Population-Based Cancer Prevention Education Intervention Through mHealth: A Randomized Controlled Trial

Cancer is currently the second leading cause of death worldwide and the projections estimate that, not only cancer deaths will likely surpass other diseases and become the leading cause of death globally, but cancer will also pose a dramatic economic burden as populations age. In 2020, 282 421 new cancer cases were diagnosed in both sexes in Spain and 113 054 people died because of cancer. Colorectal, prostate and breast cancers were the most common cancers, while prostate, colorectal and lung cancers were the most lethal ones [1]. It is well established that at least 40% of cancers are attributable to modifiable risk factors such as tobacco and alcohol consumption, overweight, or sun exposure, therefore, primary prevention would be possible by avoiding or reducing exposure to certain risk factors and modifying unhealthy behaviours [2, 3]. Likewise, secondary prevention through cancer screening to detect certain types of cancer early, followed by effective treatment, are the main tools to avoid deaths from cancer [4]. Thus, prevention offers the most cost-effective cancer control strategy in the long term and, therefore, great potential for public health [5].

Health education and health promotion aim to increase individual and societal awareness of health and its determinants, as well as to develop and strengthen the health empowerment of the individual. From this perspective, cancer prevention education would improve the knowledge about established cancer risk factors, unhealthy lifestyles, and evidence-based interventions towards reducing the cancer burden. Consequently this may affect individuals' willingness to adopt and maintain healthy attitudes as well as to uptake cancer prevention interventions such as vaccination and cancer screening [6]. With the aim to increase European Union (EU) citizens’ awareness on cancer prevention, the European Code Against Cancer (ECAC) was set by the European Commission (EC) in 1987 and it has been updated several times. The ECAC is a key educational cancer prevention tool that translates the latest scientific evidence on multiple risk factors and effective interventions, into a list of recommendations for the public to reduce cancer incidence and mortality in the EU. Currently, the ECAC 4^th edition is in force, listing "12 ways to reduce your cancer risk" [7], and it is complemented with a website containing extensive information on cancer prevention for non-specialized audiences, in a "question and answer" format [8]. In 2021, a new update of the ECAC was announced by the EC in the Europe’s Beating Cancer Plan, along with a project to develop an EU Mobile App for Cancer Prevention to encourage the uptake of the ECAC recommendations [9].

Periodic population-based surveys conducted in some European countries have studied public’s knowledge of cancer risk factors and symptoms, behaviours, and risk perceptions. However, current literature suggests that knowledge on the association of cancer and risk behaviours such as unhealthy diet or alcohol consumption, is still scarce in the general population [10‐12]. In studies recently conducted in Spain, authors concluded that the ECAC lacks real impact as it is unknown to a large part of the population (86.7%; IC95%:85.2–88.2) [13], perhaps because the channels for its dissemination were insufficient and unsuitable for the target population. Addressing this lack of awareness is crucial to empowering the public to make informed choices to reduce the risk of developing cancer and prevent death. Yet, for an educational and health promotion tool to be effective in increasing awareness and to have a significant public health impact, it needs to ensure that a large proportion of the target population is reached and exposed to such educational content [14].

Mobile phones have become the most accessible form of communication in history as they offer mobility, flexibility, convenience, and real-time communication. Currently, there are more than eight and a half billion mobile phone subscribers in the world [15]. Mobile health or “mHealth” refers to the application of mobile technologies, including phones, tablets, etc., as a strategy for supporting and improving the performance of healthcare and public health practices, as it can reduce cost and travelling time for patients and care providers. Given the global reach of mobile phones to a variety of audiences, interventions to deliver cancer prevention education based on mHealth have the potential to impact large groups of people, especially those that are difficult to reach, to produce changes in unhealthy behaviours and improvements in health outcomes. Recent literature examining the effectiveness of mHealth interventions across health topics, such as physical activity, diabetes management or antiretroviral therapy adherence, has showed that these interventions are significantly more effective at improving health outcomes than comparison interventions [16]. However, mHealth applications’ implementation is difficult and complex, involving more challenges in management, social, cultural, and organisational issues than on technical aspects. A variety of factors can influence the effectiveness of mHealth interventions ranging from participant demographics (e.g., gender, age), disease status (e.g., disease self-management, disease prevention), or perceived ease of use, usefulness, security and privacy, and reliability factors influencing adoption of mobile devices [17‐20].

With the aim to gather knowledge on the feasibility of a future cancer prevention education intervention, we conducted an implementation research pilot study to assess the adoption, fidelity, appropriateness, and acceptability of an intervention to effectively disseminate cancer prevention messages to the public through a mHealth implementation strategy. The study was carried out in the Autonomous Community of Andalusia, the largest in number of inhabitants in Spain (17.9%), which has already implemented a population-based mHealth multi-channel platform to provide healthcare services and personalized health information to its citizens [21, 22].

Beneficiaries of the SSPA from a sample of 1991, registered or not in the SR platform, were randomly allocated to receive 2 or 7 cancer prevention messages per week through an app notification (“push”) or SMS. To investigate the feasibility of a future large-scale educational intervention at population level, participation rates are important to be estimated as the result of getting a positive response at the recruitment point plus accepting to participate in the study (in other words, the willingness to receive messages among respondents). The total response rate was 67.8% (95% CI 65.7 to 69.8%), however, the percentage fell to 31.0% (95% CI 29.0 to 33.1%) when looking at the participation rate. In the randomized groups, the response rate and participation rates were, respectively: 72.0% (95% CI 67.9 to 75.8%) and 34.8% (95% CI 30.8 to 39.1%) in those using the SR app (G1), and 61.0% (95% CI 56.7 to 65.2%) and 26.6% (95% CI 22.9 to 30.6%) in those registered in the SR platform through SMS (G2); those not registered in the SR platform (G3 and G4 combined) had a response rate of 69.0% (95% CI 66.1 to 71.8%) and participation rate of 31.6% (95% CI 28.8 to 34.5%).

The average age of those who accepted participate in the study was 49 years old, 56.2% were women, 61.9% were employed, 80.4% had secondary education or more, and did not live alone (91.9%). After the second randomization, the distribution of participants across the four groups was homogeneous by sociodemographic variables, except for age, employment status and level of education. The mean age was significantly higher among individuals in G4. This group also had a significantly higher proportion of individuals who were retired and with a level of education of primary or less. The random selection resulted in beneficiaries from the 8 administrative provinces of Andalusia, being Sevilla and Malaga the most populated ones, thus obtaining a geographic representation (Table 1).

At follow-up after the four weeks of intervention, 453 (73.4%) individuals were lost (Table 2). Most common reasons for not answering the post-intervention survey were “not being reachable” (51.7%) and “not willing to participate any longer in the study” (17.8%). In addition, due to the Covid-19 pandemic some disruptions hampered conducting the post-intervention survey at the time planned, therefore, to avoid recall bias, those individuals who answered the post-intervention survey more than one month after they received the last message (N = 90, 19.9%) were excluded from the study and, thus, considered as “lost at follow-up”. Consequently, due to the small sample size at follow-up, the study team decided to merge all subjects receiving the messages through the same channel: G1 and G3 receiving messages through the app, and G2 and G4 receiving messages through SMS. A total sample of 108 (65.9%) participants in the app arm and 56 (34.1%) in the SMS arm completed the post-intervention survey. As shown in Table 2, the group distribution according to reception channel was homogeneous, since no differences were observed by socio-demographic variables nor by mHealth variables, except for age, as individuals receiving the messages through SMS were significantly older than those using the app. As regards mHealth-related variables (Table 2), almost all participants had a smartphone and permanent access to the Internet through it. As for general access to the Internet through other devices, 97 (89.8%) participants from the app group and 44 (78.6%) participants from the SMS group reported being able to access the Internet from a portable computer (PC) or tablet, mostly from home. From a scale from 0 to 10, the mean ability to browse a website of our sample was 7 points. Eighty-eight (81.5%) participants from the app group and 39 (69.6%) from the SMS group felt confident navigating through a telephone (app group), a PC, or a tablet, instinctively. Finally, 91 (84.26%) participants using the app would be willing to receive cancer prevention messages through their phone, in comparison to (78.57%) participants receiving SMS.

Implementation research aims to close the research-to-practice gap, and support scale-up of evidence-based interventions. It involves studying the challenges and strategies involved in deploying new innovations and technologies in real-world healthcare settings, by measuring implementation outcomes, which are distinct from patient outcomes (e.g., symptoms, behaviours) and health service outcomes (e.g., efficiency, safety). Proctor et al. proposed a taxonomy of eight implementation outcomes, defined as “the effects of deliberate and purposive actions to implement new treatments, practices, and services” consisting of: acceptability, appropriateness, adoption, feasibility, fidelity, implementation cost, penetration, and sustainability of evidence-based practice [26]. In our study, we aimed to measure some of these outcomes as indicators of the feasibility of deploying a cancer education intervention using mHealth as implementation strategy, through the currently in use SR platform of the SAS that covers 67% of the Andalusian’s population.

Although mHealth is a powerful implementation strategy, still most of the studies focus on the efficacy of interventions to improve complex lifestyle behaviours in the short term, instead of the impact at population level. Some studies assessing the effectiveness (most frequently through SMS-based interventions) have provided preliminary evidence of small, sustained effects to produce positive change in preventive behaviours [27]; positive results in chronic disease management, reduction of deaths and hospitalisation; or improvement of healthcare attendance rates; however, the methodological quality of most studies has been described as low [28]. Well-designed population-based implementation research studies are needed to investigate the external validity or “real-world” impact of a scaled-up intervention to reach massive dissemination effectively, and to gather information on the coverage, acceptability, and adherence to the intervention, as most mHealth research shows high self-reported acceptability to receive messages but low coverage or reach of the intervention [29]. In line with this, our results show an overall low participation rate in comparison to more than double initial positive response rate, showing that there is a considerable mismatch between the theoretical coverage that the intervention could have achieved and what happened in practice. Our study also shows generally high access to the Internet through the mobile phone or other devices, high ability and confidence to browse a website, and high willingness to receive cancer prevention messages in the phone. However, while the uptake or adoption was shown to be high in those participants that completed the study, in particularly in the SMS arm, a discrepancy with the results on the fidelity of the intervention to the implementation plans becomes apparent, as messages received and read weekly, either by app or SMS, were lower than expected. Recall bias may explain at least part of these results [30], along with social desirability, which is the tendency of respondents to distort self-reports by providing responses that they think are consistent with social norms and expectations [31]. In addition, data show a high recall bias in the higher frequency group, receiving 7 “push” notifications or 7 SMS per week, as most of them recalled having received or read a smaller number of messages than anticipated; and only one reported having received and read seven messages over the last week. In the app group, this may be explained by the fact that the “push” notification needed to be activated by the user to acknowledge reception of the messages; also, some users may have uninstalled the app during the study, without notifying the interviewers at the post-intervention survey. To overcome this, system usage data should be enhanced to quantitatively measure engagement with mHealth interventions [32].

As regards appropriateness of the implementation strategy, receiving cancer prevention messages through mHealth was shown to be appropriate in terms of the frequency and the time when the messages were received. Likewise, the messages were well understood, especially by those in the SMS group, and perceived useful, mainly by those receiving 2 SMS per week. However, the willingness to use complementary communication modalities such as website as a source of additional information, was shown to be low (below 22% in both groups). Finally, when assessing the acceptability or potential success of scaling-up the intervention, and despite the inconsistency with the fidelity outcome, the willingness to receive more messages in the future was shown to be very high (above 80% in the app group and 90% in the SMS group).

The observed data show that an intervention for cancer prevention education, using mHealth as implementation strategy, is acceptable and feasible. However, to investigate potential changes in attitudes and knowledge about cancer prevention as a prerequisite for effective behavioural changes, before scaling-up such an intervention, it would need to be adjusted to increase coverage and adoption of the intervention, and fidelity would need to be ensured, to guarantee penetration and sustainability in the healthcare system of Andalusia. The results of this study may contribute to inform the design of future interventions using mHealth at large scale, such as the future EU Mobile App for Cancer Prevention, to achieve a broad coverage across the EU as well as to ensure adoption of the cancer prevention messages as those promoted by the ECAC.