Every Doctor Tests. Psychiatrists Ask.

A field-by-field comparison of how medicine diagnoses, monitors, and measures treatment success.

We have been building continuous biometric monitoring tools for therapists for the last year. That work required us to understand how mental health care measures outcomes compared to every other branch of medicine. What we found was striking enough that we think it is worth laying out plainly, without commentary, and letting the facts speak for themselves.

Here is what a diagnosis looks like across eight medical specialties, followed by what it looks like in psychiatry. Same format. Same questions. Just the answers.

What a Diagnosis Looks Like in Veterinary Medicine

Your veterinarian orders bloodwork (complete blood count, chemistry panel), urinalysis, and often imaging (X-ray or ultrasound) before diagnosing your dog or cat with most conditions. Biopsies confirm cancer. Cultures confirm infection. Genetic testing identifies breed-specific risks. Diagnosis is confirmed against published reference ranges for the species. Treatment response is measured with repeat labs, repeat imaging, or both. When treatment works, the animal is discharged or transitioned to maintenance with continued objective monitoring.

This is the standard of care for a golden retriever.

What a Diagnosis Looks Like in Cardiology

The cardiologist orders an electrocardiogram (ECG), echocardiogram, stress test, and blood markers like troponin and BNP. For suspected coronary disease, cardiac catheterization provides direct imaging of the arteries. Diagnosis is confirmed by measurable findings: ejection fraction below 40%, ST-segment elevation on ECG, troponin above a specific threshold. Treatment response is measured with repeat echocardiography, repeat stress testing, and increasingly with continuous wearable heart rhythm monitors that track every heartbeat 24 hours a day. When a cardiologist suspects atrial fibrillation, they do not ask the patient how their heart has been feeling. They look at the data.

What a Diagnosis Looks Like in Endocrinology

For diabetes, the endocrinologist orders fasting blood glucose, HbA1c, and sometimes an oral glucose tolerance test. Diagnosis is confirmed at specific numeric thresholds: HbA1c at or above 6.5%, fasting glucose at or above 126 mg/dL. There is no subjective judgment involved. The number is the diagnosis. Treatment response is measured continuously. The Dexcom G7 and Abbott FreeStyle Libre provide a glucose reading every five minutes, 24 hours a day, 365 days a year. Patients and their doctors can see exactly what happened at 3 AM last Tuesday. The continuous glucose monitor did not just improve diabetes management. It fundamentally changed how patients and doctors understand the condition. Patterns that were invisible in periodic spot-checks became obvious in continuous data.

What a Diagnosis Looks Like in Oncology

The oncologist orders imaging (CT, MRI, or PET scan), followed by biopsy with histopathological analysis. Genomic sequencing identifies the specific mutation driving the cancer. Tumor markers in the blood (PSA, CA-125, CEA, AFP) provide additional data. Diagnosis is confirmed by tissue analysis under a microscope. There is no ambiguity. The cells are either cancerous or they are not. Treatment response is measured with repeat imaging on defined schedules, repeat tumor marker blood draws, and increasingly with liquid biopsies that detect circulating tumor DNA. When treatment works, the objective markers improve. When it stops working, the markers show it before the patient feels anything.

What a Diagnosis Looks Like in Infectious Disease

The infectious disease specialist orders blood cultures, PCR testing, or antigen and antibody tests. Diagnosis is confirmed by identifying the specific pathogen: the bacterium, virus, fungus, or parasite causing the illness. Treatment is targeted at that specific organism. Treatment response is measured by repeat testing until the pathogen is no longer detectable. The patient is then discharged. The condition is resolved. The word for this is "cured."

What a Diagnosis Looks Like in Orthopedics

The orthopedist orders X-rays, MRI, or CT scan. Diagnosis is confirmed by imaging. A fracture is visible. A torn ligament is visible. Cartilage damage is visible. Treatment response is measured with repeat imaging. The fracture heals, confirmed by X-ray. The ligament repairs, confirmed by MRI. The patient is discharged. Physical therapy may continue, but the acute condition is resolved and that resolution is objectively verified.

What a Diagnosis Looks Like in Nephrology

The nephrologist orders glomerular filtration rate (GFR), serum creatinine, blood urea nitrogen, urinalysis, and sometimes kidney biopsy. Diagnosis is confirmed by lab values against defined thresholds: GFR below 60 mL/min indicates chronic kidney disease. Treatment response is measured with repeat lab work on regular schedules. The numbers go up or they go down. The trajectory is objective.

What a Diagnosis Looks Like in Psychiatry

Diagnostic tests ordered: none.

Imaging required for diagnosis: none.

Bloodwork required for diagnosis: none. Blood tests may be ordered to rule out thyroid dysfunction, vitamin deficiencies, or other medical conditions that mimic psychiatric symptoms, but no blood test confirms any psychiatric diagnosis.

Objective criteria for diagnostic confirmation: none. Diagnosis is based entirely on a clinical interview, patient self-report, and clinician judgment applied to the DSM-5 symptom checklists. Two psychiatrists evaluating the same patient can reach different diagnoses. This happens frequently enough that it has its own research literature.

Validated biomarkers for any psychiatric condition: zero. Not for major depressive disorder, not for bipolar disorder, not for generalized anxiety disorder, not for ADHD, not for PTSD, not for schizophrenia. Decades of neuroimaging research have identified group-level statistical associations but nothing sensitive or specific enough to diagnose a single individual patient. The largest neuroimaging consortium in the world (ENIGMA, over 6,500 participants) achieved classification accuracy between bipolar disorder and healthy controls of approximately 65%. That is 15 points above a coin flip.

How treatment response is measured: the clinician asks the patient how they are feeling. Standardized questionnaires (PHQ-9, GAD-7) may be administered periodically, but these are self-reported, retrospective, and subjective. No objective measurement of treatment response is part of standard care.

Expected treatment duration: indefinite for most conditions. Major depressive disorder, bipolar disorder, generalized anxiety disorder, ADHD, and schizophrenia are all classified as chronic conditions requiring ongoing medication management. Current clinical guidelines generally recommend continuing medication indefinitely after two or more depressive episodes.

Cure rate: no psychiatric condition has an expected cure as a standard treatment outcome. The clinical framework is symptom management and "remission," defined as a reduction in symptoms below a subjective threshold. Not resolution of the underlying condition. Not discharge. Not "you no longer have this."

Objective discharge criteria: no objective definition of "recovered" exists for any major psychiatric diagnosis.

The Numbers

These are publicly available figures from government sources, peer-reviewed research, and federal court documents.

Approximately 1 in 6 American adults currently takes psychiatric medication. That is roughly 55 million people.

Among long-term antidepressant users, average duration of use exceeds 5 years. Many patients remain on medication for 10, 15, or 20+ years.

The DSM-5 contains approximately 300 diagnostic categories. The number of those categories confirmable by any laboratory test, imaging study, or biomarker is zero.

A 2006 study by Cosgrove and Krimsky in Psychotherapy and Psychosomatics found that 56% of DSM-IV panel members had financial ties to the pharmaceutical industry. For the panels governing mood disorders and schizophrenia, two categories where medications are first-line treatment, 100% of panel members had industry financial ties.

A 2012 follow-up in PLoS Medicine found that 69% of DSM-5 task force members had industry ties, a 21% relative increase despite promises of reform.

A January 2024 BMJ study found nearly 60% of DSM-5-TR panel members had industry ties, collectively receiving $14.2 million in documented pharmaceutical company payments between 2016 and 2019.

Between 1996 and 2014, every major manufacturer of psychiatric medications paid federal fraud settlements for illegal marketing practices. The combined total exceeds $9.8 billion. GlaxoSmithKline paid $3 billion. Pfizer paid $2.3 billion. Johnson and Johnson paid $2.2 billion. Abbott Laboratories paid $1.5 billion. Eli Lilly paid $1.415 billion. AstraZeneca paid $520 million. Bristol-Myers Squibb paid $515 million. Forest Laboratories paid $313 million.

The total revenue generated by the psychiatric drugs involved in these settlements exceeds $100 billion. The penalties represent 2-5% of that revenue.

What Other Fields Would Call This

In evidence-based medicine, a treatment administered indefinitely without objective outcome measurement is called empirical therapy. It is typically a temporary bridge used until diagnostic confirmation is available. It is not intended as a permanent treatment model.

In engineering, a system that operates without measuring its output is called open-loop control. It is universally recognized as inferior to closed-loop control, where output is continuously measured and fed back to adjust input. No engineer would design a critical system without feedback. The mental health system operates in open loop.

In pharmacology, prescribing a drug and adjusting dosage based solely on subjective patient report is called trial and error.

The FDA requires objective endpoints for drug approval in virtually every therapeutic area. Oncology drugs must demonstrate tumor shrinkage or survival benefit. Diabetes drugs must reduce HbA1c. Cardiovascular drugs must show measurable outcomes. In psychiatry, the FDA accepts subjective symptom rating scales as primary efficacy endpoints for drug approval.

Something Interesting Has Been Happening Since 2022

In 2022, CMS created new CPT codes (98975 through 98981) specifically for Remote Therapeutic Monitoring. These codes established, for the first time, a federal reimbursement framework for continuously monitoring patient therapeutic response between clinic visits.

In 2024, CMS added CPT 98978 specifically for cognitive behavioral therapy device and software monitoring. This code directly validates software that tracks mood, anxiety, and therapy adherence through digital tools.

In 2026 (effective January 1), the CMS CY 2026 Physician Fee Schedule Final Rule expanded the framework further. Four new codes were added: 98984, 98985, 98979, and 98986. The previous requirement of 16 or more days of data collection and 20 or more minutes of provider management time was relaxed. Providers can now bill for shorter monitoring windows (2-15 days of data) and lower time thresholds (10-19 minutes of management). The conversion factor increased to $33.40, with reimbursement increases across existing RTM codes as well. CMS also proposed optional add-on codes for Advanced Primary Care Management that eliminate time-tracking requirements for behavioral health integration services.

Year over year, the federal government is building out the payment infrastructure for exactly the kind of objective, continuous, between-visit monitoring that has never been part of standard psychiatric care. For therapists interested in the practical details, we wrote a complete guide to RTM billing for mental health practices.

The billing codes exist. The wearable hardware exists. Over 100 million Americans wear a wearable device that tracks heart rate variability, sleep stages, activity levels, and circadian rhythm around the clock. AI systems can interpret this data and deliver clinical intelligence. The measurement gap between psychiatry and every other branch of medicine is, for the first time, technically closable.

The question is not whether the tools exist. The question is who builds the clinical integration layer that connects them.

These are the facts. Draw your own conclusions.