The overall objective is to examine if participating in an SMS-transmitted health educational program targeting a cluster of dietary factors over a period of weeks during the years of adolescence can induce durable changes towards more healthful dietary habits and body weight trajectories. The investigators will examine if impact is reinforced by asking a friend of the participant to also participate; if impact is reinforced by asking the mother of the participant to also participate; and if impact is reinforced when the SMS-transmitted health promotion program is designed to target a specific dietary factor only - decided upon from participants' reporting of their dietary habits at the beginning of the program - rather than to simultaneously target a cluster of dietary issues uniform for all participants.

Groups of adolescents from the Danish National Birth Cohort (DNBC) will be offered to participate in the health educational programs. At 14 years and 14 years plus 3 months all children DNBC are invited to report comprehensively on their dietary and physical activity habits by means of a web-based questionnaire (FFQ). Eligible DNBC adolescents for the SMS-educational program will be invited among those who have completed the FFQ. Thus, once they have completed the DNBC FFQ, they may be invited to participate in the SMS-based educational programs, either addressing only one of the three target factors: sugar sweetened beverages, fruit & vegetables, or fish (SSB, F&V, or Fish) or addressing all three factors (cluster). The adolescents may also potentially be invited to include a friend, their mother or both.

Trial Organization:

• Sjurdur F Olsen, MD PhD, Lead Principal Investigator (Responsible Party), Statens Serum Institut, Copenhagen, Denmark;
• Eric Ding, ScD, Co-PI, Harvard TH Chan School of Public Health;
• Daniel Zoughbie, Co-PI, Epidemic Health Inc.

Design for Inviting Adolescents to Educational Program: The study will be conducted corresponding to a '2x2x2 factorial plus 1' design, with 9 arms. In addition to allowing testing of the effect of participating in an SMS based educational program as compared to not participating in such program, this will allow for additional testing of effect of implicating a friend v. not implicating friend; of implicating the mother v. not implicating the mother; and of delivering tailored messaging targeting only one of the three dietary factors (SSB, F&V, or Fish) v. delivering uniform messaging targeting all three factors.

In the tailored program, the participants will receive messages distributed over 4 weeks, whereas in the FULL program (addressing all three clusters) messages are distributed over 12 weeks. Messages can be divided into five categories: Information, quizzes, tips, challenges, and reminders.

Hypotheses to tested:

1. Overall effect. Hypothesis 1: SMS messaging (versus being allocated the group not offered to participate in an SMS messaging program) will improve dietary habits and body weight trajectories
2. Effect of including a friend. Hypothesis 2: Including a friend (versus not including a friend) will increase above effects
3. Effect of including the mother. Hypothesis 3: Including the mother (versus not including the mother) will increase above effects
4. Effect of tailored program v. a cluster of factors. Hypothesis 4: A tailored program will be more efficient than a full program in changing specific outcome.

For testing the hypothesis 4, which addresses one of the three specific, tailored programs (underlined below), eligible adolescents will be stratified into four strata depending on their FFQ response:

Stratum 1: Nearly all participants have a low intake of fish; whereas there will be no problem with sugar sweetened beverages and no problem with fruit and vegetables. Tailored program AND outcome measure to be tested: FISH / fish intake.

Stratum 2: All participants have problems with fruit and vegetables, and no problems with sugar sweetened beverages, but (NB!), like in stratum 1, nearly all have a problem with fish. Tailored program AND outcome measure to be tested: F&V / intake fruit and vegetable.

Stratum 3: All participants have problems with sugar sweetened beverages; whereas nearly all will have NO problem with fruit and vegetables. Nearly all have a problem with Fish (986/1195). Tailored program AND outcome measure to be tested: SSB / intake of sugar sweetened beverages.

Stratum 4: All participants have a problem with sugar sweetened beverages, fruit and vegetables and fish. Tailored program AND outcome measure to be tested: FISH / fish intake.

Strata 1 & 4 can be aggregated and, when effect on fish intake is tested, a test for interaction by intake of sugar sweetened beverages or by intake of fruit and vegetables can be made.

A detailed a priori Statistical Analysis Plan SAP (version 2.0) was filed at the Office of Research Integrity, Statens Serum Institut, on June 17, 2019. This date preceded the date of opening the data (i.e., relating the outcome measures to information about which trial arms the adolescents had been randomly allocated to).

Inclusion Criteria:

• Adolescents participating in the Danish National Birth Cohort

Exclusion Criteria:

• Lacking information from maternal interviews 1 and 2