Data collection of the participants
The community health workers (CHWs) collected sociodemographic data, such as age, family structure, religion, parents’ occupation, parents’ education, monthly family income, and KAP information, using field-tested validated survey questionnaires during school visits. The KAP questionnaire was developed and adopted by the Food and Agriculture Organization of the United Nations [41]. The KAP questionnaire has 22 questions, with 9 focusing on knowledge, 6 on attitude, and 7 on practice. Subsequently, the CHWs examined anthropometric data on height (cm), weight (kg), mid-upper arm circumference (cm), hip circumference (cm), and waist circumference (cm). To ensure data consistency and minimize self-reported bias trained CHWs performed data collection, and two individuals did the data entry on two separate computers. Additionally, blood samples were collected by a certified medical laboratory technologist during the school visits. An auto hematology analyzer was used to measure hemoglobin levels, ensuring precise and reliable data on anemia status and its severity. After assessing the eligibility of the study participants, a statistician randomized the schools for the intervention group and control group; and performed a simple random sampling to enroll participants in the intervention and control groups. Afterward, the research personnel in charge opened the participant lists for each school, which had been placed in separate envelopes, to reveal the randomly assigned intervention and control group participants and then shared their IDs with the CHWs. Although the CHWs involved in the trial were not blinded to the intervention or control group, they were blinded to baseline interviews. The CHWs then contacted the trial participants in the intervention group, inviting them to participate in the mHealth education program and counseling sessions to improve hemoglobin levels and thereby reduce anemia.
In addition to the baseline assessment, the CHWs conducted physical examinations at the end of the 4th month and 8th month and collected data during the intervention at their school visits. Laboratory technologists collected blood samples during these visits. At the end of the 8th month, the same KAP questionnaire was used for the assessment (Table 1).
KAP assessment
The KAP questionnaire consisted of 22 questions. The knowledge part comprised nine variables: (1) knowing about anemia, (2) symptoms of anemia, (3) health risks for adolescents owing to a lack of iron in the diet, (4) health risks for pregnant women lack of iron in the diet, (5) know about causes of anemia, (6) prevention of anemia (7) know the iron-rich foods, (8) foods that increase the iron absorption, and (9) beverages that reduce iron absorption. The attitude part included six variables: (1) how likely think to be anemic, (2) how serious is anemia, (3) prepare meals with iron-rich foods, (4) difficult to eat or prepare meals with iron-rich foods (5) taste of iron-rich food, and (6) confidence in preparing meals with iron-rich foods. The practice part included seven variables: (1) consumption of iron-rich foods the last week, (2) usually intake of vitamin C/citrus fruits or juice, (3) every day eating citrus fruits or juice, (4) timing of eating citrus fruits, (5), usually tea or coffee consumption, (6) everyday drink coffee or tea, and (7) timing of drinking tea or coffee. Each variable was scored as either 0 or 1, with 1 indicating a favorable response. In total, all variables’ scores were combined for the mean average.
Questionnaire validity
Following the forward-backward translation procedure, the English questionnaire was translated into the local language, Bangla. This process involved two native Bangla speakers fluent in English: a public health specialist and a Ph.D. student. Each translated the English questionnaire into Bangla independently. Their translations were then combined to create a single forward-translated Bangla version. Next, this Bangla version was translated back into English by a professional medical translator, followed by a second translation by a medical doctor not involved in the previous steps. The researcher compiled and compared these back-translated versions. To ensure validity, all translated versions were reviewed by an expert committee. The finalized questionnaire was then pretested with 5% of school-going adolescent girls not included in the main study. Feedback from the pretest was collected and incorporated into the final version of the questionnaire [31].
Questionnaire reliability
The KAP questionnaire’s reliability was measured through the test-retest method and Cronbach alpha.
Laboratory investigation
Blood samples were obtained to measure hemoglobin levels and white blood cell (WBC) and red blood cell (RBC) counts. Laboratory tests were conducted at the Lab Aid Diagnostic Center in Comilla, Bangladesh, using a Sysmex automated hematological analyzer (Model: XN-550). Blood hemoglobin levels were measured using a spectrophotometric method for whole blood samples. Quality control, in accordance with the manufacturer’s guidelines, involved running control samples with known hemoglobin levels to confirm the perfection and accuracy of the analyzer and ensure reliable results.
Intervention group
Online group counseling sessions and mHealth technology were used to provide health education to the intervention participants.
Online group counseling session: The online group counseling session was designed to promote behavioral change and create awareness regarding dietary habits, nutrition, and hygiene management, all of which can contribute to anemia. The health education counseling session was conducted for the intervention group via Skype for approximately 90 min. A public health physician conducted two sessions (baseline and midline) with adolescent girls and their guardians. All 69 participants and their guardians attended the same online session, which was conducted in a large classroom in the school. The session was live and interactive. These sessions used online videos, PowerPoint presentations, and discussions to address the causes, consequences, prevention, and recovery from anemia. The sessions covered topics such as iron-rich foods, functions and sources of iron, dietary strategies, food habits and plans to improve iron status, personal hygiene management, and healthy lifestyle practices. At the end of the session, there was a Question and Answer part, and participants asked several questions regarding anemia and recovery strategy, the public health physician gave detailed answers and suggested follow-up health education guidelines (Supplementary material 1). Following the online education sessions, the participants were provided with health education materials in the form of notebooks containing information on anemia prevention and a list of local iron-containing foods. Modified health education materials (notebooks) were provided monthly to each participant throughout the intervention period (Supplementary materials 2).
mHealth education: During the 8-month intervention period, trained female CHWs delivered mHealth education through mobile phone calls and SMS text messaging to enhance participants’ nutritional knowledge, improve behaviors, dietary habits, and raise awareness about anemia. Participants received phone calls and mobile text messages four times each month in the initial 4 months, followed by three times during the 5th and 6th months, and finally twice during the 7th and 8th months. The mobile messages delivered targeted health education on anemia prevention, iron-rich dietary practices, strategies to improve iron absorption, hygiene management, and healthy lifestyle behaviors. Additionally, the CHWs provided information to the participants about locally available iron-rich foods and their benefits for iron absorption. They also guided avoiding beverages that may hinder iron absorption when consumed with meals, as well as limiting fatty and oily foods, and emphasized the importance of thorough hand washing before and after eating. Additional instructions included washing fruits carefully before consumption, wearing shoes while using the toilet, and thoroughly washing hands after using the toilet. Participants were also advised on the importance of taking anti-helminth medication every 6 months and using sanitary napkins during menses. Individual participants, along with their guardians, engaged in a 5-minute health education conversation with CHWs over the phone. Additionally, they received regular informative health education messages that were updated weekly (Table 2).
Control group
The control group received usual care and was observed without any specific intervention throughout the trial. Physical examinations and blood sample investigations were performed on these participants at the end of the 4th and 8th months. The KAP questionnaire interviews were conducted at the end of the 8th month. However, after each blood test, we provided hard copy reports with information on their anemia stage and advised them to consult with physicians if they wished. To ensure ethical fairness, control group received all health education materials after the intervention period.
Outcomes
The primary outcome was improved hemoglobin levels to reduce anemia in school-going adolescent girls.
The secondary outcomes were as follows:
-
1.
An increase in the scores of KAP related to anemia, measured using the KAP questionnaire.
-
2.
Improvement in anthropometric indicators such as BMI, MUAC, hip circumference, waist circumference.
Outcome measures
An automated analyzer measured blood hemoglobin levels to assess anemia categories (non-anemia, mild, moderate, and severe) in adolescent girls. Adolescent girls aged 10–11 years with hemoglobin levels below 11.5 g/dL were classified as having anemia. Within this age group, hemoglobin levels falling within the ranges of 10–11.4 g/dL, 7.0–9.9 g/dL, and < 7.0 g/dL were identified as mild, moderate, and severe anemia, respectively. Similarly, nonpregnant adolescent girls aged 12–19 years with hemoglobin levels < 12 g/dL were considered to have any form of anemia. For this older age group, hemoglobin levels ranging from 10.0 to 11.9 g/dL, 7.0–9.9 g/dL, and < 7.0 g/dL were categorized as mild, moderate, and severe anemia, respectively [11, 17, 39].
After the mHealth education intervention, changes in hemoglobin levels associated with anemia were examined. These changes, including those from anemic to non-anemic, mild to non-anemic/moderate/severe, and moderate to mild/non-anemic/severe, were assessed at baseline, midline, and endline in the intervention and control group.
The KAP measurement involves a total of 22 variables, including 9 related to knowledge, 6 related to attitude, and 7 related to practice. Each variable is scored as 1 for a favorable response and 0 for an unfavorable response. Finally, all variable scores are summed up and analyzed for the mean average.
Statistical analysis
Initially, the data were entered into Microsoft Excel 2019 and subsequently imported into the Stata Software (Version 18.5). Our study was an RCT, allowing three times data collection at the baseline, midline, and endline. Those who provided incomplete data during any time point of data collection were considered missing data. After excluding missing data participants (n = 17), a per-protocol analysis was used to ensure comparability between the control and intervention groups. Descriptive statistics were analyzed as frequencies, percentages, means, and standard deviations. We performed a Box Plot analysis to check the outlier data, but it was not detected. However, in the baseline assessment, continuous variables between groups were evaluated using either the independent t-test or the Mann–Whitney U-test after checking the normality of variables, whereas chi-square tests were employed to compare categorical variables. To evaluate potential violations of normality assumptions, we conducted the Shapiro–Wilk test to assess the data distribution. A p-value > 0.05 indicated normally distributed data, while highly skewed data were analyzed using non-parametric tests. Mixed-effects model was conducted to assess various outcome variables, including hemoglobin levels and physical measurements of hip–waist circumference, MUAC, and BMI. Changes in KAP scores were assessed using the Mann–Whitney U test for non-normally distributed data and the independent t-test for normally distributed data. In addition, we performed a generalized estimating equations (GEE) model with a Gaussian family and identity link function to analyze our study outcomes. The exchangeable correlation structure was selected based on the study design and prior literature. To assess the robustness of our findings, we conducted a sensitivity analysis by refitting the GEE model using different correlation structures (independent and autoregressive). The GEE model was applied to estimate the effect of health education over time after adjusting for age, student grade, father’s education and occupation, mother’s education and occupation, and family income covariates. Statistical significance was set at p < 0.05 in all analyses.
Quality assessment
To ensure the quality of the mHealth education during the intervention, we employed several key measures. Before the intervention was implemented CHWs received comprehensive training through a public health physician and principal investigator regarding the standardized data collection, health education delivery, and follow-up process. The principal investigator’s weekly supervision and reviews of SMS and phone call content were conducted following the study protocol. CHWs maintained a Microsoft Excel for documenting all the interactions with the participants including SMS, phone calls, educational materials distribution, planning, and feedback of the participants. Furthermore, CHWs took assistance from school teachers when they encountered difficulty reaching participants. The feedback was collected monthly from participants during the in-person delivery of health education notebooks at the school, allowing for real-time adjustments to the intervention. Additionally, maintaining standard procedures during online counselling sessions and phone interactions confirmed that all educational content was delivered as planned, thereby ensuring high quality and consistency throughout the study intervention period.
Digital divide
To minimize the digital divide, this intervention was specifically designed to be delivered via phone calls and SMS, which were accessible on feature phones and smartphones. This approach ensures that adolescent girls and their families can engage with health education content without needing internet access or advanced technology.
link