Can Text Reminders Improve Uptake of Cervical Screening?

In Hillingdon specifically, where this randomised controlled trial (RCT) will be set, the uptake of the prevalent population of women (24-29 years as a proxy), was only 35% in 2012, whilst the incident population (30-64years as a proxy) was slightly higher at 51% (HSCIC2, 2013).

One strategy which healthcare has previously employed to improve uptake of both screening and routine appointments is the use of SMS or "text message" reminders. Standard reminders have often been used in healthcare and act as "primers", prompting people to make healthier choices (Blumenthal-Barby \& Burroughs, 2012). They also serve to reduce the practical barriers preventing women from attending i.e. failure to book or remember the time/location of the appointment. Previous research investigating the effectiveness of appointment reminders demonstrated that both telephone and postal reminders, although at times effective, were often impractical, unsustainable, and/or unaffordable (Gurol-Urganci et al, 2013, Perron et al, 2013).

SMS interventions have been effective at improving medication adherence, disease management and smoking cessation (Krishna et al, 2013; Abroms et al, 2014). They have also proven effective at reducing "Do Not Attends" (DNAs) within healthcare, estimated to cost the NHS £600 million per year in wasted time and resources (Dr Foster Research, 2009). SMS reminders were also found to be very successful at increasing the attendance of breast screening appointments, increasing the uptake by 17% compared to usual care (Kerrison et al, 2013). As well as improving attendance, SMS reminders have also proved effective at reducing costs, increasing the efficiency of clinics, enabling more effective booking, reducing mismatches between demand and capacity, and generally helping to increase productivity within healthcare.

SMS interventions have many benefits over alternative mediums. Firstly an SMS is cheap, roughly 3p per text whilst a letter can cost 10 times that. SMS is also capable of reaching the recipient wherever they are whilst letters can only be delivered to recipients last known address. This is particularly important, as people are more likely to retain their mobile numbers nowadays even if they change address, especially in London where the population is notoriously mobile. Mobile phone use is also ubiquitous in the UK, with a recent report estimating that over 99% of adults in London owned a mobile phone in 2013 (Ofcom, 2013). Another advantage of SMS is that there is evidence that people increasingly use electronic day planners and therefore a message direct to a person's phone encourages people to promptly record appointments and even allowing them to set alarms creating their own reminder. SMS interventions would therefore be as wide reaching as a postal intervention (if not more so) at a fraction of the cost and with greater accuracy.

Although previous evidence provides a sound rationale that SMS interventions have great potential for cervical screening, there is still a considerable amount of information required in order to introduce this intervention as standard practice.

The research also aims to evaluate whether minor differences in wording affect women's response to these messages. This idea stems from "Nudge" theory, a form of Libertarian Paternalism or "Asymmetric Paternalism" (Thaler and Sunstein, 2003). This concept is derived from a mix of cognitive psychology and behavioural science which is based on the principal that humans, due to a lack of information, experience, cognitive ability and self-control, often fail to act in a way that promotes their own welfare (Sunstein and Thaler, 2003). Libertarian paternalisms therefore encourages individuals to act in their, and/or societies, best interests whilst maintaining freedom of choice (Aggarwal et al 2014). Although the use of "nudge" has been viewed as controversial within healthcare by some, this study has justified its use by avoiding any form of coercion, and making it clear that this text is simply a reminder of the normal invitation process.

There is mounting evidence demonstrating that even small changes in the content and framing heavily influences people decisions (Haynes et al, 2013; Hallsworth et al, 2014). Whether the positive effects of small wording changes or "nudging" is reflected in women's attitudes towards cervical screening programs has yet to be evaluated. This information will be crucial in order to maximise efficiency of this intervention and knowledge gained could be applied to similar interventions in healthcare making them more cost effective. We also hope to evaluate whether people with different levels of deprivation respond differently to the different messages. This is hugely important, as tailoring the message to the population is crucial to improving the efficiency of the intervention and to reduce health inequalities within London.

In light of this evidence we have chosen seven different text message options, all of which have according to research great potential to improve the uptake of screening. These include a control of no text message, a neutral reminder and five reminders with behavioural messages to prompt uptake. The intervention arms according to research have great potential to improve the uptake of screening. It is well established that reducing the effort required to perform an action increases the likelihood of follow-through (Leventhal et al, 1965).

All text messages begin with the test message, followed by a standard reminder message that is constant across the messages and the participants GP practice phone number.

The chosen messages are no text message, a neutral message; a messenger; two social norms; a gain frame and a loss frame highlighting the benefits of cervical screening.

Neutral Reminder

The neutral message will not "nudge" but simply remind the woman to book a cervical screening appointment, acting as a direct reminder of the letter invitation sent out as normal practice. It will act as a control for the other types of message.

"Your cervical smear test is due. To book please call <xxxxxxxxxxx>"

Messenger Reminder

Some evidence suggests that people respond better to invitations from their personal practice rather than a centralised NHS organisation (Segan et al, 1998; Arbyn et al, 2010). This is especially true of young women, immigrants, urban residents and those with low socioeconomic status, which are key demographics we hope to improve uptake in through this intervention (Bosgraaf et al, 2013).

"<GP NAME>: Your cervical smear test is due. To book please call <xxxxxxxxxxx>"

Social Norms Messages

Social norms messages focus on the principal that people have a tendency to overestimate the extent to which their peers perform acts, causing themselves or others harm. Evidence suggests educating people about the actual degree to which people perform these acts can alter behaviour (Perkins et al, 2002). This theory, combined with the fact that people are strongly influenced by comparison to what others do will hopefully prompt women to take up the offer of cervical screening. We have also chosen to use a local statistics as personalised messages have a greater influence as they increase the relevance of the message to the women (Institute for government: Cabinet Office, 2010)

1. Social Norms (local, proportion) reminder: "Last year in Hillingdon 7 out of 10 women took part in cervical screening. Your cervical smear test is due. To book please call <xxxxxxxxxxx>"
2. Social Norms (local, total number) reminder:

"Last year 12000 women in Hillingdon took part in cervical screening. Your cervical smear test is due. To book please call <xxxxxxxxxxx>"

Gain and Loss Frame Messages

The role gain and loss frame messengers to encourage uptake of good health behaviour is not entirely clear. Some evidence points to humans having "have a stronger aversion to losses than we do affinity to gains" therefore highlighting losses may have a greater impact (Blumenthal-Barby \& Burroughs, 2012). On the other hand a meta-analysis found that gain frame was more likely to be effective in changing prevention behaviour. (Gallagher, 2001). To evaluate this further a gain and loss frame message were included above. Statistics included in these messages were taken from (Peto, 2004) and correlate to widely used statistics on cancerscreening.nhs.uk to maintain consistency with other information women might be seeing.

1. Gain frame "Cervical cancer screening saves 4500 lives in England every year. Your cervical smear test is due. To book please call <xxxxxxxxxxx>"
2. Loss frame

"Failing to attend cervical screening could lead to 4500 avoidable deaths in England each year. Your cervical smear test is due. To book please call <xxxxxxxxxxx>"

POTENTIAL ETHICAL ISSUES

The main ethical issue is that in order to analyse the data, the researcher will need to access minimal amounts of patient data. The researcher will specifically need to access a list of women in Hillingdon who are sent an invitation to cervical screening on a weekly basis, if they are registered with a consenting GP. All participating GPs (which to date include over 85% of those approached for the trial) have consented to share relevant data with the research team and are happy with the safeguards put in place to protect confidentiality. These include NHS secure emails, password protection on all patient data, anonymisation of data (identifiable by NHS number only) and strict adherence to all NHS information governance regulations.

On this list, each participant will be identified by their NHS number, and this will be the only unique identifier. The researcher will also need information about their age (not date of birth), participants "attended/not attended status" at the end of the 18 week follow up period and the GP practice code, to ensure we only include women registered with consenting GPs. We would also like to access the second half of participants' postcode to allow for a basic deprivation analysis.

Before accessing this information we will offer women the option to opt out of the research and therefore not share this data or receive a text message reminder. This will be achieved by the cervical screening (call-recall) team adding an insert on behalf of the GP, into the screening invitation letter to all eligible women. This insert will provide informed consent slip and contact details of the researcher if women would like further information. They will have 2 weeks to respond and be removed from the study.

As an extension of this research, the trial will also investigate how variations in the wording of the text messages affect uptake. This wording has been carefully selected with the guidance of the NHS Cervical Screening Program (NHSCSP) as well as Public Health England psychology experts and any data within the text messages is up to date and sourced from the NHSCS.