Algorithms in Practice: Habitual Application of Defined Steps to Improve Care

Algorithms, ubiquitous in today’s increasingly technical culture, are sets of rules or processes to be followed in calculations or problem solving. Underlying the computer program and applications (“apps”) that we use daily, they are really nothing more than equations and step-by-step instructions that humans have always used to achieve goals.

As a clinician, what’s the conscious algorithm you utilize to provide the very best patient care?

Clinical psychopharmacology is best practiced rationally through a routine process of ongoing, thorough, and complete assessments of our knowledge base, our patients, and the treatment progress. This discipline is attached to conscious application of certain forms of logic in formulating and adjusting diagnoses and treatment plans until they produce outcomes desired by patient and practitioner.

An algorithm, a defined set of steps followed habitually, improves overall clinical practice. The elements of an effective algorithm, informed by the therapeutic alliance, are a practitioner’s database, continuous evaluation of a treatment plan, and clinical judgment guided by critical thinking.

Understanding and accepting uncertainty and probabilities, clinicians and their patients use randomized, controlled trials (RCTs) to help guide treatment selection.

The rational practice of medicine most often utilizes abductive reasoning—using incomplete information to form hypotheses and then testing these hypotheses. Bayesian logic, using prior assumptions corrected with outcome data, is an effective method for modeling and revising, as needed, diagnosis and treatment planning.

Many of us accede to the human tendency to not use these skills on a daily basis, to not really understand the study design or statistical analysis in journals, to read non-peer-reviewed articles, and even to fail to ask enough questions to understand our patients. We respond to hearsay, anecdotal evidence, and clinical gossip (partial information that takes on an invalid life of its own). Even in peer-reviewed journals, many studies suffer from inadequate design or analysis, so practitioners need to update and maintain an understanding of the statistical methods used in RCTs and meta-analyses to find the best data. Even the best intentions must overcome many unconscious and potentially conscious barriers to understanding and using only the best knowledge. We must take into account these biases that may distort our interpretation of data and cause us to reject important information that we need in our knowledge base.

In contemporary clinical medicine, it is also very easy to get carried away with molecular mechanisms and to use errors in inductive and deductive reasoning across scales to supply an invalid rationale for therapeutic choices. Clinical observation has always preceded the development of underlying theories, and RCTs allow us to use abductive reasoning to predict the results we might obtain from any pharmacological treatment.

The best treatment planning targets syndromes over individual symptoms, identifying the smallest number of accurate diagnoses possible that fit the data and fulfill DSM-5 criteria. Adding unnecessary medications to treat isolated symptoms may only confound results.

Commitment to a consciously chosen, externally validated, and standardized routine is the beginning of success. A thorough knowledge of medicine, the scientific literature, the history of pharmacological options, supplements that may or may not help, lifestyle issues (such as alcohol, smoking, caffeine use, sleep, diet, exercise), and knowledge of potential adverse events and their management are essential in creating and monitoring effective treatment plans.

Follow-up visits, often labeled med checks, may last as little as 15 minutes, but it is entirely possible to cover all the information necessary, make rational decisions about diagnosis and treatment plan changes, answer questions, and write prescriptions in this amount of time. Practitioners must know what information is necessary to obtain from patients and hone their skills for eliciting it. Regularly scheduled follow-up sessions with patients can be designed to efficiently recapitulate much of the initial evaluation, which will then be applied to refining the practitioner’s understanding of the clinical problem and best solutions.

Commitment can be made to return to best practices learned from others that may have been forgotten, devalued, or discarded during busy and hectic practice. Identifying weaknesses in practice style and improving them where psychopharmacology is involved is an effort that may not only “do no harm,” but may also be able to do some good.

Medicines will come and go, but a consciously and regularly utilized rational algorithm, applied alongside regular literature reviews, should keep a practitioner on the cutting edge for the length of a career. This enables the practitioner to help each patient to the greatest degree possible and contributes to professional satisfaction. ■