DeCamp 2020.

Study characteristics
Methods	Study design: RCT, 2 arms Geographic location: Maryland, USA Ethical approval: yes Recruitment setting: large urban, academic general paediatrics clinic Method of recruitment: review of the clinic schedule for completed initial newborn visits; potentially eligible parents were sent an informational letter about the study. A bilingual research assistant recruited potential participants either by follow‐up phone call or during a subsequent newborn visit. Length of follow‐up: length of programme: 10 months; follow‐up survey at child age: 12 to 15 months, which was 1 to 3 months after the programme was completed Dropouts: 22 participants lost to follow‐up (7 in the intervention group (5 moved or switched clinics, 2 were unable to be contacted) and 15 in the control group (4 moved or switched clinic, 5 were unable to be contacted and 6 declined) A priori calculation of effect size/power?: yes
Participants	Description: immigrant parents or legal guardians of Latin descent with US‐born infants < 2 months of age Health topic Child health Inclusion criteria Parents or legal guardians of publicly insured, singleton US‐born infants < 2 months of age, minimum parent age of ≥ 18 years, self‐identification as Latino or Latina, preferred health care language of Spanish, 1 household cellular phone Exclusion criteria Not reported Intervention group Salud al Día, Spanish‐language interactive text messaging intervention (79 randomised and analysed for observer‐reported outcomes, for participant‐reported outcomes only 72 analysed) Control group Usual care (78 randomised and analysed for observer‐rated outcomes, for participant‐reported outcomes only 63 analysed) Note: an intention‐to‐treat analysis was performed for primary outcomes (analysed via electronic medical record (EMR)); secondary outcomes that were not abstracted from the EMR included only individuals who finished follow‐up survey. PROGRESS‐Plus Place of residence: urban, USA Time living in host country (years), mean (SD): 7.3 (5.3) Race/ethnicity: Latinos Occupation: 79.0% spouse or partner employed Gender: 100% female Education: 40.8% ≤ 8th grade, 26.1% some high school, 33.1% some high school or greater Socioeconomic status/income (annual): 42.7% < USD 20,000, 24.2% USD 20,000 to 30,000, 7.6% > USD 30.000, 19.1% did not report or unknown Health insurance: all children publicly insured Social capital: 20.3% single, 79.6% spouse or partner Age (years), mean (SD): 29.3 (6.2) Health literacy (baseline) Assessment tool, range (score): Spanish‐language version of the Newest Vital Sign (NVS), 6 items, 0 to 6, higher score is better (validated in English and Spanish) Intervention group: 46.0% limited HL (0 to 1), 39.0% marginal HL (2 to 3) 15.0% adequate HL (4 to 6) Control group: 51% limited HL (0 to 1), 37.0% marginal HL (2 to 3), 12.0% adequate HL (4 to 6) English proficiency was assessed using the US Census Bureau question, “How well do you speak English?” Intervention group: 97.0% Control group: 96.0%
Interventions	Intervention: Salud al Día, an interactive text messaging intervention to reduce ED use and increase vaccine adherence Theoretical framework: situated Information, Motivation, Behavioral Skills (sMIB) model (Amico 2011) Description: parents received interactive personalised text messages, push messages and watched an animated Spanish‐language educational video. Sequences included appointment reminders, support for obtaining medicines, support for completing referrals, and illness care monitoring and education. Interactive text messages included reminders of flu vaccine or information on parent support programmes and public benefit programmes. Certain response records generated an email to a clinic nurse who contacted participants and offered further support. Intervention provider: research staff, clinic staff Delivery method/mode: 1 individual video session lasting 9 min (plus take‐home DVD at 2‐month visit in clinic) and monthly interactive text messages for 10 months, if necessary email contact to clinic nurse Language of delivery: language concordant (bilingual) Format: tailored, algorithm‐based interactive messages Setting/location: academic general paediatrics clinic (video) Consumer involvement: evaluated with members from the community of interest Comparator Type: no health literacy intervention (usual care/no additional intervention) Description: usual care for infants in the 1st year of life
Outcomes	Outcomes assessed in the study: infant health knowledge, up‐to‐date immunisations*, well visits, parent depression, emergency department use, parent experience of care rating, change in mean parent engagement, receipt of 2 doses of the influenza vaccine, well visit no‐shows and cancellations, clinic visit provider continuity, number of sick care visits, speciality care referral completion, participant‐generated telephone encounters, electronic medical record (EMR) patient portal (MyChart) status, Supplemental Nutrition Assistance programme (SNAP)/food stamp participation Outcomes considered in this review Health‐related knowledge (infant health knowledge) Health behaviour (up‐to‐date immunisations) Health outcome (parent depression) Health service use (emergency department use) *Prioritised outcome in the category 'health behaviour' based on consensus opinion of the authors Methods of assessing outcomes Surveys were orally administered by bilingual research assistants, either in‐person (enrolment and follow‐up) or via telephone (midpoint). Responses were captured using a touchscreen tablet computer and Research Electronic Database Capture software. Infant health knowledge: questionnaire based on intervention topics: (1) fever criteria, (2) public health insurance renewal, (3) right to interpretation during medical encounters, (4) obtaining an outside care report, (5) availability of after‐hours clinic resources); 5 items, multiple choice, true/false questions, 1 point for each correct response, 0 to 5, higher score is better Up‐to‐date immunisations: assessed via EMR Parent depression: Patient Health Questionnaire (PHQ‐8), 8 items, 0 to 24, cut‐point ≥ 10 (moderate or severe depressive symptoms), lower score is better Reliability/validity: validated tool Emergency department use: assessed via EMR Language of assessment: English, Spanish Timing of outcome assessment: baseline, short‐term (at 11 to 14 months after randomisation, which was 1 to 3 months after the programme was completed)
Health literacy	Definition: not reported Health literacy components addressed by the intervention Prerequisites and tools Knowledge Motivation Steps of information processing Access Understand Appraise (unclear) Apply Health domain: disease prevention
Notes	Trial ID: NCT02647814 Funding: funding was provided by the Gordon and Betty Moore Foundation. Additional notes: authors provided additional information on request.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Random assignment was performed by computer random number generation in blocks of 10, with a 1:1 allocation ratio."
Allocation concealment (selection bias)	Low risk	Low risk of bias due to randomisation method used.
Blinding of participants and personnel (performance bias) All outcomes	High risk	"Participants and research staff were not blinded to which intervention participants were allocated to. Clinical staff and providers were not aware of group assignment unless revealed by the participant." Personnel and participants were not blinded and some outcomes of interest were subjectively measured. Therefore, results of subjective outcomes might be bias
Blinding of outcome assessment (detection bias) subjective outcome measures	High risk	"All surveys were orally administered by bilingual research assistants. Survey responses were captured simultaneously with administration using a Touchscreen tablet computer and Research Electronic Database Capture software" Participants were aware of group assignment and depression was measured via self‐reported questionnaire, which might have introduced a bias.
Blinding of outcome assessment (detection bias) objective outcome measures	Low risk	Participants and research staff were aware of group assignment. However, knowledge, health behaviour (child's up‐to‐date immunisation) and health service use (emergency department use) were objectively measured and not subject to interpretation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	"Analyses of primary outcomes were conducted per the intention‐to‐treat principle. Analyses of secondary and process outcomes that were not abstracted from the EMR included only those individuals with corresponding follow‐up survey data." Authors report numbers and reasons of dropouts separately for each study arm using a CONSORT diagram. In total, 22 participants were lost to follow‐up, n=7 (8.86%) in the intervention group and n=15 (19.23%) in the control group. The dropout rates are unbalanced. However, the differential loss between intervention and control arm is less than 15% (10.37%) and the reasons are reported transparently.
Selective reporting (reporting bias)	Low risk	All prespecified outcomes are reported in the results.