A retrospective cohort study investigating reintervention rates and pregnancy rates in women undergoing myomectomy and uterine artery embolisation in 2010-2014 in England

****AUG 2022: Imperial College London are requesting a 6-month extension to this Data Sharing Agreement to allow time for continued processing of NHS Digital data. Although v0 was signed off in August 2021 data was not released until March 2022 and thus Imperial College London have had insufficient time to conduct adequate processing of the data for the purposes outlined under this Agreement.****

Uterine fibroids have a high prevalence in women and cause considerable morbidity for many millions of women around the world. Women with uterine fibroids may experience problematic abnormal uterine bleeding, resulting in attendances to primary or secondary care health facilities. Myomectomy (surgical removal of fibroids) and uterine artery embolization are two fibroid uterine-sparing treatment options. As women opt to delay childbearing, there have been trends towards uterine-sparing treatment for fibroids. However relatively little is known regarding the medium-term outcomes after these treatments. The impression relating to health-related quality of life of uterine fibroids has been poorly characterised in studies. It is expensive and difficult to design studies of appropriate design and length of follow up after treatment which do not result in significant numbers of participant attrition (participant attrition means there will be participants lost to follow up so participants will leave the study by undergoing hysterectomy as a reintervention). It is therefore not well understood regarding medium-term outcomes following treatment for uterine fibroids. Imperial College London acknowledges that a number of these procedures are carried out within the private healthcare sector, and this will not be captured as part of the study. Although the initial cohort will not include those undertaking their index procedure in the private sector, any follow-up reintervention of the NHS cohort in the private sector in the follow-up period will be included.

Imperial College London are seeking NHS Digital data to investigate a large sample of women in a cohort size of 20,000 aged 18 years old and above undergoing uterine-sparing treatment for uterine fibroids and improve understanding regarding short term readmission, and reintervention rates, pregnancies rates and leiomyosarcoma diagnosis rates over 10 years. Imperial College London will process this data to compare medium-term and long-term outcomes for these two very important treatments. This data will be highly generalisable to future women considering to undergo these treatments.

Imperial College London is requesting up to 10 years of follow-up data to allow improved understanding regarding longer-term outcomes following fibroid treatments. There is currently a paucity of studies that have investigated these outcomes in the medium and long term. However, it is important that these outcomes are investigated as fibroids are slow-growing and therefore recurrence may take time after the initial index procedure, with the gradual development of symptoms and subsequent delayed presentation by a patient with fibroids for reintervention. Additionally, pregnancy and adverse cancer diagnosis data may not be available in the short term. By investigating longer-term outcomes it is likely that this study may allow a better understanding of the experience of the woman with fibroids.

This study will involve a retrospective cohort study investigating reintervention rates and pregnancy rates in women undergoing myomectomy and uterine artery embolisation between 2010 and 2015 in England. Specified procedure, operation, diagnosis and pregnancy codes will be searched for over a 10 year period for this cohort to determine surgical reinterventions, pregnancies and leiomyosarcoma diagnoses. These will be determined from NHS Digital Health Episode Statistics Admitted Patient care data sets over 10 years. The record level data will help to understand the real-world fibroid patient pathway and healthcare interventions over a time period given the chronicity of the condition.

Imperial College London require data on a national level in an effort to gain external generalisability for fibroid treatment outcomes, and to assess outcomes in a real-world setting. Previous high-quality fibroid research has used only small data sets and been limited to specialist tertiary centres and therefore has limited generalisability. This is unsatisfactory given the high prevalence of the condition. This data will contribute to auditable standards. In addition, Imperial College London will investigate whether there is equity in access to treatment, particularly minimally invasive treatment from a geographical perspective in England.

One other objective of the study is to identify whether outcomes vary according to surgeon/commissioners/volume/caseload of healthcare services/ethnicity. Imperial College London will develop study outputs which will look at variability of outcomes based on ethnicity as well as geographical area. This might inform a need nationally for the development of specialist centres for treatment of fibroids. There is a need for up to 10 years of follow up data as there is a paucity of data focusing on robust long term outcomes in previous research; fibroids are a chronic disease and slow-growing and therefore it might take time for them to grow following the index treatment and pregnancy outcomes are being investigated.

NHS Digital will identify a cohort using specified procedure codes for this study. Baseline characterisation fields extracted will include:

- Age,

- Parity,

- Ethnicity,

- Sociodemographic status,

- Index treatment (myomectomy or uterine artery embolisation),

- Geographical location in England,

- Co-morbidities.

Details regarding index treatment will include:

- Operation date,

- Postoperative stay,

- Treating commissioners of healthcare services.

Date and primary diagnosis of readmissions under gynaecology within 90 days of index treatment will be extracted. Surgical reintervention (for 21 specified surgical codes) will be collected for 10 years of follow up after the index treatment. Pregnancy outcome, gestational age at birth, mode of birth and birth weight be extracted for all pregnancies up to 10 years after index treatment. The 10 year follow up is limited by funding and would seek to extend this further if required to enhance this research.

This study may provide increased understanding regarding reintervention rate in women undergoing myomectomy or uterine-artery embolisation over 10 years. Both myomectomy and uterine artery embolisation are non-definitive and are associated with risk of fibroid recurrence and treatment failure. This may result in problematic symptoms and impact on health-related quality of life and considerable cost to the taxpayer. Technological advancements in both treatments may mean that the risk of reintervention quoted from Randomised controlled trials (RCT) investigating treatments in the ‘learning curve’ period of their use are no longer applicable. Additionally, RCT findings that are often performed in specialist centres may not be generalisable to all centres. Furthermore, it is difficult to follow-up women for the appropriate amount of time to understand reintervention rates in the medium and long term because of study cost and staffing limitations. NHS Digital data in terms of reintervention could allow patients and clinicians to understand outcomes better. It will provide data on non-emergency hysterectomy rates over 10 years in a large cohort of women in England, but importantly it may improve understanding of the rates of women undergoing multiple uterine-sparing treatments, of which there is sparse data. It will importantly allow a comparative assessment of both uterine-sparing treatments, of which there is sparse data. Reintervention data may also be useful in understanding the overall cost-effectiveness of treatments. A subgroup analysis will be undertaken in women under 40 given that rates of reintervention may differ than in older women.

Imperial College London are the sole Data Controller who also process data. Imperial College London have no commercial agenda for this data.

Imperial College Healthcare NHS Trust are a funding body only, who will not determine the purpose or means of the processing of the data nor will access, store or process any NHS Digital data.

Imperial College London processes this data under Article 6(1)(e) and Article 9(2)(j) of the General Data Protection Regulation, both relating to use of data in the public interest. The work is in the public interest as it may help improve patient management pathways for a highly prevalent condition. It may improve pre-procedure counselling. In England the direct total costs for fibroid related admissions is £86 million a year. This data is likely to help towards greater cost effectiveness for the NHS regarding these treatments, where resources are likely to be limited.

This study represents a standalone project.

Imperial College London are investigating a cohort of women undergoing myomectomy or uterine artery embolisation (UAE) between 2010 and 2015. This cohort has been identified by NHS Digital using a mixture of procedure and diagnosis codes. Imperial College London are interested in women of premenopausal age as this is the age group most affected by fibroids. It is anticipated to be useful to investigate outcomes in older women as these women may be at a higher risk of developing cancer.

Imperial College London are investigating whether the following surgical interventions took place within 10 years of the index treatments for each cohort individual. Imperial College London are also hoping to investigate pregnancy outcomes, looking at each pregnancy episode up to 10 years after the index treatment. As myomectomy and UAE are non-definitive treatments which presents risk of developing uterine cancer. Imperial College London are also analysing any diagnosis of uterine cancer up to 10 years from the index treatment.

The pseudonymised HES extracts are stored in a research database where researchers and Imperial College London statisticians can access the data for analyses. Descriptive statistics will be undertaken. For non‐normally distributed continuous variables, results will be described using medians and interquartile ranges. For categorical baseline characteristics, the numbers and percentages in each group will be described. Analysis will be done using multivariate logistic regression. Chi squared test of independence will be used to compare proportions between cohorts. Survival analyses will be created for reintervention.

All the researchers working on the data are substantive employees of Imperial College London. The research database is hosted in an Imperial College London managed certified secure environment. Registered users can only access the environment using registered terminals or two-factor authentication.

Users accessing the data have completed the mandatory training:

• Data Protection Awareness

• Information Security Awareness

• NHS Digital Data Security Awareness Level 1

There is no requirement by Imperial College London to identify individuals using the HES data and no attempt will be made to identify individuals.

No publications/outputs from Imperial College will present data which allow the identification of individuals. All data presentation is based on groups of subjects (often considerably large numbers). Outputs will involve aggregate data with small numbers suppressed in line with the HES Analysis Guide.

No record level data will be transferred outside of England and Wales under this Agreement.

Published work describing the readmission rates, reintervention rates, pregnancy rates and leiomyosarcoma rates is expected to be completed by Imperial College London and submitted to the American Journal of Obstetrics and Gynaecology.

Following publication Imperial College London aims to collaborate with the patient group British Fibroid Trust by providing a layperson summary of research findings. Additional work investigating the variation in outcomes after UAE and myomectomy relating to ethnicity is to be submitted to the Fertility Sterility Journal.

Work investigating the relationship between surgeon caseload/volume and commissioners of healthcare services caseload/volume and subsequent readmission and reintervention rates after myomectomy and uterine artery embolisation is to be submitted to the British Journal of Obstetrics and Gynaecology.

Additional work reporting on the variation in access to fibroid treatment according to geographical location is expected to be submitted to the British Journal of Obstetrics and Gynaecology.

Work is hoped to be further disseminated via national conferences including the British Society of Gynaecological Endoscopy, European Society of Gynaecological Endoscopy and American Association of Gynaecology Laparoscopists.

It is anticipated this data may inform future studies regarding cost-effectiveness.

Summary reports of the work and research undertaken are anticipated to be compiled and made available on the Imperial College London website.

All outputs will be aggregated with small number suppression in line with the HES Analysis Guide.

The first results for this study are expected within a year of the download of the NHS Digital data. This study may lead to recommendations which may have an impact on the following categories:

1 Clinicians

If the data demonstrates differences in short-term and long term outcomes between the two important treatments, uterine artery embolization and myomectomy, then this may lead to improved patient care by directing patients towards the treatment with the more favourable outcomes in an evidence-based manner. Clinicians may have robust data to identify best practice and to inform recommendations for fibroid treatment. This work may help to understand whether certain treatments are associated with an increased risk of hospitalisation, readmission, reintervention, pregnancy and diagnosis of uterine cancer. There may be improved pre-procedure counselling. This work may also lead to the development of specific treatment algorithms and care pathways to help guide clinicians in their recommendations. Alternative treatment modalities and optimal care can be planned to minimise adverse outcomes. This work may also be generalisable to other high-income countries.

This study may provide data on non-emergency reintervention hysterectomy rates over 10 years in a large cohort of women in England, but importantly it may allow improved understanding of the rates of women undergoing multiple uterine-sparing treatments, of which there is sparse data. It could importantly also allow a comparative assessment of both uterine-sparing treatments, of which there is sparse data.

This study may provide increased understanding regarding subsequent rates of leiomyosarcoma malignancy in women undergoing myomectomy or uterine-artery embolisation. Unfortunately there is no current effective and accepted method of diagnosing leiomyosarcoma pre-procedurally. There is significant overlap with how these cancers appears on imaging with fibroids. As such, for any women undergoing a non-resective or uterine-sparing treatment of presumed fibroids, there is a risk of undiagnosed sarcoma and potentially worse outcomes associated with missed or delayed diagnosis. Furthermore, there are concerns regarding upstaging an undiagnosed sarcoma with power morcellation during myomectomy which led to a ban on this technology in 2014. The prevalence of leiomyosarcoma varies with age, but there is poor understanding and poor consensus regarding overall rates. As it is likely an uncommon outcome in women of premenopausal age, it is likely to be met with relative rarity in randomised controlled trials. By undertaking this large cohort study, there may be improved understanding of rates of diagnosis after these fibroid treatments.

This study may provide increased understanding regarding pregnancy outcomes following myomectomy and uterine artery embolisation. Uterine fibroids are common in women of reproductive age and therefore impact of treatments on fertility are an important consideration. Myomectomy is often reserved for women intending on fertility conservation, due to theoretical concerns that embolisation affecting the endometrium or ovarian vascular supply might lead to adverse fertility or pregnancy outcomes. Given these demonstrated effects, it may be difficult to design studies investigating fertility and pregnancy outcomes in randomised controlled trials or randomized prospective studies because of ethical considerations. It is important to more fully elucidate such outcomes as uterine artery embolisation may have some demonstrable benefits when compared to myomectomy such as quicker return to normal activities and reduced risk of major complications such as blood transfusion or visceral injury. A small number of women may undergo pregnancy after UAE. This nationwide cohort study may help to capture data relating to pregnancy outcomes in such women to provide greater understanding.

In addition, this work may lead to the development of auditable standards as it may enable clinicians to compare local outcomes against national benchmarks and then improve care.

2 Patients

Through collaboration with the patient group British Fibroid Trust, Imperial College London will provide a layperson summary of research findings which may be accessed by women with fibroids. Currently little has been published regarding long term outcomes following these fibroids treatments and this data may allow patients to estimate risk of reintervention before undergoing these treatments. This work may increase satisfaction for patients as it may allow patients to understand their potential patient treatment pathway with increased clarity and improve expectations. Patients may have improved understanding of short term and medium-term outcomes, so that they can be better informed as to whether they wish to proceed with treatment. As fibroids are a serious, but benign condition, the risk-benefit profile of myomectomy and uterine artery embolization may be better understood. Patients may derive significant benefit by avoiding treatments that have a worsened outcome profile. On the other hand, patients may be reassured by more favourable risk profiles then anticipated if both treatments are shown to have low risk of readmission or reintervention. Women with uterine fibroids have been demonstrated to have reduced health-related and quality of life, and improved understanding of the treatments and associated risks may help to alleviate some concerns and anxieties. Identifying the potential risk of readmission may help patients to appropriately plan and manage their resources relating to employment and childcare.

3 Impact of uterine fibroid treatment provision

If patients are guided against treatments with increased reinterventions after recommendations from this study, then this may reduce potential future use of services for women with fibroids who would otherwise experience symptom or fibroid recurrence after index treatment. This then frees up access for other women requiring use of gynaecological services.

4 Commissioners

This work may lead to improved financial planning. There may be more robust data on hospital activity workload relating to fibroid treatment and NHS costs associated. Reintervention data could also be useful in understanding the overall cost-effectiveness of treatments. This may generate important knowledge that could inform the allocation of resources for fibroid treatments leading to better commissioning and value for money.

There may also be improved understanding as to whether there is equity in access to treatment, particularly minimally invasive treatment from a geographical perspective in England. This could enable commissioners of healthcare services commissioners to identify patients with the highest risk of poorer outcomes – leading to better care and better delivery of services. This study may provide increased understanding regarding potential variations in outcomes following myomectomy or UAE relating to ethnicity. A number of studies have demonstrated adverse healthcare outcomes following fibroid treatment in Black and ethnic minority women in the US. This may relate to a number of different factors including varying severity of fibroid burden across ethnicities, varying associated co-morbidities across ethnicities and systemic racism. This data may help to understand if outcomes such as reintervention rates and leiomyosarcoma rates vary according to ethnicity in the UK

The data may also identify whether outcomes vary according to surgeon/commissioners of healthcare services volume/caseload. This might inform a need nationally for the development of specialist centres for treatment of fibroids.

5 Impact on the NHS and society

Any recommendations for change resulting from this study may therefore be evidence-based and considered from both the health benefit and cost benefit viewpoints, whilst attending to the needs of women with fibroids. Uterine fibroids are a highly prevalent condition and a substantial proportion of women will seek treatment for fibroids, therefore this data is likely to have significant impact on a national level within the field of gynaecology.

The study has the potential to reduce the direct and indirect costs relating to uterine fibroids by increased understanding of treatments, since fibroids exert significant effect on quality of life, absenteeism, presenteeism and social interactions.

No benefits have been yielded as Imperial College London are still processing the NHS Digital data disseminated. As noted in section 5.d.ii results are expected within a year of the download of NHS Digital data. The HES APC data was only disseminated in March 2022.