Using Secure Functioning to enhance PDSA Cycle

What each one is actually for

Secure Functioning

PACT is short for the Psychobiological Approach to Couple Therapy, developed by the clinician Stan Tatkin, and the awkward first word is the load-bearing one: psychobiological means the model starts from what two nervous systems do to each other before it gets anywhere near what the two people say. On this account, a partner who goes flat and quiet mid-argument is not being difficult, they are in a body state, and no amount of better phrasing reaches them until that state shifts. Secure Functioning is the name for what the model asks you to build in response. Two people agree, out loud and in specifics, that the relationship gets protected ahead of other things they also care about, and they write the agreements that make that concrete: how you leave in the morning and how you come back at night, what happens when one of you is late, what the other is allowed to interrupt. Tatkin calls the protected space those agreements create a couple bubble. The bubble is not a feeling of closeness. It is a small set of rules two adults wrote for each other and can be held to, which is why the model keeps insisting that goodwill on its own is not a plan. Rituals are how the rules get executed on ordinary days: a fixed morning goodbye, a ten-minute check-in before bed, a standing Sunday conversation about the week.

PDSA Cycle

The Plan-Do-Study-Act cycle is a four-step loop for finding out whether a change actually helps, and it is unusually strict about one thing: you have to commit to what you expect before you find out. Plan means writing the change down along with the result you predict, in terms specific enough that you would know if you were wrong. Do means running it small and briefly, on purpose, so being wrong is cheap. Study means putting the real result next to the written prediction and looking at the gap. Act means one of three decisions, usually stated as adopt, adapt or abandon: keep the change as it stands, alter exactly one thing about it and run the loop again, or drop it. The historical detail worth knowing is that W. Edwards Deming, who popularised the loop, deliberately renamed the third step from Check to Study, because checking suggests confirming that a thing was done and he wanted people to sit with what the result actually taught them. That distinction is the whole practice. A loop where the third step degrades back into checking is the failure this method was named to prevent, and it is exactly the failure a long-running relationship ritual tends toward.

What you get out of pairing them

Three weeks in, the nightly check-in is still happening and neither of you can say whether it is doing anything. That is a specific and common place to be stuck, and it is uncomfortable precisely because nothing is wrong. Nobody skipped it. Nobody broke an agreement. And yet the ten minutes have quietly become a thing you get through rather than a thing that reaches you, and saying so out loud feels like an accusation, because the ritual was your idea or theirs and it was made in good faith. What the improvement loop hands you is a way to have that conversation without either of you being the problem. When you wrote the prediction down at the start, the thing under examination is the prediction, not your partner. You are not saying the check-in has failed. You are saying you expected it to produce something you can name, and it has not, and neither of you knew that until you looked. In the other direction, what the couples model hands the improvement loop is the reason the third step happens at all. A person running a small test alone will almost always do the change and skip the study, not out of laziness but because the study step has no deadline, no witness, and no cost when it slips. Bind it to another person the way secure functioning binds anything else — a named day, a named question, and a rule that a missed one gets rescheduled rather than dropped — and the step that usually disappears stops disappearing.

What Secure Functioning rests on

• Attunement requires active monitoring not good intentions clinical practice
  Staying attuned to a partner's internal state requires ongoing structured attention rather than goodwill alone because connection erodes by default without it.
• Explicit commitment changes daily priority weighting clinical practice
  Explicitly and mutually committing to protect the relationship above other priorities changes how partners weigh choices during ordinary daily conflict.
• Rituals of connection reduce reliance on spontaneous attunement practitioner report
  Building repeatable rituals of connection reduces dependence on spontaneously noticing a partner's needs in the moment.

What PDSA Cycle rests on

• Feedback against prediction corrects belief plausible mechanism
  Comparing the actual result of a change against the predicted result before adopting it corrects mistaken beliefs faster than adopting or rejecting changes based on intuition alone.
• Small scale testing precedes scaling observational studies
  Running a change through a small planned test cycle before wider adoption reveals real-world effects while limiting the risk of being wrong.

Where they actually meet

A ritual and a process improvement fail the same way: both keep running. The couples model says that staying tuned in to a partner's inner state has to be actively worked at, because on its account it degrades quietly whenever two people stop paying deliberate attention — nothing announces the drift, which is what makes it dangerous. The improvement loop says something structurally identical about a change made at work: a plan carried out and then judged by whether it felt about right is not the same as a plan whose real result was set beside its prediction before anyone decided to keep it. Both are describing the same trap from opposite ends of life. Something you believe is working goes on working in your head, uninspected, because inspecting it never got scheduled and nothing forced the issue. And both answer it the same way, by making an invisible thing into an inspectable one. The agreement you wrote out loud can be pointed at; the prediction you wrote down can be compared. Where they differ is which half of the problem each one solved. The couples model built the commitment and never built the review. The improvement loop built the review and has no answer at all for why a person alone would bother doing it. Each one has the piece the other is missing, which is a better reason to put them together than any resemblance between their subject matter.

The principle underneath both

Something you agreed to once will keep being performed long after it has stopped producing the thing you agreed to it for, and you will not notice, because performing it feels like keeping the agreement. The fix in both traditions is the same: separate the doing from the result, and put a scheduled moment between them where the result gets looked at on its own. Note what this is not. It is not more discipline about the doing — both practices assume the doing is already happening. It is a second, smaller commitment, aimed at the first one, and its only job is to ask whether the first one is still earning the time it takes. Neither couples work nor improvement science owns that idea. It belongs to anything that runs on a schedule.

What that makes the real goal

Call the goal this: a commitment that stays answerable to what it actually produces. Not a commitment you keep harder, and not one you hold more loosely — one that has a built-in way of reporting back. Stated like that, it stops being about marriages or hospitals and becomes a general property you could want in any standing arrangement: a weekly team meeting, a Sunday phone call to a parent, a shared practice you and a friend agreed to in January. The arrangement continues to exist, and separately, at a known interval, somebody looks at whether it is still doing its job and is allowed to change it. The reason to name the goal this way rather than as either practice's version of it is that once you can see it, you can go looking for better instruments than the two on this page — and a couple of them have research behind them that neither of these does.

Other practices that serve the same goal

Other structures aim at exactly this and come with their own track records. The military after-action review is the closest cousin: a short structured debrief run immediately after an event, asking what was supposed to happen, what did happen, what accounts for the difference, and what to change next time. It has been used at scale for decades and studied mainly through organisational and observational work rather than randomised trials, but its four questions are usable at a kitchen table tonight with no adaptation at all. In medicine, the n-of-1 trial does the same job with more rigour — a single patient cycles between a treatment and a comparison in a planned order, with outcomes recorded against a stated expectation, and this design has a real methodological literature behind it. In couples work specifically, Integrative Behavioral Couple Therapy has been tested in randomised trials against an earlier behavioural approach, and one of its ordinary components is reviewing agreed changes with a therapist rather than assuming they took. That is the piece worth noticing: the review step, which the couple bubble model leaves to the couple, is standard practice in the couples therapies that have been trialled.

A tradition that already does this

The nightly review is old. Ignatius of Loyola's Examen, written in the 1520s, asks a person to sit at the end of each day and walk it back against what they had committed themselves to that morning — not to catalogue faults, but to notice where the day and the intention parted company, and to decide what tomorrow needs. That is a review against a stated prediction, done on a fixed schedule, four centuries before anyone drew the loop as four boxes. It was built for one person's commitment rather than two people's agreement, so the couple version here is an extension of the shape rather than a citation of the practice. But the shape is not new, and it did not come from industry.

Using Secure Functioning (PACT) to enhance the PDSA Cycle

The step that dies is Study, and it dies for a structural reason: it is the only one of the four with no natural trigger. Plan happens because you are excited. Do happens because it is on the calendar. Act happens because a decision gets forced eventually. Study happens only if someone makes it happen, and alone, nobody does. The couples model's answer to exactly this problem is the explicit agreement — not a promise to try, but a written, mutual, specific arrangement that names who, when, and what happens on a miss. Borrow it whole. Four steps, and the order is doing work. First, before the change starts, write the prediction as a single sentence a stranger could mark right or wrong: not 'the new evening routine should help', but 'by the end of two weeks, on at least four nights out of seven, I will be in bed before eleven.' Second, name the person and the slot — a specific human, a specific fifteen minutes on a specific day, in the calendar, with their agreement, not a vague 'we should catch up about this.' Third, agree in advance what a missed review means: it gets rescheduled inside forty-eight hours, and the loop does not advance until it happens. That rule is the couples model's fast-repair instinct applied to a process, and it is the difference between a slip and a quiet abandonment. Fourth, at the review itself, read the prediction out loud before either of you says a word about how it went. Prediction first, always. Say how it went first and the prediction retroactively becomes whatever the result was, which feels like learning and is not.

Using the PDSA Cycle to enhance Secure Functioning (PACT)

Take one ritual — the one you would defend hardest — and run it as a test for four weeks. Plan: write down, together and before you start, what the ritual is supposed to produce, in terms you could actually observe. 'Feeling closer' is not observable. 'On most nights, each of us can say one specific thing the other told us that day, without being reminded' is. Pick a number you would count as a pass and say it out loud, because a threshold agreed in advance is the only kind that cannot be moved afterward. Do: run it unchanged for the full four weeks and keep the record to one line a night — did it happen, yes or no, and a number from each of you, zero to ten, for how connected you felt at the end of it. One line. The moment logging becomes journalling, it stops in week two. Study: sit down with the log at the end and read the prediction out first, then the log, then talk. Both of you look at the same page, which matters more than it sounds like it should — it makes the ritual the subject of the conversation instead of one of you having a complaint about the other. Act: choose one of three things out loud. Keep it as it is. Change exactly one variable — the time, the length, the location, the opening question, but only one, because two at once and you will never know which one mattered. Or stop it and put something else in the slot. Stopping is a real option here and naming it in advance is what makes the whole review honest; a review where the only permitted outcome is carry on is not a review.

Night nineteen of the check-in, and you are counting the minutes

It is Tuesday, day nineteen, and the ten-minute check-in you both agreed to at the start of the month is happening on schedule, the way it has every night. You are sitting on the end of the bed. Your partner is telling you about a thing at work that you have already heard about, and you are aware that you are waiting for it to be over, and you are aware that this is a slightly awful thing to be aware of. You do not say anything, because the ritual is working — look, it is happening. That silence is the whole problem, and it is not a communication failure, it is a missing step. There is nowhere in what the two of you built for the sentence 'I do not think this is doing what we hoped it would do.' So you make somewhere. On Thursday, in the fifteen minutes you have already put in the calendar, you start by reading back what you both said in week one — that the check-in would mean going to bed without anything unsaid. Then the log: seventeen nights out of nineteen it happened, and the numbers you each wrote sat around a four. Nobody defends anything. The pattern in the log is that the good nights were the ones that started with something specific rather than 'how was your day', so you change one thing — the opening question becomes 'what was the hardest part of today' — and run another two weeks. That is the whole move. The ritual survived. It just stopped being unquestionable.

When this is worth reaching for

You have a ritual that everyone still shows up for and nobody would call useful anymore. Or the opposite condition: you are about to add a third agreement to a relationship that already has two nobody has looked at since they were made, and some part of you knows that adding is easier than reviewing. Both are the moment for this. The other one is timing — there is no point testing a ritual you started last Thursday, because four days of anything is noise, and a review that early will mostly measure novelty. Give it three or four weeks of running before you look. And if what is actually happening is that one of you is angry about something else entirely, this is not the tool; a structured review will just give the anger a spreadsheet to live in.

Where this does not apply

The improvement loop needs a result that can be named and set against a stated prediction, and a fair amount of what secure functioning protects resists that — feeling chosen, feeling safe in someone's presence, the sense that this person would take your side. Numbers can gesture at those. They cannot hold them. Run every ritual as a formal test and you risk something a partner will feel before they can name: being audited rather than tended. There is also a failure mode worth stating plainly. The loop assumes someone is willing to change the plan when the review says to. If one of you is using the ritual to keep a harder conversation from starting — the one about money, or the affair, or whether either of you still wants this — then measuring the ritual with more rigour just measures the wrong thing more precisely. That situation is not a design problem. And if either of you would use the log as evidence in an argument, do not keep a log. The record only works when both people are sure it is not being collected against them.

What the evidence does and does not support

The improvement loop has an odd evidence profile for something so widely used. Its central claim — that setting a real result against a prediction you committed to beforehand corrects a mistaken belief faster than intuition does — is well grounded mechanistically and is the basis of most quality-improvement practice in healthcare, but reviews of published improvement work have repeatedly found that many projects labelled as this kind of cycle do not actually complete all four steps, which tells you how easily the study step gets skipped even by professionals with an institutional mandate to do it. Secure functioning's claims sit on a different footing: attunement needing active structure, and explicit mutual commitment shifting how ordinary daily choices get weighted, come out of clinical couples practice and case observation rather than randomised trials of the framework itself. So neither side of this page rests on a trial of the thing being described, and the specific move here — putting a four-step improvement loop around a couple's connection ritual — has not been tested by anyone in either field. There is no study comparing couples who review their agreements on a schedule against couples who do not. What is on this page is a reading of how the two mechanisms would meet, and the honest test of it is whether your own four weeks of log say anything you did not already believe.