Laboratory Medicine In Psychiatry And Behavioral Science

The Actual Workflow Nobody Tells You About

Before you write a single prescription, there is a whole mess of bloodwork that needs to happen first. It is not glamorous. It is not quick. But if you skip it, you are flying blind in a way that will come back to bite you later. The core panel usually includes a complete blood count with differential, comprehensive metabolic panel, thyroid-stimulating hormone, free T4, vitamin B12, folate, lipid panel, HbA1c or fasting glucose, and a urinalysis. That is eight to ten separate tests before you even consider what kind of mood stabilizer to start. Some clinics run this during intake. Some wait until after the first appointment because nobody wants to schedule two visits for one problem. Then there is the question of which lab network you use. Quest, LabCorp, hospital-affiliated, independent — they do not report the same reference ranges. I learned this the hard way when a patient came in with a borderline low TSH from a reference lab and I nearly started levothyroxine on the basis of it. The same patient two weeks later at a different lab showed perfectly normal TSH. The difference was in the assay platform, not the patient. Always check which lab ran the test and whether the reference interval matches what your institution uses.

The Role of Laboratory Medicine In Psychiatry And Behavioral Science

People outside the field think psychiatry is just talk and pills. What actually happens is a lot more technical than that. You are interpreting lab results that tell you whether a patient's psychosis might be secondary to a medical condition, whether their cognitive slowing is thyroid-driven or something else, whether a new medication is going to tank their liver enzymes before they even feel better. This is the practical engine behind every treatment decision, even if the patient never sees it. The reference ranges matter enormously here. A creatinine clearance of 58 means something very different for dosing vancomycin than it does for dosing lithium. The same number looks totally benign if you are reading it without context. That is why the work in this space is less about diagnosing from labs alone and more about filtering out the noise so you can see what is actually driving the psychiatric presentation. Pharmacogenetic testing is another area that gets talked about a lot and used incorrectly even more often. The CYP2D6 and CYP2C19 polymorphisms are real. They affect how people metabolize selective serotonin reuptake inhibitors, tricyclics, and some antipsychotics. But a poor metabolizer status for CYP2D6 does not mean you cannot prescribe a standard-dose SSRI. It means you need to watch more closely for side effects and consider a lower starting dose. The data are suggestive, not deterministic. I have seen people get scared off a perfectly effective medication because a genetic report used language that implied certainty where none exists.

Here is a specific edge case that took me three weeks to untangle. A patient presented with treatment-resistant depression and a normal metabolic panel, normal thyroid, normal B12, and negative toxicology. Everything looked clean. Then I noticed the ammonia level was slightly elevated at 58 micromoles per liter. The standard range went up to 47. We did not chase it initially because it was borderline. Three months later, on valproate for what we thought was bipolar depression, the patient developed acute encephalopathy. The ammonia was 180. The initial elevation was the signal. Valproate-induced hyperammonemic encephalopathy is a known entity but it is easy to miss when the baseline is only mildly abnormal. The workaround is simple in hindsight: check a baseline ammonia in anyone you plan to start on valproate, and recheck if they develop cognitive changes, vomiting, or lethargy. Not every patient gets dramatic symptoms. Some just get slower and more confused, which looks like their depression worsening.

What Gets Missed in Training

Copper and ceruloplasmin testing. We do not order it enough. Wilson's disease presents as psychiatric illness in young adults far more often than anyone admits. A 24-year-old with new-onset psychosis and mild abnormalities should get a ceruloplasmin checked before you commit to an antipsychotic course. The same goes for anyone under 30 with movement disorder features and psychiatric symptoms. I do not say this to scare you. I say it because you will miss it if you are not looking for it, and missing it means the patient stays on medication that is not helping while the underlying disease progresses. Another thing nobody emphasizes enough is the timing of lab draws. If you are checking thyroid function in a patient taking biotin supplements, the results will be wrong. High-dose biotin interferes with streptavidin-biotin immunoassays, which most labs still use. The TSH can read falsely low, the free T4 falsely high. A patient appears hyperthyroid when they are not. The fix is straightforward: stop biotin for at least three days before the draw. Tell the patient this at intake. They are probably already taking it for their hair and nails and do not expect you to care. Lithium monitoring deserves its own section because it is the simplest drug to manage and the easiest to mess up if you are careless. The therapeutic range is 0.6 to 1.2 milliequivalents per liter for acute mania and 0.6 to 0.8 for maintenance. You draw trough levels, twelve hours after the last dose. If the patient took their morning dose and you draw at noon, the level will be spuriously high and you may reduce the dose unnecessarily. That is a common error. I see it happen in residency and I see it happen now. Write the order so it is clear: lithium level, trough, 12 hours post-dose. Put it in the electronic health record with a note. The system will remind you if you do not specify.

Valproate monitoring is more complicated. You check the level, yes, but you also need to watch the ammonia as I mentioned, the platelet count, and the liver function tests. Valproate causes thrombocytopenia in about five to ten percent of patients. It is usually mild but it can be severe. The bleeding risk is real. I had a patient on valproate who presented with gum bleeding and easy bruising after six months of stable therapy. The platelet count was 45,000. We stopped the drug and it recovered in two weeks. No one predicted it from the labs. That is just how it goes.

Where the System Falls Apart

The biggest limitation in this field is that laboratory medicine in psychiatry and behavioral science is not standardized the way it should be. There is no universal consensus on which baseline tests every psychiatric patient needs before starting every class of medication. Different hospitals use different panels. Different states have different reporting requirements for controlled substance monitoring. The pharmacogenetic testing landscape is a commercial minefield with companies selling reports that overstate the clinical utility of findings that are, at best, probabilistic. Clozapine monitoring is a special case. The required absolute neutrophil count checks are strict. They are also expensive and inconvenient for the patient. The system works in the United States through the Clozapine Risk Evaluation and Mitigation Strategy registry. It works in Europe through national registers. It does not work well in places without centralized registries. Patients fall through the cracks. I have lost track of how many times a patient on stable clozapine for refractory schizophrenia could not get their ANC checked because the local lab did not participate in the registry or the patient moved to a state with different requirements. The medication was working. The logistics were not. Another honest limitation: most of the lab values we rely on were derived from general medical populations, not psychiatric ones. The reference ranges assume a healthy adult with normal organ function. Your patient population is often none of those things. Elderly patients with depression and mild renal impairment will have different baselines than the standard range suggests. Pediatric patients are even more complicated. There are very few pediatric reference intervals for many of the tests we run. We extrapolate and hope. Sometimes it works. Sometimes it does not.

Practical Takeaways

Order the full baseline panel before starting any psychotropic in a treatment-naive patient. Do not cherry-pick based on your clinical impression. You are not special. Your gut feeling about what is causing the depression is not a substitute for a metabolic panel. Know which lab network your patients use and what reference ranges they apply. If a patient brings outside labs, verify the methodology and the reference interval before acting on the numbers. Check biotin use explicitly. It is in multivitamins, hair supplements, and a hundred over-the-counter products. Patients do not think of it as a medication. You need to.

For lithium, always specify trough timing on the order. For valproate, check ammonia at baseline and if symptoms develop. For clozapine, the registry is mandatory and the neutrophil monitoring is non-negotiable. For any young patient with new psychosis, consider ceruloplasmin and copper studies. The work is tedious. It is also essential. The labs do not give you answers. They give you constraints. Within those constraints, you make the decisions that actually help the patient.