A practical guide to measurement-based care: which measures each program should send, how often, what the scores should trigger, and why to track the therapeutic alliance alongside symptoms.
Key takeaways
Measurement-based care (MBC) is the routine use of short, validated questionnaires to track how patients are doing and to adjust treatment based on the results. The PHQ-9 for depression and the GAD-7 for anxiety are the usual starting points. The idea is simple. The execution is where most programs struggle: paper forms that never reach the chart, scores nobody reviews, and one schedule forced on every program whether it fits or not.
This guide covers how to implement measurement-based care in a behavioral health organization: which measures to use for which program, how often to send them, the scoring bands and decision rules clinicians need, why the therapeutic alliance belongs next to symptom scores, when to use alerts, and what the Joint Commission and CCBHC requirements expect.
Verity for Patients automates the workflow in this guide, so clinicians spend session time discussing scores instead of collecting them.
Comparing options? See our guide to the best measurement-based care software.
Measurement-based care is also called routine outcome monitoring (ROM), and it underpins feedback-informed treatment. Three things separate it from occasional screening:
The evidence is consistent. A 2021 meta-analysis in Clinical Psychology Review (de Jong et al.) pooled 58 studies and 21,699 patients. Progress feedback produced a small but significant improvement in symptoms (d = 0.15), a similar effect for patients who were not on track (d = 0.17), and a modest reduction in dropout. The effect is small per patient but applies to every patient in a program, at low cost. It also explains the main failure mode: when clinicians don't look at the feedback, there is no effect to get.
The PHQ-9 asks about nine depression symptoms over the past two weeks and produces a total score from 0 to 27. The GAD-7 asks about seven anxiety symptoms over the same period and produces a score from 0 to 21. Both are short, widely studied and require no license fee, which is why most MBC programs start with them. The standard severity bands are:
| Severity | PHQ-9 score (0–27) | GAD-7 score (0–21) |
|---|---|---|
| Minimal | 0–4 | 0–4 |
| Mild | 5–9 | 5–9 |
| Moderate | 10–14 | 10–14 |
| Moderately severe | 15–19 | Not used |
| Severe | 20–27 | 15–21 |
A score of 10 or higher is the common cutoff for a positive screen on both measures. PHQ-9 item 9 asks about thoughts of being better off dead or of self-harm. Any positive answer needs a same-day clinical review under your suicide risk protocol, whether or not you use automated alerts.
Symptom scores tell you whether a patient is improving. They don't tell you why, or whether the patient is about to stop coming. The therapeutic alliance, meaning the agreement between patient and clinician on goals and tasks plus the bond between them, fills that gap. It is one of the strongest and most consistently replicated predictors of both outcomes and retention:
In practice, a weak or falling alliance score is an early warning. It often shows up before a missed appointment or a stalled PHQ-9. Validated short alliance measures make this practical to collect at every session or every few sessions. Examples include the four-item Session Rating Scale (SRS) and the 12-item Working Alliance Inventory–Short Revised (WAI-SR). Many programs pair an alliance measure with a brief satisfaction question.
Verity can send alliance and satisfaction measures on the same schedule as symptom measures, and show both on the same dashboard. A clinician can then see, for example, that a patient's GAD-7 is flat while alliance is dropping, and use the next session to talk about the fit of treatment rather than adding another intervention.
One schedule for the whole organization rarely works. An intensive outpatient program, a school-based program and a medication for opioid use disorder clinic see different patients at different frequencies. Set a cadence for each program and keep it consistent within that program, so trends are comparable over time.
| Program type | Example measures | Example cadence |
|---|---|---|
| Adult outpatient therapy | PHQ-9, GAD-7, alliance measure | Intake, every 2–4 weeks, discharge; alliance every session or monthly |
| Intensive outpatient / PHP | PHQ-9, GAD-7 | Intake, then weekly |
| Anxiety-focused program | GAD-7, alliance measure | Intake, monthly, discharge |
| SUD treatment | PHQ-9, GAD-7 plus substance use measures (for example AUDIT-C, DAST-10 or the Brief Addiction Monitor) | Intake, every 2–4 weeks, discharge |
| Youth and school-based | Age-appropriate measures (for example PHQ-A or PHQ-9 modified for adolescents, PSC-17) | Intake, monthly or each grading period |
| Collaborative care in primary care | PHQ-9, GAD-7 | Each care manager contact, often every 2–4 weeks |
| CCBHC (any program) | PHQ-9 or PHQ-9M for clients 12+ with depression | Index score, then around six months for DEP-REM-6 |
These are starting points, not rules. The test is whether a clinician will see a new score before each decision point in that program. In Verity, each program, location or clinician can have its own set of assessments and its own cadence. One program can send the PHQ-9 weekly while another sends the GAD-7 monthly, and changing a program's schedule doesn't affect the others.
See program-level MBC in action. Verity sends each program its own measures on its own schedule and posts results to your EHR. Book a 30-minute demo.
Choose instruments with established reliability, validity and sensitivity to change, since that is what the Joint Commission expects. Start with the fewest measures that answer your clinical questions. Every extra measure lowers completion.
Completion goes up when patients answer on their phone a day or two before the session instead of on a clipboard in the lobby. Send one reminder. Scores are then ready when the clinician opens the chart.
This is the step most programs skip, and the one the research depends on. Spend the first few minutes on the graph: "Your PHQ-9 went from 16 to 11 this month. What do you think helped?" Patients keep completing measures when they see them used.
Write down what each result should trigger, so decisions don't depend on who happens to be treating the patient. Common rules for the PHQ-9:
Alerts are optional. Some programs want an alert when a score worsens, when alliance or satisfaction is low, or when item 9 is positive. Others prefer clinicians review results in the chart and on dashboards, and don't want more notifications. Both are reasonable. If you turn alerts on, decide which responses trigger them and who receives each one, such as the treating clinician, a supervisor or a crisis team. If you leave them off, make sure item 9 responses still get a same-day review through another route.
Scores belong in the chart, not a separate spreadsheet. EHR posting keeps the score next to the progress note, supports treatment planning and supervision, and feeds quality reporting. Confirm whether your EHR receives results by API or by scheduled file transfer, and where they appear.
Track completion rates, the share of patients with at least two scores, response and remission rates, and how often decision rules were followed. Aggregate data is where MBC turns into outcome evidence for payers and accreditors.
Joint Commission. Since January 1, 2018, standard CTS.03.01.09 has required accredited behavioral health care organizations to assess outcomes of care, treatment or services using a standardized instrument. The Joint Commission's guidance says the instrument should have well-established psychometric properties and documented sensitivity to change. It should be given at multiple points during care, with results used to inform goals, monitor individual progress and guide care planning. Aggregate results should support performance improvement. A PHQ-9 at intake only does not meet this standard. Repeated measurement that clinicians act on does.
CCBHCs. Certified Community Behavioral Health Clinics report clinic-collected quality measures. SAMHSA lists Screening for Clinical Depression and Follow-Up Plan (CDF-AD and CDF-CH) and Depression Remission at Six Months (DEP-REM-6) as required clinic-collected measures. DEP-REM-6 tracks clients aged 12 and older with major depression or dysthymia whose first PHQ-9 or PHQ-9M score above 9 sets an index date. Remission is a score below 5 at six months, plus or minus 60 days. Without routine PHQ-9 collection there is no follow-up score to count, so MBC and CCBHC reporting are the same workflow. Learn more about Verity for CCBHCs.
Verity works with any EHR and connects to 200+ systems through API or secure file transfer, so assessments go out on schedule and results post back to your EHR.
Measurement-based care is the routine use of validated measures, such as the PHQ-9 and GAD-7, at set points during treatment, with clinicians reviewing the scores and using them to adjust care. It is also called routine outcome monitoring. Meta-analyses show progress feedback improves symptoms and modestly reduces dropout, especially when clinicians act on it.
A common schedule is intake, every two to four weeks during active treatment, and discharge. Intensive programs often measure weekly, and CCBHCs need a follow-up score around six months for DEP-REM-6. Choose a cadence each program can sustain and apply it consistently so trends are meaningful and comparable over time.
Yes, and they usually should. An SUD program may add substance use measures, a youth program needs age-appropriate measures, and an anxiety program may send the GAD-7 monthly while another sends the PHQ-9 weekly. Verity lets you configure which assessments go to each program, location or clinician, and how often.
The alliance is one of the strongest predictors of outcomes and retention. A 2018 meta-analysis of more than 30,000 patients found an alliance–outcome correlation of about .28, and a 2010 meta-analysis linked weaker alliance to dropout. Short measures like the Session Rating Scale make it practical to track every session.
No. Alerts are optional. Programs that want them can trigger alerts from responses such as worsening scores, low alliance or satisfaction, or a positive item 9, and choose who receives each alert. Programs that prefer dashboard and chart review can leave alerts off, as long as item 9 still gets a same-day review.
For many organizations, effectively yes. Joint Commission standard CTS.03.01.09 requires accredited behavioral health organizations to monitor outcomes with a standardized instrument over the course of care. CCBHCs must report clinic-collected measures, including depression screening and DEP-REM-6, which need repeated PHQ-9 scores. Payers increasingly ask for outcome data too.
Verity for Patients sends the PHQ-9, GAD-7 and 500+ validated assessments on each program's schedule, can send alliance and satisfaction measures, offers optional alerts, and posts results to any EHR. Book a demo to see it with your own programs and EHR.