Defining Improvement Science: a means to a beginning.

” Nothing new can ever be learned by analyzing definitions. Nevertheless, our existing beliefs can be set in order by the process and order is an essential element of intellectual economy , as of every other ‘

– Charles. S. Peirce.

The cumulativeness revolution in healthcare began two decades ago with the publication of Don Berwick’s article: continuous improvement as an ideal in healthcare. The cumulativeness revolution has two basic components: A. Effectiveness. B. Learning.

In 2003 , the IHI campaign sought effectiveness on an unprecedented scale. Its success marked a watershed in the evolution of global healthcare. Yet it did highlight a deficiency in healthcare which Don Berwick himself acknowledged as a future agenda: to learn systematically from the experiment. The issue becomes more acute if one were to focus on the experience of the international participants of the campaign. It would be hard to discern the extent of learning as there have been no publications from these participants in the experiment. It is quite likely that improvement occurred without learning. Traditional medicine had the problem of learning without improvement; is it possible that modern medicine has the problem of improvement without learning !

To highlight the need for greater scrutiny of this improvement- learning dyad, I propose a distinction between improvement and improvement science by defining the terms:

Improvement:cumulative effectiveness of systems change under un-certainty.How to improve work ?

Improvement Science: cumulative learning from systems change under uncertainty. How improvement works ?

Improvement science includes Improvement. While Improvement is optimization of effect , Improvement Science is maximization of learning from improvement. There is a distinction between Improvement and learning , not a dichotomy.

Improvement Theory: The assumptions underlying the science. Why Improvement works ?

Let me summarize the different emphasis between the first two entities:

Improvement science as the Study of improvement.

If Deming ’s theory of profound knowledge is the magna carta of the modern quality movement , it is time to revisit the theory in full, from a theory building point of view . Much of the achievements of the movement has be unequally harvested from the systems thinking and statistical thinking components of the theory. This paired dichotomy with its attendant preference for the systems – statistics pair has led to the mistaken point of view of systems of improvement ie as of being fixed and final in its methods, while it would benefit by being described as a process of improvement.It may be time to plough the epistemological and psychological domains as well. While there may be other definitions that may be forthcoming, as a means to a beginning we attempt to provide one that is based on the four components of Deming’s theory and the PDSA method. [Cumulative change= PDSA; Learning =Psychology + Epistemology or Cognitive Science; Systems Change= Systems Thinking; Uncertainity= Statistical Thinking ] .

Note : ‘cumulative effectiveness of systems change ‘ is systemic change. Thus one could collapse the definition as study of systemic change under uncertainity. Yet I wish to maintain that systemic change is only an objective , not a process. Further the word systemic refers to a completeness which in practice is not available to us. The word systemic is also a description of the whole, where as improvement is that of the parts and the relationship of the parts to the whole. The idea of continuity ie progress , action, collectivity and learning deserve theoretical attention and is well captured by the explicit use of the word cumulative . Thus if we define Improvement as ‘systemic effectiveness of change under uncertainty’ or as ‘ systemic effectiveness of continuous change under uncertainity ‘ to include temporality , it is still a focus on evaluation and not on ‘improvement’. My attempt to capture the essence may detract from intellectual economy, in which case the definition as study of purposeful change or study of the validity or effectiveness of change would be too broad. Similarly , the definition as study of continous improvement would be too general. Methodologically the tradition in medicine has been ‘comparative’, it is ‘improvement’ that marks the ‘cumulative revolution ‘ and I hope you will agree that this deserves to be celebrated.

While improvement and cumulative effectiveness are single shade concepts in that they concern themselves with successful change, improvement science is more secular and allows for learning from failure. ‘The profound change in the idea of failure: It is no longer analysed competitively but therapeutically’. The focus on the psychology and epistemology of improvement could even  accommodate the long cherished ideal of clinical medicine to learn from each patient, which clearly cries out for attention in the current population and group based approaches to quality.

The distinction between knowledge and method is a false and incomplete dichotomy. Improvement is more likely only if its focuses on knowledge based change. How ever what constitutes knowledge based change is again not a homogenous concept. Knowledge could be categorized as that within the system ie Practice Based Improvement / Improvement Science or what is currently more popular in the evidence based era from outside the system ie . Evidence Based Improvement/ Implementation Science . Berwick’s call to broaden the evidence precisely focuses on this variety. It is important to note that his arguments in no way supports a postmodern or relativistic point of view. He argues for pluralism not relativism. This however calls for much more modest versions of epistemic values such as truth and objectivity: realism. Critics of this view  mistakenly insist that progress through ‘improvement’ was in making the methods more rigorous. yet this may be unnecessary. Their suggestion is unlikely to be productive as evidence is a theoretical concept and can afford greater refinement, while practice is always more complicated. The case of clinical decision making and the dis-utility of decision analysis in daily practice is an good example. The real issue is about fitting the tools to the object understudy: parsimony and prudence are inherent to improvement. However, this should not be considered as a call for an atheoretical method. On the other hand ,the need and justification for epistemology in  improvement theory is to rule out any ‘substance skepticism’ about practice knowledge as a precondition to successful change and not ‘methodological skepticism’ about EBM.

In contemporary discussion , several taxonomies have used learning as an element eg. Judgment/ accountability vs learning; selection vs learning. But the essence of learning has not been thoroughly explored. The emphasis on systemic learning would get us closer to learning systems and knowledge management.

Knowing -Doing-Learning Continuum.

When doing is replaced with improvement, learning becomes easier ! Implementation Science is concerned with the knowing – doing gap; Improvement science is focused additionally on the doing -learning gap. Improvement science is ‘the structure of learning in practice’ .

To Be continued.

Keywords: Bounded Predictability, Cumulative Effectiveness, Cognitive Science, Knowing-Doing Gap, Doing-Learning Gap, Effectiveness, Learning, Theory of Profound knowledge, Improvement science , Implementation science, Cumulative Revolution.