Critical Thinking

WCHS Conference 2006 : 

WCHS Conference 2006 Using Pattern Recognition and Critical Reasoning in Homeopathic Case Analysis Karen Allen, RC, CCH www.homeopathyvirtualclinic.com

Learning Objectives for You: : 

Learning Objectives for You: Understand how assessment and decision making occurs and recognize rational and irrational decision strategies Understand how pattern recognition techniques apply to homeopathic case analysis Discuss characteristics of effective differentials Identify simple differential techniques

Homeopathic Concepts : 

Homeopathic Concepts Potato-ness of a remedy Granularity of data: Boulder, Rock, Sand, Silt Bias of Keynotes Relationship of Materia Medica to Repertory

Pattern Recognition Basics : 

Pattern Recognition Basics “Like Cures Like” is dependent upon identification of similarity: a matching of patterns, so we must understand pattern recognition Each client has a pattern. We identify this through case taking and analysis Each remedy has a pattern. We identify this through provings and clinical experience.

We See Patterns Often : 

We See Patterns Often 3, 6, 9, 12, ___ ABC, CBA, DEF, FED, GHI, ______ 0XX0, 0X0X0, 0X00X0, ______ red, blue, green, blue, yellow, blue, pink, _______ Pattern matching is a simple task when patterns are rigidly definable without ambiguity

What Completes The Pattern (1) ? : 

What Completes The Pattern (1) ?

What Completes The Pattern (2)? : 

What Completes The Pattern (2)?

What Completes The Pattern (3) ? : 

What Completes The Pattern (3) ?

Complexity Matters : 

Complexity Matters These example patterns have some collection of elements that define them. The elements can be fixed points or predictable variations (repetition, alternation, periodicity, sequence). The relationships between elements such as proportion, interaction or causality can also define the pattern. As the complexity of the pattern increases, the challenge of matching it increases.

We Need to Understand: : 

We Need to Understand: How is a pattern defined? How is a pattern recognized? How much of the pattern must be present? How do we account for pattern variation?

We Are Using The Terms: : 

We Are Using The Terms: Pattern: a form, template or model, or an abstract set of rules that define an entity. A pattern allows the entity to be built, recreated or perceived. The built or perceived entity 'exhibits' the pattern. It can be named. Pattern Recognition: identifying the defining elements of a known pattern. More complex than pattern matching because the patterns are less well defined; test case data ambiguous.

Patterns Are Defined By: : 

Patterns Are Defined By: Rule: a regulation, law or guideline that is always true for a pattern. Within a pattern definition it represents an invariable fixed point. Good for chess, less useful for homeopathy. Example: This is how remedy states are learned - e.g. all Aconite provings and clinical experiences. A known case can serve as a template for a pattern. A definition can then be built up through the accumulation of many examples. (Not exact enough for chess.) Analogy and Reference: A description of something else that has the same features or character as the pattern being defined. Not useful at all in chess, but used in homeopathy to understand comparisons of remedies - Aristolochia = Sepia + Pulsatilla

Patterns Facilitate: : 

Patterns Facilitate: Recognition: we can identify those entities that exhibit the pattern. E.G. Pattern = stripes. We recognize American flag, zebra, tiger. Generalization: based on the defined pattern, we can form a hypothesis of any other entities we might find which exhibit the pattern. E.G. Pattern = ¾ timing = waltz. We recognize ¾ timing and identify that The Blue Danube is a waltz. Abstraction: We can separate the pattern from the entities that have it, and use the concept separate from the entity. E.G Pattern = potato-ness of remedies.

Formal Patterns : 

Formal Patterns Formal Patterns like Chess game: rule oriented Complex with known variation Fixed rules apply and are non-intersecting (no royal pawn or rookie king - each piece is only itself, never another piece) Complete pattern always present All data in each test case is useful (which piece, in which square) This requires pattern matching

Informal Patterns : 

Informal Patterns Informal Patterns are like homeopathy: referential Complex; unpredictable variation No specific rules, no fixed points Partial patterns always present Test case data usefulness is random; subject to interpretation / distortion Test case always contains multiple pattern fragments within it This requires pattern recognition

What is or is not Included in the Pattern? : 

What is or is not Included in the Pattern? Assume a pattern called 'Greet Friend' Involves recognizing a friend and saying 'Hi'. The important data is identity of friend: Friend can be dressed in anything; in any location; in frontal or profile view. These background details are not part of the relevant pattern. In a variant pattern called 'Greet Customer', The important data is the location of the person entering a store. Clothing and angle of view unimportant. In this case, the identity is also unimportant, but location is now important. Separation of pattern from background is crucial.

Chess vs. Remedy Examples : 

Chess vs. Remedy Examples White Queen checks Black King . Decide an action: There are many (finite) options Every option is known Value of each option can be compared No spurious data to distract decision-making Multiple reasonable solutions are equally acceptable Man falls down stairs; contusions. Decide an action: The remedy state may be acute or chronic; multiple therapy options Not every remedy option is known, and case presents only part of remedy There is much data unrelated to the presenting stimulus/response that is relevant for remedy selection Multiple reasonable solutions are tightly interwoven, but not equally successful

Remedy States are Complex and Ambiguous : 

Remedy States are Complex and Ambiguous “To understand is to perceive patterns.” Isaiah Berlin Inventor of Value-Pluralism: “The idea that there are several values which may be equally correct and fundamental, and yet in conflict with each other. An example of value-pluralism is the idea that the moral life of a nun is incompatible with that of a mother, yet there is no purely rational measure of which is preferable.” This complexity and ambiguity is similar to the difficulty of recognizing remedy patterns where the validity of one choice does not imply invalidity of another choice.

Remedy states are like this!!!!! : 

Remedy states are like this!!!!!

Slide 20: 

Remedy state is a definition of all entities that exhibit a representative pattern of symptoms Remedy states overlap with each other Same symptom in multiple remedies (Lyc and Chel: < R side) Same symptom, increasing intensity (Fear in Ars and Acon) Same miasmatic pattern in multiple remedies (Calc and Sulph) Remedy states contain remarkable variation Imagine a Platina shy little 4 year old with recurrent UTI Imagine a Platina 42 y.o. gay man with trigeminal neuralgia Imagine a Platina 17 y.o. teen boy with eating disorder

Remedy “Patterns” : 

Remedy “Patterns” Remedy patterns are defined on various levels Physical symptom details - modality, location, sensation Idiosyncratic presentation of SRPs Miasmatic affinity Themes of mental / emotional state Concomitant These complex and extensive patterns will never be seen in full in any one case. The pattern exhibits the full remedy state.

Pattern Recognition : 

Pattern Recognition Recognition is based either on prior knowledge of a pattern or statistical similarity to a pattern. A science within the field of machine learning: 'The act of taking in raw data and taking an action based on the category, class or state of the data' Many models used. Some Useful in Homeopathy: Bayesian Likelihood Ratios Occam's Razor Decision Trees

Bayes: Success Probability : 

Bayes: Success Probability Thomas Bayes (1701-1761) Bayesian probability: a measure of the degree of belief a person has in some proposition. Likelihood ratio: we make valid future predictions based on successful past experiences. This is why SRP symptoms are useful remedy state indicators.

Occam's Razor: Simplicity Wins : 

Occam's Razor: Simplicity Wins William of Ockham (1285-1349) Occam's Razor: Do not introduce any unnecessary entities in explanations. Stick with the simplest theory that adequately accounts for the data. Also known as Principle of Parsimony, KISS Useful in homeopathy: The best pattern is the one with the fewest unnecessary assumptions. Not Useful in homeopathy: Reductionism to a minimal subset of the pattern (think of the 3 legged stool analogy)

Decision Trees: Facilitating recognition with high variability : 

Decision Trees: Facilitating recognition with high variability 20 Questions is a Decision Tree Example Decision Trees allow mapping of test case data to conclusions about the case's state. Each branching point marks a symptom; each subsequent limb represents a presentation of that symptom. A leaf represents the remedy state matching the values of the symptoms all along the tree. This is a realistic simile of the way that we perceive remedies, and provides a decision process that can be articulated, reviewed and improved.

Remedy Recognition Tips : 

Remedy Recognition Tips Referential, not Rules: Fixed points are unrealistic expectations Characteristic features occur at levels of Boulder, Rock, Sand and Silt. Most of our data about remedy states is sand. Directionality: We are determining if the case exhibits the pattern, not if the pattern exhibits the case Multiplicity: Cases will contain indications for multiple remedies; differentiation must be rigorous and free of unnecessary assumptions

Pattern Variation : 

Pattern Variation How do we accommodate pattern variation? Expect a remarkable variability in presentation; “forewarned is forearmed”. Eliminate expectation of fixed points; It's not Chess Learn variation within remedy states. The variation exists just as much in the remedy state as it does in the client case presentations. In variable patterns it is crucial to stick with facts and avoid assumptions

Recognizing Partial Patterns : 

Recognizing Partial Patterns Recognition threshold: How much of the pattern must be present to be recognizable? Experience informs this. Some symptom patterns present more often allowing quicker recognition. How much of an elephant do we need to see in order to recognize that it is an elephant? (Fable of the elephant and the blind men)

Bias Prevents Recognition : 

Bias Prevents Recognition Recognition must be based on the match between the elements in the case and the known pattern of the remedy Expectation, bias, agenda, conjecture all interfere with valid pattern recognition. These problems often result from confusion about which data are “useful” and which are “irrelevant” Research consistently confirms that expectation influences outcome

Requirements of an acceptable solution: : 

Requirements of an acceptable solution: Consider bell curve of response. Similimum will act most profoundly, but several remedies will act. Find a reasonable remedy. Match enough elements to decrease uncertainty to an acceptable level; decision is made on preponderance of data. Is there enough of a match to feel confident this remedy will act? Prioritize individualizations: modalities, SRPs, sensations, most limiting. Focus on the pattern, not the background. Conflicts may arise when “boulders” match one remedy and “sand” matches another remedy. Some cases are indeterminate based on the data provided. Ask more questions. Get the full case.

Conclusions for Case Work : 

Conclusions for Case Work Remedy patterns are complex and extensive There is remarkable variability in case presentation Individual case pattern will always be a partial presentation of the full remedy pattern Expectation of or adherence to fixed points interferes with partial pattern recognition Pattern recognition requires thoughtful consideration, rather than conjecture

Critical Thinking : 

Critical Thinking Cognitive skills Interpretation: accurately perceive evidence Analysis: identify salient and relevant info Evaluation: asses information without bias Inference: draw reasonable conclusions Explanation: in coherent way present results Self-regulation: double check self

SIX STEPS TO EFFECTIVE CASE WORK : 

SIX STEPS TO EFFECTIVE CASE WORK IDENTIFY the problem: what are we looking at in this case? DEFINE the context: what evidence of limitation and pathology? ASSES choices: identify 3 - 4 likely remedies CHOOSE options: what is the best remedy choice? LIST reasons explicitly: exactly why are we making this choice? SELF CORRECT: looking again, what did we miss?

Constructs of Critical Reasoning : 

Constructs of Critical Reasoning Complete, mature judicial process Accurately perceives evidence Identifies salient and relevant considerations Thoughtfully analyses and evaluates possible options without bias Decision based on warranted, judicious, valid conclusions Justifies decision, explaining considerations and assumptions Fair-mindedly follows where the evidence leads Imagine Albert Einstein in problem solving mode

Constructs of Critical Thinking : 

Constructs of Critical Thinking Incomplete, immature judicial process Biased interpretation of data, evidence, statements Failure to identify or hasty dismissal of relevant considerations Ignore or superficially evaluate obvious alternatives Decision based on fallacious / irrelevant argument or unwarranted claims Does not justify results or articulate reasons for decision Regardless of evidence, maintains or defends preconceived viewpoint Evidences close-mindedness or hostility or apathy to reason Imagine Homer Simpson in problem solving mode

Human Thinking Intuitive and reflective : 

Human Thinking Intuitive and reflective Integrative functioning of two mutually supportive systems. Two systems of thinking are present in each of us and can act in parallel to process cognitively the matters over which we are deciding. Capable of monitoring and over riding each other. System 1: more intuitive, reactive, quick and holistic. System 2: more deliberative, reflective, computational, and rule governed.

System 1 Thinking : 

System 1 Thinking Relies on cognitive shortcuts, key situational characteristics, readily associated ideas and vivid memories. Used to arrive quickly and confidently at a judgment. Helpful in familiar situations when time is short and immediate action required.

System 2 Thinking : 

System 2 Thinking Useful for making judgments in unfamiliar situations and when we have more time to figure things out. Allows us to process abstract concepts, to deliberate, to plan ahead, to consider options carefully, to review and revise our work. Relies on well articulated reasons and more fully developed evidence. Reasoning based on what we have learned through careful analysis, evaluation, explanation and self correction. Values maturity of judgment, fair-mindedness, elimination of bias, and truth-seeking.

Why bother being a critical thinker? : 

Why bother being a critical thinker? Improving our critical thinking skills and reinforcing our disposition to use those skills is perhaps the best way to develop and refine system 2 reasoning.

Thinking shortcuts –we all do this : 

Thinking shortcuts –we all do this Humans decide in a way that preserves cognitive resource- use cognitive short cuts Availability: Recent information is more readily recalled Association: One word of keynote supports a decision Affect: Positive or negative emotional reaction anchors perception Satisficing: Done when a 'good enough' solution identified Temporizing: This choice is 'good enough' for now Dominance Structures: One good enough option diminishes others; the resulting investment in that option can motivate subconscious reframing or redefinition of the problem or options to support the 'to-be-chosen' option.

Effective Differential Characteristics : 

Effective Differential Characteristics Factual; opinions welcome when noted as such Relevant to the case, showing how the remedy applies Congruent with the prioritization and repertorization Use all levels of granularity of data Well referenced to MM and Repertory Free of bias, conjecture, situation, fiction, negative logic

7 Simple Differentiation Strategies : 

7 Simple Differentiation Strategies Use 2 levels (crucial sx’s then confirmatory) Reference the Repertory with MM as back up Fill in the gaps to confirm or decline a symptom Compare all of the symptoms, from boulders to silt Ignore MM remedy data that is not relevant to this case Be persuaded by preponderance of data, not by bias or expectation of remedy presentation Be diligent, thorough, procedural, and disciplined

The Proverbial Example: Elise : 

The Proverbial Example: Elise Handout: Review this case An effective differential An uninformed differential An unsubstantiated differential A biased differential A random differential A materia medica discussion differential