This study adopted a cross-sectional research design. From November 2021 to August 2022, 9 months were utilized for the completion of this study. The target population of this study, as specified in the inclusion criteria, was school-going Malaysian adolescents between the ages of 13 to 17 living in Selangor, Malaysia. Non-Malaysians and individuals with diet-related or chronic diseases were excluded from the study.

The sample size was calculated using Cochran's single proportion formula for sample size calculation (25). The sample size calculation for this study includes a confidence interval (z) of 95% = 1.96, with an estimated prevalence of the population that consume fruit and vegetables (p) at only 23.5% (7) and a margin of error (e) of 0.05. The sample size calculated amounted to 277 participants.

This study adopted convenience sampling as the sampling method. An online questionnaire (Google form) on vegetable and fruit intake was designed and advertised on various social media platforms (such as WhatsApp, Telegram, Facebook, and Instagram) and aimed to recruit participants in the state of Selangor. Non-Selangor citizens were excluded from this study. Eligible participants' who have signed the consent form were included as part of the study.

This study was conducted with ethical approval (Code: UKM PPI/111/8/JEP-2022-226) from the Human Research Ethics Committee of Universiti Kebangsaan Malaysia.

The main means of data collection was a four-part self-reporting online questionnaire. Part 1 and Part 2 ask for sociodemographic factors and anthropometry measurements. Part 3 and Part 4 ask questions on the frequency of fruit and vegetable intake and the factors that influenced it. This version of the questionnaire was validated by Rohin et al. (11) and De Bourdeaudhuij et al. (26), respectively for the Malay and English version.

Part 1: Sociodemographic factors

This section explains the collected data on the adolescents' sociodemographic factors, including age, sex, ethnicity, parents' education level, and household income.

Part 2: Anthropometry measurements

The respondents self-reported their body weight and height. Their body mass indexes (BMI) were then calculated and categorized according to the BMI-for-age z-score (BAZ). Respondents were categorized into groups, which included “thin” if their BAZ values were <−2 SD, “normal” if their BAZ values were between ≥ −2 SD and ≤ +1 SD, and “overweight” if their BAZ values were > +1 SD and “obese” if their BAZ values > +2 SD.

Part 3: Frequency of vegetable and fruit intake

This questionnaire comprised questions on the frequency and amount of vegetable and fruit intake. There are two separate sections: frequency and amount of intake. For frequency, the participants were asked how often they consumed vegetables and fruits in a week. The response categories were never, 1 day, 2–4 days, 5–6 days, and every day. For amount, the participants were also asked how many servings of vegetables and fruits they consumed per day, with the response categories being No, one, two, three, four and five or more servings. The amount of participants' vegetable and fruit intake was assessed using the Malaysian Dietary Guidelines for Children and Adolescents 2023 (MDGCA 2023) (27) to determine whether they met the recommended servings, which are five servings (three servings of vegetables and two servings of fruits) daily.

Part 4: Factors associated with vegetable and fruit intake

This section has questions regarding the personal, social-environmental, and physical-environmental correlates of fruit and vegetable intake, as adapted from De Bourdeaudhuij et al. (26) and Malay version of De Bourdeaudhuij et al.'s questionnaire (26) adapted and validated by Rohin et al. (11). Validation of this questionnaire revealed that the internal consistency and test-retest reliability of the scales range between Cronbach's α values of 0.52 to 0.89 and Intraclass Correlation Coefficient (ICC) values of 0.52 to 0.85, which were comparable between all five countries involved in the testing (26). Permission to use the questionnaires was obtained from both authors. For this study, different components were included to fully encompass the possible domains of the possible associated factors, including personal factors (knowledge, attitude, liking, self-efficacy, intention, and habit), social-environmental factors (parental modelling, peer modelling, parental encouragement, parental demand, and parental allowance), and physical-environmental factors (availability of fruits and vegetables at home, school, friends' homes, and leisure places). The response category for all these questions was in the form of Likert scales: from Fully Agree (+2) to Fully Disagree (−2) and Very Often (+2) to Never (−2).

Statistical Package for the Social Sciences Statistics (SPSS version 28) was used to analyze the collected data. This study employed four main statistical analysis to fulfil the research objectives. Descriptive statistics, including frequency, percentage, mean, and standard deviation, were used to present the demography and characteristic of the study population. Data on adequacy of vegetable and fruit intake was obtained from computing frequency of intake and number of daily servings into new dichotomous variables. Cases with intake of 7 days and 3 or more daily number of servings (two or more for fruit intake) were denoted as 1, while other entries were put as 0, following the Malaysian Dietary Guidelines for Children and Adolescents 2023 (27). Frequencies and percentages were then used to present the prevalence of vegetable and fruit intake within the population. The chi-square test was used to identify the associations between sociodemographic factors such as age, gender, ethnicity, parent's education level, household income status, and BMI-for-Age with vegetable and fruit intake. Generalized linear model (GLM) with Poisson loglinear distribution and the robust estimator was employed to model the adolescent vegetable and fruit intake based on potential associated factors, including personal factors (knowledge, attitude, liking, self-efficacy, intention, and habit), social-environmental factors (parental modelling, peer modelling, parental encouragement, parental demand, and parental allowance), and physical-environmental factors (availability of fruits and vegetables at home, school, friends' homes, and leisure places). These variables were treated as scores of −2 to +2, and included into the model as continuous variables. A p-value of <0.05 was applied as the cut-off for significance.

Table 1 shows the sociodemographic information of the participants in this study. A total of 277 adolescents aged 13–17 years (mean age and SD) participated in this study. More than half of the participants are male (n = 146, 52.7%). In terms of ethnicity, most are Chinese (n = 125, 45%) and have parents with at least secondary school education (n = 127, 45.8%). Meanwhile, 40.1% of the participants have moderate family income (n = 111). Most of the participants were either overweight or obese (n = 143, 51.6%), while others had normal BMI or were thin (n = 134, 48.4%).

Table 2 shows the prevalence of vegetable and fruit intake by BMI-for-age status. In this study, the overall prevalence of adolescents who met the recommended intake of vegetables and fruit were 23.5% and 57.8%, respectively. A lower percentage of overweight (25.2%) and obese (15.0%) adolescents consumed more than three servings of vegetables. However, the lowest prevalence was observed among thin adolescents (14.8%), while 27.1% of normal BMI adolescents contributed to the overall prevalence. As for fruit consumption of more than two servings, up to 57.8% of all adolescents consumed the recommended amount of fruits, which includes 51.9% of thin adolescents, 50.5% of adolescents with normal BMI, and 64.1% and 65% of overweight and obese adolescents, respectively.

The associations between sociodemographic factors and vegetable and fruit intake of the adolescents are shown in Tables 3, 4. Fisher's exact test was applied when the expected frequencies assumption was violated for ethnicity in Table 4. The findings observed a significant association between sociodemographic factor, BMI-for-age (χ² _{(1, N = 277)} = 5.236, p = 0.022), with fruit intake among overall adolescents. However, no significant association was found between sociodemographic factors and vegetable intake of all adolescents. Similarly, when observing the association between all the sociodemographic factors and the vegetable and fruit intake of overweight and obese participants, no significant association was found. Intake specified among overweight and obese participants; no significant association was found.

Tables 5, 6 show the associations among personal, social-environmental, and physical environmental factors and the prevalence of vegetable and fruit intake among the adolescent participants, particularly adolescents with overweightness and obesity.

This study observed significant association between physical-environmental factors with adequate vegetable intake among the overall adolescent sample. Adolescents with positive reports of availability at friend's house (PR: 0.804, 95% CI: 0.669, 0.967, p = 0.020) and at places of leisure (PR: 0.755, 95% CI: 0.590, 0.967, p = 0.026) both revealed 19.6% and 24.5% lower prevalence of vegetable consumption. Meanwhile, personal and social-environmental factors did not portray any significant influence on the population vegetable consumption.

On the other hand, the study highlighted self-efficacy (PR: 1.102, 95% CI: 1.013, 1.198, p = 0.024), and intention (PR: 1.126, 95% CI: 1.013, 1.251, p = 0.028) as significant correlates of fruit intake among the general adolescent population, with positive self-efficacy and intention fostering higher prevalence of fruit consumption within the population. In contrast, habit (PR: 0.909, 95% CI: 0.830, 0.996, p = 0.41), and the physical-environmental factor, availability at places of leisure (PR: 0.874, 95% CI: 0.793, 0.963, p = 0.006), significantly lower prevalence of fruit intake among the general adolescent population, while no significant association were found among social-environmental factors and fruit intake within the overall adolescent sample.

For overweight and obese adolescents, the findings indicated no significant association between the personal, social-environmental, and physical-environmental factors with vegetable intake. However, overweight and obese adolescents reporting positive intention (PR: 1.146, 95% CI: 1.002, 1.310, p = 0.047) and parental allowance (PR: 1.125, 95% CI: 1.011, 1.252, p = 0.030) were observed to have 14.6% and 12.5% higher prevalence of fruit consumption, respectively, while availability at home (PR: 0.849, 95% CI: 0.731, 0.987, p = 0.033) showed significantly lower prevalence of fruit intake with more reports of home availability.

The findings of this study revealed a low consumption of fruits and vegetables observed among overweight and obese adolescents in Selangor. From the results, only 23.5% of the adolescent participants have met the recommended daily intake of vegetables, set at equal to or more than three servings per day, as suggested by the Malaysian Dietary Guideline for Children and Adolescents 2023, in line with standards of the World Health Organization (27, 28). In comparison, the national prevalence of vegetable intake, as of 2022, was at 27.1%, which is like the prevalence found in this study (9). In retrospect to both rates, the NHMS 2017 observed a higher percentage of adolescents meeting the vegetable intake recommendation, which had been at 36.0% (7). At the same time, Rohin et al. (11) found that vegetable intake among adolescents in rural Terengganu showed that only 9% had adequate vegetable intake as per the recommendation, which had been substantially lower than that of this study as well as the national prevalence. However, rurality may play a part in these contrasting results. As studies suggest, rurality and sociodemographic factors do affect health literacy (29). Nevertheless, despite the fluctuating rates observed, the findings across these studies suggest that the prevalence of adequate vegetable intake remains low in the country, especially accounting for the national prevalence across the years.

57.8% of the adolescents were reported to have two or more servings of fruit per day, as per the recommendation of the MDGCA 2023 (27). This was significantly higher than the national prevalence of NHMS 2022, which was at 37.3% (9). Meanwhile, the NHMS 2017 reported up to 46.8% of adequate fruit intake among adolescents (7). This suggests that both reports of the national prevalence were lower compared to the findings of this study. Jamaludin et al. (30) found that in Marang and Hulu Terengganu, lower rates of prevalence were observed, with adequate fruit intake among adolescents was 21.7%. Similarly, a study in Ghana reported only 35.7% of adolescents had adequate fruit intake (31). Although this study suggests a higher prevalence of fruit intake of more than 50% among adolescents in Selangor, other studies, as well as the national prevalence, ultimately suggest a low prevalence of adequate fruit intake among adolescents nationally.

Among the participants with overweight and obesity, the findings showed that 25.2% of overweight and 15.0% of obese adolescents had met the recommended serving of vegetables per day, indicating that only a few of the participants achieved the recommended intake suggested in the Malaysian Dietary Guideline for Children and Adolescents (MDGCA) 2023 (27). In NHMS 2017, adequate vegetable intake among overweight and obese adolescents was reported at 9.1% and 5.9%, respectively (7). Similarly, Selamat et al. (32) reported only 3.6% of overweight and obese adolescents had met the recommended intake for vegetables. The current study showed a greater number of adequate vegetable intakes among overweight and obese adolescents as compared to the national prevalence reported in NHMS 2017 (7). Seidu et al. (31), in comparison, revealed a higher prevalence, with 28.7% of overweight and 33.1% of obese adolescents had adequate vegetable consumption, and higher than our study prevalence.

On the other hand, our study reported a total of 64.1% of overweight and 65.0% of obese adolescents to be having adequate fruit intake daily. In contrast, lower percentage of overweight (27.5%) and obese (32.8%) adolescents consumed adequate amounts of fruits were reported in NHMS 2017 (7). Comparable to the findings on vegetable intake, this study showed a higher number of adequate fruit intake among overweight and obese adolescents compared to the percentage reported in NHMS 2017 (7). Seidu et al. (31) also found comparable results in Ghana with 33.0% of overweight and 26.4% of adolescents with obesity have met the recommended intake for fruits. Meanwhile, Selamat et al. (32) also found relatively similar results with only 32.6% of overweight and obese adolescents having adequate intake of fruits.

In contrast, the present study suggests higher intake of fruits among overweight and obese adolescents compared to those of normal weight, which may contradict the general consensus. This could potentially be attributed to societal pressure. For the past several years, a rise in knowledgeable social media influencers (SMI) covering various range of topics including healthy lifestyles and diets has shifted the societal expectations among Malaysian youth (33). It has then led a social contagion mimicking the healthy diet and lifestyle of SMIs, which became prevalent among adolescents, due to their susceptibility to peer influence and social media (34). On top of that, Farhat et al. (35) revealed that adolescents with overweight and obesity are significantly influenced by stigmatization and social comparison, which was seen to have significant associations with social media usage (36). With such trends circulating, it can be postulated that the higher intake of fruits among overweight and obese adolescents could be the result of societal conformity.

Overall, our study shows evidence that the prevalence of vegetable intake among overweight and obese adolescents remains low, reflective of the national prevalence. Meanwhile, the rates for adequate fruit intake were relatively high compared to all the previous studies (7, 31, 32). However, there are generally limited studies which focus on fruit intake as an individual entity, especially among adolescents with overweight and obesity.

The chi-square test revealed no associations between the sociodemographic factors and vegetable and fruit intake. Likewise, when examining the relationship between sociodemographic factors and vegetable and fruit intake among overweight and obese participants, no significant associations were identified. Similar findings were observed in previous research studies, where sociodemographic factors, particularly age, gender, and ethnicity, were not associated with the vegetables and fruit intake of their samples (36–39).

In this context, our findings align with the result from Jongenelis et al. (40), which showed no significant association between sociodemographic factors of age, location, socio-economic position and adequate vegetable intake, however the study instead revealed significant differences in vegetable intake between sexes. In the same study, Jongenelis et al. (40) also revealed that age significantly influence fruit intake among their sample, but the study suggested otherwise for sex, location, and socio-economic position. Albani et al. (41) then suggested that age and sex are not significant influence of vegetable intake, while fruit intake was observed to be influence by age only. This constitutes a relatively similar stand with the findings of Jongenelis et al. (40) for fruit intake, but contrasts with their findings on vegetable intake. The findings also contrast with the present study, where sociodemographic factors do not significantly influence with fruit and vegetable intake (41). In comparison to the present study, Amini et al. (42) suggested that socio-economic status and location positively affects vegetable and fruit intake, contradicting the findings of the present study as well as the of Jongenelis et al. (40).

On another note, parents' education level and household income status were also seen to not have a significant influence on vegetable and fruit intake among adolescents, as reported by Safdar et al. (43) and Drouillet-Pinard et al. (44). However, other studies like Antonogeorgos et al. (45) and Longacre et al. (46) have found that both parents' education level and household income status do have a significant influence on vegetables and fruit intake in adolescents. The studies suggest that adolescents who had parents with higher education levels and higher household income status would have increased consumption of vegetables and fruits, possibly due to the higher health and nutrition literacy that comes with a higher education tier. Moreover, Hamzah et al. (47) showed that a higher level of health literacy in individuals was observed in higher-income households, compared to the other socioeconomic tiers, for up to 53.4% (47). Parents and familial health and nutrition literacy act as a precursor for facilitating healthy eating behaviors among the children of the household. This notion was supported by researchers Chang et al. (48), who highlighted the importance of dietary literacy as a promoter of health status and good eating behaviors of school-age children and adolescents as evidence supported by their research (48).

On the other hand, this study has found physical characteristics, particularly BMI-for-Age (p = 0.022), as a significant factor in fruit intake among adolescents. This study found that individuals with obesity and overweight had higher adequate fruit intake compared to adolescents with thinness and normal weight, revealing a positive association. Similarly, Ham and Kim (49) revealed a significant positive association between fruit intake and BMI among adolescents. However, our findings constitute a different standing from the ISAAC Global Research, which revealed an inverse relationship between BMI and the intake of fruits and vegetables among adolescents in their cohort (50). This notion aligned with the study by Sharma et al. (51) which revealed that adequate intake of fruits was observed to have paradoxical effects, the anti-obesity factor to the prevention of adiposity. The findings of our study may not have reflected this concept, and a higher percentage of overweight and obese adolescents had adequate intake of vegetables and fruits compared to those with normal weight and thinness.

The findings of this study emphasized several personal, social-environmental, and physical-environmental factors as significant predictors of vegetable and fruit intake. The general adolescent sample found self-efficacy, intention, habit, and availability at places of leisure potentially determining adolescent fruit intake behavior, while vegetable intake of this population was shown to be influenced by availability at a friend's house or at leisure places.

Meanwhile, the overweight and obese adolescent population found factors such as intention, parental allowance, and home availability, associated with the population's fruit consumption, whereas vegetable intake was found not affected by the all potential determinants.

The findings indicate significant influence of personal factors such as self-efficacy, intentions, and habits on adolescent vegetables and fruits consumption. Cross-sectional studies by Rohin et al. (11) and Sato et al. (52) revealed similar findings, with correlations between higher consumption of vegetables with positive attitudes and personality attributes toward vegetable and fruit consumption among Malaysian and Japanese adolescents. This was consistent with the study conducted in Bangladesh and the US, whereby adolescents with greater intentions are more likely to consume vegetables and fruits compared to those with poor intentions (53). Moreover, González-Gil et al. (54) found that female European adolescents with healthy dietary patterns or habits consumed a significantly greater number of fruits. Habits were also considered as a significant predictor of vegetable and fruit intake among UK adolescents. However, the result from this European study has a possibility of cofounders such as health purposes or the awareness of the important effects of a sufficient number of vegetables and fruit consumption. Even so, positive attitude, good intentions and good habits were still suggested as crucial elements to be implemented in the adolescents' dietary practices (55). Contrarily, studies suggested liking as a personal indicator of fruit intake (56). Preferences or liking are personal aspects nurtured through physiological growth such as sensory development and sensitivity transitioning throughout age. A past study emphasized that a child with higher sensory sensitivity toward a certain taste, such as bitterness, has been shown to be less accepting of vegetables. However, it was enunciated that the sensory sensitivity has the potential to decrease throughout the years and thus may increase acceptability toward bitterness and hence can be nurtured to increase vegetable intake (57). This study has also shown a higher intake of vegetables going into adulthood compared to during childhood.

The findings of our study also demonstrated a significant association between social-environmental factors, namely parental allowance, with fruit consumption among overweight and obese adolescents. This is consistent with findings by Fu et al. (58) and Bailey-Davis et al. (19) who revealed that vegetable and fruit intake among adolescents are significantly associated and can be facilitated by positive parent modelling and family rules such as demand and allow, but contrasts with Rohin et al. (11), who reported parental modelling, parental encouragement, parental demand, and parental allowance were not associated with neither vegetable nor fruit intake of the Malay adolescents in rural Terengganu.

In a way similar to Rohin et al. (11), the present study finds parental modeling, parental encouragement, parental demand, and parental allowance were not significantly associated with vegetable. However, parental allowance was found to influence fruit intake among our overweight and obese adolescents. Nevertheless, Rohin et al. (11) postulated that their findings can be attributed to the potential dominance of personal character. This notion was supported by the findings revealed by Deliens et al. (10), which suggested better effectiveness in increasing vegetable intake seen through improvement of personal or individual subject norms. In addition, the increase in autonomous functioning—also a facilitator of self-efficacy—present during adolescence constituted a strive of freedom from dependence on parents, which consequently leads to less time spend with parents and siblings and may diminish familial-motivated decision-making (59, 60).

Vegetables and fruit intake of adolescents have also shown significant association with physical-environmental factors. This study highlighted that greater availability of fruits in different environments would increase adolescents' intake of fruits. The availability of fruits at home appeared as a significant factor in the fruit intake of overweight and obese adolescents. The current outcome is similar to the findings from the Family Life, Activity, Sun, Health and Eating (FLASHE) study, where adolescents' fruit intake was significantly affected by home availability (22). Home availability of fruits also showed a significant association with fruit intake of adolescent girls in a Brazilian study (42). The findings of this study also suggest that the availability of fruits at friends' houses and leisure places as significant predictors of vegetable and fruit intake. In contrast, some findings also found that availability at a friend's houses was not significantly related to the daily fruit intake of school children in European countries (61). However, availability at leisure places was found to be a significant determinant for adolescents' vegetable intake, although not fruit intake, in the study by Rohin et al. (11). The possible explanation for these factors could be due to the time and frequency spent in certain locations, namely home or school, and their fruit consumption mainly depends on the availability of fruits at those places (62). A study by Gustafson et al. (23) and Azeredo et al. (61) also reported findings in contrast to ours, whereby school environments with higher availability of fruits promote the fruit consumption of adolescents compared to those with lower fruit availability.

Several limitations have been accounted for in this study. First, due to the COVID-19 pandemic, data collection was carried out using self-reported online questionnaires, which might increase the risk of over- or underreporting the participants' vegetable and fruit intake. It must also be accounted that the use of technology as the primary medium for data collection may skew the demographic representation within the sample, as it favors more tech-savvy individuals and those with internet accessibility. Adolescents with limited access to the internet may not be able to participate and inadvertently be excluded from the sample. However, due to the COVID-19 restrictions enforced by the government such as strict bans on gatherings and temporary shut-down of education institutions, an online data collection was the viable option at that point in time. In turn participants also had to self-report their weight and height, which increases risk of inaccurate BMI estimation.

To counter these limitations, future research should employ physical forms of data collection, through printed questionnaires and guided data collection sessions at the target institutions by trained enumerators and researchers. Anthropometric data should also be collected on-site and be conducted by trained personnel using the right anthropometric equipment such as the SECA weighing scale for measuring body weight and a stadiometer for measuring height.