Coping with complexity: educating for capability
BMJ 2001; 323 doi: https://doi.org/10.1136/bmj.323.7316.799 (Published 06 October 2001) Cite this as: BMJ 2001;323:799
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Dear Sir,
I enjoyed the paper by Paul Plsek and Trish
Greenhalgh(1)
with its sweeping overview of the world of medical activity seen from
the perspective of complexity science, however I think the treatment of
non
linearity was potentially misleading.
All chaotic systems are non linear but not all non linear systems are
chaotic(2). The sensitivity to initial conditions of a system is a
hallmark
of chaos rather than non-linearity. Most deterministic relationships
in medicine are non linear. For example the dose-response curve for a
drug is
non linear but,generally,the effect is not chaotic.
Yours faithfully,
Gerald Partridge M.A, M.R.C.G.P.
Principal,
Holycroft Surgery,
Keighley BD21 1SA
gpartridge@bradford-ha.nhs
1 Plsek P.E,Greenhalgh T. BMJ 2001;323;625-8
2 Hilborn R.C."Chaos and Nonlinear Dynamics" O.U.P. 2000
NB. Competing interest: NONE
Competing interests: No competing interests
Recent articles about complexity theory applications are of great
interest to us as, like Plsek & Wilson, Fraser & Greenhalgh, we
believe complexity science has important benefits in the management of
organisations. Only in the last few years, however, have we had the
computing capacity and skills to build robust models of social
organisations. This enables us to simulate real events in real contexts
through a dynamic model, calibrated against real data.
However, not surprisingly, few social organisations use this kind of
modelling yet - it's very new science (a hard one, based on mathematical
and computational modelling), so there are few people with the skills and
experience to do it.
As these recent BMJ articles say, complexity science does have real
relevance for the NHS. But what's the nature of this relevance exactly?
The articles aren't 100% clear. Plsek & Wilson present the need to use
simple rules to run the system, yet give no real hint as to how to find
those simple rules, and how to differentiate between simple rules which
work and those which don't. This is exactly where mathematical and
computational modelling, firmly grounded in complexity theory, can provide
a precise, exact relevance through clear identification of what the simple
'rules' are that are creating seemingly intractable problems, and can make
valid, tested recommendations about what the new 'rules' are to mitigate,
or even occasionally eliminate the identified problem.
There are many successful examples (in many different fields) of
this: For example, the multi-agency Nasdaq wanted to know what would
happen when the SEC changed the rules of the game and 'decimalised the
markets'. The Bios Group, an consulting off-shoot of the reknowned Santa
Fe Institute, built a scientific model of the market and developed 6
precise predictions of what would change in the behaviour of the market as
a whole. Of these, 5 have come true.
The robustness of these intractable problems, called 'emergent
phenomena' in the literature, are a feature of all complex adaptive
systems. What Bios, the Santa Fe scientists and people like Brookings'
Axtell and Epstein do, is create a different, equally robust set of
outcomes, based on a tweaking or changing of the 'rules'.
But what rules should you be changing? Moreover, too, in the real
world, you might not even 'see' the problem, even though you're mightily
concerned about an aspect of the system (waiting lists, a week of a flu
crisis every year, for example!) Where does one start with some of the
hard problems the NHS is facing? Are managers indeed asking the correct
questions? No amount of brainstorming, ideas generation, hard work by
dedicated staff or collecting of data will tell you.
But models and simulations will let you design many new simple
policies, and test them over and over again within the model. Running lots
of 'what-if?' scenarios through a calibrated model allow you to identify
which 'rules' produce which outcomes; i.e. those 'rules' which you want
embedded in the system so x,y,z happen instead of a,b or c.
It's like this: before this kind of modelling were possible, who
would've thought you could reduce Southwest Airlines overnight cargo
transfers by 70%? (as described in the Harvard Business Review article
Swarm Intelligence last May) Once the model was built, (in this instance,
Bios scientists used agent based modelling), and explained the results,
someone could then come along and retrospectively kind of explain why the
simple rules put in worked. But there were equally simple rules there
beforehand which didn't work.
In conclusion, scientifically valid and replicable methods for
discovering 'simple rules' within a complex system do now exist. Their
application will back up what many NHS managers are talking about, and
provide them and others with scientifically robust examples for the
future.
Competing interest: Eurobios UK is a consulting company applying
complexity science to business and organisational problems -
www.eurobios.com
Competing interests: No competing interests
There were some interesting points in the series of articles on
complexity. The diabetes example from the second article in the series
(1) suggests the sort of results which the mathematics of chaotic systems
may be capable of when they are applied to healthcare. The point that
some things are not necessarily predictable even when they are described
quantitatively is a valuable one. Plsek and Wilson (2) develop this
further when they say that variation is inevitable in complex systems and
should therefore not always be considered to be a bad thing when we
encounter it in health: it is not a sign of failure which must be
eliminated. This is a timely reminder to be sceptical about a naïve
approach to variability, and becomes an important issue in health systems
with an emphasis upon performance indicators and league tables.
But some aspects of the series were less encouraging. There is a
persistently vague use of metaphor from the language of the physical
sciences. For example, ‘non-linear equation’ is a well defined concept
sometimes associated with complex properties such as sensitive dependence
upon initial conditions. But they don’t always go together (think of the
equation for a circle, for instance). Does ‘non-linear’ really apply to
story telling?(3) Sadly, I suspect that concepts from mathematics are
useful when applied mathematically, not as metaphors. The sheer use of
mathematical and physical terms does not alone lead us to knowledge: as
with any claim for insight the terms must be applied in a precise and
carefully reasoned fashion. Worse, such superficial use of metaphor can
distract from genuinely original ideas.
It is true that the mathematics of chaos is very generally
applicable. Even planetary motion, the paradigmatic example of a
predictable Newtonian system, has been found to be chaotic.(4) However
this example is a reminder that, like much else in physics, chaos is not
necessarily a direct antithesis to an earlier theory. A new way of
looking at phenomena doesn’t necessarily mean that everything we thought
before was wrong, which is why the Newtonian equations of gravity are
still good predictors of planetary motion for most practical purposes.
Chaos theory presents interesting new ways of understanding health
care, but we should not accept vague, unspecific metaphor and extravagant
claims that it holds all the answers in opposition to traditional
approaches.
I have no competing interests.
1. Wilson T, Holt T, Greenhalgh T. Complexity science: Complexity and
clinical care. BMJ 2001;323(7314):685-688.
2. Plsek PE, Wilson T. Complexity science: Complexity, leadership,
and management in healthcare organisations. BMJ 2001;323(7315):746-749.
3. Fraser SW, Greenhalgh T. Complexity science: Coping with
complexity: educating for capability. BMJ 2001;323(7316):799-803.
4. Lissauer JJ. Chaotic motion in the Solar System. Reviews of Modern
Physics 1999;71(3):835-845.
Competing interests: No competing interests
Dear Sir,
We have enjoyed your last four reviews on the subject of complexity in
medicine. We have also, though painful at times, found much of interest
throughout your series so far in the many references made to governmental
statements.
We also find, like the authors of your latest review, the Times Higher
Educational Supplement to be a great source of vernacular reportage. We
will endeavour, in our centre of evidence-based dermatology, to seek out
reference number twelve in this weeks review "Intuition and evidence-
uneasy bedfellows." Most important of all though, is that we would like to
express our incandescent excitement at the prospect of your fifth and
final review in next weeks BMJ where you will no doubt get to grips with
the evidence (reference 1) for such intuition. Thank you once again for
such insightful and scientifically well-argued post-modern irony.
Yours faithfully,
Dominic Smethurst.
1)Power laws: Are Hospital waiting lists self-regulating? Smethurst DP,
Williams HC. Nature 410: 652-3. Apr 4 2001.
Competing interests: No competing interests
A Different Approach
I have enjoyed this fresh look at the world of medicine (1) but I am
not sure that swapping the old rules for the mathematics of complexity
theory are right. Maybe the rules of complex systems are simpler and more
fundamental than we think.
Frijof Capra in his seminal work "The Tao of Physics" (2) identifies 6
parameters that should govern scientific
thinking:
1. Knowledge of the structure does not predict function
2. Process is primary and determines structure.
3. The observer is part of the whole system
4. There are no fundamental equations.
5. All descriptions are approximations.
6. Co-operation not dominance should prevail.
These paradigms can be applied to the world of medicine (2) and have
provided me with a different perspective for my clinical practice. I
suspect they are applicable to all other specialities. If they are
not,then Capra's paradigms are flawed!
I was sorry that Plsek, Fraser and Greenhaulgh (1) did not address
the issue of research in their paper though they indicated that they
planned to.
When I applied Capra's paradigms to the area of research into chronic
pain (3), I was able to understand why such research is so difficult to
undertake. Classical approaches to clinical trials (RCTs for
example)flounder when trying to assess the effects of drugs with complex
neuro-chemical effects in patients who's pains are a complex of
biological,
psychological, social and spiritual elements.
A visit to the leading pain journals will reveal the rarity of
classical clinical trials. Yet chronic pain afflicts about 1 person in 12.
In complexity lies the reason why it is so difficult to evaluate the
effects of Beta Interferon or Cannabinoids in Multiple Sclerosis. Current
mainstream clinical trial methodology does not overcome the problem of
complexity in patients. We need new approaches to the evaluation of
therapies that not only move away from the analytical reductionist
approach but also remain rigorous and acceptable.
Yours sincerely
References
1. Plsek PE and Greenhalgh T (2001) "Complexity science: The challenge
of complexity in health care" BMJ 323:625-628 to 799-803.
2. Capra F. Tao of Physics. London: Flamingo 1992, 3rd Edition
3. Notcutt WG. The Tao of Pain. Pain Reviews 1998;5 203-215
Competing Interests: None
Competing interests: No competing interests