Watson et al.

BMC Pregnancy and Childbirth (2019) 19:68

https://doi.org/10.1186/s12884-019-2210-1

STUDY PROTOCOL Open Access

EQUIPTT: The Evaluation of the QUiPP app

for Triage and Transfer protocol for a
cluster randomised trial to evaluate the
impact of the QUiPP app on inappropriate
management for threatened preterm
labour
Helena A. Watson* , Naomi Carlisle, Katy Kuhrt, Rachel M. Tribe, Jenny Carter, Paul Seed and Andrew H. Shennan

Abstract
Background: Accurate diagnosis of preterm labour is needed to ensure correct management of those most at risk of
preterm birth and to prevent the maternal and fetal risks incurred by unnecessary interventions given to the large
majority of women, who do not deliver within a week of presentation. Intervention “just-in-case” results in many
avoidable admissions, women being transferred out of their local hospital unnecessarily and most women receiving
unwarranted drugs, such as steroids and tocolytics. It also precludes appropriate transfers for others as neonatal cots
are blocked pre-emptively, resulting in more dangerous ex-utero transfers. We have developed the QUiPP App which is
a clinical decision-making aid based on previous outcomes of women, quantitative fetal fibronectin (qfFN) values and
cervical length. It is hypothesised that using the QUiPP app will reduce inappropriate admissions and transfers.
Methods: A multi-site cluster randomised trial will evaluate whether the QUiPP app reduces inappropriate
management for threatened preterm labour. The 13 participating centres will be randomly allocated to receive either
intervention or control. If the QUiPP app calculates risk of delivery within 7 days to be is less than 5%, clinicians are
advised that interventions may be withheld. Women’s experience of threatened preterm labour assessment will be
explored using self-completed questionnaires, with a subset of participants being invited to semi-structured interview.
A health economics analysis is also planned.
Discussion: We hypothesise that the QUiPP app will improve identification of the most appropriate women for
admission and transfer and ensure that therapies known to reduce risk of preterm neonatal morbidities are offered to
those who need them. We will determine which women do not require these therapies, thereby reducing over-
medicalisation and the associated maternal and fetal risks for these women. The findings will inform future national
guidelines on threatened preterm labour. Beyond obstetrics, evaluating the impact of an app in an emergency setting,
and our emphasis on balancing harms of over-treatment as well as under-treatment, make EQUIPTT a valuable
contribution to translational medicine.
Trial registration: The EQUIPTT trial was prospectively registered on 16th January 2018 with the ISRCTN
registry (no. 17846337).
Keywords: Preterm birth, Prediction, Fetal fibronectin, Cervical length, QUiPP app

* Correspondence: [email protected]
Division of Women’s Health, Faculty of Life Sciences & Medicine, King’s
College London, 10th Floor, North Wing, St Thomas’s Hospital Campus,
London SE1 7EH, UK

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Watson et al. BMC Pregnancy and Childbirth (2019) 19:68 Page 2 of 8

Background a number of on-call obstetric, neonatal and midwifery

Women with symptoms of preterm labour have long teams and takes an average of 340 min to complete [12].
posed a diagnostic challenge for clinicians concerned with This does not include the return journey time of the
managing the risks of preterm birth when the reality is transferring midwife. Paradoxically, unwarranted ante-
that most women will not deliver imminently [1]. Various natal transfers may increase more dangerous postnatal
prediction methods are available to direct interventions transfers by impairing efficient management of neonatal
that delay or ameliorate the consequences of preterm cots (i.e. cots reserved for babies that do not actually de-
birth (e.g. in-utero transfer, tocolysis, antenatal corticoste- liver preterm). With these neonatal cots blocked, infants
roids). However these interventions have psychological, of women in true preterm labour are only able to be
economic and clinical implications if given unnecessarily. transferred postnatally.
In light of recent UK guidance to treat all women in
threatened preterm labour (TPTL) prior to 30 weeks [2], Maternal risks
most of these women are likely to be admitted. However, Hospitalisation also increases the risks of venous
the majority of women presenting with symptoms of thromboembolism, hospital-acquired infections and the
preterm labour will not deliver in seven days [1]. psychological strain for the mother.
Our research group has developed the QUiPP App,
Risks associated with antenatal corticosteroids which improves prediction of preterm labour by com-
Clinicians are familiar with steroid-induced glucose in- bining clinical parameters and calculates risks more pre-
tolerance in the mother, necessitating insulin/dextrose cisely from continuous variables, to better assess risk
infusions and prolonged admissions. However, fetal ex- [13–15]. It utilises quantitative fetal fibronectin (qfFN)
posure to steroids has become the over-arching concern. (a protein released into the vagina at high concentra-
There is a significant reduction in infant birthweight of tions during preterm labour), the use of which has been
those women exposed to antenatal corticosteroids who developed by our group [13–15]. In a prospective obser-
deliver more than seven days after the first dose, com- vational study of 300 women, we reported a negative
pared to those receiving no treatment (mean difference predictive value (NPV) for qfFN at < 10 ng/mL of 98.2%,
147 g, 95% CI -291.97 to − 2.05 g) [3]. Infants exposed to and a positive predictive value (PPV) for delivery < 34
steroids are at increased risk of neonatal hypoglycaemia weeks at a 200 ng/mL threshold of 37% [13]. The value
(1.61 CI 1.38–1.87) and are significantly more likely to is limited by these arbitrary thresholds, and we showed
be in the lower quartile of academic ability (p = 0.01) that improved prediction can be obtained by interrogat-
(ARR 9.2–17.7% to 8.5%) [4]. There is biological plausi- ing the data across the whole qfFN range i.e. better sen-
bility supporting the latter finding given evidence of de- sitivity and “rule out” at low levels and improved
creased brain growth in preterm and term infants specificity and “rule in” at high levels. We combined
exposed to antenatal corticosteroids in animal studies qfFN and cervical length measurements (also as a con-
[5]. There is also evidence that fetal exposure to gluco- tinuous variable) to improve the prediction further in
corticoids modifies fetal hypothalamic-pituitary-adrenal 1249 asymptomatic women [14]. We then created a pre-
function, which is associated with increased risk of car- dictive model from a prospective observational dataset
diovascular and metabolic diseases in adulthood [6, 7]. of 382 women with TPTL to provide an individualised risk
Structural development of the brain may also be affected of delivering within 1/2/4 weeks and before 30/34/37
by steroid exposure, such as impaired cortical folding [8] weeks [15]. The accuracy of the model was further con-
and cortical thinning [9]. firmed using a validation set and performance reliability
demonstrated by comparison of expected and observed
Risks associated with unnecessary in-utero sPTB rates (p > 0.05). Receiver operating characteristic
transfer areas in the validation set differed from the training set by
Appropriate antenatal in-utero transfer (as opposed to amounts between − 0.04 and + 0.02 [15]. The model was
ex-utero) is essential to avoid the significant increases in subsequently updated on a larger dataset comprising 1032
neonatal morbidity and mortality associated with post- women presenting with symptoms of TPTL, and this is
natal transport of preterm infants [10]. In some units a the analysis that is incorporated into the latest version of
treat-all policy for threatened preterm labour would dra- the QUiPP app.
matically increase the number of such transfers. The im- In terms of guiding management decisions, the major-
plications of this far exceed the ambulance transfer ity of cases are assigned a very low-risk (< 1%) by QUiPP
costs: being transferred to another hospital unit (out of app, which clearly indicates a “no admit” strategy. In
area) is associated with a significant stress and expense EQUIPTT, we provide a guidance threshold of 5% risk
for the parents and family [11]. There is an immediate of delivery within seven days for intervention, based on
clinical burden of transfer arrangements, which involves our previous published work [16]. However, given the
Watson et al. BMC Pregnancy and Childbirth (2019) 19:68 Page 3 of 8

other multiple variables that contribute to this decision power calculation and analysis, the effectiveness of the
(such as gestational age), and our intention for the intervention may be exaggerated [18].
QUiPP app to be a tool to help clinicians make their All 13 centres will provide data related to the principal
own decisions, we have refrained from dogmatic recom- outcome under their current practice in a six-week
mendation based on absolute thresholds. pre-intervention data collection period. Following random-
We also want to establish if use of the QUIPP app re- isation, the centres will use intervention (using the QUiPP
duces women’s anxiety and/or helps her to come to de- app) or control (routine management) for a nine-month
cisions about her care that she is happy with. Women analysis period. In the final phase of the trial, the interven-
may have different preconceptions about the significance tion will be introduced in the 6 control units. Following
of symptoms of TPTL and varying perspectives on con- this, data will be collected for a further six weeks across all
sequences of admission and or transfer, such as financial 13 sites, in order to equally incentivise all sites to partici-
anxieties and separation from family, friends and other pate. The QUIPP app approach will be adopted as standard
children [11]. Furthermore, the use of an app to guide practice in all sites if successful. A list of the study sites is
clinical management is new to healthcare practice. Given available from the authors on request.
the importance of shared decision-making emphasized An health economics analysis will also be conducted; this
by the Montgomery ruling [17], women’s views regard- will involve a cost-minimisation and cost-consequences
ing its application in threatened preterm labour will be analysis of the use of the QUIPP app for triage and transfer
central to the future development of the QUIPP app. compared to current practice. We hypothesise that there
will be no negative impact on the health and wellbeing of
women and neonates, as a result of the intervention, par-
Methods/design
ticularly given its non-invasive nature and that the inter-
Trial objectives
vention will be cost saving.
Our aim is to evaluate the ability of the QUIPP app to re-
Women’s experience of TPTL assessment will be ex-
duce inappropriate management of TPTL. This will exter-
plored using self-completed questionnaires, with a subset
nally validate the prediction model, assess its success as a
of participants being invited to semi-structured interview.
clinical decision making tool, and measure its potential
cost-saving impact in the emergency obstetric setting.
Sample size
For inappropriate admission decisions, a total sample size
Hypothesis: The implementation of the QUIPP app and of 580, approximately 50 recruits per site, was calculated.
management algorithm will decrease inappropriate Data from our group’s ongoing prospective observational
management for TPTL study into the ability of qfFN to predict preterm birth in
symptomatic women (PETRA REC Ref. 14/LO/1988) has
Trial design allowed us to estimate the likely treatment effect for the
EQUIPTT is a cluster randomised controlled trial (incorp- intervention reducing inappropriate admissions from 25
orating a parallel group design) across 13 obstetric centres. to 10% and intra-class correlation 0.030. Based on 13 clus-
This design was chosen as for this scale of protocol inter- ters this treatment effect requires approximately 580 for
vention (in an emergency situation), it would be challen- 80% power. We are aware that a larger number of centres
ging to randomise at participant level, affecting uptake would also be preferable, but believe 13 is adequate and at
and generalisability of the findings, and individual clini- the limit of what is currently feasible.
cians cannot be randomised to varying decision-making For our qualitative questionnaires, we have estimated a
with individual patients. target recruitment of 300 participants, 25 recruits per site,
The pragmatic approach was taken to introduce the allowing for non-compliance and data collection errors.
QUIPP app to entire hospital antenatal units, as stand- Based on previous research involving visual analogue scale
ard practice for all clinicians, and all affected pregnan- (VAS) scores [19], a total sample size of 272 was required
cies. Using this parallel cluster studies have the added at standard levels of significance (alpha = 0.05 two-sided) to
value of limiting the time-bias which occurs with a achieve 90% power to detect a 10% difference in mean VAS
stepped wedge design. Also, unlike in a crossover design, scores between hospitals with and without the intervention.
the intervention does not need to be retracted from the
cluster, avoiding contamination bias. Individuals within a Randomisation
cluster tend to have more similar outcomes than across Randomisation is at the cluster level. Centres have been
clusters. The similarity in the outcomes of individuals randomly allocated to receive either intervention or con-
within a cluster within a time period is typically mea- trol. Due to there being an odd number of centres, seven
sured by the within-cluster within-period intra-cluster sites were allocated to one group and six to the other.
correlation coefficient. If this is not factored into the Minimisation of the randomisation list was not appropriate
Watson et al. BMC Pregnancy and Childbirth (2019) 19:68 Page 4 of 8

for this design but the imbalances between three levels of the decisions made. Anxiety levels and decisional conflict
NICU facility at intervention and control sites were small will be ascertained by visual analogue scale (VAS), and the
(4:3, 3:2, and 0:1). TPTL rates were known for 10 sites and Decisional Conflict Scale (DCS) questions [19]. The ques-
the difference between the average rates at control and in- tionnaire booklet has been designed in collaboration with
terventions sites were within chance (72 and 89%). the Women’s Health Academic Centre’s Preterm Birth
Studies public and patient involvement (PPI) panel. Differ-
Data analysis ence in VAS anxiety scores between women before and
Cluster randomised trial after assessment, and between hospitals with and without
Data analysis will follow the intention to treat principle, QUIPP app in use at the time of assessment, will be calcu-
according to the planned intervention. Data will be ana- lated (paired and independent samples t-test). Difference
lysed using Stata software Version 14 or later (StataCorp, between DCS scores after assessment, between hospitals
College Station, Texas) to estimate the size and test for with and without QUIPP app in use, will also be calcu-
statistical significance of any effects of the intervention lated (paired and independent samples t-tests as appropri-
on primary and secondary, and analysis will be super- ate). If the DCS and VAS data is found to be not normally
vised by the trial statistician, Mr. Paul Seed. A per proto- distributed, non-parametric equivalent tests will be used.
col analysis will also be performed, excluding those who A subgroup of participants (approx. 20–30) will be invited
are incidentally admitted or treated for reasons unrelated to provide a more in-depth account of their experiences
to preterm labour. Treatment effects for binary end- during one-to-one face-to-face or telephone interviews. The
points will be expressed as risk ratios (relative risk) with interview schedule will be determined following interim
95% CIs, using binomial regression and adjusting for analysis of the data and free text box comments. The
variables used in the minimization process. Risk differ- Framework approach [20] will be used to analyse the
ences will also be calculated for the primary endpoint. qualitative data obtained from the interviews. This
The analysis model will include a random effect for clus- method of qualitative data analysis is suitable for use
tering and adjusted standard errors for clustering. Adjust- with the anticipated findings from this study. Data will
ments will also need to be made for differences between be analysed using NVivo software.
cluster populations, such as ethnicity and maternal age.
Our primary outcome of inappropriate decisions will be Data collection
measured per 1000 deliveries. Local research staff will prospectively collect data on all
available eligible episodes of threatened preterm labour.
Health economics evaluation Key data for all patients (at control and interventions sites)
For the health economics analysis, a multi-level general
linear model using appropriate family and link function,  Hospital number
will be used to calculate the average cost per participant of  Gestation at presentation
use of the QUIPP app for triage and transfer compared to  Quantitative fetal fibronectin values (if available)
current practice. Costs will also be reported alongside sec-  Cervical length measurements (if available)
ondary clinical outcomes as part of a cost-consequences  Current NICU availability
analysis. We are not conducting a cost per quality adjusted  IUT attempt planned.
life year (QALY) analysis (also called cost-utility analysis)  Patient characteristics (age, parity, ethnicity, risk
given the hypothesis that the intervention is cost saving factors for preterm labour)
and does not result in any health decrement. These as-  Outcome data as outlined above (5.1 and 5.2).
sumptions make a cost per QALY analysis redundant. Any
health improvement will be captured by adverse incident Anonymised participant data will be transcribed on to
reporting and secondary maternal and neonatal clinical the study database prepared for the trial by MedSciNet
outcomes, which will be reported alongside costs as part (Stockholm, Sweden).
of the cost-consequences analysis. Anonymous data will be stored on the Preterm Clinical
Network (PCN) Database (REC Ref. 16/ES/0093). Identi-
Women’s experience study (EQUIPTT-Q) fiers (initials, date of birth, hospital and NHS number) re-
When feasible (i.e. participants identified prior to assess- quired to ensure accurate outcome data collection will be
ment and staff available to obtain consent), participants in stored on the separate, but linked, PCN Patient Details
the cluster randomised trial will be invited to complete a Database which is only accessible to local site users.
questionnaire booklet. This includes questions about her Women will be followed up until postnatal discharge.
symptoms, anxiety levels before and after her clinical as- Neonates will be followed up to discharge or 28 days
sessment, the tests she received, her options for further (whichever is sooner). Local research staff will monitor
management and the extent to which she is content with expected dates of delivery and ensure timely collection
Watson et al. BMC Pregnancy and Childbirth (2019) 19:68 Page 5 of 8

of pregnancy outcomes. Outcome data (medical and may be via telephone or face-to-face. The interview
economic) will be collected by review of obstetric hand- schedules will be designed following interim analysis of
held/electronic notes. Outcome data for women who are the questionnaire data and free text comments, and in
transferred outside their local unit will be obtained by consultation with the preterm birth PPI panel. Inter-
contacting the referral hospital. The Emergency Bed Ser- views will be recorded on digital audio equipment and
vice and Neonatal Transport Service are supportive of then transcribed and prepared for analysis.
the study and can assist us in tracking patients who are At the end of the study and once outcomes have been
not accommodated in their local hospital. collected by individual site users, the anonymous data will
For the health economics analysis data will be col- be extracted for analysis by HW from the PCN Database
lected from routine medical records on antenatal admis- onto an Excel spreadsheet. The spreadsheet will be stored
sions, length of stay, outpatient appointments, day cases on a secure password protected computer for review by
and scans for women identified as participants in the two researchers to remove duplicates and participants
trial from the time of first being identified as part of the with missing data that should not be included in the ana-
trial until discharge following delivery. Healthcare re- lysis (HW and NC). Data cleaning will also be performed
source group codes for each event will be collected and using Stata software Version 14 or later (StataCorp, Col-
costs from National Reference costs applied. We will lege Station, Texas) by our trial statistician (PS) prior to
conduct a time and motion study of clinician time and analysis. The data extraction method will be piloted dur-
resources spent identifying a suitable transfer location ing the data analysis plan in the 9-month study period to
on a sample of transfers. We will also conduct a micro ensure any potential pitfalls are identified.
costing of the cost of a transfer. We will also collect and
report statistics and costs per patient for mode of deliv- Data management
ery for women identified as participants in the trial and For the randomized cluster trial anonymised data only
ex-utero transfers and NICU stays for their neonates will be stored on a secure study-specific internet data-
(adjusting for non-singleton births). We will include the base (MedSciNet). All Investigators and study site staff
cost of the intervention which will include the cost of any must comply with the requirements of the Data Protec-
training and the cost of clinician time associated with tion Act 1998 with regard to the collection, storage, pro-
using the app. This information will be collected from the cessing and disclosure of personal information and will
app, patient notes and a time and motion questionnaire uphold the Act’s core principles.
administered to clinicians for a subset of patients. Local principal investigators will ensure that the
For the women’s experience study (EQUIPTT-Q) partici- confidentiality of the patients’ data is preserved. As
pants, following consent, the researcher will enter the permitted by all applicable laws and regulations, lim-
woman’s initials, hospital number, date of birth and date of ited participant attributes such as, age or date of birth
signing consent against the next consecutive EQUIPTT-Q may be used to verify the participant and accuracy of
ID number on the EQUIPTT-Q register. This ID number the participant’s unique identification number. Indi-
will be entered on the consent form and alternative pages vidual participant data will not be disclosed outside of
of the questionnaire booklet. The participant will be asked study staff and will not appear on any reports pro-
to complete part 1 (questions regarding baseline demo- duced by the sponsor. The following people may also
graphics, her symptoms and anxiety levels prior to assess- access these records:
ment) straight away. She will be asked to keep the booklet
with her and to complete part 2 (tests, planned interven-  Study monitors and auditors (including the data
tions, anxiety post assessment and decisional conflict scale monitoring committee), who may work for the
questions) after her clinical assessment. She will then hand sponsor or its authorised representatives, who check
the completed booklet back to staff before leaving the unit. that the study is being performed correctly and that
The researcher checks the booklet for completion and en- the information collected is accurate.
ters data on EQUIPTT-Q datasheet (spreadsheet). The  National and international regulatory authorities
datasheet (which contains no identifiable data) is emailed involved in keeping research safe for participants.
to trial management team on a weekly basis. Data will be
merged into the EQUIPTT-Q SPSS database which is sep- Participant information provided by the EQUIPTT-Q
arate from the main trial database. questionnaire will be labelled with a unique EQUIPTT-Q
Participants selected for interview will be contacted identification number. It will not include and patient-identi-
and given further verbal and written information. If will- fiable information. In order to arrange the interviews, the
ing, arrangements will be made at a time and place con- contact details (email, phone number) of willing partici-
venient to the participant and they will be asked to sign pants will be obtained from hospital records only if the par-
a consent form prior to the interview taking place. This ticipant has been selected for interview.
Watson et al. BMC Pregnancy and Childbirth (2019) 19:68 Page 6 of 8

Published results will not contain any personal data team and local research support teams. Inclusion and ex-
that could allow identification of individual participants. clusion criteria are defined above.
A data monitoring committee (DMC), independent All sites will receive training on the use of the QUIPP
from the sponsor, will evaluate outcome data (e.g. ad- App and management guidance prior to its introduction,
verse neonatal outcomes and maternal hospitalisation) and periodically throughout the trial at peaks of staff
in a blinded fashion at regular intervals through the trial turnover. This training will be delivered by the trial team
with the ability to report any clinical concerns that may at different times and locations to suit each site, for ex-
arise from the blinded data. Because of the parallel clus- ample, at departmental audit meetings and staff induc-
ter design, data will only be available on the unmasked tion. Training will be delivered via clinical vignettes
results at the end of the trial. Depending on the DMC’s which provide opportunity for the clinicians to use the
requests, the closed DMC session will be able to con- app in a training environment. Aide-memoires such as
sider the results of the primary maternal and neonatal lanyards and pens are provided to each site to encourage
endpoints in both arms of the trial and any Serious Un- use of the app. The doctor/midwife assessing the woman
expected Adverse Events and form a view on whether it inputs the gestation, previous history of late miscarriage
is ethical for the trial to continue. or preterm birth, quantitative fetal fibronectin value
and/or ultrasonic cervical length into the app and the
Selection of participants app provides % risk of delivery within 1/2/4 weeks and
Main study inclusion criteria before 30/34/37 weeks. The exact thresholds for admis-
sion or treatment are gestation-dependent and may need
 Pregnant women with symptoms of TPTL to be tailored to individual circumstances. However as a
(contractions or abdominal pain) recommendation, our guidance will suggest a 5% risk of
 Between 23+ 0 and 34+ 6 weeks delivery within 7 days as threshold for antenatal cortico-
 Capacity for quantitative assessment of fetal steroid administration, admission, and/or in-utero trans-
fibronectin and/or transvaginal ultrasonic cervical fer. Ultimately however the management decisions
length if randomised to the intervention group. following the knowledge provided by the QUIPP app will
remain the clinicians’ responsibility. A feedback option is
Main study exclusion criteria available on the app to directly contact the trial team with
specific non-clinical queries. Incentives and competitions
 Definitive diagnosis of labour (i.e. regular painful will also aim to increase recruitment for the qualitative
contractions with cervical change > 3 cm on recruitment.
speculum or digital examination)
 Confirmed ruptured membranes (on speculum
examination) Outcomes
 Significant vaginal bleeding Primary outcomes
Our primary outcome of inappropriate management for
Women’s experience (EQUIPTT-Q) study inclusion criteria threatened preterm labour is defined by:

 Pregnant women with symptoms of threatened  number of inappropriate admission decisions:

preterm labour (contractions or abdominal pain) admitted and delivery interval > 7 days OR not
 Between 23+ 0 and 34+ 6 weeks admitted and delivery interval < 7 days

Women’s experience (EQUIPTT-Q) exclusion criteria and

 Unable or unwilling to give informed consent  number of inappropriate in-utero transfer decisions/
 Under 16 years of age actions: in-utero transfers that occurred or were
 Unable to understand English language sufficiently attempted > 7 days prior to delivery, and ex-utero
to complete the questionnaire booklet transfers within 24 h that should have been in-utero
(attempted and non-attempted)
Implementing the QUiPP app intervention
In each phase, each cluster will recruit consecutive women
with singleton or twin pregnancies presenting between Secondary outcomes
23+ 0 and 34+ 6 weeks presenting with symptoms of pre- Secondary outcomes will include:
term labour. Eligible participants will be identified at
labour ward and day assessment triage by the direct care  All components of primary outcome individually
Watson et al. BMC Pregnancy and Childbirth (2019) 19:68 Page 7 of 8

 Maternal clinical outcomes (e.g. new onset odds of 5:95 which is 1:19. This estimates the harm of a
gestational diabetes, thromboembolic disease and missed imminent delivery (a false negative) is 19 times
confirmed sepsis) as great as the harm of overtreatment (a false positive).
 Neonatal clinical outcomes (e.g. neonatal death prior It was found to be an acceptable threshold for interven-
to discharge, gestational age at delivery, birthweight, tion in our Delphi survey of preterm birth specialists. In
days of supplemental oxygen) the context of national guidance which advocated more
 Process measures (days of maternal hospitalisation, intervention and disregarded the risk of harm this incurs
steroid, tocolytic and magnesium sulphate to the majority [2], this study’s comprehensive evaluation
administration, neonatal intensive-care admissions, of the true clinical impact of the QuiPP app will deter-
ex-utero transfer within 24 h of delivery) mine and justify an appropriate threshold for TPTL
 Compliance to management recommendations intervention for the first time.
If, as we hypothesise, the QUIPP app improves selec-
Information will also be collected on the number of tion of the appropriate women for admission and trans-
preterm deliveries before 34, and 37 weeks, and on the use fer, it will ensure that therapies known to reduce risk of
of treatments to prevent prematurity [antibiotics, cerclage preterm neonatal morbidities are offered to those who
(by site), progesterone]. The authors may be contacted for need them and reduce the associated maternal and fetal
samples of the data collection forms. risks to women who do not. Conversely if a treat-all pol-
icy is justified, we can help define resource use and
Health economic outcomes clinical need to inform policy makers. Beyond obstetrics,
Cost-savings will be generated as a result of: evaluating the impact of an app in an emergency setting,
and our emphasis on balancing harms of over-treatment
 Reduction in the number of inappropriate admission as well as under-treatment, make EQUIPTT a valuable
decisions. contribution to translational medicine.
 Reduction in the number of inappropriate in-utero
Abbreviations
transfer decisions. CI: Chief investigator; DCS: Decision conflict score; EQUIPTT: The Evaluation of
the QUIPP app for Triage and Transfer; EQUIPTT-Q: EQUIPTT Women’s
Women’s experience study (EQUIPTT-Q) Experiences Study; NHS R&D: National Health Service Research &
Development; NPV: Negative predictive value; Participant: An individual who
Our key outcome is difference in VAS anxiety scores be- takes part in a clinical trial; PI: Principle Investigator; PPI: Patient and Public
tween women before and after assessment, between hos- Involvement; PPV: Positive Predictive Value; QALY: Quality Adjusted Life Year;
pitals with and without QUIPP app in use at the time of qfFN: Quantitative fetal fibronectin; REC: Research Ethics Committee;
TPTL: Threatened Preterm Labour; TSC: Trial Steering Committee; VAS: Visual
assessment. We will also analyse difference between Analogue Scale
DCS scores after assessment, in hospitals with and with-
out QUIPP app in use. Acknowledgements
The authors would like to acknowledge contributions from Rachael Hunter
(Health Economist, Priment Clinical Trials Unit, University College London),
Discussion Russell Hughes, (Genetic Digital) and the Preterm Birth Patient and Public
Management of TPTL has represented a clinical conun- Involvement Group, Guy’s and St Thomas’ Hospital.
This work is supported by Biomedical Research Centre at Guy’s and St
drum for decades. Intervention “just-in-case” results in Thomas’ NHS Foundation Trust and King’s College London. The views
many unnecessary admissions, women being transferred expressed are those of the authors and not necessarily those of the NHS, the
out of their local hospital unnecessarily and most women NIHR or the Department of Health and Social Care.
receiving unnecessary drugs, such as steroids and tocoly- Funding
tics. It also prevents appropriate transfers as neonatal cots The development of the QUiPP app and the EQUIPTT study are funded by the
are blocked unnecessarily, resulting in more dangerous Guy’s and St Thomas’ Charity (Registered Charity No. 1160316) and Tommy’s
(1060508). The EQUIPTT protocol was peer-reviewed by the Guy’s and St
ex-utero transfers. One reason for the current trend for Thomas’ Charity. Guy’s and St Thomas’ Charity reviewed the design of the study
over-treatment of TPTL is concern that the false-negatives and made recommendations (in particular recommended additional health
any triage system entails. This study will determine the economics evaluation). Neither funder has been involved in the collection,
analysis, interpretation of data, or in writing the manuscript. As it is a portfolio
risks and benefits of limiting management and treatment study, recruitment is supported through local Clinical Research Networks. King’s
to those most at risk of delivery. College London and Guy’s and St Thomas’ NHS Foundation Trust are co-
The exact thresholds for admission or treatment are sponsors of the study but neither sponsor are providing funding for the trial.
gestation-dependent, may need to be tailored to individ-
Availability of data and materials
ual circumstances and will remain the clinicians’ respon- Not applicable.
sibility. As a recommendation, our guidance will suggest
a 5% risk of delivery within seven days as threshold for Authors contributions
AHS and HAW conceived and designed the study. HAW and AHS drafted the
antenatal corticosteroid administration, admission, and/ original grant proposal and trial protocol with input from PS, RMT and JC.
or in-utero transfer. A 5% threshold is equivalent to an During the trial PS provides methodological and statistical expertise. NC, KK,
Watson et al. BMC Pregnancy and Childbirth (2019) 19:68 Page 8 of 8

HAW and AHS provide day to day oversight of participant recruitment, day 11. Porcellato L, Masson G, O'Mahony F, Jenkinson S, Vanner T, Cheshire K,
to day running of the trial, data collection and liaison with other sites. HAW Perkins E. ‘It's something you have to put up with’—service users’
drafted the manuscript with additional input from NC and KK. All authors experiences of in utero transfer: a qualitative study. BJOG Int J Obstet
critically reviewed and approved the final version of the manuscript. Gynaecol. 2015;122(13):1825–32.
12. Gale C, Hay A, Philipp C, Khan R, Santhakumaran S, Ratnavel N. In-utero
Ethics approval and consent to participate transfer is too difficult: results from a prospective study. Early Hum Dev.
EQUIPTT was granted a favourable ethical opinion (REC reference 17/LO/1802) 2012;88(3):147–50.
by the London Bridge Research Ethics Committee on 21 November 2017. 13. Abbott, D. S., S. K. Radford, P.T. Seed, R. M. Tribe, and A. H. Shennan. Evaluation
For the cluster trial, individual consent was not applicable; however consent of a quantitative fetal fibronectin test for spontaneous preterm birth in
was obtained for each study site. In the women’s experiences (EQUIPTT-Q) symptomatic women. Am J Obstet Gynecol 2013; 208, no. 2: 122 e1–5.
part of the study, participants will be approached individually by the local 14. Abbott, D. S., N. L. Hezelgrave, P. T. Seed, J. E. Norman, A. L. David, P. R.
research teams for inclusion, information leaflet provided and informed Bennett, J. C. Girling, and others. Quantitative fetal fibronectin to predict
consent obtained. If appropriate, they will be given verbal and written preterm birth in asymptomatic women at high risk. Obstet Gynecol 2015;
information about the questionnaire and given the opportunity to ask 125, no. 5: 1168–1176.
questions. If they are happy to proceed they will be asked to sign a consent 15. Kuhrt K, Hezelgrave NL, Foster C, Seed PT, Shennan AH. Development and
form and to indicate if they would be happy to be contacted, at a later date, validation of a tool incorporating quantitative fetal fibronectin to predict
should they be selected for interview. Those that are selected will be given spontaneous preterm birth in symptomatic women. Ultrasound Obstet
an additional information sheet and asked to sign another consent form Gynecol. 2016;47(2):210–6.
prior to the interview taking place. 16. Watson HA, Carter J, Tribe RM, Seed PT, Shennan AH. The QUiPP app: a safe
alternative to a treat-all strategy for threatened preterm labor. Ultrasound
Consent for publication Obstet Gynecol. 2017. https://doi.org/10.1002/uog.17499.
Not applicable. 17. Sokol DK. Update on the UK law on consent. BMJ (Clinical research ed)
2015. 350:h1481.
Competing interests 18. Arnup SJ, Forbes AB, Kahan BC, Morgan KE, McDonald S, McKenzie JE. The
In the last 3 years, AS has been a consultant for GSK (unpaid) and UK CI for a use of the cluster randomized crossover design in clinical trials: protocol for
tocolytic study, funds paid to institution. He has advised Hologic and Medix a systematic. Syst Rev. 2014;3:86.
biochemica on clinical studies (unpaid). AS is a consultant to Owen 19. O’Connor AM. Validation of a decisional conflict scale. Med Dec Making.
Mumford, unrelated to the topic of this paper. NC received financial 1995;15(1):25–30.
assistance to cover expenses, paid to institute, to provide an educational talk 20. Richie J, Spencer L. Qualitative data analysis for applied policy research. In:
on preterm birth from Hologic, USA. The other authors did not report any Analysing Qualitative Data. Bryman and burgess ed. London: Routledge; 1994.
potential conflicts of interest.

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

Received: 6 July 2018 Accepted: 31 January 2019

References
1. Iams JD, Newman RB, Thom EA, Goldenberg RL, Mueller-Heubach E,
Moawad A, Sibai BM, Caritis SN, Miodovnik M, Paul RH. Frequency of uterine
contractions and the risk of spontaneous preterm delivery. N Engl J Med.
2002;346(4):250–5.
2. National Collaborating Centre for Women's and Children's Health (UK).
Preterm labour and birth. In: National Institute for health and care
excellence: clinical guidelines. London: National Institute for health and care
excellence (UK); 2015.
3. Roberts D, Dalziel S. Antenatal Corticosteroids for Accelerating Fetal Lung
Maturation for Women at Risk of Preterm Birth (Review). Cochrane Database
Syst Rev. 2006;19. https://doi.org/10.1002/14651858.CD004454.pub2.
4. Stutchfield PR, Whitaker R, Gliddon AE, Hobson L, Kotecha S, Doull IJM.
Behavioural, educational and respiratory outcomes of antenatal
betamethasone for term caesarean section (ASTECS trial). Arch Dis Child
Fetal Neonatal Ed. 2013;98(3):F195–200.
5. Jobe AH, Wada N, Berry LM, Ikegami M, Ervin MG. Single and repetitive
maternal glucocorticoid exposures reduce fetal growth in sheep. Am J
Obstet Gynecol. 1998;178(5):880–5.
6. Kemp, M W, J. P. Newnham, J. G. Challis, A. H. Jobe, and S. J. Stock. The
Clinical Use of Corticosteroids in Pregnancy. Hum Reprod. 2016; update 22,
no. 2: 240–259.
7. Moisiadis VG, Matthews SG. Glucocorticoids and fetal programming part 1:
outcomes. Nat Rev Endocrinol. 2014;10(7):391–402.
8. Modi N, et al. The effects of repeated antenatal glucocorticoid therapy on
the developing brain. Pediatr Res. 2001;50:581–5.
9. Davis EP, Sandman CA, Buss C, Wing DA, Head K. Fetal glucocorticoid
exposure is associated with preadolescent brain development. Biol
Psychiatry. 2013;74:647–55.
10. Shlossman PA, Manley JS, Sciscione AC, Colmorgen GH. An analysis of
neonatal morbidity and mortality in maternal (in utero) and neonatal
transports at 24-34 weeks' gestation. Am J Perinatol. 1997;14(8):449–56.