How PelvicEx designs and personalizes its training programs

Short answer

PelvicEx does not personalize a programme by guessing muscle function from age, sex, body size, fitness, or a single “beginner” label. It uses a versioned, deterministic rule system that considers:

• the user’s selected goal or no-goal starting route;
• safety and treatment context;
• whether contraction and full release are recognized;
• an optional, self-reported capacity check;
• completed technique lessons;
• recent session response; and
• available time and presentation needs.

Safety rules run before training goals. Pain, inability to release, an active catheter restriction, urgent symptoms, or instructions from a healthcare professional can pause, redirect, or narrow the plan.

This is explainable self-guided planning, not diagnosis or an objective muscle test. Program selections are explainable planning decisions based on recorded answers and safety rules. The implemented timed trainer runs the selected session, but it cannot verify the movement or establish that the plan is clinically suitable.

The design question

The planner answers one bounded question: “Given the information the person chose to provide and the recorded response to previous sessions, what is the most conservative appropriate next PelvicEx session?”

It does not answer:

• What condition does this person have?
• Is the pelvic floor objectively weak, tight, injured, or correctly contracting?
• Has surgery or birth tissue healed?
• Will this programme cure a symptom?
• What would a clinician find on examination?

Keeping the question narrow prevents a questionnaire from being presented as medical assessment.

Evidence and product rules are separate layers

The program model starts with evidence-supported principles: a pelvic-floor contraction should include closure and inward lift rather than bearing down; complete release and normal breathing matter; repeated training can address different capabilities; progression depends on performance and response; and condition-specific guidance may require supervision.

Research does not provide every value a software planner needs. A product still has to decide what happens when answers are skipped, how modules from two compatible goals fit into one session, how long to wait before considering progression, and how to record a stopped workout.

PelvicEx labels those implementation choices as product extrapolations. They are designed conservatively, stored in versioned catalogs, and tested as software rules. They are not described as the one clinically proven programme. The evidence overview explains why the optimal PFMT dose and delivery remain uncertain.

The training building blocks

PelvicEx separates content into layers so one timer component does not contain hidden clinical decisions.

1. Action: one timed instruction, such as contract, release, gradual build, or gradual release.
2. Technique: a named movement pattern with teaching points and stop rules.
3. Block: a technique repeated with an authored load and recovery.
4. Module: blocks aimed at one capability, such as technique, release, strength, endurance, faster control, functional timing, urgency control, or bowel-related control where appropriate.
5. Session: the ordered modules selected for today.
6. Program: the longer sequence, review points, progression boundaries, and maintenance route for a goal.

This structure allows the same release module to support several compatible programmes without copying its clinical rules into every screen. It also lets evidence IDs, prerequisite lessons, and stop conditions travel with the content they govern.

What the user is asked before a plan

PelvicEx tries to collect only information that changes safety, instruction, or session selection.

Goal

A primary goal selects the main programme emphasis. Compatible secondary goals can add supporting modules, but the app does not stack two complete workouts into one session. If no goal is selected, an authored general-basics route is available.

A goal is not a diagnosis. Selecting “bladder control,” for example, does not establish the type or cause of leakage.

Safety and treatment context

Questions cover urgent symptoms, professional restrictions, pregnancy or postpartum context, pelvic or prostate treatment, catheter status, known neurological context, pain or painful sex, difficulty releasing, bowel emptying difficulty, and other factors that alter a self-guided route.

Withheld or unresolved safety information is not silently treated as “no.” The planner can provide education, ask for clarification, or require external guidance instead of loading exercise.

Experience

Prior Kegel experience changes tutorial depth, not assumed capacity. “I practise regularly” does not automatically create longer holds or more repetitions. A person taught by a professional should be able to preserve that plan rather than have it silently replaced.

Optional self-check

The self-check records perception, not measurement. It may ask whether a closing and lifting movement feels clear, whether release feels complete, what remains comfortable while breathing normally, and whether pain, pressure, urgency, or unusual fatigue appears.

Skipping it leaves capacity fields unknown. Unknown is not converted into an advanced score and does not automatically mean unsafe. PelvicEx can begin with an authored gentle or balanced starter and learn from session response.

Safety routing happens first

Program rules are applied in a fixed order:

1. Urgent or external-care route: possible emergency signs, an active restriction, or a treating professional’s instructions override exercise goals.
2. Treatment-stage gate: catheter, surgery, cancer treatment, pregnancy, postpartum, neurological, or other contexts can require clearance or a different pathway.
3. Relaxation-first route: pain, worsening with contraction, difficulty releasing, or relevant emptying difficulty prevents automatic strengthening progression.
4. Technique route: uncertainty about the movement leads to locating, coordination, and release teaching rather than a harder workout.
5. Compatible starter: only after those gates does the planner select a gentle or balanced authored loading block.
6. Response-based adaptation: later sessions can maintain, regress, redirect, or move to an adjacent authored level.

The order matters. A strength goal cannot outvote a safety route, and a streak cannot override pain.

Independent capabilities, not one global level

A person can have good contraction endurance and poor release, or recognize a movement but lose normal breathing after a few repetitions. PelvicEx therefore keeps separate self-reported fields for:

• contraction recognition;
• technique confidence;
• full relaxation;
• comfortable hold control;
• controlled repetition capacity;
• quick contraction and release control;
• comfort and symptom response; and
• confidence in the assessment itself.

The word “comfortable” is deliberate. Calibration is not a maximum-effort challenge, and a higher number is not always a better result.

Building a session from one or more goals

Each programme gives modules priorities and compatibility rules. The planner resolves the primary goal first, then adds only compatible supporting work. It also respects a total session-time boundary and avoids duplicating the same module merely because two goals request it.

For example, general technique and release may support both continence and a broader control goal. They should appear once, not twice. A relaxation-first safety route may replace loading modules entirely. Some treatment-specific or bowel modules remain unavailable without the right route.

The resulting session stores the programme, module, catalog, and ruleset versions that produced it. Reason codes record why a route, tutorial, level, or overlay was selected. This creates an audit trail instead of a black-box recommendation.

How the next session changes

After a session, PelvicEx asks a small number of response questions: whether the user finished, perceived difficulty, whether full release remained possible, and whether unwanted symptoms occurred.

The planner then follows conservative transitions:

• redirect or stop for a safety-significant response;
• regress after incomplete control, inability to release, a stopped session, or repeated excessive difficulty;
• maintain when the session was appropriate or evidence for change is insufficient;
• progress one adjacent step only after a consistent pattern of controlled, comfortable sessions without unwanted response;
• recalibrate after a long break, new goal, treatment change, repeated mismatch, plateau, or symptom change.

Completion alone is insufficient. Streaks and rewards are engagement records and are not permitted to change dose, safety, progression, or clinical route.

Explainability

The interface should show the actual reasons that affected a plan, such as:

This session keeps the same level because you completed it comfortably but were unsure about full release.

It should not list every answer, expose sensitive details unnecessarily, or invent an explanation after the fact. A plan decision is reproducible from the stored profile, recent applied responses, catalog version, ruleset version, and time.

No machine-learning model selects a PelvicEx programme in the current design. The logic is authored and deterministic, which makes rule precedence and edge cases testable.

Privacy and local progress

The current free design stores the compact profile, plan state, and session history in the browser through a versioned storage interface. The clinical planner does not depend directly on browser storage, so a future cloud adapter could be added without changing the rules.

There is currently no account or cloud sync. Sensitive goals, symptoms, treatment context, and workout history must not enter advertising or analytics events. A future sync service would require explicit opt-in and a continued local-only option. Read the privacy approach for present behaviour and limitations.

What the current product can demonstrate

The repository contains the authored action, tutorial, module, programme, frequency, personalization, evidence, and example-decision catalogs. It validates their references and runs deterministic safety, planning, and progression tests. The current user interface can collect onboarding information, resolve a plan, save progress locally, and expose development simulations.

The public timed trainer runs prepared sessions with explicit timing, pause, resume, restart, end, interruption handling, reduced-motion support, optional sound and vibration cues, and responsive mobile controls. It records technical completion locally, while user feedback and the safety rules determine whether a result may affect later planning. It does not observe the body or turn a completed timer run into evidence of correct technique.

Clinical and scientific limitations

The ICS assessment terminology report describes observations and measurements that a questionnaire cannot reproduce. PelvicEx cannot palpate movement, measure force or tone, inspect tissue healing, identify tenderness, establish a diagnosis, or verify that requested effort occurred.

Research also does not establish one optimal PFMT dose. The 2024 Cochrane comparison reinforces the need to keep exact program choices appropriately qualified.

The catalog is scoped to adults and mostly non-neurogenic self-guided use. Children, known neurological bladder or bowel conditions, complex postoperative care, severe pain, and other higher-risk contexts need condition-appropriate professional pathways.

Review and change control

Catalogs carry version numbers. A change to dose, safety precedence, eligibility, stop conditions, review timing, or a clinical cue requires tests, source review, and the appropriate professional review. Historical sessions retain their original versions so a later rule change does not rewrite what happened.

Programme governance covers clinical wording, every numeric transition, goal boundaries, treatment overlays, locale adaptations, and examples. Changes to these rules are recorded internally and tested against the safety-routing requirements.

Sources and methodology basis

• Bø K, Berghmans B, Mørkved S, Van Kampen M, editors. Evidence-Based Physical Therapy for the Pelvic Floor. 3rd edition. Elsevier; 2023. Chapters 5, 6, and 17. Official publisher record.
• Frawley HC, Shelly B, Morin M, et al. An International Continence Society report on pelvic-floor muscle assessment terminology. Neurourology and Urodynamics. 2021;40(5):1217–1260. PubMed.
• National Institute for Health and Care Excellence. Pelvic floor dysfunction: prevention and non-surgical management. NG210. NICE recommendations.
• Cochrane. Which approaches work best to treat involuntary urine leakage in women? Updated 2024. Cochrane evidence summary.
• The ordering, catalog architecture, reason codes, authored starter modes, progression thresholds, and local implementation are PelvicEx product decisions. They are documented internally as evidence-informed extrapolations with explicit limitations and deterministic tests.