Artificial sciences are in-separable from human interests. One can speak of purpose of an event only because we are [ or someone is ] involved in the beginning, middle and end of the event. Natural events are usually understood as only purposeful in a theological sense and thus ‘science’ itself gave up its ‘faith’ in teleological or final causes since the time of Francis Bacon at least.
We can speak of a purpose or aim or plan only from the point of view of who we are [in case of individual action] or in the case of improvement science, what the practice is. Thus there is no exclusive theoretical interrogation of the current state. We can think of what we want to accomplish only from a reasonable understanding of the current state of the practice. A practice that has effectiveness of 80% [but doesn’t know it] cannot meaningfully claim to a target an output of 70% [ under the assumption that it is currently less than that]. Thus, it is impossible to do the right thing, if we don’t know the current state. The surprising possibility is that it is still possible [ actually it is often the case ] to ‘do things right’ i.e do the wrong things the right way. The PDSA cycle actually guards against such a practical predicament. The PDSA is in a sense a cognitive mechanism or heuristic to do more good than harm.
If you are starting out to improve something, then it is presumed you know where you are. ‘Mindless change’ is like trying to use a map without knowing your current location on the map. It is no good having a good aim without knowledge of your current state. Proper prediction always require greater information about premises than the conclusion. Such an understanding of current state is ‘practice knowledge’. No piece of know-how [practical knowledge] will result in improvement if it is not compared to current practice.
Any number of PDSA cycles, without clear understanding of the current state will be a turtle on its back. A Heraclitan understanding of practice does not deny the possibility of learning. The turtle on its feet, however slow, is wiser than its frantic cycling on its back. The PDSA requires knowledge of the [particular] practice as much as the practice requires PDSA. All the same, knowledge of practice alone will not result in improvement.
The scientific concept [ as opposed to marketing concept ] of improvement can be modeled thus: PRACTICE SCIENCE MODEL.
Artificial science is limited to what is within human experience. Natural science tends to extend human perception and understanding about the hidden secrets of nature. Improvement science is an artificial science and its method is largely concerned with the technique and experience of improvement. What I have posited in this essay is a requirement not to consider the PDSA or for the matter the scientific method as something that exists out there on the world independent of the human mind. It is very much a mental construct. [ it may not even be limited to humans: watch a money reach a banana out of its arm reach]. We are talking about turtle s and monkeys and an ethnology of improvement is not impossible in the near future .
Years ago Berwick posited a reference to the ontogenetic development of improvement in the learning behavior of children. Anyway my intention is not to offer a pylogenetic alternative but point out that in the realm of improvement the PDSA can be implicit and not merely an explicitly staged systematic procedure that we come to be told through various models of improvement. Whether ontogenetic or phylogenetic, what both share is the possibility of ‘improvement thinking’ as an epistemic entity. This is something that is has not been acknowledged in the literature. The domains of epistemology and psychology in Deming’s ‘system of profound knowledge’ can be combined to explore future evolution of the cognitive science of improvement. Improvement thinking need not be limited to our intention to improve.
The model in its outer loop [learning loop] explains the generation of ‘Improvement based evidence’ or the growth of practice knowledge. Improvement is not the ‘growth of knowledge’ but specifically the growth of practice knowledge or growth of know- how. Improvement science is the growth of improvement knowledge. It is the science of practice.
Research is legitimately the growth of knowledge. There is no distinction between research and improvement on this count but only regarding the object of study. While research may encompass all knowledge with varying degree of efficiency and cost, Improvement is limited to the study of ‘know-how’. It has no locus-standi in the growth of ‘ know –that’.
‘Research based evidence’ is Know-that. Research is growth from current state of knowledge. ‘Practice based evidence’ is Know- how. Improvement is growth from current state of practice. The experience and learning loops generate ‘improvement based evidence’. This second- order ‘knowledge’ about know- how, is improvement science.
Thus evidence is of three varieties: Research based, Practice based and improvement based. The neglect of practice based evidence and the promotion of improvement as ‘applied science’ instead of as a practice science, leads to confrontation with research. Applied science is the practice of theory. Practice science is a theory of practice.
Both, Berwick and Batalden treat improvement science differently. IHI calls improvement science as applied science. Berwick would explicitly use the phrase practical [ meaning applied ] improvement science instead of the more accurate “practice improvement science”. This error is largely the result of mis-nomination of Deming’s system of profound knowledge as the ‘science of improvement’ by Nolan et al . This is a historical anomaly as improvement had to be adapted from outside. There is no theory-practice distinction in Deming’s writing.
Batalden himself swerves close to this anomalous characterization in his model of improvement when he includes ‘ generalized ‘ knowledge with ‘ particular’ knowledge . That is not a theory of improvement but of clinical practice as such. However, Batalden’s overall oeuvre promotes improvement as a ‘general theory of work’.
The above Practice Science Model is not perturbed by evidence based medicine. The model treats ‘ research based evidence’ as ‘ improvement’. But limits the idea of ‘application’ to the experiential loop with use of an implicit PDSA. However, Improvement is far more ‘grounded’ in practice than the notion of research utilization. There is a certain idea of perfection that is not synonymous with ‘updated’ knowledge and its application. In the implementation or application model there is learning and experience but not necessarily learning from experience.
There is a story told by the enlightenment philosopher Immanuel Kant: He tells us about a physician who makes daily visits to his patient. Each day his patient asks him: how am I doing doctor. The physician gives the same answer daily : You are improving. This continues only till the patient looses his ‘patience’ and on the fourth day when given the same encouraging reply says : doctor please don’t kill me with improvement.
Kant like his predecessors Plato and Descartes is a standard bearer of an epistemology that seeks certainty in Knowledge, yet his foundationalism [if at all one could call him that without offending Kant scholars, though he titled some of his non ‘critical’ works Grundlegung] is on grounding the objective in the subjective. Karl Popper the philosopher of science [ the favorite, at least of scientists] calls his demarcation project a Kantian project. While improvement science may be weary of explicit foundationalism or have difficulty in explicating profound knowledge’s foundational relationship [though Deming unjustifiably calls it so and Dr Berwick would fret about the ‘turtles all the way down’ analogy from Indian cosmology] , Popper offers a lifeboat: his critical realism that provides ‘Truth’ only a regulative role in science. Unlike Kuhn he separates criticism[critical rationality] from justification. This allows Popper a non- dogmatic version of ‘normal science’ that is based on ‘continual revolution’. Popper is anti-foundationalist but retains a ‘base’ in a pragmatic sense as ‘piers’ going down into a swamp[uncertain base] just enough to stabilize the edifice’.
Regardless of the patient’s lament in Kant’s story , healthcare is a stochastic endeavor that can offer only ‘improvement’ as a pragmatic foundation. While ‘improvers’ move forward in our attempt to ‘found’ healthcare on ‘improvement’ we would still need to retain an ‘oslerian’ equanimity or a Stoic sagacity to answer our patients.
The idea of improvement in relation to ‘know-that’ is rudders without the boat. Research based evidence without practice based evidence is boat without rudders. Improvement based evidence that is combined with research based evidence and practice based evidence is boat with rudders. Our predicament is to repair the ship while at sea, a epistemic metaphor first offered by Otto Neurath for science in general. Practice knowledge is the only object of improvement science and thus requires a thick version of the metaphysics and epistemology of practice: Improvement science is a practice science. It is neither an implementation science nor an applied science.1