Rapid Responses to:

ANALYSIS: Alan Shiell, Penelope Hawe, and Lisa Gold Complex interventions or complex systems? Implications for health economic evaluation BMJ 2008; 336: 1281-1283 [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Measuring and reporting the process of inpatient care

Peter H Millard, Brendon Rae   (10 June 2008)

Dynamic vs static aspects of nonlinearities

John Urquhart, MD   (10 June 2008)

Measuring and reporting the process of inpatient care 10 June 2008
Peter H Millard, Emeritus Professor of Geriatrics, St. George's University of London 12, Cornwall Road, Cheam, Surrey, SM2 6DR, Brendon Rae Send response to journal: Re: Measuring and reporting the process of inpatient care	New methods of measuring and reporting the outcome of change in health and social care systems are needed, because the average stay of discharged patients masks the process of care. Not simply because the distribution of length of stay is skewed, but also because internal and external factors influence the outcome of inpatient care. Seeking to explain the outcome of change in a New Zealand acute medical service we mapped the pattern of daily bed occupancy and used this to show the changing pattern of bed closure and bed borrowing. We also used the percentile distribution of length of stay and destination discharge to explain and explore the outcome of fast-tracking the process of care. The data analysed included all admissions between 1992 and 2003. During these years there have been a series of bed crises with different explanations as to cause. Now, in 2008, there is a prevailing bed crisis which still needs to be resolved The first bed crisis followed bed closure. In 1996 fast-tracking acute medical care resolved that crisis. Percentile analysis showed that comprehensive change in medical staff behaviour established a new stable state in the process of care,in which geriatric medicine also played a part, as a secondary recovery and rehabilitation service. Whether this new style of behaviour service would have been sustainable is unknown as further bed closures and relocation of geriatric medical services to the teaching hospital site led to an ongoing bed crisis that still needs to be resolved. Also length of stay is increasing as the problem faced now concerns availability of services in the community. This is out with the control of the hospital staff. Full details of the study and methods are in the Australian Health Review 1. 1. Rae B, Busby W, Millard PH. Fast-tracking acute hospital care — from bed crisis to bed crisis. Australian Health Review 2007;31(1):50-62. Competing interests: None declared
Dynamic vs static aspects of nonlinearities 10 June 2008
John Urquhart, MD, Chief Scientist, AARDEX Ltd 975 Hamilton Ave, Palo Alto, CA 94301 USA Send response to journal: Re: Dynamic vs static aspects of nonlinearities	Shiell et al. use too simple a definition of non-linearity in their discussion of complex systems, which, as they say, 'behave in a non-linear fashion'. Their characterization of nonlinearity is 'change in outcome [that] is not proportional to change in input'. This definition concerns static properties only, neglecting dynamic aspects that may alter the very nature of the 'change in outcome'. Dynamic changes are typically driven by certain temporal patterns of input. Examples from therapeutics are: (a) the pro-arrhythmic effects of leading anti-arrhythmic drugs when doses are escalated too rapidly; (b) excessive hypotensive effects of peripheral vasodilators when doses are escalated too rapidly; (c) rebound hypersensitivity to catecholamines when widely-used beta adrenergic receptor antagonists are stopped abruptly; (d) the nullification of nifedipine's anti-hypertensive action when the rate of increase of its concentration in plasma is too high; (d) the loss of anti-depressant action of paroxetine within 48 hrs after a last-taken dose, in stark contrast to the fortnight required for this drug's antidepressant actions to develop fully after dosing starts; (e) the shift from agonism to antagonism of pituitary gonadotropin secretion when the temporal pattern of administration of gonadotropin-releasing hormone shifts from a short pulse repeated hourly to continuous infusion. In physiology, to cite but one example, the pupillary reflex to light is inverted in the eye's paradoxically constrictive response to a darkness 'flash', i.e.,a brief reduction then return of light intensity to its previous level. In the physical world, the phenomenon called 'stall' nullifies an airplane wing's aerodynamic lift when the angle of climb exceeds a theshold value. These and other dynamic non-linearities tend to lurk unrecognised within certain complex systems, to appear only during or shortly after their encounter with specific temporal patterns of input. Dynamic non-linearities change familiar rules of operation into a different set of rules of operation that can be confusingly unfamiliar and sometimes hazardous. References and a more full discussion of these matters can be found in: Chapter 13 of "Simulation for designing clinical trials". Eds: Ko HC, Duffull S. New York: Marcel Dekker, 2002, pp 245-69. ISBN 0-8247-0862-8 Competing interests: shareholder of AARDEX Ltd, which makes & sells electronic medication event monitors to compile drug dosing histories in ambulatory patients.