12-phase flow (12)

1. 1FRAME
   Menopause management is a person-centred BENEFIT/RISK ELIGIBILITY decision over symptom domains (vasomotor, GSM, sleep, mood, long-term bone/CV), staged by STRAW+10 and gated by the MHT timing hypothesis × route × progestogen rule — not a single disease (NAMS HT 2022 PMID 35797481; Harlow 2012 PMID 22344196). BAYESIAN CONDITIONAL #1 (VMS-cause prior | age + menstrual pattern): the pre-test probability that midlife VMS = the menopausal transition is NOT fixed — it is conditioned on age band and STRAW+10 menstrual pattern. Prior ≈0.85–0.90 when age ≥45 y AND ≥60-day amenorrhoea/late-transition pattern AND typical hot flushes (VMS ~80% prevalence, persist median >7 y — Wang 2025 PMID 40049806); prior falls to ≈0.30–0.40 when age <40 y (POI competes — route OUT) or when menstrual cycles remain regular with isolated sweats (thyroid / drug-induced / anxiety mimics rise). The clinical-diagnosis LR is therefore evaluated AGAINST this conditioned prior, not a flat prior
   inputs: age, years_since_final_menstrual_period, menstrual_pattern_and_straw_stage
   actions: calc.straw10
   advance: STRAW+10 stage assigned and the symptom-domain inventory framed against the timing-hypothesis structure
2. 2ENTRY
   Vasomotor symptoms; GSM; perimenopausal cycle change; established-MHT review/de-prescribe; midlife sleep/mood; or amenorrhoea+VMS before 40 (POI question — route out) (NAMS HT 2022; NAMS GSM 2020)
   inputs: vasomotor_symptom_burden, age
   actions: workup.menopause
   advance: Engine entered via a recognised trigger and the menopause workup launched
3. 3CONTEXT
   Capture age, years since FMP, STRAW+10 stage, uterus status, VMS/GSM burden, contraindication inventory (estrogen-sensitive cancer, prior VTE/stroke/CHD, active liver disease), VTE-risk/migraine route modifiers, current contraception/medications, and reproductive-transition status (NAMS HT 2022; NICE NG23)
   inputs: age, years_since_final_menstrual_period, uterus_present, estrogen_sensitive_cancer_history, vte_stroke_chd_history, current_medications_and_contraception
   advance: Full eligibility profile captured and mapped to the timing/route/progestogen conditionals
4. 4RED_FLAGS
   Unexplained / postmenopausal vaginal bleeding → investigate endometrial cancer BEFORE any MHT (hard stop). New focal neurology / first migraine-with-aura on estrogen, acute VTE symptoms, or breast mass → stop/withhold estrogen and evaluate. Amenorrhoea + VMS before 40 → POI pathway, not standard menopause (NAMS HT 2022; ACOG; Harlow 2012)
   inputs: unexplained_vaginal_bleeding, estrogen_sensitive_cancer_history
   actions: workup.menopause
   advance: Bleeding/cancer/VTE red flags screened and escalated if present; POI<40 routed out
5. 5INITIAL_WORKUP
   Clinical diagnosis if ≥45 y with typical VMS + cycle change — NO routine FSH (NICE NG23). Targeted labs only for mimics/eligibility: TSH (thyroid mimic), lipid (CV context for route + long-term risk), vitamin D + bone context (cross-ref osteoporosis), baseline BP, and the hormone panel only when biochemistry is actually indicated (Harlow 2012; NAMS HT 2022)
   inputs: tsh, blood_pressure
   actions: panel.tsh, panel.lipid, panel.vitamin_d
   advance: Mimics excluded, CV/bone context obtained, and clinical menopause diagnosis confirmed without unnecessary biochemistry
6. 6BRANCHING_WORKUP
   FSH/AMH ONLY in the defined scenarios: suspected POI age <40 (then route OUT — different estrogen-replacement imperative until ~51 y); post-hysterectomy (no menstrual landmark); perimenopausal contraception overlap (when to stop CHC). BAYESIAN CONDITIONAL #2 (FSH interpretability | cycle status × contraceptive use): the diagnostic value of an elevated FSH is NOT a fixed LR — it is conditionally DESTROYED by exogenous estrogen and DAMPENED by perimenopausal cycle-to-cycle variability. (a) On combined hormonal contraception, FSH is uninterpretable — exogenous estrogen suppresses pituitary FSH → LR+ collapses to ≈1 (test is non-informative; do NOT order on CHC — NICE NG23). (b) Off contraception but still cycling (late transition, STRAW −1), a single elevated FSH (>25–40 IU/L) has LR+ only ≈2–3 because FSH fluctuates wildly within and between cycles — adds little over the clinical pattern; LR rises only with persistent elevation on repeat ≥4–6 wk apart. (c) Age ≥45 y + ≥12-mo amenorrhoea: the CLINICAL diagnosis already carries LR+ ≈8–10 against the conditioned FRAME prior, so FSH adds negligible incremental value (post-test probability barely moves) — the quantitative justification for NICE NG23 "do not routinely measure FSH". FSH is decision-changing ONLY when the clinical landmark is unavailable (post-hysterectomy) or competing (POI <40). Model this as an explicit test|context conditional dependence (NICE NG23; Harlow 2012 PMID 22344196)
   inputs: fsh_amh, current_medications_and_contraception
   actions: panel.hormone, calc.us_mec
   advance: Biochemistry resolved where indicated; POI<40 routed out; contraception overlap addressed
7. 7DIFFERENTIAL
   MECE: perimenopause/menopause vs thyroid dysfunction (TSH) vs premature ovarian insufficiency (age <40 — route to POI) vs anxiety/depression (mood-predominant, VMS-poor) vs medication-induced flushing/sweats (SSRI, opioids, niacin, tamoxifen, GnRH agonist). VMS + cycle change + age ≥45 with normal TSH = menopause; isolated GSM is treated locally regardless of the systemic differential (NICE NG23; NAMS GSM 2020)
   inputs: menstrual_pattern_and_straw_stage, age
   advance: Terminal label assigned (menopause vs a routed mimic) and symptom domains prioritised
8. 8RISK_STRATIFICATION
   The MHT timing hypothesis IS the risk-stratification calculator: (<60 y OR <10 y from FMP) AND no contraindication → systemic MHT benefit/risk FAVOURABLE; (>10 y FMP OR >60 y) → UNFAVOURABLE → prefer non-hormonal. Route sub-stratification: VTE/stroke/migraine/obesity → transdermal over oral. BAYESIAN CONDITIONAL #3 (MHT benefit/risk | time-since-menopause × route — timing hypothesis modeled as a CONDITIONAL effect, not a main effect): the same drug carries opposite net effects depending on the conditioning variables. WHI CEE+MPA at mean age 63, >10 y from FMP (PMID 12117397): CHD HR 1.29 (1.02–1.63), stroke 1.41 (1.07–1.85), PE 2.13 (1.39–3.25), invasive breast Ca 1.26 (1.00–1.59), hip fracture 0.66 (0.45–0.98) — absolute excess +19/10,000-py global index. CONDITION ON AGE: the 50–59 vs 70–79 all-cause-mortality HR ratio 0.61 (0.43–0.87) (Manson 2017 PMID 28898378) is the quantitative interaction term — the strongest wired effect-size anchor (~39% relative mortality-hazard reduction by being in the early-initiation stratum). CONDITION ON TIME-SINCE-FMP (mechanistic): ELITE (Karim 2022 PMID 35474254) — early-postmenopause (<6 y) oral E2 slowed carotid IMT progression to −0.50/y (−0.82,−0.18) vs placebo −1.47/y (−1.81,−1.13), p<0.0001; late-postmenopause (>10 y) NO effect (p=0.28): the benefit is conditional on the time stratum, not a property of the drug. CONDITION ON ROUTE: transdermal estrogen carries a materially lower VTE/stroke signal than oral (no first-pass hepatic thrombotic activation) — the route conditional modifies the VTE/stroke arm of the WHI risk set. Net MHT eligibility tier is the joint posterior over {age band × years-since-FMP × route × contraindication set}, not any single variable (NAMS HT 2022)
   inputs: age, years_since_final_menstrual_period, estrogen_sensitive_cancer_history
   actions: calc.straw10, calc.us_mec
   advance: Net MHT eligibility tier (favourable / unfavourable / contraindicated) + preferred route documented
9. 9TREATMENT
   Bothersome VMS within the timing window, no contraindication → systemic MHT first-line: estrogen (transdermal preferred if VTE/stroke/migraine/obesity risk) + a progestogen MANDATORY if uterus (micronized progesterone or LNG-IUS preferred). MHT contraindicated/declined → non-hormonal ladder: fezolinetant (NK3R antagonist) → SSRI/SNRI (paroxetine 7.5 mg only FDA-approved; venlafaxine; escitalopram) → gabapentin → oxybutynin → CBT/clinical hypnosis. GSM (parallel, independent of systemic MHT): low-dose vaginal estrogen → vaginal DHEA → ospemifene → non-hormonal moisturiser/lubricant; vaginal estrogen acceptable in many breast-ca survivors WITH oncology input. BAYESIAN CONDITIONAL #4 (GSM vaginal-estrogen safety | breast-cancer status × oncology co-management × aromatase-inhibitor use): low-dose vaginal estrogen is unconditionally appropriate for moderate–severe GSM in women WITHOUT estrogen-sensitive cancer (minimal systemic absorption, no added progestogen needed — NAMS GSM 2020 PMID 32852449); but in a breast-cancer survivor the safety LR is conditional — data are insufficient to confirm safety, so the decision is gated on (a) tumour ER status, (b) concurrent aromatase-inhibitor therapy (theoretical concern that even minimal systemic estradiol antagonises near-zero AI target levels), and (c) explicit oncology co-management. Non-hormonal moisturiser/lubricant first; vaginal estrogen ONLY if refractory AND oncology-endorsed. BAYESIAN CONDITIONAL #5 (endometrial-bleeding work-up threshold | menopausal status × hormone exposure): any bleeding on/while considering systemic estrogen with an intact uterus changes the endometrial-cancer pre-test probability and the transvaginal-ultrasound endometrial-thickness threshold (post-menopausal ≤4 mm reassuring vs the irrelevance of a thickness cutoff on sequential MHT) — route to gyn.abnormal-uterine-bleeding.core.v1 with the menopausal-status + hormone-exposure context carried over. Lowest effective dose, periodic re-evaluation, shared decision-making (NAMS HT 2022; NAMS Nonhormone 2023; NAMS GSM 2020)
   inputs: vasomotor_symptom_burden, uterus_present, estrogen_sensitive_cancer_history
   advance: Symptom-domain-specific regimen initiated (systemic MHT / non-hormonal / GSM) with route + progestogen + duration plan documented
10. 10DISPOSITION
    Outpatient in essentially all cases. Refer: gynaecology/menopause specialist for complex contraindications or treatment-resistant VMS; oncology for vaginal-estrogen decisions in cancer survivors; route POI<40 to the dedicated pathway; route perimenopausal-contraception questions to gyn.contraception-management.core.v1 (NAMS HT 2022; NICE NG23)
    advance: Treatment delivered or referral booked; cross-engine routing set
11. 11MONITORING
    Re-evaluate benefit/risk at ≤3 months after starting then at least annually: VMS response, GSM response, BP, weight, breast awareness/age-appropriate mammography, any new bleeding (investigate), VTE/CV symptom screen, and a deliberate duration/de-prescribing conversation. Achieved estrogen exposure varies with BMI/alcohol/smoking — individualise (ELITE serum-E2 PMID 32925623; NAMS HT 2022). Non-hormonal: efficacy + tolerability (SSRI/SNRI ~50 % vs ~30 % placebo VMS reduction — MsFLASH PMID 31977667)
    inputs: vasomotor_symptom_burden, unexplained_vaginal_bleeding
    actions: panel.lipid
    advance: Response, safety, and a duration/de-prescribing plan reviewed at interval
12. 12FOLLOWUP
    Long-term: there is no fixed mandatory stop date — continue for documented persistent VMS with shared decision-making and periodic re-evaluation; attempt taper/withdrawal trials and reassess. Bone health → cross-ref endo.osteoporosis.core.v1 (MHT is bone-protective but not first-line solely for bone if >60 y). Perimenopausal contraception transition (still fertile until 12 mo amenorrhoea / age ~55) → gyn.contraception-management.core.v1. CV risk modification; breast-cancer surveillance; return precautions (new bleeding, breast change, focal neurology, calf swelling) (NAMS HT 2022; Heiss 2008 PMID 18319414)
    advance: Long-term duration/de-prescribing plan, bone + CV + breast surveillance, and cross-engine transitions booked