This study was designed as a 2-arm, parallel-group, superiority randomized controlled trial with individual randomization and a 1:1 allocation ratio. The trial was conducted in 2 municipal refugee shelters in Freiburg, Germany, between November 2023 and February 2024. Outcomes and assessment time points were defined a priori, and no protocol modifications were made after trial initiation.

The study was approved by the ethics committee of the University of Freiburg prior to participant enrollment. Due to administrative delays, registration in the German Clinical Trials Register (DRKS00032017) was completed after recruitment and follow-up had been completed; therefore, the trial is classified as retrospectively registered. No changes were made after study initiation to the primary objective, primary end point, eligibility criteria, intervention, allocation procedure, or core study design.

Eligible participants were adults (aged ≥18 years) residing in the participating shelters who were able to understand at least 1 of the languages in which the intervention was available (Arabic, Kurdish, Persian [Dari and Farsi], or Ukrainian) and had at least 10 teeth per jaw to allow reliable clinical index assessment. Individuals with acute dental emergencies or those requiring antibiotic prophylaxis were excluded. All eligible residents present during the recruitment period were invited to participate. A total of 86 participants met the inclusion criteria, provided written informed consent, and were enrolled. Participants with clinically relevant findings during examination received individual treatment recommendations and, if necessary, were referred to the university dental clinic to avoid delays in necessary care.

Randomization was performed after baseline assessment using a stratified allocation procedure to balance groups by age and native language. The allocation sequence was generated independently by the statistician (KV), who was not involved in recruitment, intervention delivery, or clinical outcome assessment. Allocation was concealed from the examining dentist until baseline assessments had been completed.

The intervention was developed through an iterative participatory process involving dental professionals, translators, and individuals with refugee experience recruited through collaborating shelter networks and community contacts. Development was informed by oral health literacy principles emphasizing plain language, visual demonstration, multilingual narration, and accessibility across different literacy levels. Draft scripts and visual materials underwent repeated refinement based on feedback regarding comprehensibility and usability within refugee shelter settings.

The GlobeSmile intervention [29] is publicly accessible online via YouTube (Alphabet Inc) and consists of a short, multilingual educational video providing simple instructions on daily oral hygiene practices. The full script is provided in Multimedia Appendix 1.

The educational content focuses on key preventive behaviors, including toothbrushing technique, the use of fluoride-containing toothpaste, brushing duration and sequence, tongue cleaning, and rinsing, and was informed by international oral hygiene recommendations [27,28]. The video has a total length of approximately 4 minutes and combines simple animations with multilingual voice-over explanations to ensure accessibility across different literacy levels. The final version was made available in Arabic, Kurdish, Dari, Farsi, and Ukrainian. The control group received access to the video after completion of the follow-up to ensure ethical equipoise.

The primary outcomes were changes in the PI (according to Silness and Loe [30]) and GI (according to Loe and Silness [31]). Secondary outcomes were assessed using a study-specific multilingual questionnaire developed for this field setting, addressing oral hygiene behaviors as well as oral health literacy–related cognitions.

Follow-up was scheduled approximately 2 months after baseline. This interval was chosen pragmatically, as longer retention in communal shelter settings could not be assumed, while still allowing sufficient time for the adoption of oral hygiene behaviors and the detection of short-term changes in plaque accumulation and gingival inflammation.

Clinical examinations were conducted at baseline and follow-up under standardized field conditions. At both time points, the PI and GI were assessed.

All examinations were conducted by a single trained and calibrated examiner (FS) using portable dental equipment and headlamps to ensure comparable examination conditions across time points. Prior to data collection, the examiner underwent calibration with an experienced practitioner to ensure measurement reliability (interrater reliability: PI Cohen κ >0.8 and GI Cohen κ >0.8).

Caries experience (Decayed, Missing, and Filled Teeth index according to Klein and Palmer [32]) was recorded at baseline as a descriptive characteristic.

Participants in the intervention group viewed the video on a tablet computer immediately after the baseline clinical examination in one of the available intervention languages that they reported understanding. They also received a printed QR code for continued access, allowing them to rewatch the video as often as desired during the 2-month study period (Figures S1 and S2 in Multimedia Appendix 1). Participants in the control group received no intervention during the trial but were provided with access to the video via a QR code after completion of the follow-up examination.

Questionnaires were administered in the participants’ native languages at baseline and follow-up, with assistance from trained translators to ensure comprehension and cultural appropriateness. The survey included items on demographic information (age, sex, country of origin, native language, and education); refugee experience; oral health knowledge, attitudes, and perceived self-efficacy; oral hygiene behavior (brushing frequency and utensils); nutrition (according to Woelber et al [33]); physical activity; and tobacco and alcohol use (Tables S1 and S2 in Multimedia Appendix 1). The questionnaire was developed specifically for this multilingual field study and included behavioral, lifestyle-related, and oral health literacy–related items informed by oral health behavior literature, anti-inflammatory dietary recommendations, and the Internal-External Locus of Control Short Scale–4. Translation was supported by trained translators and reviewed for comprehensibility and cultural appropriateness. The questionnaire was pilot-tested informally prior to implementation and was not intended as a formally validated oral health literacy instrument.

Participant blinding was not feasible due to the nature of the intervention. At follow-up, clinical assessments were performed before participants were asked about intervention use, such that the examiner was not routinely aware of group allocation. However, due to the shared shelter setting, complete assessor blinding could not be guaranteed. Examinations were conducted by a single calibrated examiner under standardized conditions, and data analysis was performed using coded group labels.

No formal a priori sample size calculation was performed; the sample size was determined pragmatically by the number of eligible shelter residents who were present during the recruitment period and willing to participate. All individuals who met the inclusion criteria and provided written informed consent were enrolled. On the basis of the observed variability of the PI, which represented the primary outcome expected to change most rapidly, 40 participants per group were sufficient to detect moderate between-group differences in PI change with 90% power at a 2-sided significance level of α=.05.

Continuous outcomes are reported as mean (SD) to aid interpretability. For univariate analyses, within-group changes in PI and GI were assessed using Wilcoxon signed-rank tests. Between-group comparisons of change scores were conducted using 2-sided t tests as the primary contrasts. Multivariable linear regression models were used to examine associations of age, sex, smoking status, and group allocation with changes in PI and GI and additionally served as sensitivity analyses for the nonparametric results. Analyses were performed in Stata (version 19; StataCorp), with a significance level of α=.05.

Ethics approval was obtained from the Ethics Committee of the University Medical Center Freiburg (23‐1392-S1) prior to participant enrollment. All participants provided written informed consent before participation. Study information and consent materials were provided in relevant participant languages, with translator support where needed to ensure comprehension. Participant data were pseudonymized at the time of collection and analyzed in deidentified form to ensure confidentiality and data protection. No financial compensation was provided. Participants with clinically relevant findings during examination were informed and referred for appropriate dental care.

A total of 86 residents of the 2 participating shelters were assessed for eligibility, all of whom met the inclusion criteria, consented, and were enrolled in the study. Of these, 42 (48.8%) participants were randomized to the intervention group and 51.2% (44/86) to the control group. Follow-up at 2 months was completed by 83 (97%) participants. In total, 3.5% (3/86) participants were lost to follow-up (2/42, 4.8%) in the intervention group and 1 (2.3%) in the control group. No missing clinical or questionnaire data were observed among participants who completed follow-up (Figure 1).

Baseline sociodemographic and clinical characteristics were broadly comparable between groups (Table 1). Participants had a mean age of 35.4 (SD 12.6) years, and 54.7% (47/86) of participants were female. The largest subgroups originated from Ukraine, Iraq, and Afghanistan, with native languages reflecting this distribution. All participants reported sufficient understanding of at least 1 of the intervention languages.

Educational level, lifestyle factors, and baseline clinical indices (PI; GI; and the Decayed, Missing, and Filled Teeth index) showed no relevant differences between groups.

Mean PI score decreased from 1.21 (SD 0.48) to 1.01 (SD 0.46) in the intervention group (mean change −0.21, SD 0.27; P<.001), whereas the control group showed only a small change (mean change −0.04, SD 0.17; P=.23). The between-group difference in PI change was significant (P=.002).

GI decreased in both groups (intervention group: mean change −0.11, SD 0.23; P=.002 and control group: mean change −0.07, SD 0.16; P=.006), but the between-group difference was not significant (P=.55). Table 2 summarizes all primary outcomes.

Questionnaire findings were mixed (Multimedia Appendix 1). Brushing frequency improved significantly, while selected cognitive items showed favorable but nonsignificant trends. Other items remained largely unchanged.

Exploratory adjusted regression analyses were conducted to assess whether participant characteristics were associated with changes in clinical outcomes. In these models, group allocation remained significantly associated with PI change, confirming greater plaque reduction in the intervention group (β=0.1662, 95% CI 0.0697-0.2626; P=.001). No significant intervention effect was observed for GI change (β=0.0361, 95% CI −0.0504 to 0.1226; P=.41).

Age was significantly associated with greater improvements in both clinical outcomes. With each additional year of age, participants showed slightly greater reductions in PI (β=−0.0046, 95% CI −0.0087 to −0.0006; P=.02) and GI (β=−0.0042, 95% CI −0.0078 to −0.0005; P=.02). Sex and smoking status were not significantly associated with changes in PI or GI. These findings were consistent with the primary analyses and supported by nonparametric sensitivity analyses (Multimedia Appendix 1).

Reported engagement with the video was generally high: 76.2% (32/42) of participants in the intervention group rewatched the video via the QR code at least once during the study period. Participants who reported repeated viewing tended to show larger reductions in PI, suggesting that repeated exposure to the content may be beneficial.

This randomized controlled trial evaluated whether a brief microlearning video could improve oral hygiene outcomes and oral health literacy among refugees living in communal shelters in Germany. The findings indicate that a simple digital intervention may improve plaque control and selected cognitive determinants of oral health behavior, even in settings with limited access to preventive care and substantial linguistic diversity. Oral health served as a pragmatic model for addressing health information barriers in an underserved population.

The intervention was associated with a greater reduction in PI than that observed in the control condition, suggesting improved oral hygiene performance independent of brushing frequency. In addition to a general Hawthorne effect, the baseline dental examination itself may have acted as a behavioral prompt, increasing oral hygiene awareness in both groups. The script recommends a brushing duration of approximately 3 minutes, which is slightly longer than commonly cited minimum recommendations. This was intentionally chosen as a behavioral strategy to encourage sufficiently thorough cleaning, acknowledging that actual brushing duration in everyday practice is often shorter than recommended. Video-based educational interventions have previously been shown to influence health behaviors across a range of contexts by facilitating observational learning and modeling of desired behaviors, supporting the plausibility of such mechanisms in this study [34]. One possible interpretation is that the intervention affected brushing technique rather than brushing frequency. Toothbrushing is often a habitual and partly automatic behavior, and previously learned routines may interfere with the acquisition of new oral hygiene skills [35]. Because PI and GI involve examiner judgment, the absence of fully blinded outcome assessment may have influenced estimates of effect size and should be considered when interpreting the findings. However, the examiner was not routinely aware of group allocation during follow-up assessments, which may have reduced the risk of observer bias.

Questionnaire findings were mixed; only brushing frequency improved significantly, while selected cognitive items showed favorable but nonsignificant trends. These findings suggest that broader oral health literacy–related effects should be interpreted cautiously. This interpretation is consistent with recent meta-analytic evidence indicating that psychologically informed interventions can improve oral hygiene behaviors, plaque levels, and self-efficacy in the short term by targeting cognitive and motivational determinants of oral health behavior [36]. These findings are consistent with established health literacy frameworks, in which increased knowledge, enhanced confidence, and a stronger internal health locus of control are key mechanisms linking educational interventions to improvements in oral health behaviors and clinical outcomes [18,24,34]. Therefore, the data support a plausible mechanism whereby culturally adapted microlearning enhances both oral health literacy and the application of brushing techniques, leading to reduced plaque accumulation.

Participants demonstrated high digital engagement. Three-quarters of the participants (32/42, 76.2%) rewatched the video at least once, and individuals who reported repeated viewing tended to show larger PI reductions. Although the frequency of repeated viewing was insufficient to support a formal dose–response analysis, the observed pattern suggests that repeated exposure may reinforce skill acquisition and message internalization. This finding is consistent with behavior change theory and supports the potential value of microdose repetition in digital health literacy interventions [23,36]. The study thereby highlights an important practical implication: even minimal recurrency of exposure may meaningfully enhance effectiveness.

The results complement current international evidence documenting pronounced oral health inequities among refugees and asylum seekers, driven by structural access barriers, linguistic limitations, inconsistent entitlements to care, and fragmented service pathways [6,10,12,17]. Consistent with reports from Germany and other European countries, participants exhibited high baseline levels of untreated caries and low oral health literacy, factors strongly associated with reduced service use and poorer health outcomes [10,18,24].

Digital health interventions have been proposed as a promising strategy to reduce inequities in forced migrant populations, particularly where conventional education is difficult to deliver due to low literacy, language heterogeneity, or limited access to health professionals. Similarly, behavioral interventions using visual cues and motivational messaging have demonstrated improvements in plaque control among schoolchildren with refugee and immigrant backgrounds, highlighting the potential of simple visual communication strategies for promoting oral self-care in underserved populations [37]. These findings are consistent with previous work suggesting that digital health literacy interventions can be beneficial in forced migrant populations [23]. These findings add experimental evidence to this literature and suggest that ultralow-threshold, multilingual microlearning tools can produce clinical improvements without requiring complex infrastructure or advanced digital skills.

Compared with multicomponent mobile health programs that demand extensive engagement, device access, and continuous coaching [26], the simplicity of the present intervention may offer greater scalability for settings such as shelters, reception centers, or transit facilities. Importantly, however, digital education alone cannot compensate for structural constraints, including the emergency-focused care entitlements that limit preventive dental treatment for refugees in Germany [10]. Therefore, the intervention should be viewed as a complementary public health strategy rather than a substitute for system-level reform.

The observed effects may in part reflect reinforcement of previously known oral health information rather than entirely novel knowledge acquisition. However, this distinction is less relevant from a public health perspective: even when basic knowledge exists, accessible, linguistically tailored microlearning may facilitate translation into daily practice and measurable clinical outcomes. Therefore, in populations facing structural and informational barriers, reinforcement and accessibility may represent key mechanisms of effective prevention.

Strengths of this study include its randomized controlled design, the use of calibrated clinical assessments, the high follow-up rate of 97%, and the naturalistic implementation within refugee shelters, which enhances ecological validity. The participatory development of the video, involving translators and individuals with refugee experience, likely contributed to cultural acceptability and comprehension.

Several limitations should be considered. First, the sample size was modest and determined pragmatically rather than through a formal a priori power calculation. Therefore, the study was primarily exploratory and may have been underpowered to detect smaller between-group differences, particularly for secondary outcomes and subgroup analyses. Second, the study population was heterogeneous with regard to country of origin, language, migration history, and social circumstances. Although this heterogeneity reflects the real-world composition of refugee shelters and strengthens the practical relevance of the findings, it may also have increased variability in responses to the intervention, limited the precision of effect estimates, and reduced the interpretability of subgroup patterns. Third, the 2-month follow-up period restricts conclusions regarding long-term behavioral change and the sustainability of effects. Fourth, some degree of contamination between groups cannot be entirely ruled out, given the shared living environment. Fifth, gingival outcomes are difficult to influence over a short period without professional prophylaxis, which may partly explain the absence of a significant between-group effect for GI. Complete assessor blinding could not be guaranteed, and some risk of observer bias remains. Questionnaire outcomes were assessed using a pragmatic study-specific instrument rather than a formally validated oral health literacy scale, which limits precision, reliability, and comparability with other studies. Finally, adherence was assessed by self-report, and objective use metrics for repeated video viewing were not available.

This trial provides actionable evidence supporting the implementation of multilingual digital microlearning as part of preventive strategies for displaced populations. Although evaluated here in the context of oral health, the intervention model may be transferable to other preventive health domains in which multilingual, low-threshold access to information is limited. The intervention was feasible to deliver, well accepted, and associated with a meaningful improvement in plaque control and limited favorable changes in selected self-reported outcomes. Because it is low cost, language adaptable, and accessible via QR code, it could be readily integrated into refugee shelters, reception centers, community health programs, and nongovernmental organization–led services.

From a health equity perspective, the results indicate that well-designed tools can reduce information barriers, thereby supporting the overarching aims of Inclusion Health, which focuses on improving health outcomes and access to care for socially excluded populations. Nevertheless, digital education should be embedded within broader multisectoral approaches that address restricted care entitlements, financial and administrative barriers, and limited access to preventive dental services for refugees [6,10,18]. Without structural reforms, improvements in knowledge and technique may not translate into long-term disease reduction. Because the intervention requires minimal infrastructure and can be distributed via QR codes or standard video platforms, it represents a highly scalable digital public health strategy that could be adapted to other preventive health domains and multilingual populations.

Detailed information on previous dental attendance and prior exposure to oral health education interventions was not systematically collected. These factors may have influenced baseline behaviors and responsiveness to the intervention. Future trials should assess prior dental care use and previous preventive education exposure as potential moderators of intervention effects.

This study illustrates how even ultrabrief digital interventions can produce measurable clinical change when delivered in linguistically adapted formats. The model may be transferable to other preventive domains where information barriers limit health equity. The GlobeSmile video was developed without external funding and relied on deliberately simplified, self-produced visual materials to ensure rapid and low-cost implementation. Although this didactic reduction enhances accessibility and scalability, future iterations could explore more professionally designed or visually enriched formats to assess whether enhanced aesthetic and pedagogical features further increase engagement and effectiveness. Future studies should evaluate long-term outcomes, assess cost-effectiveness, and determine whether combining video-based education with in-person demonstrations or support from community health workers provides additional benefits. Research across diverse refugee settings and language groups could assess generalizability and inform implementation strategies. Additionally, investigating the mechanisms linking improved oral health literacy to clinical change could support theory-based optimization of digital microlearning interventions.

Brief multilingual digital microlearning may represent a feasible and scalable strategy to improve plaque control and reduce health information barriers in refugee shelter settings. Larger, prospectively powered trials with longer follow-up periods are warranted.