Mechanisms Behind Benefits Framing

Patients in this trial will be randomized to one of four conditions. In all conditions, patients will receive a WhatsApp message from a clinic network, encouraging them to schedule a preventive dental check-up. The conditions vary in message length and the specificity of the benefits emphasized. Previous pilot studies run by the investigators suggest that adding health- or appearance-related meaning to preventive dental checkup invitation messages may reduce uptake relative to a basic WhatsApp invitation. The present study is designed to replicate this pattern and explore the mechanism(s) underlying this pattern by isolating the role of message length and the references to benefits. Specifically:

• Basic Outreach Message: Patients receive a brief outreach message inviting them to schedule a preventive dental check-up and providing basic information about price and availability.
• Expanded Neutral Outreach Message: Patients receive a longer outreach message matched in length to the benefit-focused messages (see below), which adds only neutral wording to the Basic Outreach Message without introducing any additional information.
• Expanded Generic Benefit Outreach Message: Patients receive a longer outreach message that emphasizes general benefits from a preventive dental check-up, without referencing specific outcomes related to dental health or appearance.
• Expanded Specific Benefit Outreach Message: Patients receive a longer outreach message that highlights specific benefits of a preventive dental check-up. Within this condition, patients are further randomly assigned to receive one of four message variants that independently vary (1) whether the benefit emphasized relates to dental health or appearance and (2) whether the benefit is framed as occurring immediately or in the long term.

The researchers primarily ask the following research question: Does emphasizing specific benefits in an expanded outreach message affect dental check-up uptake relative to a basic outreach message? The researchers hypothesize that patients receiving the Expanded Specific Benefit Outreach Message will have lower uptake of dental checkups than patients receiving the Basic Outreach Message. The investigators will also assess the roles of message length and language referencing benefits.

Analysis:

The investigators will run ordinary least squares regressions (OLS) with robust standard errors to predict outcome variables.

• To test whether emphasizing specific benefits in an expanded outreach message affect uptake relative to a basic outreach message, the investigators will pool (collapse) the four variants of the Expanded Specific Benefit Outreach Message condition into a single treatment indicator and compare outcomes for patients assigned to this pooled condition versus patients assigned to the Basic Outreach Message condition.

In addition, the investigators will estimate a model that includes indicators for each of the four specific-benefit variants and conduct a joint F-test of whether the four variant indicators are jointly equal (i.e., whether outcomes differ across the variants).

• The investigators will conduct additional comparisons to understand why the Expanded Specific Benefit Outreach message may reduce uptake relative to the Basic Outreach message. Specifically, the investigators will compare outcomes between the Basic Outreach Message condition versus the Expanded Neutral Outreach Message condition to assess the role of message length. If there is no significant difference between the two treatments, they will be pooled and compared to the Specific Benefit (pooled) treatments. Then the investigators will compare outcomes between the Basic Outreach Message condition versus the Expanded Generic Benefit Outreach Message condition to assess the role of message length and language referencing benefits.

These regressions will be run with and without control variables. Control variables include:

• Time fixed effects (since patients will be enrolled into the trial over time)
• Participant age
• Participant gender (male, female, other/unknown)
• Whether patients had had at least one dental visit at this hospital in the past
• Region or clinic indicators, if available

For robustness, the investigators will conduct logit models as well.

The investigators will examine whether the following factors moderate the differences in uptake between the Expanded Specific Benefit Outreach Message condition and Basic Outreach Message condition:

• Whether the patient had completed any dental visit prior to the experiment (based on the data the hospital can provide)
• Whether the patient had completed any preventive medical exam (e.g., mammogram) prior to the experiment (based on the data the hospital can provide)
• Whether participants attended a non-dental medical appointment the day before contact (vs. have a non-dental medical appointment scheduled after the contact day)
• Number of prior visits to the clinic network that were not for dental purposes (based on the data the hospital can provide) prior to the contact (excluding the visit that some of the participants had the day before the contact)
• Insurance type (public vs. private)
• Patient age
• Patient gender (male vs. female)

For exploratory purposes, the investigators will look at longer time windows (e.g., 90 days) for dental appointment attendance and will also examine behaviors related to the dental appointments (e.g., cancellation, rescheduling, no-show).

Inclusion criteria

• Patients aged 18 and above
• Have a valid mobile phone number that can receive WhatsApp messages, recorded in the hospital network's patient database
• Attended a non-dental medical appointment the day before the intended contact day OR have a non-dental medical appointment within the week after the intended contact day
• Have not been contacted in the past two months about preventive dental checkups
• Have not attended a dental appointment at the hospital network in the past 6 months

Note: The partnering clinic agreed to run the study until a concurrent experiment registered at NCT07121673 is finished. The investigators aim to reach 100,000 patients by then, but the final sample size will vary depending on the actual outreach efforts by the end of the study period.