How are you taking part in this consultation?

You will not be able to change how you comment later.

You must be signed in to answer questions

    Guidance assessment consultation document for HTG10877 Technologies for diagnosing endometriosis in primary care

    Open for comments Open until   Request commenting lead permission

    We have updated this service so that members of the same organisation can now collaborate on a joint online response.

    Read our blog to learn more.

    3 Committee discussion

    The diagnostics advisory committee considered evidence on technologies for diagnosing endometriosis in primary care from several sources. This included evidence submitted by the companies, a review of clinical and cost evidence by the external assessment group, and responses from stakeholders. Full details are available in the project documents for this guidance.

    The condition

    3.1

    Endometriosis is a chronic inflammatory disease in which tissue similar to the endometrial lining is present outside the uterus (womb). It is estimated that 10% of women and those assigned female at birth who are of reproductive age in the UK have endometriosis. The most common symptom of endometriosis is chronic pelvic pain. Other symptoms include:

    • period-related pain that affects daily activities and quality of life

    • deep pain during or after sexual intercourse

    • period-related or cyclical gastrointestinal symptoms, in particular painful bowel movements

    • period-related or cyclical urinary symptoms, in particular pain passing urine or blood in the urine

    • subfertility or infertility associated with 1 or more of the above.

    Current practice

    3.2

    NICE's guideline on endometriosis includes the following recommendations for diagnosis and referral in primary care:

    • When endometriosis is suspected, offer an abdominal and pelvic (internal vaginal) examination.

    • Regardless of the examination findings, offer a transvaginal ultrasound to all people with suspected endometriosis.

    • If transvaginal ultrasound is declined or unsuitable, consider a transabdominal ultrasound of the pelvis.

    • Ultrasound scans should be organised by the general practice.

    • A normal abdominal or pelvic examination or ultrasound does not exclude endometriosis.

    • Pharmacological treatment can start alongside the diagnosis and referral process if endometriosis is suspected.

    3.3

    NICE's guideline on endometriosis recommends referral to gynaecology services if:

    • symptoms of endometriosis have a detrimental impact on activities of daily living

    • initial treatment is not effective, not tolerated or contraindicated

    • there are persistent or recurrent symptoms of endometriosis, or

    • there are pelvic signs of endometriosis on examination but deep endometriosis is not suspected.

    3.4

    NICE's guideline on endometriosis recommends referral to a specialist endometriosis service (endometriosis centre) if there is suspected or confirmed:

    • endometrioma

    • deep endometriosis (including that involving the bowel, bladder or ureter), or

    • endometriosis outside the pelvic cavity.

    3.5

    Specialist pelvic ultrasound or pelvic MRI (to assess the extent of deep endometriosis) may inform diagnosis in some cases. But if imaging findings are normal and endometriosis is still suspected, surgical laparoscopy should be considered to diagnose endometriosis.

    Unmet need and innovative aspects

    3.6

    There is often a long wait for endometriosis to be diagnosed. There are delays in referral pathways and long waiting times for gynaecology services. Expertise in transvaginal ultrasound in primary care varies. This may result in lower diagnostic accuracy, leading to a missed or delayed diagnosis. There is also an unmet need for alternative, non-invasive diagnostic options.

    3.7

    The technologies may enable earlier diagnosis of endometriosis by providing an additional option in primary care. They do not rely on operator expertise for accuracy. Earlier decision making in the diagnostic pathway informed by the results of these tests may reduce delays to diagnosis and avoid unnecessary investigations. Earlier diagnosis and treatment in primary care could help prevent the endometriosis from becoming more severe.

    Clinical effectiveness

    Patient perspectives

    3.8

    Submissions from 2 patient organisations were presented to the committee. Common themes were that non-invasive tests were generally more acceptable and that the technologies may enable faster diagnosis, treatment and referrals.

    3.9

    The patient experts highlighted the importance of early and accurate diagnosis of endometriosis in primary care. They stated that a presumptive diagnosis could lead to earlier access to effective treatment and improved quality of life. The technologies are not a complete solution, but they are a step towards enabling earlier diagnosis.

    Healthcare professional perspective

    3.10

    The clinical experts highlighted that increased diagnosis of endometriosis in primary care could lead to more referrals to secondary care gynaecology or specialist endometriosis services. This could increase the burden on secondary care services. But, some referrals could be avoided if a diagnosis is made in primary care and the endometriosis can be subsequently managed there.

    3.11

    Imaging does not identify all types of endometriosis. This was also highlighted by the patient experts. Additional diagnostic options in primary care could support more informed decision making, particularly when people do not want to start treatment without a diagnosis.

    3.12

    The clinical experts thought the technologies could be used in women's health hubs but highlighted that service provision varies between hubs. Also, it was not clear if the services provided by women's health hubs would be standardised.

    3.13

    The clinical experts thought it was reasonable to expect the technologies to perform similarly in detecting endometriosis in primary care and in secondary care. But differences in endometriosis prevalence between settings may affect diagnostic accuracy. Despite this, the clinical experts considered that people could benefit from a diagnosis in primary care, even if it is presumptive.

    Evidence base

    3.14

    The evidence base comprised:

    • 9 published studies, including 1 cost-effectiveness study, 1 market analysis study and 1 study reporting clinician-reported outcomes

    • 5 conference abstracts, including 1 that reported patient acceptability

    • 5 unpublished studies.

      The committee noted that the quantity and quality of studies for each technology varied. Diagnostic accuracy was the primary outcome in 16 of the studies. No studies reported any clinical outcomes, such as time to diagnosis, time to treatment, reduction in laparoscopy rates or quality of life. All peer-reviewed studies were done outside the UK.

    3.15

    The committee noted that the evidence for diagnostic accuracy for all 3 tests came from secondary or tertiary care, where the prevalence of endometriosis is higher than in primary care. There were multiple studies from different centres, as well as validation studies, for EndoSure and Endotest. There was a single-centre study with no validation for DotEndo.

    EndoSure
    3.16

    There were 10 diagnostic accuracy studies for EndoSure (2 peer-reviewed, 3 conference abstracts, 5 unpublished), with sample sizes ranging from 8 to 286. Two peer-reviewed, prospective studies (of 50 and 154 people) reported sensitivity ranging from 91% to 96% and specificity from 95% to 96%. These sensitivity and specificity values were based on AI‑derived modelling, incorporating area under the curve, symptom scores and age. The studies recruited women with suspected endometriosis who were awaiting laparoscopy in specialist centres. Both studies used diagnostic laparoscopy, with visual identification of endometriosis and histological biopsy, as the reference standard. The tests were done before the diagnostic laparoscopy. The healthcare professionals and electroviscerography technicians were blinded to each other's results and to surgical findings.

    3.17

    A published conference abstract of 100 people reported that the majority (60%) preferred EndoSure over transvaginal ultrasound. But the committee noted that the setting was a specialised endometriosis centre, not primary care.

    Endotest
    3.18

    There were 5 diagnostic accuracy studies for Endotest (3 peer reviewed, 2 conference abstracts), with sample sizes ranging from 200 to 1,000. Three peer-reviewed, prospective studies (of 200, 200 and 971 people) reported sensitivity ranging from 96% to 97.3% and specificity from 94.1% to 100%. The studies recruited women with signs and symptoms suggestive of endometriosis. All studies used diagnostic laparoscopy with histology or MRI as the reference standard. The external validation study used transvaginal ultrasound as a reference standard, as well as diagnostic laparoscopy or MRI. Saliva samples were collected before the MRI or diagnostic laparoscopy, and microRNA signature was assessed blinded to the surgical or imaging findings.

    3.19

    The committee acknowledged that a peer-reviewed survey of 141 German gynaecologists found limited support for using Endotest. But only 15% of those surveyed had actually used the test.

    DotEndo
    3.20

    There was 1 peer-reviewed diagnostic accuracy study for DotEndo. This prospective study of 100 women reported sensitivity of 83% (specificity‑optimised threshold) and 90% (balanced threshold), and specificity of 96% (specificity‑optimised threshold) and 90% (balanced threshold). The study recruited women with symptoms suggestive of endometriosis. Diagnostic laparoscopy, with visual confirmation and histological pathology, was used as the reference standard. Serum samples were collected before the diagnostic laparoscopy, and microRNA analysis was done blinded to the surgical findings.

    Cost effectiveness

    Conceptual model

    3.21

    The external assessment group developed a de novo health economic model in which the tests were positioned in primary care after a negative or inconclusive ultrasound, or when ultrasound was declined. It compared the technologies with current care in the NHS. A Markov model was used to represent the care pathway from initial presentation through to diagnosis and management. Decision trees were used to determine outcomes for people receiving either a single diagnostic test or multiple diagnostic tests, applied to the relevant proportions of the population. The model used a 60‑year time horizon and monthly cycles.

    Model parameters

    3.22

    The key clinical inputs to the model were the diagnostic accuracies of the technologies, informed by the most comprehensive evidence available for each technology.

    3.23

    The external assessment group explained that health-state utility benefits from having a presumptive or definitive diagnosis after a positive test result were based on assumptions because there was no available evidence. Baseline utility was derived from a French tertiary-care study, and utility changes associated with diagnosis were based on symptom severity rather than the effect of receiving a diagnosis. The committee recalled that having a diagnosis would be beneficial to the person, even if it is presumptive.

    Model results

    3.24

    The external assessment group's assessment showed that all technologies were cost effective at a willingness-to-pay threshold of £20,000 per quality-adjusted life year (QALY) gained in both the deterministic and probabilistic base cases and over the 60‑year time horizon. The committee noted that the incremental QALYs (0.10 to 0.11) and net health benefit (0.07 to 0.11) appeared modest and were not realised in the short term. All the technologies also reduced the time to presumptive diagnosis compared with current practice (reduction of 1.44 years to 1.82 years) and the time to definitive diagnosis by 0.97 years to 1.22 years.

    3.25

    Results from one‑way sensitivity analyses and scenario analyses changed the incremental cost‑effectiveness ratios but not the overall interpretation. Lower prevalence of endometriosis was identified as a key cost driver, with lower prevalence associated with higher costs across all 3 technologies. Test specificity was a key driver of QALYs, with a lower specificity leading to a higher QALY gain in the model. The EAG explained that a lower specificity will produce more false-positive results. In the model, people with a positive test result were assumed to have a presumptive diagnosis, which is associated with higher utility than having no diagnosis. The committee discussed this assumption as a limitation.

    Equality considerations

    3.26

    The committee was aware of potential equality issues related to endometriosis diagnosis, use of services and access to tests in certain population groups. These include:

    • people with female reproductive organs who do not identify as women, including trans men and non-binary people

    • young people (aged 12 to 17)

    • women, trans men and non-binary people:

      • from ethnic minority backgrounds

      • who have a learning disability, or

    • who are uncomfortable with invasive diagnostic testing.

      No evidence was identified that addressed any of the equality issues.

    3.27

    The committee was also aware of geographical variations in service delivery.

    3.28

    The committee noted that some populations may particularly benefit from using the technologies, including:

    • women, trans men or non-binary people who find transvaginal ultrasound unacceptable, or for whom transvaginal ultrasound is not suitable

    • women, trans men or non-binary people who do not have access to healthcare professionals in primary care with high levels of expertise in transvaginal ultrasound.