Most measures are better sent before the session than during it. A PHQ-9 and a GAD-7 filled in the day before means you start the hour already knowing where the client is, instead of spending six minutes of a fifty-minute session watching someone tap through a form in silence.

Administer live when the filling is the work. A measure you're walking through item by item because the wording needs explaining. A risk screen where you want to see which items they pause on. A worksheet you're building together. A first-session measure with a client who won't do homework and you both know it.

If you are sending one mid-session, four things need to be sorted before you hit send: whether their device can actually render it, whether they can see the scoring, where the finished result lands, and what happens if the connection drops halfway through. None of those is difficult, and all four are unpleasant to discover live.

When is live administration actually better?

When the client won't do it otherwise. A form emailed in advance is a form with a completion rate. In session it is a form with a completion rate of one.

When you want the process, not just the score. How long they sit on an item, which one they ask about, whether they answer quickly without reading — that information does not survive an emailed link, and for some clients it is more useful than the total.

When the wording needs you. Plenty of standard measures contain items that land differently depending on how they're read. A client who interprets "trouble concentrating" as a question about their phone habits gives you a number that means nothing.

When something has just changed. A measure taken right after a disclosure is a different measure from one taken the day before it.

When should I send it beforehand instead?

Intake stacks, routine outcome monitoring, and anything long. If your standard opening is three questionnaires, sending them with the intake paperwork buys back most of a session and lets you open with what the answers mean rather than with the answers.

The trade is that you lose the observation, and you get a form filled at eleven at night in whatever state that implies. For routine monitoring that is fine. For a first risk screen it may not be.

What should the client's screen show?

Their job is to answer. Yours is to interpret. Anything on their screen that blurs that line is worth removing.

Specifically, the client should generally not be looking at:

• The numeric weight of each option. Once a client can see that "nearly every day" is worth three, some of them start doing arithmetic instead of answering.

• A running total. The same problem, faster.

• A severity band. "Moderately severe" is a clinical label with a conversation attached. Delivered by a form, unaccompanied, it can land as a diagnosis.

• Your annotations. Notes you make about how they're answering are for the chart, not for them.

Scores belong to the interpretation, and the interpretation belongs to you and the client together, at a moment you choose.

Where does the finished measure end up?

This is the question people skip, and it is the one that costs an hour later.

Some tools save the completed measure into the client's record automatically. Some hand you a copy to paste. Some leave it in a browser tab that closes with the call, and the score is gone.

Before you administer a measure live, know which of those you have, because the answer determines whether you can leave the record until later or whether you need to capture it before you close the window. If the export is a copy or a download, do it before you end the call — not after, when the client's next message is already in front of you.

The same question applies to your own annotations. Notes attached to items are worth very little if they don't travel with the responses.

What if the connection drops mid-fill?

Ask what your tool does before you find out. Some hold the in-flight answers and resume where the client left off; some start again from item one.

If yours starts again, that changes what you should administer live. Asking a distressed client to answer twenty items twice is not a neutral event. On an unreliable connection, shorter measures live and longer ones in advance is the sensible split.

Either way, tell the client at the start what happens if they drop. "If you lose me, stay where you are and I'll call you" covers the assessment as well as everything else.

Do I need consent for this?

You need the same consent you'd need on paper, which most practices cover in their general consent to treatment. The parts that are specific to doing it electronically are worth a sentence in your communications consent: that measures may be administered through the platform, that the responses form part of the clinical record, and where those records are held.

If a client asks to do it on paper instead, that is a reasonable request and worth honouring rather than negotiating.

What about the awkward silence?

It is real, and it is the main reason practitioners avoid live administration. Four minutes of watching someone tap is four minutes of nothing.

Two things help. Say what you're doing and how long it will take — "there are nine items, it's about three minutes, take your time and I'll be here." And then actually be there: stay on camera, stay quiet, don't check your other screen. Clients notice, and a therapist visibly reading something else while they answer questions about wanting to be dead is a rupture you will spend the rest of the hour repairing.

The short version

Send it in advance unless the filling is part of the work. If you're doing it live, keep the scoring on your side, know where the finished measure lands before you start, agree what happens if the connection drops, and stay present while they answer.

About Rivet

Rivet is a Canadian practice line built for therapists — a separate number for calls, texts and voicemail, with video sessions and clinical tools in the same place. Your data stays in Canada, and transcription runs on Rivet's own hardware rather than a third-party AI service.

One plan, $65 CAD a month, everything included. Fourteen-day trial, no card.

See what's included · Start the trial