Psychiatry

The Psychopharmacology Workup Nobody Talks About in Training

Most residents learn the algorithms. SSRI first, then SNRI, then augmentation strategies. What they don't teach you is how to actually execute a medication workup for a patient who has already failed three trials without anyone properly figuring out why. I spent about four years doing outpatient psychiatry before moving into consultation roles, and the gap between textbook algorithms and real-world outcomes is where most people get stuck.

Here is what actually matters when you are trying to untangle a case that isn't responding. The first thing to check is whether the initial diagnosis was correct. This sounds obvious but it accounts for more treatment failure than anything else. A patient diagnosed with bipolar II who has been on SSRIs alone will cycle worse, not better. They present as "treatment-resistant depression" when they actually need mood stabilization. I had a patient last year who had been through four antidepressant trials over six years with zero response. Her PHQ-9 scores stayed in the high 20s. We pulled her old records and noticed she had two hospitalizations in her early twenties that were never properly evaluated. She was bipolar II, misdiagnosed as MDD. Started her on lamotrigine, tapered the sertraline. Her scores dropped to single digits within eight weeks. That is not uncommon. Maybe one in five "treatment-resistant" cases I see turns out to be a diagnostic issue, not a pharmacological one. Once you have the diagnosis right, the next bottleneck is usually comorbidity. Medical conditions that mimic or worsen psychiatric symptoms get missed constantly. Thyroid dysfunction, B12 deficiency, sleep apnea, hypogonadism in men. I run a standard lab panel before escalating psychotropics: TSH free and total, B12, folate, CMP, CBC, testosterone in males, and a sleep apnea screening questionnaire. If someone has untreated OSA and you are giving them mirtazapine, you are making things worse. Sedation, weight gain, metabolic issues - the whole package compounds the underlying problem. Then there is substance use. Alcohol is the big one. People will tell you they drink two glasses of wine on weekends. They are not going to volunteer that they are drinking a fifth of vodka daily until you ask directly and create space for an honest answer. I use CAGE-AID now instead of just CAGE. The "A" stands for amnesia and it catches people who black out regularly. Even if they don't acknowledge dependence, the amnesia component flags problematic use. Untreated substance use will undermine any psychotropic regimen within weeks. I have seen it repeatedly. Patients who seem to fail medications perfectly on paper are actually metabolizing them through induced CYP450 enzymes from chronic alcohol or cannabis use.

Pharmacokinetics matter more than people think. CYP2D6 poor metabolizers make up about 7-10% of Caucasian populations and 1-2% of Asian populations. If you prescribe escitalopram or venlafaxine to someone who is a poor metabolizer at standard doses, they will sit in side effects with no therapeutic benefit. A CYP2D6 genetic test costs around $150-200 through LabCorp or similar platforms and takes about two weeks for results. It changes your dosing strategy immediately. I do this for anyone who has failed two adequate trials. The return on investment is significant because it prevents the trial-and-error cycle that burns out both patients and prescribers. Therapeutic drug monitoring is another tool that gets underutilized. TDM for tricyclics and some second-generation antipsychotics is well established. Clozapine levels, for example, should generally stay above 350 ng/mL for therapeutic response. Below that threshold, you are likely not getting the full benefit regardless of the dose you think you are giving. I check clozapine levels every three months on stable patients and after any dose change. It takes the guesswork out of augmentation decisions. Same with valproate levels - knowing the actual serum concentration tells you whether a "failure" is really just subtherapeutic dosing. Psychiatric formulations like extended-release versus immediate-release matter clinically. Switching from immediate-release bupropion to extended-release can reduce the peak-related side effects like anxiety and insomnia without losing efficacy. It is a small adjustment but it explains why some patients who couldn't tolerate one formulation do fine on the other. I also pay attention to timing. SSRIs that are activating - fluoxetine, sertraline - go in the morning. Sedating ones like mirtazapine or trazodone go at night. Getting this wrong is a common reason patients report "it didn't work" when it was really just poor tolerability due to timing. I find patients will often self-adjust on their own and then present with vague complaints about the medication.

The biggest mistake I see in practice is premature switching. People switch antidepressants after four to six weeks when the full trial should be eight to twelve weeks at a therapeutic dose. That is a whole extra month of suffering for no reason. A proper adequate trial means the right dose for the right duration. Sertraline needs to be at least 100mg daily for eight weeks. Venlafaxine needs 150mg minimum for the same period. Escitalopram at 10mg is often insufficient - 20mg is the evidence-based therapeutic dose for many patients. I track this with a simple chart note template that flags inadequate trials so I don't repeat the mistake.

When Pharmacology Is Not Enough

There is a real limit to what medications can do, and pretending otherwise is dishonest. For treatment-resistant depression, the data supports augmentation strategies like adding aripiprazole or quetiapine, or switching to an MAOI like phenelzine. MAOIs have about a 50% response rate in true treatment-resistant cases where SSRIs and SNRIs have failed. They are underutilized because of dietary restrictions and drug interactions, but the efficacy data is solid. I use them selectively when other options are exhausted. The key is starting low and going slow with dietary counseling. Tyramine-rich foods like aged cheeses and cured meats need to be avoided, but the restriction is more manageable than most patients fear once they understand the actual risk threshold. ketamine and esketamine represent a different category entirely. Response rates around 60-70% for acute treatment resistance, but the effects are temporary without maintenance. Esketamine nasal spray requires clinic administration and monitoring for at least two hours post-dose. It is expensive - roughly $900-1,200 per treatment session before insurance. Insurance coverage varies wildly. I coordinate prior authorizations early because the administrative burden is real and can delay treatment by weeks. TMS is another option worth understanding. Repetitive transcranial magnetic stimulation has FDA approval for treatment-resistant depression and works best for patients who have failed at least one antidepressant trial. Response rates are around 50-60%, remission around 30%. The course is five sessions per week for six weeks. It is not a quick fix. Insurance usually covers it but prior authorization is standard. I find the best candidates are younger patients with a single prior medication failure and no psychotic features. Older patients with multiple failures and medical comorbidities respond less well.

ECT remains the most effective intervention for severe treatment-resistant depression, with response rates of 80% or higher. The cognitive side effects are real but usually transient. Memory disruption during the course typically resolves within weeks to months after treatment ends. I recommend ECT when a patient is at imminent risk - active suicidality, catatonia, psychosis, or inability to maintain nutrition and hydration. The stakes justify the tradeoffs. For anyone else, it should be a discussion, not a default. Psychotherapy combined with pharmacotherapy consistently outperforms either alone. This is not a feel-good observation - the data is robust across depression, anxiety, and PTSD. I prescribe medications for symptom reduction and refer for CBT or ACT for skill building and cognitive restructuring. The medication handles the biology, therapy handles the patterns. When I try to do it all alone, outcomes are worse. I have a referral list I maintain and update regularly. Not having one is a professional oversight. The monitoring framework matters too. I use PHQ-9 and GAD-7 at every visit for measurable tracking. Baseline scores, then every two to four weeks during active treatment. If there is less than a 25% reduction in score after six weeks of an adequate trial, I am not continuing the same approach. The numbers tell you when to pivot. I also ask patients to track side effects on a simple scale from one to ten. This gives me objective data instead of vague complaints like "I feel weird." "Weird" is not actionable. "My anxiety went from seven to four but my libido is a two" is.

What Psychiatry Cannot Fix

I want to be clear about the limitations. Medications do not solve poverty, trauma, isolation, or chronic stress. A patient on the perfect medication who is still working three jobs and sleeping four hours a night is not going to get better. Social determinants of health are not adjunctive - they are foundational. I assess housing stability, food security, and social support as part of every new patient intake. These are as important as any lab value or rating scale. Medication side effects are real and sometimes unacceptable. Weight gain from second-generation antipsychotics affects roughly 25-40% of patients significantly. Metabolic monitoring with fasting glucose and lipid panels every six months is standard but compliance is variable. Some patients simply cannot tolerate the side effect profile and need alternatives even if the alternative is less effective. That is a legitimate clinical decision, not a failure. Relapse is common. About 50% of patients who achieve remission from a first depressive episode will relapse within a year. After two episodes, the relapse rate jumps to 70-80%. Maintenance treatment is usually indicated after the second episode. I discuss this explicitly during the first remission so there are no surprises. Patients who discontinue medication abruptly after feeling better have the highest relapse rates. Tapering over three to six months reduces that risk substantially.

Prescribing psychiatry is not about finding the perfect medication. It is about systematic troubleshooting, accurate diagnosis, appropriate timelines, and knowing when to escalate or pivot. The algorithm is a starting point, not a destination. Most of the work is in the details - the history, the lab values, the medication levels, the timing, the comorbidities, and the willingness to revisit assumptions when something is not working.