tient safety programmes (whether re- factors (especially cultural factors) that tor identified as causal in some 6.5% of
search or managerial) have to be highly may affect willingness to report error surveyed incidents in nuclear power
connected to the organisations they seek ( plants.7 Nor do all errors originate at the
to influence and require a deep under- However, such denominator free data operator level: the literature is littered
standing, not only of scientific matters, underestimate many errors. This is im- with examples of failures attributable to
but of the policy environment in which portant when management action is organisational and cultural factors—
they work. predicated, not just on the existence of despite two similar incidents, manage-
Patient safety agencies and research problems, but on their incidence. Thus, it ment at Three Mile Island nuclear power
programmes have a special duty to does not matter in terms of policy if drug plant had done nothing to prevent its
reduce single acts which have serious calculation errors have been recurrence8; at Bhopal the plant superin-
consequences. Note that although the underestimated—there are far too many tendent was untrained for his job9; NASA
disaster can be traced directly to a single anyway and we need to act. But deciding top management cleared Challenger to
act, that act itself will have multiple whether to divert national resources to launch because they were unaware of a
antecedent “causes”. This, then, is where improve “pain to needle times” for launch constraint put in place by the
patient safety interventions get their patients with a heart attack or to reduce NASA booster project manager10; the
bite; they intervene in the chain of delay in operating on fractured neck of bosun of the Herald of Free Enterprise did
events where the probability of the unto- femur would require more accurate not close the bow doors because “it
ward event is the product of the prob- measurement of the scale of each prob- wasn’t part of his job”, even though ear-
abilities of (independent) antecedent lem. Unbiased measurement of error is lier he had relieved from duty the
events. This leads us into consideration also needed for comparative purposes— crewman responsible for doing so11; and
of the forms that research into patient for instance, when monitoring the per- so forth. Qualitative research has proved
safety might take. formance of healthcare providers or invaluable in helping to unravel the

complex social dynamics which deter- drugs), and education (e.g. simulations .....................
mine safety in health services,12 and to teach procedures). Culture seems to be
Correspondence to: Professor R J Lilford,
behavioural interventions have reduced improved by introducing specific meas- Director of the Patient Safety Research
accident rates in many industries.13 ures of this sort (which then have Programme, Department of Health, and
beneficial knock on effects) rather than Professor of Clinical Epidemiology, University of
Birmingham, Edgbaston, Birmingham B15 2TT,
DEVELOPING AND EVALUATING by non-specific exhortation.13 14 However, UK;
INTERVENTIONS bringing about meaningful directed
Basic research into the antecedents of change requires resources and large scale REFERENCES
injury should lead to development of managerial action. A large challenge for 1 Jencks SF, Daley J, Draper D, et al.
interventions designed to reduce risk. patient safety research is to work with Interpreting hospital mortality data: the role of
clinical risk adjustment. JAMA
The potential effects can be modelled managers to introduce change around an 1988;260:3611–6.
from the epidemiology and the degree of evaluation framework (preferably in- 2 Dubois RW, Rogers WH, Moxley JH, et al.
volving before and after measurements Hospital inpatient mortality: is it a predictor of
confidence in the intervention. If the quality? N Engl J Med 1987;317:1674–80.
desired outcome is an almost inevitable in both control and intervention sites). 3 Hayward RA, Hofer TP. Estimating hospital
consequence of an intervention, then The role of patient safety research is to deaths due to medical errors. JAMA
agencies should simply act. For example, get the evidence about what is likely to 4 Neale G, Woloshynowych M, Vincent C.
a number of recent deaths in England work, then to proselytise for change Exploring the causes of adverse events in NHS
have followed the administration of based on that evidence and, above all, to hospital practice. J R Soc Med
undiluted potassium chloride. So, get the encourage managers to innovate in such 5 Reason J. Actions not as planned: the price
stuff off the wards and we will prevent a way that the whole world may learn. of automation. In: Underwood G, Stevens R,
these deaths. The effects of other inter- However, we should be wary of inadvert- eds. A speech of consciousness. London:
Academic Press, 1979.
ventions may be less certain and control- ently creating the problems we wish to 6 NUREG. Loss of Main and Auxilliary
led before and after studies may be avoid through an overzealous campaign: Feedwater Event at the Davis-Besse Plant on
educational interventions designed to June 9, 1985. NUREG-1154. Washington,
needed to provide really convincing DC:US Nuclear Regulatory Commission,
evidence. In that case, a proof of princi- promote road safety awareness among 1985.
ple study (analogous to a phase 1 drug school leavers have the consistent and 7 Rasmussen J. What can be learned from
human error reports? In: Duncan K,
trial) may be needed to refine the apparently perverse results of increasing Gruneberg M, Wallis D, eds Changes in
intervention. For example, I have put out road deaths (mediated by the unex- working life. London: Wiley, 1980.
a call for a study to determine the effects pected effect of prompting early acquisi- 8 Kemeny J. The need for change: the legacy
of TMI. Report of the President’s Commission
of various types of simulation and drill to tion of driving licences). on the Accident at Three Mile Island. New
improve the management of acute ob- Managers and policy makers beware; York: Pergamon, 1979.
stetric emergencies. A large trial ran- the road to hell is paved with good 9 Morehouse W, Subramaniam MA. The
Bhopal tragedy. New York: Council on
domising all the labour wards in the intentions. So, promulgate plausible International and Public Affairs, 1986.
country must await the results of these service delivery interventions, but first 10 Report of the Presidential Commission on
liaise with those who commission re- the Space Shuttle Challenger Accident.
initial studies. Washington DC: Government Printing
In the end, patient safety will be search so that epistemologically sound Agency, 1986.
enhanced by automating procedures prospective evaluations can be built in 11 Sheen, Mr Justice. MV Herald of Free
from the start. The Patient Safety Re- Enterprise. Report of Court No. 8074 Formal
that can be automated (e.g. interpret- Investigation. London: Department of
ation of heart rate traces in labour ward/ search Programme in England will work Transport, 1987.
automated dispensing of drugs), trap- closely with the NPSA and others to 12 Firth-Cozens J, Mowbray D. Leadership and
the quality of care. Qual Health Care
ping errors before they occur (e.g. online ensure this happens. 2001;10(suppl II):113–7.
reminders), reducing pitfalls at the inter- 13 Krause TR, Seymour KJ, Sloat KEM.
faces between care settings (e.g. by link- Long-term evaluation of a behaviour based
method for improving safety performance: a
ing hospital and community prescribing ACKNOWLEDGEMENT meta-analysis of 73 interrupted time-series
systems), improving the design of proce- The author thanks Rachel Anderson (Bir- replications. Safety Sci 1999;32:1–18.
dures and equipment (e.g. delivery sys- mingham, UK) and Paul Barach (Chicago, 14 Nasanen M, Saari J. The effects of positive
USA) for helpful advice on earlier drafts. feedback on housekeeping and accidents at a
tems which preclude inadvertent in- shipyard. J Occup Accidents
trathecal administration of neurotoxic Qual Saf Health Care 2002;11:113–114 1987;8:237–50.