The perfect is the enemy of the good.
We have, we hope, covered a large but not unwieldy swath of territory of practical relevance for the everyday clinician trying to make pharmacological decisions informed by evidence. As illustrated throughout the preceding pages, the availability of empirical data to guide treatment decisions varies greatly within and across disorders. It probably matters more that clinicians know how to think empirically – that is, knowing when, where, and how to look up information pertinent to a given case – rather than try to tackle the impossible task of comprehensively knowing the ever-changing clinical trials database for all disorders. Wisdom equally involves recognizing when evidence is lacking, prompting reliance on opinion, extrapolation, and plausible rationales – but not conflating those guideposts with an empirical database.
It is time now to cull the principles we have tried to illustrate and summarize what we would consider to be basic maxims for practical psychopharmacology.
Change Only One Variable at a Time
Apart from making good general sense, this axiom also serves as an excellent starting point in any complex clinical case. Among its virtues, such a strategy imposes an element of reflective delay and deliberation, requiring both clinician and patient to resist their own limbically driven urges and temptations to make critical decisions with undue haste, or unwittingly inflict iatrogenic sources of confusion by making messy situations even messier. Changing one variable at a time also means allowing enough time to elapse in order to judge the effect of what one has just done before causing additional changes to the system. In any multivariate equation, holding all but one variable constant also can be one of the most powerful ways to leverage clarity out of complexity, letting ambiguities unfold and resolve without the added confusion of desperate and possibly capricious meddling.
However, there are times when changing only one variable at a time may not be feasible. In real-world practice, patients are seldom if ever treated under tightly controlled environments where all key moderators and mediators of outcome can be held constant so as to allow for more certain inferences about probable cause-and-effect relationships. Sometimes clinical urgency demands altering more than one pharmacological parameter at a time – as in the case of stopping a drug that has become inappropriate (such as an antidepressant during mania) or deleterious (such as a stimulant or MAOI in the setting of a hypertensive crisis). From a purely scientific perspective, it would be informative to stop an antidepressant in the setting of emerging suspected mania and simply observe whether signs of psychomotor activation spontaneously attenuate (suggesting a direct adverse drug effect, as opposed to a catalytic effect that should persist despite removal of the triggering agent); but, on a practical level, disaster could result from withholding an appropriate remedy for here-and-now severe symptoms. Therein lies the trade-off between gathering evidence and providing sensible and compassionate care. The pragmatic compromise often involves simply acknowledging to oneself and one’s patient that cause-and-effect relationships may be hard to know with confidence, yet one must persevere nevertheless.