Automating PHQ-9 and GAD-7 screening means sending the questionnaires before the visit, scoring them automatically, and routing each result to a named reviewer. PHQ-9 item 9 needs its own route to same-day clinician follow-up. Automation removes manual arithmetic and repeated delivery work. It does not choose the measure, interpret the score, or replace clinical review.
The automated workflow
A typical sequence starts when an appointment is booked. The patient receives a link by email or text before the visit and completes the PHQ-9 and GAD-7 on a phone or computer. The platform scores each questionnaire and places the result where the clinician will see it before the session. A PHQ-9 with any answer above zero on item 9 goes to a separate safety route instead of the ordinary review queue.
After the visit, the clinician documents the result, and the next questionnaire is scheduled under the care plan. Set the follow-up interval as a practice decision, write it down, and keep it consistent within an episode of care. Both questionnaires ask about the past two weeks, so results collected less than two weeks apart share part of the same recall window. The automated scheduling guide covers how to document the interval and what to do when a result does not arrive.
What to automate
The PHQ-9 and GAD-7 suit routine delivery at intake and at the follow-up points the care plan sets. For screening, the PHQ-2 and GAD-2 can serve as a first step. A score of 3 or more on either is a positive screen and should lead to the full PHQ-9 or GAD-7, or to a clinical interview.
Some measures belong to a defined group of patients rather than everyone. The PCL-5 measures PTSD symptoms and fits patients with a trauma history or a positive trauma screen. The AUDIT-C is a brief alcohol-use screen. Send these from a clinical flag, a referral reason, or a clinician request.
Implementation options
Your EHR may include questionnaire features. Check whether it supports portal completion, automatic scoring, item-level storage, and a view of earlier results. These features and the patient-facing screens differ between systems, so test them with real staff and patients before committing.
A dedicated assessment platform needs an interface to the EHR, such as HL7, FHIR, or a vendor API. Manual re-entry of scores adds a transcription step and removes much of the time saved. Some practices keep documentation and billing in the EHR and use a separate platform for patient-facing questionnaires. The practice scaling guide covers how these systems fit together as a practice grows. The EHR integration guide lists what an interface needs to carry, including questionnaire version, completion state, and review state.
Configuration essentials
Clean up patient contact data before launch. Bounced emails and wrong numbers mean questionnaires never arrive, and silence can be mistaken for a low score. Record each patient's consent for email and text contact. Keep message content minimal because notifications can appear on shared devices or lock screens.
Configure delivery with one of three triggers:
- Time: send the PHQ-9 a set number of hours before each appointment.
- Interval: send the GAD-7 at the interval written in the care plan.
- Event: send the AUDIT-C when a clinician flags a need for alcohol screening.
Start with one trigger type. One workable reminder pattern is an initial request, one follow-up reminder, and a final reminder on the day of the visit. Choose the timing so a completed result still reaches the reviewer before the session. Then adjust it using your own completion data. The no-show guide covers what the evidence shows about reminders and attendance.
Alerts and item 9 responses
PHQ-9 item 9 asks about thoughts of death or self-harm. Any answer above zero needs same-day clinician follow-up under the practice's safety protocol, regardless of the total or trend. The protocol should name who contacts the patient, how quickly, and how to escalate if the patient cannot be reached. It should also say who covers evenings, weekends, and absences. Do not promise patients real-time monitoring unless someone actually watches the queue at those hours.
On the patient side, show crisis resources as soon as item 9 is endorsed and tell the patient how to reach the practice. Do not rely on the patient reading that screen. On the clinical side, flag the record, notify the responsible clinician, create a follow-up task, and document the contact.
If you need help now. In the US, Call 988 or text 988. Call 911 if you are in immediate danger. Outside the US, contact your local emergency number or find support in your country.
Keep other alerts few and deliberate. A total in a band the practice has chosen, such as moderately severe or higher on the PHQ-9, can notify the treating clinician. A rise between comparable results can be flagged for review before the next session. In a 2004 study of 434 older adults in the IMPACT trial, a five-point PHQ-9 difference was the estimated minimum clinically important change. That supports a review signal in a similar population, not a universal response rule. The PHQ-9 tracking guide explains how to compare results before acting on a change.
Repeated exposure to alerts desensitizes clinicians, a pattern AHRQ describes as alert fatigue. Route each alert to the person who can act on it, and separate urgent alerts from informational ones.
Patient experience
Make the form work on a phone, with large tap targets and few screens. Tell patients why they are receiving the questionnaire, how many questions it has, and that their clinician will review the answers before the appointment. The PHQ-9 has nine scored questions, and the GAD-7 has seven.
After submission, confirm that the clinician will review the answers. If you show the patient a score, show the band in plain words and state that a screening result is not a diagnosis. Some patients prefer not to complete forms online. Offer a tablet in the waiting room or a paper form, and record the administration mode so later results can be compared fairly.
Billing considerations
CPT 96127 describes a brief emotional or behavioral assessment with scoring and documentation, reported per standardized instrument. For a patient with a commercial plan who completed a PHQ-9 and a GAD-7, AAFP coding guidance advised one unit of 96127 for each instrument. Payer rules vary. Check the payer's current policy for unit limits, covered settings, who may administer the instrument, and how it treats electronic completion.
For documentation, AAFP's editors list the date, the patient's name, the informant if not the patient, the instrument, the score, and the name and credentials of the person who administered it. The physician should also document review of the score in the context of the patient's presentation. An automated system can capture the instrument, score, date, and mode. Your note adds the interpretation and the action. Write the interpretation as a symptom band, for example "PHQ-9 total 14, moderate symptom band", because a screening score is not a diagnosis.
Measuring success
Track each stage separately: questionnaire due, delivery attempted, delivery succeeded, completed, completed before the visit, and reviewed. A high send count can sit beside low completion or late review. Measure a baseline in the first weeks, set targets from your own data, and find where the drop-off happens before changing reminder timing.
Compare screening coverage with your written protocol. The useful question is whether eligible patients receive the intended questionnaire and whether a clinician reviews it in time to use it. If you later report clinical change across patients, the practice-level outcome measurement guide covers cohort and denominator rules.
Common problems
Plan for integration failures: results that do not reach the chart, duplicate results, or results attached to the wrong encounter. Test with realistic cases before go-live, keep a manual fallback ready, and name the person who works with the vendor when something breaks.
Introduce the workflow with a small group of clinicians, fix what they find, and then extend it. Train each role on its part, including what to do when an item 9 alert arrives. The staff training guide covers competency checks for administration, scoring, and safety scenarios. Review alert volume and alert criteria at a fixed interval, and remove alerts nobody acts on.