Why Structured Assessment Precedes Protocol Choice
For many years the reproductive-medicine literature applied the label “poor responder” to a heterogeneous mixture of patients, which made studies difficult to compare and teaching difficult to standardise. An ESHRE working group addressed this with the Bologna criteria, published in Human Reproduction, which defined poor ovarian response through a combination of advanced age, a previous poor response to conventional stimulation and an abnormal ovarian reserve test. The criteria were a step towards a common language, and subsequent commentary in the peer-reviewed literature — while noting that the resulting group remains clinically heterogeneous — treats them as the starting point for more granular frameworks.
ASRM committee guidance on ovarian reserve testing supplies the second discipline trainees need: reserve markers measure oocyte quantity and are interpreted as predictors of the ovarian response to gonadotrophin stimulation; they are not, on their own, measures of oocyte quality, and they are not a verdict on a patient's fertility. Holding that distinction firmly is what allows a trainee to use the tests for what they do well — anticipating response and informing dose and protocol — without over-reading them. The taught sequence is therefore deliberate: measure, interpret, classify, and only then select a protocol.
The Assessment Toolkit: Anti-Müllerian Hormone
Trainees are taught to build the assessment from four elements — AMH, the antral follicle count, age and the prior stimulation record — and to treat no single one of them as decisive.
AMH is produced by the granulosa cells of preantral and small antral follicles, which is why circulating levels track the size of the recruitable follicle pool. Its practical attractions are relative stability across the menstrual cycle and wide availability. Its caveats are equally practical, and the peer-reviewed literature on assay performance documents them: results differ between assay platforms and between laboratories, and levels are suppressed by hormonal contraception, so a single low value in a contraceptive user is interpreted with particular care. Trainees learn to read AMH as a continuous marker of the expected response to stimulation — consistent with ASRM guidance on what reserve tests do and do not measure — rather than as a diagnosis in itself.
Antral Follicle Count
AFC is the transvaginal ultrasound count of small antral follicles, conventionally performed in the early follicular phase. It offers something AMH cannot: a direct, per-ovary picture of the recruitable cohort at the moment of planning. It is also operator- and equipment-dependent, which is why the ultrasound literature describes standardised counting technique — systematic sweeps of each ovary, defined follicle size ranges, attention to image quality. In training, AFC is treated as a skill to be practised and audited rather than a number to be transcribed, and discordance between AFC and AMH is treated as a prompt to re-examine both.
Age and the Prior Stimulation Record
Age contributes information the biomarkers do not carry: the peer-reviewed literature on embryonic aneuploidy documents that the proportion of chromosomally abnormal oocytes rises with maternal age, so two patients with identical AMH and AFC values but different ages present different clinical problems. The prior stimulation record — protocol used, gonadotrophin dose, days of stimulation, oocytes retrieved — is the only element of the assessment that is directly observed rather than predicted, and trainees are taught to retrieve and read it in full before proposing anything new. A documented response to a known dose is a datum that no biomarker replaces.
The taught habit across all four elements is triangulation. Concordant findings make classification straightforward; discordant ones — a reassuring AFC beside a low AMH, an adequate biomarker profile beside a previous unexpectedly poor response — are precisely the cases the classification frameworks below were built to handle.
POSEIDON: From “Poor Responder” to “Low Prognosis”
The POSEIDON classification (Patient-Oriented Strategies Encompassing IndividualizeD Oocyte Number), published by the POSEIDON group in the peer-reviewed literature in 2016, reframes the field's question. Instead of asking whether a patient is a “poor responder”, it asks what kind of low-prognosis patient sits in front of the clinician, and it stratifies along two axes: age (below 35 years, or 35 and above) and ovarian reserve markers, using the group's published cut-offs of an AFC below five or an AMH below 1.2 ng/mL.
Four groups result. Groups 1 and 2 hold patients with adequate reserve markers — younger and older respectively — whose previous stimulation nonetheless produced an unexpectedly poor or suboptimal oocyte yield; the group's publications further subdivide by whether fewer than four, or between four and nine, oocytes were retrieved. Groups 3 and 4 hold patients whose reserve markers are low, again split by age. The teaching value of the framework is that it separates problems the single “poor responder” label collapses together: the unexpected under-response of a patient with normal markers raises different questions — about dose adequacy, protocol suitability and the conduct of the previous cycle — than the anticipated low yield of a patient with a depleted follicle pool.
POSEIDON also supplies a planning construct that trainees meet early: working backwards from the number of oocytes judged relevant to the individual treatment aim, as an intermediate and countable endpoint, rather than forwards from a habitual protocol. Taught carefully, this is presented as the framework's own planning logic — a way of structuring the strategy discussion — and not as a prediction or promise about any individual cycle.
Translating Classification into an Individualised Protocol
With the patient assessed and classified, protocol selection becomes an exercise in matching a finite set of adjustable variables to a defined problem. Trainees learn the variables as a list: the analogue regimen (GnRH antagonist or GnRH agonist); the gonadotrophin preparation (recombinant FSH, with or without additional LH activity from preparations such as human menopausal gonadotrophin); the starting dose and any ceiling placed on it; the choice of trigger; and cycle-level strategy — including reduced-dose approaches, dual stimulation within a single ovarian cycle (follicular-phase and luteal-phase stimulation with two retrievals), and the accumulation of oocytes or embryos across more than one cycle — all of which are described and debated in the peer-reviewed literature as strategies for the low-prognosis patient.
The discipline attached to that list is attribution. ESHRE's guideline on ovarian stimulation issues graded recommendations — each carrying a stated strength and a stated certainty of evidence — on questions including the analogue regimen, the gonadotrophin starting dose, mid-cycle dose adjustment and upper dose limits, and it appraises adjuvants (androgen pre-treatment and growth hormone among them) against an evidence base it grades in many instances as of limited certainty. The trained habit is therefore to justify each protocol choice by naming its evidence source and its certainty, to present options where the guideline presents options, and to resist converting a personal preference into a rule. Dose escalation in particular is taught as a reasoned decision made against a guideline-discussed ceiling, not a reflex response to a low biomarker.
None of this is memorisation for its own sake. The reasoning becomes durable only when it is applied case by case: given this age, this AMH, this AFC and this prior record — which POSEIDON group, which regimen, which dose, and which evidence from ESHRE, ASRM or the primary literature supports each step. The pharmacology and mechanics underlying each regimen are set out in our companion overview of ovarian stimulation protocols explained for clinicians.
Pitfalls Trainees Are Taught to Avoid
Training in reserve assessment gives as much attention to characteristic errors as to frameworks. Five pitfalls recur often enough to be taught by name.
- Reading a single AMH value as a verdict — assay variability and suppression by hormonal contraception are documented in the literature; an unexpected result prompts context-checking and, where appropriate, repetition, not an immediate change of plan.
- Conflating reserve with fertility — ASRM guidance is explicit that reserve tests measure oocyte quantity and predict stimulation response; extrapolating beyond that misleads the clinical discussion.
- Reflex dose escalation — the ESHRE guideline addresses upper dose limits directly; escalation is a decision to be argued from evidence, not a default.
- Neglecting the prior record — a documented previous response is retrieved and read, not re-derived from biomarkers.
- Presenting a preference as settled evidence — where the guideline grades certainty as low, the trainee's habit is to say so, in the notes and in the multidisciplinary discussion alike.
Developing These Skills Under Supervision
Frameworks are learned from papers; judgement is learned from cases. Assessment and protocol selection in diminished ovarian reserve become reliable when a trainee has worked through real records — discordant markers, ambiguous prior cycles, borderline classifications — with an experienced clinician checking the reasoning at each step.
The fine-ART Masterclass, a two-day case-based training programme delivered with Centrum Clinic in Ankara for physicians and embryologists, is built around that format: participants work on real cases under supervision, applying structured assessment and individualised protocol logic of the kind described here. Alongside any course, the durable habit remains the one this article has modelled — attribute every clinical statement to ESHRE, ASRM or the peer-reviewed literature, and let the evidence grade travel with the recommendation.