Can social media encourage diabetes self-screenings? A randomized controlled trial with Indonesian Facebook users

Manuela Fritz, Michael Grimm, Ingmar Weber, Elad Yom-Tov, Benedictus PradityaNpj Digit. Medicine 2024 · DOI · OpenReview

TL;DR

This study investigated whether Facebook ad campaigns could effectively encourage diabetes self-screenings in Indonesia, a middle-income country with a high diabetes burden. They found that "consequences" and "shock" framed ads significantly outperformed others in driving both clicks and completed screenings, suggesting that loss-framed messaging is a powerful tool for public health interventions on social media. The campaign also proved to be cost-effective, with an estimated cost of $9 per newly diagnosed case.

What Problem Does This Paper Address?

Non-communicable diseases (NCDs), particularly diabetes, are a growing global health crisis, especially in middle-income countries (LMICs) where awareness and screening rates are often low. While mass media campaigns are promoted for NCD awareness, social media advertising has been largely overlooked. This paper addresses the challenge of motivating individuals, especially those without obvious symptoms, to undergo diabetes self-screening in an LMIC context. Specifically, it explores whether Facebook, a widely used platform in Indonesia, can be leveraged for such campaigns and whether specific ad framings are more effective in driving engagement and action.

Methods

The researchers conducted a randomized controlled trial (RCT) using Facebook's advertising platform. They designed five different ad creatives with varying framings: "consequences" (highlighting health impacts), "shock" (emphasizing deadly outcomes), "family" (focusing on community impact), "geography" (linking to local prevalence), and "religion" (using a religious quote). These ads were randomly distributed to Facebook users aged 35 and above in Jakarta and Yogyakarta, Indonesia. The campaign objective was set to optimize for "conversions," defined as completed screening questionnaires. Users who clicked on an ad were directed to a campaign website where they could complete a diabetes risk screening questionnaire. A follow-up survey was administered to a subset of participants to assess their compliance with recommendations for professional screening.

Datasets

The study utilized Facebook user data for targeting and ad delivery. The primary dataset for analysis consisted of 1469 completed diabetes risk screening questionnaires collected from Facebook users in Jakarta and Yogyakarta. A follow-up survey dataset of 53 participants was also collected. The study also references existing datasets and literature for comparison, including Indonesian health statistics and previous studies on social media campaigns and health behaviors.

Key Insights and Ideas

  • Loss-framed ads are highly effective: The "consequences" and "shock" ads, which highlight the negative health outcomes of diabetes, significantly outperformed other ad types in driving both clicks and completed self-screenings. This supports the theoretical premise that loss-framed messages are more potent in motivating health behaviors, especially when individuals might perceive the behavior as risky or uncertain.
  • Gender-specific responses: While loss-framed ads were generally more effective, the study observed that women responded more differentially to the various ad framings compared to men. For men, the "consequences" ad was still superior, but other ads showed more similar performance. This highlights the importance of considering gender in tailoring public health messaging.
  • Reaching at-risk populations: The campaign successfully reached individuals who were at high risk for diabetes, with a significant portion of those identified as high-risk being unaware of their status. This suggests that social media campaigns can be effective in identifying and engaging individuals who could benefit most from early intervention.
  • Cost-effectiveness is promising: The campaign demonstrated a low cost per person screened (US$0.75) and a projected cost of approximately US$9 per newly diagnosed case if scaled up. This makes it a potentially highly cost-effective public health intervention, especially for resource-constrained LMICs.

Key Results

  • The campaign reached 286,776 individuals and generated 758,977 impressions in three weeks with a budget of US$1060.
  • 1469 individuals completed the diabetes self-screening questionnaire.
  • The "consequences" ad was the most effective for conversions, making users 57% more likely to complete the screening compared to the "family" ad. The "shock" ad was also significantly effective.
  • The "consequences" and "shock" ads led to a higher average diabetes risk score among completers, suggesting they resonated more with individuals who might have a higher underlying risk.
  • A projected scaled-up campaign in Java could lead to the diagnosis of almost 170,000 previously undetected diabetes cases at an estimated cost of US$9 per case.

Limitations and Future Work

The authors acknowledge several limitations:

  • Self-reported data: Compliance with professional screening recommendations and existing diagnoses were self-reported, introducing potential desirability bias. While they attempted to mitigate this, actual compliance remains unverified. Future work could involve direct collaboration with health centers to track actual appointments.
  • Sample representativeness: The follow-up survey sample was voluntary and may not be representative of the entire participant pool. Similarly, the overall sample of completers was slightly skewed towards older individuals and those at higher risk, though this is interpreted as a success in reaching the target audience.
  • Long-term effects of loss-framed ads: While effective in the short term, the long-term psychological impact of loss-framed ads (e.g., increased anxiety, information avoidance) was not assessed. Future research should explore these potential adverse effects.
  • Generalizability: While the study was conducted in Indonesia, the authors suggest the findings are transferable to other LMICs with similar diabetes burdens and social media usage.

Future work could focus on measuring actual compliance rates, investigating the long-term impacts of different ad framings, and exploring the specific psychological mechanisms behind gender-specific responses.

Summary updated

Abstract

Nudging individuals without obvious symptoms of non-communicable diseases (NCDs) to undergo a health screening remains a challenge, especially in middle-income countries, where NCD awareness is low but the incidence is high. We assess whether an awareness campaign implemented on Facebook can encourage individuals in Indonesia to undergo an online diabetes self-screening. We use Facebook’s advertisement function to randomly distribute graphical ads related to the risk and consequences of diabetes. Depending on their risk score, participants receive a recommendation to undergo a professional screening. We were able to reach almost 300,000 individuals in only three weeks. More than 1400 individuals completed the screening, inducing costs of about US$0.75 per person. The two ads labeled “diabetes consequences” and “shock” outperform all other ads. A follow-up survey shows that many high-risk respondents have scheduled a professional screening. A cost-effectiveness analysis suggests that our campaign can diagnose an additional person with diabetes for about US$9.
BibTeX
@article{fritz2024can,
  author = {Manuela Fritz and Michael Grimm and Ingmar Weber and Elad Yom-Tov and Benedictus Praditya},
  title = {{Can social media encourage diabetes self-screenings? A randomized controlled trial with Indonesian Facebook users}},
  journal = {Npj Digit. Medicine},
  year = {2024},
  doi = {10.1038/s41746-024-01246-x},
}
Figure