DOI: 10.1111/1471-0528.

16668 Original Article

www.bjog.org
General gynaecology

Effectiveness, safety and acceptability of no-test

medical abortion (termination of pregnancy)
provided via telemedicine: a national cohort
study
ARA Aiken,a PA Lohr,b J Lord,c N Ghosh,d J Starlinge
a
LBJ School of Public Affairs, University of Texas at Austin, Austin, TX, USA b British Pregnancy Advisory Service, Stratford upon Avon, UK
c
MSI Reproductive Choices, London, UK d National Unplanned Pregnancy Advisory Service (NUPAS), Birmingham, UK e Mathematica
Policy Research, Cambridge, MA, USA
Correspondence: J Lord, MSI Reproductive Choices, 1 Conway Street, Fitzroy Square, London W1T 6LP, UK. Email: [email protected]

Accepted 9 February 2021. Published Online 24 March 2021.

This article includes Author Insights, a video abstract available at https://vimeo.com/bjog/authorinsights16668

Objective To compare outcomes before and after implementation P = 0.557) and incidence of ectopic pregnancy (0.2% versus
of medical abortion (termination of pregnancy) without 0.2%, P = 0.796) were not different between models. In the
ultrasound via telemedicine. telemedicine-hybrid model, 0.04% were estimated to be over
10 weeks’ gestation at the time of the abortion; all were
Design Cohort analysis.
completed safely at home. Within the telemedicine-hybrid
Setting The three main abortion providers. model, effectiveness was higher with telemedicine than in-person
care (99.2% versus 98.1%, P < 0.001). Acceptability of
Population or sample Medical abortions at home at ≤69 days’
telemedicine was high (96% satisfied) and 80% reported a
gestation in two cohorts: traditional model (in-person with
future preference for telemedicine.
ultrasound, n = 22 158) from January to March 2020 versus
telemedicine-hybrid model (either in person or via telemedicine Conclusions A telemedicine-hybrid model for medical abortion that
without ultrasound, n = 29 984, of whom 18 435 had no-test includes no-test telemedicine and treatment without an ultrasound
telemedicine) between April and June 2020. Sample (n = 52 142) is effective, safe, acceptable and improves access to care.
comprises 85% of all medical abortions provided nationally.
Keywords Abortion, induced [E04.520.050], ambulatory care
Methods Data from electronic records and incident databases facilities [N02.278.035], health planning [N03.349], mifepristone
were used to compare outcomes between cohorts, adjusted for [D04.210.500.365.415.580], misoprostol [D23.469.700.660.500],
baseline differences. pregnancy complications [C13.703], telemedicine
[N04.590.374.800], termination of pregnancy.
Main outcome measures Treatment success, serious adverse
events, waiting times, gestation at treatment, acceptability. Tweetable abstract Compelling evidence from 52 142 women
shows no-test telemedicine abortion is safe, effective and improves
Results Mean waiting time from referral to treatment was
care.
4.2 days shorter in the telemedicine-hybrid model and more
abortions were provided at ≤6 weeks’ gestation (40% versus Linked article This article is commented on by A Glasier and
25%, P < 0.001). Treatment success (98.8% versus 98.2%, L Regan, pp. 1475–1476 in this issue. To view this mini
P > 0.999), serious adverse events (0.02% versus 0.04%, commentary visit https://doi.org/10.1111/1471-0528.16740.

Please cite this paper as: Aiken ARA, Lohr PA, Lord J, Ghosh N, Starling J. Effectiveness, safety and acceptability of no-test medical abortion (termination of
pregnancy) provided via telemedicine: a national cohort study. BJOG 2021;128:1464–1474.

healthcare systems and the women who use them. There is

Introduction
clear evidence that restricting access to abortion does not
Improved access to care for induced abortion (termination reduce abortion rates, it simply makes the procedure less
of pregnancy) would deliver significant advantages for both safe.1,2 Improving access is likely to benefit those who are

1464 ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited.
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Effectiveness, safety and acceptability of no-test medical abortion

most vulnerable,3 especially in resource-poor settings or medications to terminate a pregnancy without primary sur-
where care has to be self-funded. In its 2019 guideline on gical intervention and ‘EMA’ applies to these procedures,
abortion care, the National Institute for Health and Care commonly within the first trimester.13 The recommended
Excellence (NICE) stated that improving access to abortion EMA regimen uses the anti-progestogen mifepristone in an
was a key priority.4 oral dose of 200 mg, then, after 24–48 hours, 800 mcg of
Telemedicine, the use of information and communica- the prostaglandin analogue misoprostol by the sublingual,
tion technologies to improve patient outcomes by increas- vaginal or buccal route. In the UK, an additional dose of
ing access to care and medical information,5 has been misoprostol is recommended if expulsion has not occurred
noted to decrease costs and increase convenience and 3–4 hours after the first dose.8,14 Prior to telemedicine,
safety.6 It is an established service delivery model for abor- patients would have returned to clinic to receive this, but
tion care in many settings7 and it is recommended to in the new model the additional dose was provided in the
improve access.4 The COVID-19 pandemic required urgent treatment pack. Patients are advised to call the abortion
action to ensure delivery of essential health services, with provider on their 24-hour phone service in the event of
the Royal College of Obstetricians and Gynaecologists any problems, if bleeding is unexpectedly light or heavy, if
(RCOG) publishing guidelines to safeguard abortion care pregnancy symptoms fail to resolve quickly or if a low-sen-
in the UK.8 These guidelines profoundly changed the way sitivity pregnancy test (1000 IU) is positive 3 weeks after
medical such care is delivered in Great Britain. Prior to the using misoprostol. This information is conveyed in a vari-
emergence of COVID-19, all patients seeking medical abor- ety of formats, including verbally at consultation, in writing
tion were required to attend in-person to receive an ultra- and through on-line resources. Women are offered self-
sound scan and have mifepristone administered within the assessment with a low-sensitivity pregnancy test to deter-
clinic. Under the new guidelines, consultations were mine success of the abortion, with instructions to report
encouraged to take place by telephone or video call; an back to the abortion provider if there are any ongoing
ultrasound scan was required only if indicated. By 30 issues or a positive test, in line with national guidelines15
March 2020, all the governments in Great Britain had and high quality evidence that self-assessment is safe, effec-
issued emergency legal orders to allow mifepristone to be tive and preferred by women.16
used at home along with misoprostol up to 10 weeks’ ges- Our dataset consisted of information on EMAs extracted
tation.9–11 These approvals permitted abortion providers to directly from each provider’s electronic records and
implement a fully telemedical service delivery model, included fully de-identified patient clinical and demo-
including ‘no test medical abortion’ and direct-to-patient graphic characteristics. Each providers’ clinical incident
delivery of abortifacient medications. database was cross-referenced to patients in each cohort to
Great Britain’s new pathway for no-test medical abortion determine rates of unsuccessful medical abortion and
is unusual among existing models because it is fully remote: adverse events. All data were extracted 6 weeks after the
no clinic visit, tests or ultrasound scan are performed and end of the study period to ensure the reporting of compli-
both mifepristone and misoprostol are delivered by mail or cations was as complete as possible. We also consulted with
collected from a clinic for use at home. This new service regulators and national agencies to ensure that we
model thus presents an important opportunity to evaluate a accounted for incident reports made directly to them. The
potentially better way to provide medical abortion that independent regulator of all health and social care services
could improve access and reduce the barriers posed by in- in England, the Care Quality Commission (CQC), con-
person care.12 We examined and compared the effectiveness, firmed that all cases reported directly to them through var-
safety and acceptability of medical abortion provided up to ious routes, for example, statutory notifications and the
10 weeks’ gestation before and after the widespread imple- central NHS database of patient safety incident reports (the
mentation of no-test telemedicine. National Reporting and Learning System [NRLS] and the
Strategic Executive Information System [StEIS]), were
known to the providers.
Methods
Two cohorts were defined. The ‘traditional’ cohort com-
Population and cohorts prises all patients having an EMA between 1 January and 1
The study population comprised all patients who accessed March 2020, prior to service model change. All patients in
an early medical abortion (EMA) at the three largest abor- this cohort received an in-person assessment and an ultra-
tion providers in England – British Pregnancy Advisory sound scan, had mifepristone administered in the clinic,
Service (BPAS), MSI Reproductive Choices (MSUK) and and were supplied with misoprostol for use at home. The
the National Unplanned Pregnancy Advisory Service ‘telemedicine-hybrid’ cohort comprises all patients access-
(NUPAS) – 2 months before and after the service model ing an EMA between 6 April and 30 June 2020, in a 2-
change. Medical abortion is defined as the use of month period after the service model change at each

ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd. 1465
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Aiken et al.

provider. Patients in this cohort were offered a consultation although a proportion of these will be pregnancy of
via phone or video call, during which an assessment of eli- unknown locations (PULs), which includes failed early
gibility for treatment via telemedicine was made. Patients intrauterine pregnancies. We analysed the proportion of
were deemed eligible for no-test medical abortion via tele- cases where treatment was reported to have occurred at
medicine if they had a low risk of ectopic pregnancy and ≥10 weeks’ gestation in the telemedicine-hybrid cohort.
their self-reported last menstrual period (LMP) indicated a
gestation of <10 weeks. Medications were then delivered to Acceptability
patients via post or were made available for collection from Given the constraints of delivering healthcare during
a clinic for use at home. Those deemed ineligible for no- COVID-19, it was not possible to follow up all patients to
test medical abortion via telemedicine had an in-person capture patient-reported outcomes. Two of the providers
assessment with ultrasound as per the ‘traditional’ model. (BPAS and MSUK) collected patient feedback during the
After the in-person assessment, medications for this group study period. Patients were invited to provide feedback 1–3
were then provided from the clinic for home use. The weeks after treatment, either by telephone using a structured
medications provided to both groups in the hybrid cohort interview tool (MSUK)18 or using an online form (BPAS).19
were the same. Providers followed organisation-specific evi- We analysed data on questions reporting on satisfaction,
dence-based policies informed by the RCOG8 and associ- future preference and privacy. Although the questions were
ated decision aid17 (Figures 1 and 2) and earlier RCOG similar, only the MSUK survey included questions on the
and NICE guidelines.14,15 privacy of teleconsultation. All contact was by a non-clini-
cian who had not been involved in the patient’s care.
Outcome measures
Analysis
Access The primary analysis was to assess whether the telemedi-
Waiting time and gestation at treatment were used to assess cine-hybrid model was non-inferior to the traditional
how the two models impacted access. Waiting time was model. We compare effectiveness and safety in the two
defined as the interval from first contact with the abortion cohorts by testing the hypotheses that the telemedicine-hy-
provider to when medication was dispensed (either in-per- brid model is less effective and has a higher complication
son in clinic or posted). Gestation was recorded as that rate.
applicable on the date mifepristone was dispensed or pro- We first compared patient demographic and clinical
vided. Analysis was of mean gestation and the proportion characteristics between the cohorts to assess the need to
of abortions performed at ≤6 weeks’ gestation. covariate-adjust our hypothesis tests for systematic differ-
ences in the two groups that might affect outcomes of
Effectiveness abortion. All hypothesis tests were covariate adjusted for
Effectiveness was defined as the proportion of medical patient age, race/ethnicity, gestational age, parity and prior
abortions that were successful. Success was defined accord- abortions using logistic regression and weighted risk differ-
ing to the MARE Guidelines as successful expulsion of an ences.20
intrauterine pregnancy without need for surgical interven- We evaluated effectiveness by testing the alternative
tion,13 using the sub-categories as listed in Table 2. hypothesis that the rate of successful medical abortion in
the telemedicine-hybrid cohort is lower than in the tradi-
Safety tional cohort using a covariate-adjusted test of difference in
Safety was defined according to the proportions of medical proportions. We also performed a Chi-square test to evalu-
abortions that involved one or more significant adverse ate whether the distribution of unsuccessful abortion sub-
events. We defined significant adverse events as haemorrhage categories differed between the cohorts. We evaluated safety
requiring transfusion, significant infection requiring hospital by testing the alternative hypothesis that significant adverse
admission, major surgery and death. We also examined the events occurred at higher rates in the telemedicine-hybrid
incidence of ectopic pregnancy and when it was diagnosed in cohort than in the traditional cohort using a covariate-ad-
the care pathway. For the analysis of ectopic pregnancies, we justed hypothesis test for difference of proportions. We
included all abortion consultations rather than all abortions reported one-sided P-values for both of these tests. We also
provided because some patients with suspected ectopic preg- evaluated whether the prevalence of ectopic pregnancies
nancies were referred to Early Pregnancy Assessment Units managed before EMA and after EMA were different
(EPAU) and did not proceed to abortion. All patients between the traditional and telemedicine-hybrid cohorts
referred for further diagnostics (e.g. serial human chorionic using Chi-square difference of proportion tests.
gonadotropin [bhCG] monitoring) but who had no further The secondary analysis was to compare effectiveness
treatment are included in the ectopic pregnancy group, and safety of medical abortion for patients who received

1466 ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd.
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Effectiveness, safety and acceptability of no-test medical abortion

Aboron is essenal health care

Services should be organised so as to minimise delays in care

Woman requests aboron

Provide ultrasound if:

• Unable to provide date of concepon or LMP of reasonable certainty

within thresholds of eligibility or skill of provider
Offer remote consultaon (e.g. via video or telephone)
• History or symptoms suggesve of high risk of ectopic pregnancy such as:

- Unilateral abdominal pain and vaginal bleeding/spong in last 5 days

Provide wrien informaon prior to consultaon - Intrauterine device in-situ

(e.g. via email or link) - History of tubal damage

- Prior ectopic pregnancy

Remote consultaon – same standards as face-to-face:

Enough informaon and me should be given to allow the opportunity

for quesons and to give informed consent

Consent can be given verbally, but discussion must be recorded in notes

Safeguarding assessment should be individualised

Collect (with minimal contact) / pack & post: Clinicians must be confident the woman can speak privately without
coercion
• Aboron medicaon
• Any addional medicaon (e.g. analgesia,an-emec)
• Low sensivity pregnancy test
• Wrien advice and informaon
• Plan for remote follow-up

If STI screening indicated or Chlamydia test recommended, use remote

service (e.g. web-based home tesng) No need for roune blood tesng

Offer remote or self-assessment of outcome using low

sensivity pregnancy test three weeks aer medical
aboron

Offer discussion of contracepon opons – if appropriate

include contracepon in pack

Figure 1. Summary of early medical abortion care management during COVID-19 pandemic (adapted with permission from RCOG Coronavirus
[COVID-19] Infection and Abortion Care – Information for Healthcare Professionals8).

fully remote no-test telemedicine versus in-person care in adverse events in the telemedicine versus in-person
the telemedicine-hybrid cohort, primarily to assess groups.
whether any differences between the cohort service mod- All analyses were performed using R, version 3.6.2 (R
els were driven by one particular group. We performed Foundation for Statistical Computing, c/o Institute for
covariate-adjusted hypothesis tests under the null hypoth- Statistics and Mathematics, Vienna, Austria). Statistical sig-
esis of equal effectiveness and equal rates of significant nificance was defined using an alpha level of 0.05.

ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd. 1467
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Aiken et al.

Decision Aid for Early Medical Abortion without Ultrasound

Do you know the date your

last menstrual period began? Book scan if:
• unable to provide a date or estimate
1 (first day of proper bleeding –
• periods less frequent than 6 weekly
don’t count light spotting
prior to your full period)

If especially light, or just spotting, take the date of the last

normal LMP as accurate
Was it a normal period, or
2 was it especially light or If particularly heavy, ensure they have repeated a pregnancy test
heavy? and take the date as accurate

In the last three months, have

you been taking the • Have you been having regular periods until the last missed
period?
3 contraceptive pill, other
hormonal contraception, or • When did you start to get feelings of being pregnant?
been breast feeding?

If uncertain, book scan

Have you had any of the following:

• Abdominal (“tummy”) or pelvic pain which is more on one side, and vaginal bleeding / spotting in
the last five days?
• An intrauterine device (e.g. “contraceptive coil”, “Mirena”, “Jaydess”, “Kyleena”, “Levosert”) in
4 place when you conceived?
• A previous ectopic pregnancy?
• Been told, following an operation or scan, that your Fallopian tubes (which connect the ovaries to
the womb) are damaged?
• Had surgery on your Fallopian tubes?

If “Yes” to any, book scan

Offer Early Medical Abortion without

5
Ultrasound

Figure 2. Decision aid for early medical abortion without ultrasound (adapted with permission from RCOG Coronavirus (COVID-19) infection and
abortion care – information for healthcare professionals; 2020-06-04-decision-aid-for-early-medical-abortion-without-ultrasound.pdf [rcog.org.uk]17)

1468 ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd.
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Effectiveness, safety and acceptability of no-test medical abortion

Ethics approval Table 1. Clinical and demographic characteristics in the traditional

The study was reviewed by the Institutional Review Board and telemedicine-hybrid cohorts (n = 52 142). n (%)
(IRB) at the University of Texas at Austin, and a determi-
nation was made that the research did not meet the crite- Patient characteristics Traditional Telemedicine- P-value
(n = 22 158) hybrid
ria for human subject research as defined in the Common
(n = 29 984)
Rule (45 CFR 46) or FDA Regulations (21 CFR 56). Each
provider ensured compliance with their own internal ethics
Mean gestational age 6.4 (1.3) 6.0 (1.4) <0.001
and governance systems. This study is classed as service
in weeks (SD)
evaluation by the NHS Health Research Authority21 and Gestational age at treatment*
therefore did not require review by an NHS Research ≤6 weeks 5582 (25.2) 11 947 (39.8) <0.001
Ethics Committee (REC).22 The principles of the STROBE >6 weeks 16 576 (74.8) 18 037 (60.2)
statement, and the MARE supplement, were followed.13 Mean age in years (SD) 27.8 (6.6) 28.5 (6.7) <0.001
Patients were not involved in the development of the Ethnicity
Asian 2038 (9.2) 2652 (8.8) <0.001
study.
Black 1656 (7.5) 2282 (7.6)
Multiracial 1004 (4.5) 1361 (4.5)
Results White 15 840 (71.5) 20 910 (69.7)
Other 489 (2.2) 638 (2.1)
In total, 52 142 medical abortions were provided during Unknown 1131 (5.1) 2141 (7.1)
the study period: 22 158 in the traditional cohort and Previous abortions
29 984 in the telemedicine-hybrid cohort. Among those 0 13 098 (59.1) 16 741 (55.8) <0.001
1+ 9060 (40.9) 13 243 (44.2)
that took place in the telemedicine-hybrid cohort, 18 435
Parity
(61%) were provided via telemedicine and 11 549 (39%)
0 10 133 (45.7) 11 741 (39.2) <0.001
in-person. Our sample represents 85% of the total number 1+ 12 025 (54.3) 18 243 (60.8)
of medical abortions performed in England and Wales dur- Mean waiting time in 10.7 (19.9) 6.5 (13.5) <0.001
ing the study period.23 The clinical and demographic char- days (SD)*
acteristics of patients in the two cohorts are described in
*After checking for normality, these variables were non-parametric
Table 1.
and therefore two-sample Wilcoxon tests were used.

Access
Mean waiting time to treatment declined from 10.7 days evidence that significant adverse events were higher in the
(SD 19.9) in the traditional pathway to 6.5 days (SD 13.5) telemedicine-hybrid cohort (P = 0.557).
in the telemedicine-hybrid cohort (P < 0.001). Mean gesta- The overall incidence of ectopic pregnancy was equiva-
tional age at treatment also declined in the telemedicine- lent in both cohorts – 39 (0.2%) in the traditional cohort
hybrid cohort, resulting in 40% of abortions performed at and 49 (0.2%) in the telemedicine-hybrid cohort,
6 weeks’ gestation or less versus 25% in the traditional P = 0.796 (Table 4). The proportions managed after EMA
cohort (P < 0.001). were not significantly different between the cohorts (0.01%
in the traditional pathway and 0.03% in the telemedicine-
Effectiveness hybrid pathway, P = 0.123). There were 11 cases (0.04%)
Rates of successful medical abortion were high under both in the telemedicine-hybrid cohort where the gestational age
service delivery models (Table 2) – 98.2% in the traditional after abortion was reported by the patient or an admitting
cohort versus 98.8% in the telemedicine-hybrid cohort. We hospital as being greater than the expected 10 weeks. In all
found no evidence of a lower success rate with the teleme- these cases, the abortion was completed at home without
dicine-hybrid pathway (P > 0.999). The distribution of the additional medical complications.
different sub-categories of unsuccessful medical abortion
did not differ between the cohorts (P = 0.268). Acceptability
Patient-reported outcome data were available from 2453
Safety respondents: 96% were ‘satisfied’ or ‘very satisfied’ with
Significant adverse events in both cohorts were rare their care, or rated their experience as ‘good’ or ‘very
(Table 3). Haemorrhage requiring transfusion was reported good’; 80% reported that they would choose telemedicine
in eight (0.04%) cases in the traditional cohort and in in the future or that it was their preferred option, with
seven (0.02%) cases in the telemedicine-hybrid cohort. No 13% choosing in-person care and the remainder being
cases of significant infection requiring hospital admission, unsure. No patient reported that they were unable to con-
major surgery or death were reported. We found no sult in private using teleconsultation (n = 1243).

ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd. 1469
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Aiken et al.

Table 2. Comparison of effectiveness of medical abortions conducted in the traditional and telemedicine-hybrid cohorts (n = 52 142). n (%)

Outcome Traditional Telemedicine-hybrid P-value

n = 22 158 n = 29 984

Successful medical abortion 21 769 (98.2) 29 618 (98.8) 1.0

Unsuccessful medical abortion 389 (1.8) 366 (1.2)
Continuing pregnancy: treated with surgical management 161 (0.7) 150 (0.5) 0.268
Continuing pregnancy: opted to continue or unknown 3 (0.01) 8 (0.03)
Retained products treated with surgical management (ERPC) 225 (1.0) 208 (0.7)

As explained in the methods section, the P-value for successful medical abortion is the co-variate-adjusted P-value (i.e. all differences in patient
clinical and demographic characteristics, including gestational age, are controlled for) and was calculated using a hypothesis test where the null
hypothesis is that the traditional cohort has the same effectiveness rate as the telemedicine-hybrid cohort and the alternative hypothesis is that
the traditional cohort has a higher effectiveness rate than the telemedicine-hybrid cohort. The P-value for unsuccessful medication abortion is the
Chi-square test of whether the distribution of types of failure differ between the cohorts.

Table 3. Comparison of significant adverse events following Table 4. Significant outcomes among patients presenting for
medical abortions conducted in the traditional and telemedicine- medical abortion in the traditional and telemedicine-hybrid cohorts
hybrid cohorts (n = 52 142). n (%) (n = 52 218). n (%)

Outcome Traditional Telemedicine- P-value Outcome Traditional Telemedicine- P-value

(n = 22 158) hybrid (n = 22 197) hybrid
(n = 29 984) (n = 30 021)

Haemorrhage 8 (0.04) 7 (0.02) 0.557 Ectopic managed 37 (0.17) 39 (0.13) 0.796

requiring transfusion pretreatment
Infection requiring 0 (0.0) 0 (0.0) Ectopic managed 2 (0.01) 10 (0.03) 0.123
hospital admission post-treatment
Major surgery 0 (0.0) 0 (0.0) Gestational age later 0 (0.0) 11 (0.04) N/A
Death 0 (0.0) 0 (0.0) than expected*

As explained in the methods section, the P-value was calculated The column numbers include patients who presented for an EMA
using a hypothesis test where the null hypothesis is that the but did not receive one because their ectopic pregnancy was
traditional cohort has the same rate of adverse events as the identified pretreatment.
telemedicine-hybrid cohort and the alternative hypothesis is that the *The column numbers for the gestational age later than expected
traditional cohort has a lower rate of adverse events than the category are the same as those in Table 3, (all EMAs performed in
telemedicine-hybrid cohort. the two cohorts, i.e. n = 52 142).

Secondary outcome measures and access to abortion care is better, with both waiting
Comparison of clinical outcomes for the telemedicine ver- times and gestational age at the time of the abortion signif-
sus in-person groups in the telemedicine-hybrid cohort is icantly reduced. Although gestation could be influenced by
shown in Tables S1–S3. Rates of successful medical abor- different behaviour during the pandemic, resulting in ear-
tion were higher in the telemedicine group (99.2% versus lier presentation, in most other areas of healthcare, access
98.1%, P < 0.001), but rates of significant adverse out- has been severely impacted, with waiting times increasing
comes were not significantly different between the two substantially,24 and so reductions in both seem relevant.
groups – three (0.02%) for telemedicine versus four Even small reductions in waiting time are significant –
(0.03%) for in-person (P = 0.532). NICE noted that a reduction of 1 day resulted in annual
savings of £1.6 million to the health services in England
owing to reduced complications and fewer needing to opt
Discussion
for a surgical abortion.4 Further evidence that the new tele-
Main findings medicine-hybrid model improves access comes from a
We found that no-test medical abortion via telemedicine study showing that the rate of women seeking abortion
without routine ultrasound up to 10 weeks’ gestation is an medication outside the formal healthcare setting was signif-
effective, safe and acceptable service model. Clinical out- icantly reduced in the UK following its implementation.25
comes with telemedicine are equivalent to in-person care The implication is that those previously too vulnerable to

1470 ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd.
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Effectiveness, safety and acceptability of no-test medical abortion

attend in-person have been able to access care through tele- findings, the consequences for most are unlikely to be med-
medicine, potentially benefitting from the safeguarding, ically significant.35 The 10-week gestation limit in the Eng-
counselling and contraceptive services provided by regu- lish government’s approval order is arbitrary and is not
lated providers.26 based on evidence of safety or effectiveness. The Scottish
Our study confirms previous literature that medical government did not stipulate a limit, leaving the decision
abortion is safe and effective. Our findings of low rates of to the discretion of the clinician in consultation with their
significant complications and failure are similar to those patient. Moreover, the reported success of self-managed
reported from other high quality studies.27,28 The slight terminations of pregnancy at >12–24 weeks’ gestation is
increase in effectiveness we observed in the group that 93%, with safety similar to that expected in earlier gesta-
received telemedicine – even after controlling for lower tions.36,37
average gestational age compared with the in-person group
– may be due to the ability of patients to control better the Strengths and limitations
time at which they took the medication. Although the study is not a clinical trial, we were able to
The telemedicine-hybrid model resulted in very low rates evaluate the outcomes of both the telemedicine-hybrid and
of undiagnosed ectopic pregnancy and later than expected traditional in-person services as they operate in the real
gestations. Although the rate of ectopic pregnancy in the world, and we were able to adjust for key covariates. A
general population in the UK and USA is reported as 1– key strength of the study is the generalisability of our find-
2%,29,30 the rate reported among patients having a termina- ings, given that our sample included 85% of all medical
tion of pregnancy is 10 times lower,31 which is consistent abortions provided in England and Wales during the study
with our findings. Ultrasound is not used to screen for period.
ectopic pregnancy in the general population – it is only The main limitation of this study is that we were unable
used where signs and symptoms suggest a need.30 Routine actively to follow up patients after their abortion. There is
screening of symptom-free women is associated with a high a potential gap in the consistency of reporting incidents,
false-positive rate when the prevalence of ectopic pregnancy due to some complications not meeting the threshold of
is low, as is the case in women seeking abortion, and there- serious incidents, multiple routes of entry into the NHS
fore it is unlikely there would be significant benefits.32 and informal communication between the NHS and abor-
There is no clinical justification for maintaining this incon- tion providers. Although it is possible that some patients
sistency in care between women wishing to continue their presented to other providers and a significant adverse event
pregnancies and those choosing EMA.33,34 was not reported in our dataset, the risk management and
However, given that over 200 000 people access abortion reporting systems within the NHS are well defined, with
care each year in the UK alone, some will inevitably have serious incidents being routinely shared. The governing
an asymptomatic ectopic pregnancy and so will proceed body of the NHS in England alerted all commissioners of
with having mifepristone and misoprostol either through the need to report incidents relating to telemedicine and
telemedicine or after a false-negative scan. The essential there were review meetings of key stakeholders to ensure
issue for safety is that these are detected prior to causing compliance. No additional cases were identified from regu-
harm rather than prior to beginning the medical abortion lators that were not already recorded by the providers’ clin-
treatment; treatment with mifepristone and misoprostol in ical incident processes. More importantly, there is no
itself will have no effect on an underlying ectopic preg- reason that any under-reporting would be systematically
nancy. Indeed, the reduction in waiting times afforded by more likely in either cohort to introduce bias. Although
the telemedicine model may facilitate earlier detection than patient behaviour may have been altered in the pandemic,
traditional pathways where women present later or are sent it seems unlikely patients would not have reported prob-
away to give additional time to visualise an intrauterine lems to their provider given that there is immediate access
pregnancy on scan. Proceeding with early medical abortion to help via 24-hour telephone services. It is also possible
without a scan may permit earlier diagnosis of a developing that with NHS acute services (e.g. early pregnancy units)
ectopic pregnancy owing to increased surveillance and harder to access, patients would be more likely to engage
index of suspicion, for example where there is minimal with their abortion provider first. Finally, evidence from an
bleeding after misoprostol.8,15 equivalent population to ours in Scotland gave almost
The proportion of cases where gestational age was later identical results. This was a smaller cohort study that fol-
than expected based on LMP was low, as might have been lowed up all patients (n = 663) and cross-checked NHS
expected given the evidence that women can determine the hospital records (reporting successful medical abortion
gestational age of their pregnancy with reasonable accuracy 98%, haemorrhage requiring transfusion 0%, infection
by LMP alone.33 Nevertheless, inadvertent treatment of ges- requiring hospital admission 0%, choosing telemedicine
tations over 10 weeks is inevitable and, consistent with our again in future 71%).39

ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd. 1471
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Aiken et al.

Interpretation disclosure of interests forms are available to view online as

NICE conducted a systematic review and recommended supporting information.
using telemedicine to improve access to medical abortion
care.3 Several models for using telemedicine to facilitate Contribution to authorship
medical abortion have been described but most existing trials AA – Initial concept, developing study protocol, data analy-
are small, and many required attendances to have medicines sis, writing first draft and subsequent revisions, verification
administered, an ultrasound scan or blood tests.7,38 Our of data. Overall responsibility for data analysis and principal
study is the first to assess a real-world no-test telemedicine investigator. Guarantor; NG – Developing study protocol,
abortion care pathway in a national population. This new revision of draft paper. Overall responsibility for conduct of
national model demonstrates how a permissive framework study including data collection at NUPAS; PL – Initial con-
for medical abortion can deliver significant quality improve- cept, developing study protocol, writing first draft and subse-
ments to those needing to access abortion care. quent revisions. Overall responsibility for conduct of study
including data collection at BPAS; JL – Initial concept, devel-
oping study protocol, writing first draft and subsequent revi-
Conclusion
sions, co-ordination and liaison. Overall responsibility for
This large study of 52 142 medical abortions demonstrates conduct of study including data collection at MSUK; JS –
that incorporating no-test telemedicine into the care path- Data analysis including statistical advice, writing first draft
way is not inferior to the traditional pathway where all and subsequent revisions, verification of data.
patients are seen in person and have an ultrasound scan.
There are advantages – waiting times and gestation at abor- Details of ethics approval
tion are reduced and it is highly rated by patients. There was The study was reviewed by the Institutional Review Board
no evidence of worse outcomes in failure rate, haemorrhage, (IRB) of the University of Texas at Austin and a determi-
need for surgery or failure to detect ectopic pregnancy. In nation was made on 21 April 2020 that the research did
the 0.04% of cases where the abortion appeared to have been not meet the criteria for human subjects research as
provided at over 10 weeks’ gestation, these were all com- defined in the Common Rule (45 CFR 46) or FDA Regula-
pleted at home without additional medical complications. tions (21 CFR 56). Each provider ensured compliance with
Given the advantages of improving access to care, especially their own internal ethics and governance systems. The
in vulnerable groups and in resource-poor healthcare sys- principles of the STROBE statement, and the MARE sup-
tems or where patients have to fund their own care, the evi- plement, were followed.13
dence is compelling that no-test telemedicine should become
routine in the provision of abortion care. Funding
This study was supported in part by the Eunice Kennedy
Disclosure of interests Shriver National Institute of Child Health & Human Devel-
All authors have completed the ICMJE uniform disclosure opment of the National Institutes of Health (NIH) through
form at www.icmje.org/coi_disclosure.pdf and declare: no Center Grant P2CHD042849, awarded to the Population
support from any organisation for the submitted work; no Research Center at The University of Texas at Austin. The
financial relationships with any organisations that might content of this article is solely the responsibility of the
have an interest in the submitted work in the previous 3 authors and does not necessarily represent the official views
years; no other relationships or activities that could appear of the NIH.
to have influenced the submitted work beyond that
described under the roles held by each individual. AA – Acknowledgements
Member of the Council of the British Society of Abortion We would like to thank the Care Quality Commission
Care Providers (BSACP); NG and JS – nothing to disclose; (CQC), NHS England and NHS Improvement (NHSE & I),
PL – Co-author/committee member of national guidelines and the National Child Mortality Database (NCMD) for
cited in this paper from RCOG and NICE; member RCOG their advice and assistance in ensuring all complications
abortion taskforce; council member of British Society of reported within NHS systems have been included. We are
Abortion Care Providers (BSACP); Medical Director of also enormously grateful to the many people who helped
BPAS; JL – Co-author/committee member of national to ensure access to essential healthcare for women needing
guidelines cited in this paper from RCOG and NICE; co- an abortion not only continued, but was improved during
chair of RCOG Abortion Taskforce and British Society of the COVID-19 pandemic. There are too many to list indi-
Abortion Care Providers (BSACP); primary employment vidually, but we would like to especially thank the follow-
includes work as termination of pregnancy care provider as ing who not only helped to implement a complete service
a consultant gynaecologist and medical director. Completed change within a few weeks, but also made this paper

1472 ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd.
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Effectiveness, safety and acceptability of no-test medical abortion

possible by ensuring the data were collected and accessible: 2 Regan L, Glasier A. The British 1967 Abortion Act – still fit for
BPAS – Rebecca Blaylock, Steve Cheung, Pam Field, Ste- purpose? Lancet 2017;390:1936–7.
3 National Institute for Health and Care Excellence (NICE). Abortion Care.
phen Franklin, Jeanette Taylor, Katherine Whitehouse; Evidence Review A – Accessibility and Sustainability of Abortion Services.
MSUK – Cat James, Andrew Lord, Kay Newey, Abigail London: National Institute for Health and Care Excellence (NICE); 2019
Storan; NUPAS – Aaron Flaherty, Linda Leach, Leanne [www.nice.org.uk/guidance/ng140/documents/evidence-review].
MacCaffrey. 4 O’Shea LE, Hawkins JE, Lord J, Schmidt-Hansen M, Hasler E,
Cameron S, et al. Access to and sustainability of abortion services: a
systematic review and meta-analysis for the National Institute of
Data-sharing statement
Health and Care Excellence – new clinical guidelines for England.
The collated datasets, which include participant data with Hum Reprod Update 2020;26:886–903.
anonymised identifiers, are held by AA at the University of 5 WHO Global Observatory for eHealth. Telemedicine: Opportunities
Texas. Consideration will be given to sharing this with and Developments in Member States: Report on the Second Global
bona fide researchers on application. The original data Survey on eHealth. Geneva: World Health Organization; 2010.
6 Dorsey ER, Topol EJ. Telemedicine 2020 and the next decade.
reside with the co-authors’ own institutions. Although the
Lancet 2020;395:859.
data are de-identified, some relate to very rare events and 7 Endler M, Lavelanet A, Cleeve A, Ganatra B, Gomperts R, Gemzell-
could therefore result in identification. Therefore data on Danielsson K. Telemedicine for medical abortion: a systematic
complications, and data arising from clinical incident review. BJOG 2019;126:1094–102.
reports, will be subject to the same access restrictions as 8 Royal College of Obstetricians and Gynaecologists (RCOG), Royal
College of Midwives (RCM), Faculty of Sexual & Reproductive
those of the organisation supplying it.
Healthcare (FSRH), British Society of Abortion Care Providers
(BSACP). Coronavirus (COVID-19) Infection and Abortion Care.
Transparency statement London: Royal College of Obstetricians and Gynaecologists; 2020.
The authors affirm that the manuscript is an honest, accu- 9 Department of Health & Social Care (DHSC). The Abortion Act –
rate and transparent account of the study being reported; Approval of a Class of Places. London: English Government; 1967.
10 Scottish Government. Abortion – Covid-19 – Approval for
that no important aspects of the study have been omitted;
Mifepristone to be Taken at Home and Other Contingency
and that any discrepancies from the study as originally Measures. Edinburgh: Scottish Government; 2020.
planned (and, if relevant, registered) have been explained. 11 Welsh Government. The Abortion Act 1967 – Approval of a Class of
Place for Treatment for the Termination of Pregnancy (Wales) 2020.
Data availability Cardiff: Welsh Government; 2020.
12 Aiken ARA, Guthrie KA, Schellekens M, Trussell J, Gomperts R.
Data available on request due to privacy/ethical restrictions.
Barriers to accessing abortion services and perspectives on using
mifepristone and misoprostol at home in Great Britain.
Contraception 2018;97:177–83.
Supporting Information
13 Creinin MD, Chen MJ. Medical abortion reporting of efficacy: the
Additional supporting information may be found online in MARE guidelines. Contraception 2016;94:97–103.
14 Royal College of Obstetricians and Gynaecologists (RCOG), Faculty
the Supporting Information section at the end of the
of Sexual & Reproductive Healthcare (FSRH), British Society of
article. Abortion Care Providers (BSACP). Clinical Guidelines for Early
Table S1. Patient clinical and demographic characteris- Medical Abortion at Home – England. London: Royal College of
tics in the in-person versus telemedicine groups for the tel- Obstetricians and Gynaecologists (RCOG); 2019 [www.rcog.org.uk/
globalassets/documents/guidelines/early-medical-abortion-at-home-
emedicine-hybrid cohort (n = 29 984). Number (%).
guideline-england.pdf].
Table S2. Comparison of effectiveness of medical termi- 15 National Institute for Health and Care Excellence (NICE). Abortion
nations of pregnancy conducted in the in-person versus tel- Care. London: NICE; 2019. Report No.: NG140 [www.nice.org.uk/
emedicine groups for the telemedicine-hybrid cohort guidance/ng140].
(n = 29 984). n (%). 16 Baiju N, Acharya G, D’Antonio F, Berg R. Effectiveness, safety and
acceptability of self-assessment of the outcome of first-trimester
Table S3. Comparison of significant adverse events fol-
medical abortion: a systematic review and meta-analysis. BJOG
lowing medical terminations of pregnancy in the in-person 2019;126:1536–44.
versus telemedicine groups for the telemedicine-hybrid 17 Royal College of Obstetricians and Gynaecologists (RCOG). Decision
cohort (n = 29 984). n (%). Aid for Early Medical Abortion Without Ultrasound. London: Royal
College of Obstetricians and Gynaecologists (RCOG); 2020 [www.
Video S1. Author insights. & rcog.org.uk/globalassets/documents/guidelines/2020-06-04-decision-aid-
for-early-medical-abortion-without-ultrasound.pdf].
18 Porter Erlank C, Lord J, Church K. Acceptability of no-test medical
References abortion provided via telemedicine: analysis of patient-reported
€ Beavin C,
1 Bearak J, Popinchalk A, Ganatra B, Moller A-B, Tuncßalp O, outcomes. BMJ Sex Reprod Health 2021. bmjsrh-2020. https://doi.
et al. Unintended pregnancy and abortion by income, region, and org/10.1136/bmjsrh-2020-200954.
the legal status of abortion: estimates from a comprehensive model 19 Meurice M, Whitehouse K, Blaylock R, Chang J, Lohr P. Client
for 1990–2019. Lancet Global Health 2020;8:e1152–61. satisfaction and experience of home use of mifepristone and

ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd. 1473
 14710528, 2021, 9, Downloaded from https://obgyn.onlinelibrary.wiley.com/doi/10.1111/1471-0528.16668 by Cochrane Colombia, Wiley Online Library on [28/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Aiken et al.

misoprostol for medical abortion up to 10 weeks’ gestation at 30 National Institute for Health and Care Excellence (NICE). Ectopic
British Pregnancy Advisory Service: a cross-sectional evaluation Pregnancy and Miscarriage: Diagnosis and Initial Management.
[preprint]. Authorea 2020. https://doi.org/10.22541/au.160691768. Report No.: NG126. London: NICE; 2019 [www.nice.org.uk/guida
87050587/v1. nce/ng126].
20 Ge M, Durham LK, Meyer RD, Xie W, Thomas N. Covariate-adjusted 31 Shannon C, Brothers LP, Philip NM, Winikoff B. Ectopic pregnancy
difference in proportions from clinical trials using logistic regression and medical abortion. Obstet Gynecol 2004;104:161–7.
and weighted risk differences. Drug Inform J 2011;45:481–93. 32 Mol BW, van der Veen F, Bossuyt PM. Symptom-free women at
21 Research Ethics Service. Defining research table 2017 [www.hra-dec increased risk of ectopic pregnancy: should we screen? Acta Obstet
isiontools.org.uk/research/docs/DefiningResearchTable_Oct2017-1.pdf]. Gynecol Scand 2002;81:661–72.
Accessed 2 March 2021. 33 Bracken H, Clark W, Lichtenberg ES, Schweikert SM, Tanenhaus J,
22 NHS Health Research Authority. Do I need NHS REC review? Barajas A, et al. Alternatives to routine ultrasound for eligibility
London: NHS Health Research Authority; 2020 [www.hra-dec assessment prior to early termination of pregnancy with
isiontools.org.uk/ethics/]. Accessed 2 March 2021. mifepristone-misoprostol. BJOG 2011;118:17–23.
23 Department of Health & Social Care (DHSC). Abortion Statistics 34 Clark WH, Gold M, Grossman D, Winikoff B. Can mifepristone
During COVID-19: Data Tables. London: Department of Health & medical abortion be simplified? A review of the evidence and
Social Care (DHSC); 2020 [www.assets.publishing.service.gov. questions for future research. Contraception 2007;75:245–50.
uk/government/uploads/system/uploads/attachment_data/file/915816/ 35 Raymond EG, Grossman D, Mark A, Upadhyay UD, Dean G, Creinin
COVID_tables_FINAL.ods]. MD, et al. Commentary: No-test medication abortion: a sample
24 Griffin S. Covid-19: waiting times in England reach record highs. protocol for increasing access during a pandemic and beyond.
BMJ 2020;370:m3557. Contraception 2020;101:361–6.
25 Aiken ARA, Starling JE, Gomperts R, Scott JG, Aiken C. Demand for 36 Harris LH, Grossman D. Complications of unsafe and self-managed
self-managed online telemedicine abortion in eight European abortion. N Engl J Med 2020;382:1029–40.
countries during the COVID-19 pandemic: a regression discontinuity 37 Moseson H, Bullard KA, Cisternas C, Grosso B, Vera V, Gerdts C.
analysis [preprint]. MedRxiv 2020:2020.09.15.20195222. Effectiveness of self-managed medication abortion between 13 and
26 Moreau C, Shankar M, Glasier A, Cameron S, Gemzell-Danielsson K. 24 weeks gestation: a retrospective review of case records from
Abortion regulation in Europe in the era of COVID-19: a spectrum accompaniment groups in Argentina, Chile, and Ecuador.
of policy responses. BMJ Sex Reprod Health 2020:bmjsrh-2020- Contraception 2020;102:91–8.
200724. https://doi.org/10.1136/bmjsrh-2020-200724. 38 Gambir K, Garnsey C, Necastro KA, Ngo TD. Effectiveness, safety
27 Chen MJ, Creinin MD. Mifepristone with buccal misoprostol for medical and acceptability of medical abortion at home versus in the clinic: a
abortion: a systematic review. Obstet Gynecol 2015;126:12–21. systematic review and meta-analysis in response to COVID-19. BMJ
28 Upadhyay UD, Desai S, Zlidar V, Weitz TA, Grossman D, Anderson Global Health 2020;5:e003934.
P, et al. Incidence of emergency department visits and complications 39 Reynolds-Wright JJ, Johnstone A, McCabe K, Evans E, Cameron S.
after abortion. Obstet Gynecol 2015;125:175–83. Telemedicine medical abortion at home under 12 weeks’ gestation:
29 Centers for Disease Control and Prevention (CDC). Ectopic a prospective observational cohort study during the COVID-19
pregnancy – United States, 1990–1992. MMWR Morb Mortal Wkly pandemic. BMJ Sex Reprod Health 2021. bmjsrh-2020. https://doi.
Rep 1995;44:46–8. org/10.1136/bmjsrh-2020-200976.

1474 ª 2021 The Authors. BJOG: An International Journal of Obstetrics and Gynaecology published by John Wiley & Sons Ltd.