eighteen month writeup on owner training a psychiatric service dog for combat related PTSD, the public access training gap nobody talks about, what twenty one hundred hours of structured training time actually looked like, and the things i would tell a veteran at month zero of the same journey

Granite is our 26 month old standard poodle, 52lb black male, sourced as a puppy from a breeder with a multigenerational track record of placing dogs into service work, with the intentional plan from week 8 to owner-train him as a psychiatric service dog (PSD) for my combat related PTSD. we are now 18 months into the active training, he passed his PAT (public access test) at 22 months with our independent evaluator, and i want to write up what this 18 months has actually looked like because the PSD owner-training community on this forum and others is genuinely helpful but undersells what i now think is the central truth of this work, which is that the task training is the legible part that everyone focuses on and the public access training is the invisible 70% of the work that determines whether you end up with a functional team or a project that quietly dies at month 14.

the headline framing. you can teach a smart dog the task list (DPT, block, cover, interrupt nightmare cue, retrieve medication bag) in 4-6 months of focused training. the dogs learn the tasks faster than the handlers, this is a known phenomenon in the SD community. what takes 18+ months is the neutrality training that allows the dog to perform those tasks reliably in a chick-fil-a at lunch rush while a child runs past trailing a balloon and a service worker drops a tray. the tasks are the resume, the public access neutrality is the actual job. i wish someone had said this to me in plain language at month zero because i spent months 3-9 over-investing in task polish and under-investing in the boring environmental work that turned out to be the load bearing piece.

the hour breakdown. i tracked my training time across 18 months using a spreadsheet because my therapist asked me to and it ended up being the most useful artifact of this process. roughly 2,100 total hours, broken down as follows. 340 hours of foundation obedience (sit stay down recall heel loose leash) primarily months 0-6. 220 hours of task training (the 7 specific tasks Granite performs, all positively reinforced, broken into ~30 hour blocks per task with maintenance time across the rest of training) primarily months 4-15. 1,150 hours of public access work (graduated exposure to retail, restaurants, transit, medical settings, crowded outdoor venues) across months 6-18. 290 hours of handler skill work (managing my own physiology so my cues stay clean, learning to read Granite's micro-stress signals before they cascade, the unsexy practice of being a good half of the team) across the full timeline. 100 hours miscellaneous (vet visits, grooming desensitization, gear introductions). the public access piece was 55% of total training time. nothing in the SD literature i had read before starting prepared me for that ratio.

what the public access training actually looks like at high resolution. week 1 of public access work was not a store, it was a parking lot. specifically the parking lot of our local petsmart, parked 200 feet from the entrance, watching shopping carts roll by from the safety of the car with treats for calm orientation. week 4 was the lobby of a bank branch with a cooperative manager. week 7 was a 6 minute walk through a low traffic period at home depot. month 4 was a 12 minute lunch with two known friends at a restaurant patio. month 9 was the same restaurant patio at saturday lunch rush. month 13 was a 45 minute round trip on light rail. month 17 was a 6 hour flight. the cadence is graduated environmental exposure where each new environment is introduced at the easiest version of that environment first and the difficulty stack is added one variable at a time. the day i tried to add three new variables at once at month 8 (new venue, new time of day, new high distraction event happening unexpectedly) was the day Granite had his only real public access regression and we lost about 6 weeks to remediation. the discipline of one new variable at a time is the actual skill.

the tasks Granite performs and the calibration on what is and is not a task. DPT (deep pressure therapy, the cued behavior where he places his front quarters across my lap during a panic episode, weight distributed for grounding, hold for the duration of the episode). block (the trained position where he places himself perpendicular in front of me to create personal space buffer in lines or crowded waiting rooms). cover (the rear facing version of block, used in situations where i need to know nothing is approaching from behind, taught with a specific cue and a specific position). nightmare interrupt (a trained behavior where the alarm cue from my watch triggers him to come to the bed and physically interact with me until i am conscious, this took 9 months to fully proof). medication retrieval (he brings my benzo case from a designated location on cue). guided exit (he leads me to the nearest exit of a building on cue, used when a panic episode is escalating and i need to leave a venue calmly without me having to navigate while disregulated). brace (a stationary position that lets me use his harness handle to ground myself during balance compromise from dissociation, the only "mobility-adjacent" task he performs and one we use carefully). things that are not tasks despite being useful. being calm in public is not a task, that is neutrality training. comforting me when i am sad is not a task, that is being a dog. lying at my feet quietly during dinner is not a task, that is foundation manners. the distinction matters because the ADA-protected public access right is tied to the trained tasks not to the dog's general presence.

the things i would tell a veteran starting this journey at month zero. one, your prospect puppy selection matters more than your training program. an unsuitable prospect will wash out at month 10-14 and you will be devastated and starting over. the wash out rate for owner trained PSDs is around 50-65% by month 18 and the single biggest predictor is whether the prospect had the temperament to do the work. work with a breeder who has placed dogs into service work and who will pull you aside and tell you honestly if the puppy you fell in love with at week 6 is the puppy who should not go home with you. two, find a mentor who has a working SD team and is willing to let you tag along on 5-10 outings in your first 6 months. you cannot learn the public access skill from videos, you have to watch someone do it. the SD community has a generosity culture about this, ask. three, the dog cannot heal you, the dog can be part of a treatment plan. my therapist and my SD trainer were in conversation across the 18 months and the integration of the clinical and the training side is the part that made this work as a treatment intervention not just as a dog project. four, budget realistically. our 18 months cost about $14,500 including the puppy, food, vet care, training equipment, gas to and from public access sessions, the trainer consults (we did 1-on-1 with a CCPDT trainer at $135/hour for about 60 hours total), the PAT evaluation, and the working gear. that is approximately what a program-trained dog costs as the application fee at the major SD organizations, the hidden cost of owner training is your time not your money. five, the public access training is the work. i am repeating this because i did not internalize it at month zero. the tasks are the resume, the neutrality in chaos is the job. budget your hours accordingly, my regret is the 4-5 months of mistraining where i polished his DPT to perfection while neglecting the boring environmental work that was the actual constraint. happy to answer specific questions about any of this, from the breeder evaluation conversation forward