A comparison of the carbon footprint of alternative sampling approaches for cervical screening in the UK: A descriptive study

To understand whether self‐sampling can reduce carbon emissions (CO2e) from the NHS cervical screening programme (NHSCSP) by comparing the carbon footprint of three sampling strategies: routine cervical sampling, vaginal self‐sampling and first‐void (FV) urine collection.

used to express the sum of direct and indirect greenhouse gas emissions which are attributable to a given process or product, thereby representing the best estimate of a unit on climate change.
Sustainability is an important consideration for healthcare systems, with the NHS committing to a target of net zero emissions by 2045. 4 As part of a comprehensive strategy to tackle emissions, clinical guidelines increasingly account for the environmental burden of healthcare. 5For example, prescribing guidelines in primary care promote dry powder inhalers over metered dose inhalers given their equivalent effectiveness but considerable disparity in carbon footprint. 6,7he NHS Cervical Screening Programme (NHSCSP) has reduced mortality from cervical cancer by 70% since its inception. 8The current method requires people to have a cervical sample taken at speculum examination.This clinician-collected sample is tested for the causative organism, high risk human papillomavirus (HPV). 9,10Cytology is performed on virus-positive cervical samples for colposcopy triage, providing the opportunity for early intervention for cervical cancer prevention.
Participation is key to the success of the cervical screening programme. 11Screening uptake has declined in recent years, with routine screening attendance at 56.2% in 2021-2022. 12,13Barriers to screening include difficulty accessing appointments, and fear of discomfort or embarrassment associated with speculum examination. 14,154][35][36] Urine self-sampling shows higher acceptability than current cervical screening and vaginal sampling 35 and a promising test accuracy, 23,24,37 but requires further optimisation and testing.Studies in this area are ongoing, including the Alternative CErvical Screening (ACES) suite of studies in Manchester, UK (ISRCTN58683340).
An HPV-positive self-sample currently requires a clinician-collected cervical sample to enable cytology triage for colposcopy referral.In future, technologies such as HPV DNA methylation testing of virus-positive urine samples may mitigate the need for this. 38Nevertheless, self-sampling could reduce carbon emissions associated with the NHSCSP by eliminating the need for patient travel, healthcare facility resources and clinician time for the HPV-negative majority.
Environmental considerations in healthcare matter to patients, with a survey finding high public support for actions that reduce the NHS carbon footprint, including those requiring behavioural changes. 39A future cervical screening programme that provides a choice of screening options, 40 including less resource-intensive approaches, has potential to improve screening uptake while simultaneously reducing its associated carbon footprint on a population scale.
In this study we aimed to calculate and compare the carbon footprint of three cervical screening sampling approaches: a clinician-collected cervical sample, vaginal self-sampling and FV urine.

| M ET HODS
A carbon footprint analysis of three cervical screening collection methods was performed.The unit used to express carbon footprint is the carbon dioxide equivalent (CO 2 e), which harmonises the contributions of greenhouse gases with different global warming potentials.
Life cycle assessments are used to evaluate the environmental impact of a product or process over its lifespan, from production through to disposal.These were conducted in accordance with The Greenhouse Gas Accounting Sector Guidance for Pharmaceutical Products and Medical Devices. 41

| Scope and boundary setting
The system boundary was set to encompass all processes and products associated with each cervical screening sampling method, from the first invitation to screening through to preparation of samples for delivery to the laboratory (Figure 1).The functional unit was defined as one screening event: a visit to the GP for a routinely collected sample or a self-sample using either the FLOQ® swab (COPAN Diagnostics, Italy) or the Colli-Pee® 10 ml with urine conservation medium (UCM) (Novosanis, Belgium).
Production of materials for all equipment required to carry out sampling, including any associated primary packaging, was accounted for.For routine cervical sampling, clinic room products were also accounted for.Consumables such as suspension solutions and lubricating gel were excluded due to either confidentiality of constituents or lack of emission data, but their packaging was included.Emission factors accounted for raw material extraction, manufacturing of materials and transportation to the factory gate.Database emissions do not take into account the distance travelled to transport this material to the manufacturing location of the final product, or the distribution to the end user, for example to GP facilities in the UK.Emissions associated with manufacture of multi-component items beyond production of constituent materials were also excluded due to lack of availability of data, for example the assembly of the Colli-Pee® from its constituent funnel and floater.
Factors associated with delivery of care were included, such as gas and electricity, as well as business services and non-medical procurement required to run a cervical screening appointment for routine sampling.Twoway patient and staff travel to the healthcare facility was also included.Household energy was excluded from the self-sampling methods, as it was assumed that this would occur whether or not cervical screening was performed at home.
Postal journeys accounted for sending out letters of invitation to everyone eligible for routine screening.For those who choose urine or vaginal self-sampling, two further postal journeys were involved in delivery of kits and return of samples.Patient travel to post boxes for return of self-sampling kits was excluded as there are around 120 delivery points per square kilometre in the UK. 42Transport from healthcare facilities to the laboratory for the NHSCSP was omitted, as samples are transported in bulk along with all other clinical samples from primary care and hospital sites via zero tailpipe electric vans.
Determining emissions from laboratory processing was beyond the scope of this study.Analytic techniques are currently not standardised for self-sampling and may evolve to become automated if scale-up is required for future integration into the national screening programme.Furthermore, external laboratory processing for all investigations from a primary care practice accounted for 1.2% of their carbon footprint, which suggests that differences associated with alternative screening methods would have likely led to negligible differences in CO 2 e between methods. 7ata on waste disposal accounted for transportation from hospitals to disposal handling sites and energy for pretreatment and processing of waste, including high temperature incineration.

F I G U R E 1
The system boundary for carbon footprint of alternative sampling methods for cervical screening.

| Inventory analysis
We included three cervical screening sampling methods: routine clinician-collected cervical sampling, vaginal self-sampling using the FLOQswab® 43 and urine self-sampling using the Colli-Pee®. 44The carbon footprint inventory was based on consultation with stakeholders including staff running a primary care cervical screening clinic, principal investigators and project managers of trials investigating the implementation of self-sampling cervical screening methods in the UK.Materials used in a single primary care appointment were directly observed.Materials used for vaginal and urine self-sampling were observed utilising the ACES study at-home packs as a guide.
The material composition of items was determined through manufacturer information where available or through contacting companies directly.Where items were composed of more than one material but percentage breakdown was not supplied, estimations of contributing proportions were made.We weighed individual material components using the AccuWeight Digital Pocket Scale Model IC255 300 g, accurate to 2 decimal places.Multicomponent items were deconstructed before weighing.

| Impact assessment
The carbon footprint of each component for the three sampling methods was determined through the use of a combination of primary and secondary data.A bottom-up approach was applied to collection of primary data, which was used for all items specific to cervical screening.This method calculates emissions from individual components for each sampling method separately.The primary source for emission factors was the Inventory of Carbon and Energy (ICE) database (version 3), 45 which uses average data for materials supplied to the UK.Secondary data were sourced from relevant literature and the use of already published CO 2 e estimations was used where practical for nonspecific items such as personal protective equipment (PPE) and materials not available on the ICE database.
Estimated energy expenditure associated with a cervical screening appointment in a primary care facility was based on emission factors from government conversion factor data 46 applied to previously collected bottom-up data from 229 GP practices in England, across a sample of over 9000 patients. 47The sum of energy associated with a single appointment was based on the assumption that patients attend an average of 6.07 appointments per year. 48ata on travel were from the Department for Transport statistics (via the Health Outcomes of Travel Tool). 48For average staff commutes to primary care practices, annual total emissions were divided by annual total appointments.Patient travel modality was grouped into rail, bus other public transport, private transport (car), walk and cycle.Average distance travelled was multiplied by emission factors for alternative modes of transport based on government conversion factors. 46missions associated with waste disposal were calculated from data on UK hospital waste streams. 49In the primary care screening setting and following analysis of samples in the laboratory, all equipment and packaging are disposed of in clinical waste.For letters and waste components of self-sampling kits that are disposed of at home (such as the Colli-Pee apparatus), we assumed that 44.4% of these would be recycled and the rest would be disposed of in domestic waste, reflecting government data on household waste streams. 50Where secondary data were used for the routine cervical sampling method, we disaggregated data using individual study methods and included their waste disposal calculations in our totals.

| R E SU LTS
The total carbon footprint of a single routinely collected cervical sample for HPV testing within the NHSCSP was 3670 g CO 2 e.By comparison, the total carbon footprint for a vaginal self-sample was 423 g CO 2 e, and for urine 570 g CO 2 e.The breakdown of carbon footprint attributable to individual components across each of the different cervical sampling methods is presented in Table 1.
Equipment required for each sampling method is described in Table 1.For routine cervical sampling, 273 g CO 2 e was attributable to PPE 52 and cleaning materials 53 used in a primary care clinic room.The equipment required to conduct sampling and package it securely for transport to the laboratory was accountable for 296, 116 and 243 g CO 2 e for cervical, vaginal and urine sampling, respectively.Patient samples are transported in at least three components, including a primary receptacle (the sampling tube), sealable secondary packaging and cardboard outer packaging.For liquid urine samples, the secondary packaging is slightly more carbon-intensive owing to the additional requirement for absorbent material within leakproof casing. 54 total of 2768 g CO 2 e was attributable to carbon footprint associated with running a cervical screening appointment in primary care.This encompasses 584 g CO 2 e (21%) for gas and electricity, 492 g CO 2 e (18%) for essential business services and 112 g CO 2 e (4%) for other non-medical procurement.Medical equipment procurement was calculated separately.Patient and staff travel accounted for 1318 g CO 2 e (48%) and 262 g CO 2 e (9%), respectively.
The average distance travelled for a two-way trip was 6.64 km for patients and 26.72 km for staff.The most popular modality of travel was private transport (84.8% of patients and 73.5% of staff), followed by public transport (12.2% and 22.4%); relatively few travelled on foot or by bicycle (3.0% and 4.1%).
The carbon footprint associated with waste disposal in cervical, vaginal and urine sampling was 288, 136 and 156 g CO 2 e, respectively.
The carbon footprint associated with one-way postage of a letter or self-sampling kit is 25 g CO 2 e, which accounts for transport to mail distribution centres, processing and final delivery to the end user. 55Transfer of cervical samples to the laboratory involves transport from primary care practices to local hospital hubs with all other clinical samples, and onward transportation to laboratory sites via zero tailpipe electric vans.
Automated laboratory analytic steps are the same for urine and vaginal self-samples and liquid-based cytology (LBC).The pre-analytic stage which prepares samples for HPV testing is different for each sampling method.For self-samples it currently involves manual transfer into 10-ml tubes, with the additional step of re-suspension and vortexing of the dry vaginal swab in preservative liquid solution.For LBC, it involves manually loading and unloading sample pots onto a pre-analysis machine which automatically prepares up to 80 samples at once.

| Main findings
Routine cervical sampling has the largest carbon footprint of the three screening methods presented in this study, which is largely attributable to the emissions associated with running a single primary care-based appointment; 8.7-fold and 6.4fold more carbon-intensive than vaginal and urine selfsampling, respectively.The CO 2 e of urine self-sampling was slightly higher than vaginal self-sampling, owing to differences in the size of sampling devices and the layers of packaging required to transport liquid urine samples as opposed to dry vaginal samples.Even so, this difference is relatively small, equivalent to the CO 2 e of less than two antibacterial wipes.

| Interpretation
The Intergovernmental Panel on Climate Change report outlines that a limit of 2.3 tonnes of CO 2 e per capita per annum is required to keep the global temperature rise under 1.5°C in line with the 2015 Paris Climate Change Agreement commitments. 1 This equates to 6.3 kg CO 2 e per person per day, which would allow for one face-to-face cervical screen, 14 vaginal or 11 FV urine self-samples.
T A B L E 1 Carbon footprint associated with components of three alternative cervical sampling methods.

Clinician collected cervical sampling
Vaginal self-sampling FV urine self-sampling Self-sampling methods have the potential to reduce overall carbon emissions associated with cervical screening.The current method of cervical sampling resulted in 9208 tonnes of CO 2 e in the routinely screened 2021-2022 cohort in the UK (Figure 2).Studies suggest that up to 51.4% of patients in the UK may choose self-sampling if offered, with it being a particularly popular choice among irregular and never-attenders to screening. 24,56,57Preferences for self-sampling modalities within the Belgian VALHUDES study were 53% and 38% for urine and vaginal self-sampling, respectively, with 9% of participants having no preference for screening method. 35If a simple switch were to occur based on these expressed preferences holding true and those with no preference being split equally between the two groups, 4076 tonnes of CO 2 e could have been saved on cervical sampling of the routinely screened 2021-2022 cohort in England by including a choice of vaginal and urine self-sampling.
An alternative scenario would see the introduction of self-sampling increase the uptake of cervical screening among non-attenders (Figure 3).In a London primary care pilot study, 45% (292/652) of patients overdue for screening who were opportunistically offered a vaginal self-swab in a GP practice accepted and returned the swab, and 85% attended clinician sampling follow-up for a HPV-positive result. 20If an additional 45% of the unscreened routinely invited population of 2021-2022 12 were to take up self-sampling, this would increase participation rates to 76% overall and result in 880 356 more people tested in a year.If this increase occurred in combination with a choice of sampling methods with equivalent efficacy as described above, the programme would still save 2405 tonnes of CO 2 e per year, assuming an HPV prevalence of 11.5% 58,59 and an 85% attendance at clinician-collected sampling follow-up for HPVpositive self-samples.The economic benefits of self-sampling have been modelled in a similar scenario, where half of the routinely screened population choose self-sampling and uptake is increased among non-attenders.Huntington et al. found that the NHS could save £19.2million if FV urine sampling were used, or £16.5million if vaginal self-sampling were used in the NHSCSP each year. 59However, the parameters used to inform these hypothetical alternative scenarios are vulnerable to an intention-behaviour gap.The uptake of self-sampling was 8% when offered to all people eligible for screening in the Netherlands and it did not increase overall participation rates of the national screening programme, 60 although this is likely to differ between settings.
Although an increase in colposcopy services would likely result in additional carbon emissions, there may also be downstream carbon emissions mitigated by cervical cancer cases prevented with these extra screening opportunities.
Disease prevention programmes reduce carbon footprint by reducing the overall demand for healthcare. 2 For example, a hysterectomy for surgical management of cervical cancer is associated with a carbon footprint of 285-814 kg CO 2 e, 61,62 depending on approach.Preventing cervical cancer by optimising screening programme accuracy and engagement may therefore be an effective way of reducing healthcare carbon emissions associated with cervical cancer.Self-sampling could help overcome practical and patient-related barriers to screening uptake. 14,15Additionally, as we have demonstrated, the selection of less resource-intensive practices where clinically appropriate, can also aid in reduction of carbon emissions. 2n a scenario where a proportion of the screening population choose to use at-home self-sampling methods, the emissions associated with a screening appointment in primary care represent an opportunity cost; while emissions from cervical screening may decrease, it is likely that a different service will take its place.This activity could be of lower carbon intensity, for example telemedicine, or higher, such as prescription of a metered dose inhaler during a respiratory review.The emphasis going forwards should therefore be on reduction of emissions associated with the delivery of care, through reviewing procurement choices and energy sources, and facilitating 'greener' methods for patients and staff to access health facilities. 4As an adjunct to this, we propose a strategy to reduce the carbon footprint of cervical screening   with policy changes that have the potential to improve population health by providing patient choice and increasing screening coverage.

| Strengths and limitations
To our knowledge this is the first study of its kind to compare the carbon footprint of alternative equivalent public health interventions, with a view to informing health policy makers and the screening population about the impact of different options on the environment.Our methods maintained a broad scope while using maximally accurate and traceable bottom-up calculations.We compare both urine and vaginal self-sampling methods, which are approaching readiness for implementation in real-time, against our current standard of care.Our data therefore complement efficacy and health economic studies 59 to support decisions about how we can achieve net zero targets within healthcare.
As with other life cycle assessment studies, exclusions and necessary assumptions will impact on the accuracy of our data.We were also limited by our reliance on hypothesising the steps required to implement self-sampling for screening, as opposed to directly observing products and processes involved, as we could for the current NHSCSP.Furthermore, by using a functional unit of one screening event, we do not capture the waste associated with non-response or loss to follow-up; for example, self-sampling kits that are manufactured and posted but not returned, or clinic space and clinician time for primary care appointments not attended.To understand where waste occurs, future research could evaluate real-world data for how patients engage with a cervical screening programme that provides a choice of screening options.This hinges on whether self-sampling is offered at the point of invitation or to particular groups such as non-attenders.

| CONCLUSION
Our carbon footprint analysis finds that the current cervical screening sampling method is up to 8.7-fold more carbonintensive than self-sampling alternatives.This difference is mainly attributable to emissions associated with the delivery of care, including staff and patient travel to healthcare facilities.We find negligible differences in the carbon footprint of the two self-sampling methods, supporting the need for an informed choice of screening options for participants, which includes sharing information on their environmental impacts.Self-sampling methods for cervical screening may not only increase participation in the screening programme and save lives from cervical cancer, but also have a role in delivering a more sustainable healthcare system.

AU T HOR C ON T R I BU T ION S
MW led the study design and analysis with guidance from MS, and wrote the original draft of the paper.JCD provided critical input to study design and content of the first draft of the paper.AS contributed insight into laboratory processing techniques.MS provided critical input to the methods and content of the article.EJC initiated the article and oversaw the project, providing critical input on structure, content and data interpretation.All authors have reviewed the article and approved the final version for publication.

AC K NO W L E D GE M E N T S
We would like to thank the Manchester University NHS Foundation Trust Eco Team who signposted us to data and resources.

C ON F L IC T OF I N T E R E S T S TAT E M E N T
MS runs a sustainability consultancy, SEE Sustainability.His contributions to the article are his own.

DATA AVA I L A BI L I T Y S TAT E M E N T
The data that support the findings of this study are available in Inventory of Carbon and Energy database (version 3) at https:// circu larec ology.com/ embod ied-carbo n-footp rintdatab ase.html.

E T H IC S A PPROVA L
Ethical approval was not required for this study.

ORC I D
Maya Whittaker https://orcid.org/0000-0002-1752-9037Emma J. Crosbie https://orcid.org/0000-0003-0284-8630 Transport to end-user e.g.GP pracƟces in the UK PaƟent travel to post box Energy to run a GP clinic room PaƟent travel to the clinic Clinician commute to facility Return postage of kit from paƟent

F I G U R E 2
Carbon footprint associated with cervical sampling in the 2021-2022 routine call and recall screening cohort in England, aged 25-64 years.

F
U N DI NG I N FOR M AT ION JCD is supported by Cancer Research UK (CRUK) via funding to the Cancer Research UK Manchester Centre [C147/ A25254] and their Clinical Academic Training Award [C19941/A28707].EJC is supported by a National Institute for Health and Care Research (NIHR) Advanced Fellowship (NIHR300650) and the NIHR Manchester Biomedical Research Centre (NIHR203308).The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care.
Carbon footprint associated with different cervical sampling methods in an alternative scenario for the 2021-2022 routine call and recall screening cohort in England, where self-sampling is available as an option and increases uptake among non-attenders.