19.4 Clinical Data, PROs, and Treatment Efficacy

Beginning October 30, 2026, the FNP-BC Test Content Outline names a Domain V skill: analyze clinical data and patient-reported outcomes to assess treatment efficacy. The current outline already scores test-result evaluation, management, and reporting. Teach both so the guide survives the TCO date change. The exam item will hand you a stack — a lab, a vital-sign log, a PHQ-9, and a sentence about work — and ask whether the treatment worked.

Efficacy is not a single number crossing a line. It is whether the combination of physiologic data and the patient’s reported life says the plan is helping, hurting, or doing nothing — and whether anyone closed the loop.

Do not treat a number in isolation

An A1c of 6.4% looks like ADA-adjacent success until you notice weekly hypoglycemia, a recent motor-vehicle near-miss, or a frail 82-year-old on insulin who lives alone. That is not a win. Loosen the regimen. A BP of 118/70 is not success if the patient is dizzy, falling, or has an eGFR that fell from 58 to 31. A TSH of 0.1 on levothyroxine in an older adult is overtreatment even if they asked to “feel thinner.” A PHQ-9 of 4 after eight weeks of an SSRI is excellent unless the family describes new insomnia, racing thoughts, and spending sprees — that is possible hypomania, not remission.

Conversely, an A1c of 7.6% with no hypoglycemia, improving home glucoses, and a patient who just started metformin 4 weeks ago is on-track, not a failure (the A1c clock is about 3 months — Section 19.2). Judge the number against the clock, the goal you set with this person, and the harm signal.

A normal chest radiograph does not control asthma. A reassuring knee radiograph does not cancel a hot joint. A “reassuring” TSH does not explain new atrial fibrillation if the free T4 is high and you drew the labs too soon after a dose change.

Patient-reported outcomes you should actually use

A patient-reported outcome (PRO) is the patient’s account of symptoms or function, preferably with a validated instrument you can repeat. The October 30 skill is not “ask how they feel.” It is analyze the instrument with the labs.

PHQ-9. Nine items, score 0–27. Use it to stage severity, to track treatment, and to catch item 9 (death thoughts). A drop of about 5 points is a commonly taught reliable change; a score under 5 is the usual remission language. Repeat at follow-up — a single intake PHQ-9 is assessment, not evaluation. Always still ask suicide questions in the room; a circled 0 on item 9 is not a safety plan.

GAD-7. Seven items, score 0–21. Same logic for anxiety disorders and for SSRI, SNRI, or CBT follow-up.

ACT (Asthma Control Test). Five items, score 5–25. ≥20 generally indicates well-controlled asthma; 16–19 is not well controlled; ≤15 is very poorly controlled. Pair with exacerbations, oral-steroid bursts, and nighttime symptoms. An ACT of 12 after a “step-up” means the step-up did not happen, was not adhered to, or the diagnosis is wrong (cardiac, inducible laryngeal obstruction, eosinophilic, or occupational).

AUA Symptom Index. 0–35. Mild / moderate / severe bands guide whether watchful waiting, medication, or urology is matching the residual burden. A man whose AUA-SI improved but who now has overflow dribbling needs a post-void residual, not congratulations.

WOMAC and KOOS are the conceptual orthopedic pair: pain, stiffness, and function of the knee (or hip). You may not hand-score a 42-item KOOS in a 15-minute visit; you must still evaluate stairs, walking distance, and work, not only “pain is a 6.”

Numeric pain scales are legitimate PROs and are also how people get overtreated. Always pair pain with function and red flags.

Days of work or school missed, caregiver hours, and sleep are PROs that quality programs underuse. A child whose ACT is 21 but who missed nine school days is not controlled — re-check technique, allergens, and whether the parent is reporting a polite score. An adult whose PHQ-9 is 8 but who has not left the house in a month needs a closer look at function and safety.

Lifespan PRO examples

An infant’s “PRO” is the caregiver: wet diapers, feeds, and the Edinburgh or PHQ-9 on the parent. A preschooler may point to a faces scale; you still need play and gait. A school-age child should help complete an ACT or a pain-and-function sentence. An adolescent needs a confidential PRO (PHQ-A, ACT, a sexual-symptom question) the parent does not fill in. A frail adult may need a proxy for some instruments; write that you used a proxy and still examine the person.

Combine the stack — a worked method

When the item gives you more than one data point, write a one-line synthesis before you act:

1. Goal — what did this patient and I set (A1c <7.5, ACT ≥20, PHQ-9 <5, walk a block, sleep without a rescue inhaler)?
2. Clock — has enough time passed for this intervention?
3. Physiologic data — vitals, labs, imaging, device download.
4. PRO — instrument plus one function sentence.
5. Harm — hypoglycemia, AKI, suicidality, bleed, fall, C. difficile.
6. Implementation — was the treatment taken or attended?
7. Decision — continue, intensify, switch, stop, refer, or rescue.

Example. A 58-year-old on metformin and a low-dose GLP-1 receptor agonist. A1c 7.1% at 3 months (from 8.4%), no hypoglycemia, weight −4 kg, but she is still missing work for knee pain and a KOOS-function score is unchanged because PT was never authorized. Glycemic efficacy is acceptable. MSK efficacy is not evaluated yet — you still owe her the PT, not a third glucose drug.

Example. A 34-year-old with asthma. ACT 13, peak flow 55% personal best, albuterol daily, and a normal chest radiograph from last month. The radiograph does not prove control. Step up per GINA/NAEPP logic, check inhaler technique and adherence, and look for allergic or occupational drivers.

Example. A 70-year-old on tamsulosin. AUA-SI fell from 22 to 12, but he has a post-void residual of 350 mL and rising creatinine. The PRO improved; the clinical data say obstruction and possible chronic retention. Refer, do not refill and smile.

Closed loop: who saw it, what changed, when is the next look

Test-result evaluation on the current TCO is a system skill. A critical potassium of 6.4 at 4:50 p.m. is not “evaluated” because it sits in the EHR. Evaluation requires:

• Who received the result (named clinician, not “the pool”).
• When (time stamp).
• What changed (hold the ACE inhibitor and spironolactone, repeat, EKG, send to the ED).
• Whether the patient was notified.
• The next look (repeat K+ tomorrow; clinic in 3 days).

Critical values (the laboratory’s defined list — commonly a potassium in the mid-6s or higher, a glucose in the 30s, a hemoglobin incompatible with safe discharge, a positive blood culture, a markedly elevated troponin, an INR that is dangerously high) require same-day, often immediate action and documentation. Do not batch them with Friday routine lipids. Do not send a portal message and go home. Speak to the patient or a capable adult, or activate EMS if you cannot reach them and the value is immediately dangerous.

Non-critical but actionable results (new A1c 10.8%, positive FIT, TSH 18, biopsy “atypical,” a new eGFR of 28) need a timed plan, not an inbox that waits until the patient happens to call. The October 30 language (“analyze…to assess treatment efficacy”) still includes this plumbing: a result nobody acted on cannot demonstrate efficacy.

Tracking systems — overdue-result queues, cancer-screening registries, a named covering FNP on Fridays — are Evaluation infrastructure. A clinic that “usually catches those on Monday” is describing a known harm pathway.

Population evaluation and HEDIS-type measures

Quality measures (HEDIS and cousins) are how a panel is evaluated: A1c poor control, BP control, depression follow-up PHQ-9, asthma controller ratio, cancer-screening completion, 7-day follow-up after a mental-health admission. They are useful population mirrors and dangerous individual tyrants.

Use them to find who you never evaluated — the patient with no A1c in 18 months is an evaluation failure. Do not use them to force a frail nursing-home resident’s A1c to 6.2 or to order a colonoscopy the patient with metastatic cancer will not survive to benefit from. That is the justice problem from Section 19.1 in measurement clothing.

When a measure and a person conflict, document the individualized goal, the shared decision, and why the numerator should not drive harm. When a measure and a person agree, fix the gap: outreach, standing orders, and a closed-loop lab process are Domain V practice, not only “quality department” work. A 90th-percentile dashboard built by excluding no-shows is a fiction (Section 19.1).

Reporting that is part of evaluation

Some evaluated results leave the building. Reportable diseases (Chapter 18.3) are an evaluation output: the positive gonorrhea NAAT is not finished when you treat — you report per state rule and arrange partner management. A critical INR in a patient who is bleeding is both a clinical rescue and a systems review (was the interacting antibiotic the one you started last week?). Abnormal imaging that suggests cancer needs a tracked referral, not a hope. Positive blood cultures called after hours are critical values, not Monday problems.

Document the synthesis in language another FNP can follow: “Home BP average 138/84 after 4 weeks of chlorthalidone 25 mg; K+ 3.9; no dizziness; continues DASH with high-sodium workplace lunches. Plan: continue dose, RD referral, repeat home average and BMP in 4 weeks.” That paragraph is Domain V. “Labs reviewed, continue current meds” is not analysis.

If you remember one efficacy sentence: combine the number with the PRO and the harm, name who closed the loop, and do not let a quality dashboard talk you into injuring this person.