SMS Behaviour Change Communication and eVoucher Interventions to Increase Uptake of Cervical Cancer Screening in the Kilimanjaro and Arusha Regions of Tanzania: A Randomised, Double-blind, Controlled Trial of Effectiveness

Cervical cancer, although almost entirely preventable through cervical cancer (CCa) screening (CCS) and human papillomavirus (HPV) vaccination, is the leading cause of cancer deaths among women in Tanzania. Barriers to attending CCS include lack of awareness of CCS, affordability concerns regarding screening, and travel cost. This study compared the effectiveness of SMS (short message service) behaviour change communication (BCC) messages and of SMS BCC messages delivered with a transportation electronic voucher (eVoucher) on increasing uptake of CCS.

Door-to-door recruitment was conducted between February 1 and March 13 2016 in randomly selected enumeration areas in the catchment areas of two hospitals, one urban and one rural, in Northern Tanzania. Women aged 25-49 who were able to access a mobile phone were randomised using a computer-generated 1:1:1 sequence stratified by urban/rural to receive either (1) 15 SMS BCC messages, (2) the identical 15 BCC messages sent to the SMS group plus an eVoucher for return transportation to CCS, or (3) one SMS informing about the nearest CCS clinic (control group). Fieldworkers and participants were masked to allocation. The study used a customised, open-source Medic Mobile software platform to send the SMS and for data collection. The primary outcome was attendance at CCS within 60 days of randomisation.

Sensitisation methods commonly used in the regions were employed in the study area, including sensitisation via church announcements and posters and sensitisation from fieldworkers recruiting participants.

Draft SMS were informed by literature about barriers towards CCS in Sub-Saharan Africa (SSA) and low- and middle-income countries (LMICs) and were developed based on the Health Belief Model (HBM) after collaboration with laypeople and medical experts. A motivational tone (vs. an informational tone) was used for the BCC SMS based on research on SMS in increasing uptake of HIV testing in South Africa. A medical translator translated the SMS into Swahili. To ensure the content validity and cultural sensitivity of the SMS, the researchers conducted 5 focus groups with community members and CCS nurses. The qualitative data were used to modify the SMS and to select the final group of SMS.

For the intervention groups, the message schedule was designed to send a total of 15 unique SMS, 3 at enrolment and 1 sent every 1 or 2 days thereafter until day 21. Two additional SMS containing the location and dates of screening services were added following the mini-process evaluation, conducted a quarter of the way through data collection, which found that 25%-50% of the SMS were not reaching participants.

Participants (n=866) were randomly allocated to the BCC SMS group (n=272), SMS + eVoucher group (n=313), or control group (n=281), with 851 included in the analysis (BCC SMS n=272, SMS + eVoucher n=298, control group n=281). By day 60 of follow-up, 101 women (11.9%) attended CCS. "Due to the wide variety of barriers to CCS, and the fear and misconceptions that surround CCa, it is unsurprising that, although the SMS intervention was effective, the change in uptake was relatively modest." Intervention group participants were more likely to attend than control group participants (SMS + eVoucher odds ratio (OR): 4.7, 95% confidence interval (CI) 2.9 to 7.4; SMS OR: 3.0, 95% CI 1.5 to 6.2).

There were larger ORs in the rural areas versus the urban area for both the SMS and SMS + eVoucher interventions; three factors might explain this. First, in a previous study conducted in the same region, greater perceived travel barriers were reported in the rural areas versus the urban areas. Second, the rural area exhibited a lower level of baseline knowledge (9%) than the urban area (18%). This difference could have amplified the effects of the SMS behaviour change in the rural area. Finally, a lower baseline rate of screening in the rural area could also help to explain the greater impact of both interventions (SMS and SMS + eVoucher) in that area.

Due to the mobile health (mHealth) nature of the interventions, the researchers were unable to enrol women in the lowest socioeconomic status. As mobile phone ownership becomes increasingly pervasive, this inequity may be ameliorated. The researchers stress that attention must be devoted to equity in order to avoid marginalising disenfranchised groups.

Previous SMS behaviour change interventions have demonstrated that two-way SMS interventions are more effective than one-way interventions. In the present study, participants called/texted the phone number sending the messages. A fieldworker was responsible for answering these questions/concerns either via text or voice and for logging the communication. The researchers indicate that when this intervention or similar interventions are scaled up, it should be ensured greater capacity is available to facilitate two-way communication between participants and study officials.

In conclusion: "Health/cancer educational initiatives involving inperson contact are required to address the more fear-inducing and sensitive aspects of education and counselling surrounding CCa. Harnessing the potential of these mHealth interventions requires a multifaceted, equity-focused approach which includes interpersonal elements."