Assessing, Hypothesizing, and Diagnosing
Building a systemic hypothesis
A hypothesis is a provisional, falsifiable, circular statement about how symptoms fit the system. Milan-style hypothesizing requires circularity (relational questions), neutrality (no member blamed or allied with prematurely), and curiosity. Good hypotheses are testable by the questions you ask next; poor ones simply restate the complaint.
Assessment toolkit: presenting problem and its history, genogram across three generations, developmental stage, sequence of interaction around the symptom, attempted solutions, family strengths, culture and context, substance use, medical contributors, and risk. Use circular and reflexive questions, not interrogation.
Diagnosis
Know DSM-5-TR structure and criteria at a working level: differential diagnosis, comorbidity, specifiers, duration and impairment thresholds, and the distinction between a mental disorder and a relational problem (V-code/Z-code conditions such as relational distress or parent-child relational problem). Screening for depression, anxiety, trauma, ADHD, psychosis, and substance use prevents the classic error of reading individual pathology as purely relational, or vice versa. Cultural formulation is required, not optional.
Traps
- Diagnosing from one member's report alone; validating a label the family has weaponized.
- Missing medical or substance-induced causes of behavior change.
- Proceeding to conjoint work without screening for intimate partner violence, which changes the entire treatment design.
- Recording a diagnosis for reimbursement that is not supported by documented criteria.
Rule of thumb
A hypothesis must be circular and falsifiable: "If X, then we should see Y in session." If nothing could disconfirm it, it is a bias, not a hypothesis.
Worked example: A 14-year-old labeled "oppositional." Before accepting the label, screen for depression, anxiety, ADHD, trauma, and bullying; map the escalation cycle; check sleep, school, and substances. The finding of untreated depression plus a coercive escalation pattern yields comorbidity, not blame.