Several conditions that predominantly affect women have long average diagnostic delays. Studies of endometriosis in particular have found average intervals between symptom onset and diagnosis measured in years across multiple national samples, with the longest delays in patients whose symptoms began in adolescence.
Similar patterns appear for a number of autoimmune conditions and chronic pain syndromes.
Understanding how a delay of that length is produced is more useful than attributing it to any single failure, because no single failure of that magnitude occurs. It is accumulated.
Stage one: the patient does not present
Delay begins before any clinical contact.
Where a symptom is believed to be normal, it is not reported. The normalisation of menstrual pain discussed elsewhere in this section produces years of delay before a first appointment in a substantial proportion of cases.
Adolescents are also less likely to initiate healthcare contact independently, and where a parent shares the belief that the symptom is normal, the appointment is not made.
Studies asking patients retrospectively when symptoms began consistently find substantial intervals before first presentation.
Stage two: the presentation is attributed
At first presentation, a symptom that is common, non-specific and consistent with normal physiology receives the most probable explanation.
This is correct clinical reasoning. Most adolescents with pelvic pain do not have endometriosis. Most adolescents with fatigue do not have an autoimmune condition. A clinician applying base rates will be right most of the time.
The problem is not the initial attribution but what follows it. If the attribution is made and no review is scheduled, the patient with the less probable condition has no route to reconsideration except making another appointment and starting again — frequently with a different clinician who has no record of the first consultation's reasoning.
Stage three: the psychological attribution
Where symptoms persist and investigations are normal, an attribution to stress or anxiety is common.
This is where the evidence on differential treatment is most concerning. Studies of pain assessment have repeatedly found that women's pain reports are more likely than men's to be attributed to psychological causes, and this pattern has been found in both simulated and real clinical scenarios.
The attribution has a particularly damaging property: it is unfalsifiable within a consultation, and it changes how subsequent presentations are received. Once anxiety is recorded in a file, later symptoms are read through it.
It is worth being fair here. Anxiety does produce physical symptoms, the attribution is frequently correct, and clinicians face genuine diagnostic uncertainty. The problem is the asymmetry in how readily it is reached and the absence of a mechanism for revisiting it.
Stage four: the specialist threshold
Referral to specialist care requires a referring clinician to judge that the probability of significant pathology justifies it.
Thresholds vary by system, by clinician and by resource availability. Where waiting lists are long, thresholds rise, and conditions requiring specialist diagnosis are systematically delayed.
For endometriosis specifically, definitive diagnosis has historically required laparoscopy — a surgical procedure with its own threshold. Imaging has improved and guidelines have moved towards clinical diagnosis, but the historical requirement contributed substantially to the delay figures.
What the delay costs
Beyond the years of untreated symptoms, delay has specific consequences.
For progressive conditions, disease advances during the interval. For endometriosis this may include effects on fertility, which is one of the more consequential outcomes of late diagnosis.
Chronic pain that persists untreated is associated with central sensitisation — changes in pain processing that make the pain harder to treat subsequently. Early treatment is not only earlier relief; it may alter the trajectory.
Educational and occupational effects accumulate. Years of unmanaged symptoms during secondary and tertiary education have consequences that do not resolve when the diagnosis eventually arrives.
And there is a documented effect on subsequent engagement with healthcare. Patients whose symptoms were repeatedly dismissed report reduced willingness to present with later symptoms.
What reduces delay
Clinical guidelines recommending diagnosis on symptom pattern rather than requiring surgical confirmation have shortened pathways where implemented.
Structured symptom questionnaires at first presentation improve detection compared with open consultation, because they ask about symptoms patients do not spontaneously report.
Scheduled review rather than open-ended reassurance addresses the stage-two problem directly. "Come back in six weeks if this has not settled" creates a route that otherwise does not exist.
Education aimed at both patients and primary care about what is not normal has been the focus of several campaigns, with mixed evaluation but sound logic.
What a patient can do
Keep a dated record. A symptom diary covering several months is the single most useful thing to bring to a consultation, because it converts a vague history into a pattern.
Ask for the differential explicitly: what else could cause this, and what would distinguish them. This is a collaborative question rather than a challenge, and it frequently produces a more careful answer.
Ask for the reasoning to be recorded, including a review point. A recorded plan with a return date is much harder for the system to lose than an informal reassurance.
Where a psychological attribution is made, ask what has been excluded. Anxiety and physical pathology are not mutually exclusive, and treating one does not require abandoning investigation of the other.
And bring someone. Consultations attended by a second person produce more complete histories and more follow-up, which is unfair but true.