Sociodemographic, lifestyle, and psychological factors as controllable predictors of academic self-efficacy after reforming a medical education system; the Egyptian Nationwide experience

Background

Academic self-efficacy (ASE), influenced by various sociodemographic, lifestyle, and psychological factors, is a critical determinant of academic performance in medical students. In Egypt, the medical education system underwent recent drastic changes. Therefore, we aim to explore the effects of these changes on the variables above and, subsequently, on ASE among a national sample of Egyptian medical students.

Methods

We conducted a cross-sectional study across 30 Egyptian medical colleges using a validated electronic Arabic survey to evaluate sociodemographic, lifestyle behaviors, psychological factors, and ASE. Categorical variables were described using frequencies and percentages, while Continuous variables were summarized as mean and standard deviation. Bivariate analysis with post-hoc testing, Pearson’s correlation coefficient, and multivariate linear regression were employed to assess the associations between the study variables and ASE. Data analysis was performed using R Statistical Software.

Results

A total of 1446 participants were included, with a mean age of 20.59 ± 1.68. Of these, 61.5% (n = 890) were Females, while 94.3% (n = 1363) were single. ASE was significantly associated with certain financial status, smoking, sleeping hours, breakfast patterns, diet changes, and fluids. Stress, anxiety, and depression were associated with and correlated to poor ASE. Only depression, oversleeping, and smoking were significant predictors of worse ASE.

Conclusions

Multiple variables are significantly associated with, correlated to, or predictors of ASE after the recent reform in the Egyptian medical education system. Subsequently, targeted interventions to specific sociodemographic, lifestyle, and psychological factors can be adopted and assessed through future longitudinal pre-post interventional studies.

Academic self-efficacy (ASE), defined as students’ confidence in their ability to succeed and their persistence in completing tasks and learning new information, is a crucial predictor of academic success [1, 2]. Sociodemographic factors such as family influence, gender, age, finances, and personality traits can shape a student’s ASE [3]. Likewise, positive lifestyle behaviors such as adequate sleep, a balanced diet, and regular breakfast consumption can positively affect cognitive function and ASE [4]. In contrast, poor mental health and a sedentary lifestyle, characterized by low physical activity and energy use, negatively impact academic performance [5, 6]. Unfortunately, sedentary behaviors are becoming more common worldwide, with many people not meeting recommended activity levels [7]. Likewise, mental health issues are common among college students, particularly in academic settings [8]. Medical students face additional pressures to excel academically, making them even more vulnerable to these issues [9].

Unhealthy behaviors and sedentary lifestyles are widespread in Egypt, a lower-middle income country with recent mounting economic pressures, including rising inflation, currency devaluation, and increasing unemployment rates, leading to higher levels of health and psychological problems among university students compared to other developed countries [10,11,12]. Moreover, the medical educational system in Egypt, resembling many different countries, has followed the British model, a teacher-centered and discipline-based program consisting of six academic years followed by one year of internship [13, 14]. In 2013, the educational program in the United Kingdom, followed by Egypt in 2018, underwent a fundamental transformation into a student-centered and integrated competency-based program of five academic years followed by two years of internship, allowing earlier clinical exposure [14, 15] This drastic transition could further exacerbate the lifestyle and psychological factors undermining ASE and academic performance. Therefore, in this nationwide cross-sectional study, we aim to thoroughly investigate the effects of this significant educational change on students’ lifestyle behaviors and psychological domains and, subsequently, on their ASE. After that, we would provide insights into possible considerations and interventions that could secure a more appropriate transition with better outcomes.

Study design, setting, and participant eligibility

We conducted a multi-centric cross-sectional study between February and June 2024 in 30 Egyptian governmental and public medical schools from four regions: Upper Egypt, Suez Canal, Delta, and Greater Cario and Alexandria regions. A complete list of involved universities is outlined in Table S1. We invited a group of data collectors (The ASE-Egypt Group) representing the included universities to manage the data collection. Inclusion criteria included medical students from the first to the fifth grade, regardless of gender. Meanwhile, non-Egyptian students, interns and those who refused to participate were excluded. The study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist in its entirety [16]. Clinical trial number is not applicable.

Sample size

Epi info Statistical Software (version 7.2.6) was used to calculate the sample size using the formula [n = z2P(1-P)/d2] with a margin of error of 5%, a confidence level of 95%, and a 50% response distribution. The estimated sample size was at least 382 participants, predicted to be a population of around 70 thousand. To increase the study’s power and generalizability, we increased the sample size to 1446.

Study instruments (questionnaire design)

The questionnaire used for data collection was formulated based on previously validated Arabic indices. The questionnaire consisted of four sections. The first section encompassed the sociodemographic and lifestyle characteristics. The second section assessed sedentary behaviors using the sedentary behaviors questionnaire (SBQ). The SBQ included nine weekend and weekday behavioral patterns and was deemed reliable in the previous studies [17, 18]. All the times were converted into hours, then the Arabic SBQ’s total sitting time (hours/week) was computed using the formula [ (sedentary behavior on weekdays × 5) + (sedentary behavior on weekends × 2)], and the result was divided by 7 to obtain the daily average. Based on previous literature, we categorized sedentary behavior into four levels based on the daily sitting time calculated before as follows:1) Less than 4 h per day, 2) 4 to 6 h per day, 3) 6 to 8 h per day, and 4) 8 h or more per day [19,20,21]. The third section utilized the eight-item mixed Depression Anxiety Stress Scale-8 (DASS-8) to evaluate the psychological domains [22]. Finally, the fourth section encompassed the ASE scale, a rapid assessment tool developed to determine respondents’ self-efficacy regarding intellectual abilities, such as time management, note-taking, test-taking, and overall academic competence linked to academic success [23]. The modified Arabic version of ASE is a nine-item with a seven-point Likert scale, where one represents very untrue, and seven indicates very truthful. A diagrammatic representation of the conceptualization of the questionnaire is visualized in Fig. 1. The complete questionnaire is depicted in Table S2.

Questionnaire conceptualization: a consort diagram detailing the steps of questionnaire development, including the different sections with the corresponding number of questions

Full size image

Data collection

We implemented snowball and convenience sampling techniques for data collection. Electronic self-administered Google Forms representing the questionnaire were published through common social media platforms and official medical student channels. At the beginning of the questionnaire, the study’s goal and the voluntary, anonymous nature of participation were described, followed by informed consent. To limit duplicate responses and information bias, we activated the restriction option in the Google form setting for only one response per email and excluded incomplete responses.

Ethical considerations

Contribution to this study was voluntary, and it followed the principles of the Declaration of Helsinki (1964, last revised in 2013) [24]. Ethical approval was obtained from the institutional review board committee at Al Azhar University, Faculty of Medicine (approval code: 0000003/2/24). Every participant provided their informed consent electronically before participation. The online questionnaire link did not collect any personal information. The collected information was accessible only to the research team.

Statistical analysis

Data analysis was conducted using R Statistical Software (v4.1.3; R Core Team, 2022) [25] Categorical variables were summarized as frequencies and percentages. Continuous variables were summarized as mean and standard deviation (SD). Due to the large sample size, normality tests would be unnecessarily oversensitive to even small deviations from normal distribution, even though this deviation will not affect the results of parametric tests [26]. Therefore, after fulfilment of the size requirements in each group for every single analysis [26], we used parametric tests including the independent t-test and one-way ANOVA test to compare group means. Following a significant result from the ANOVA, post hoc pairwise comparisons were performed using the Bonferroni correction to adjust for multiple comparisons. Pearson’s correlation coefficient was employed to assess correlations between continuous variables. A multiple linear regression analysis was performed to identify predictors of ASE. Independent variables that demonstrated statistical significance in the bivariate analysis were included in the final regression model. A p-value of less than 0.05 was considered statistically significant.

Participants demographic

In total, 1509 participants completed the questionnaire, of whom 1446 were included in the final analysis (Fig. 2) with a mean age of 20.59 ± 1.68. Of the total sample, 61.5% (n = 890) were Females, 94.3% (n = 1363) were single, and 54.6% (n = 790) described their current financial situation as sufficient without saving. A summary of all other baseline characteristics is shown in Table 1. Answers to the questionnaire are illustrated in detail in Table S3.

Participants’ eligibility: a consort diagram detailing the process of participants’ inclusion and exclusion with the total numbers in every step

Full size image

Association of categorical variables with ASE

The ASE score showed a significant reduction with worse financial status, smoking, sleeping ≥ 9 h, not having breakfast, worse diet changes, and fluid intake < 1 L. We found no significant association between the ASE and sedentary scores. Moreover, increasing anxiety, stress, or depression score was associated with worsening ASE.

Stress scores

Significant differences in ASE were observed based on stress scores. Post-hoc comparisons revealed that individuals with severe stress scored significantly lower than those with moderate (Mean Difference = −2.34, p = 0.002, 95% CI: −3.98, −0.70) and normal stress (Mean Difference = −4.38, p < 0.001, 95% CI: −6.22, −2.53). The effect size for stress score on ASE was moderate (Eta-squared = 0.023, 95% CI: 0.009–0.039).

Anxiety score

A significant effect of anxiety on ASE was found p < 0.001 (Eta-squared = 0.039, 95% CI: 0.021–0.059). Post-hoc comparisons revealed significant differences between the anxiety groups. Individuals with normal anxiety had significantly higher ASE compared to those with moderate anxiety (Mean Difference = 3.60, p < 0.001, 95% CI: 1.99, 5.22), and higher than individuals with severe anxiety (Mean Difference = 5.42, p < 0.001, 95% CI: 3.59, 7.25).

Depression score

A significant effect of depression on ASE was found p < 0.001 (Eta-squared = 0.086, 95% CI: 0.059–0.113). Post-hoc tests revealed significant differences: individuals with moderate depression scored lower than those with normal depression (Mean Difference = −3.25, p < 0.001, 95% CI [−4.91, −1.59]). Additionally, individuals with severe depression scored lower than those with normal depression (Mean Difference = −8.23, p < 0.001, 95% CI [−9.94, −6.53]), and lower than those with moderate depression (Mean Difference = −4.98, p < 0.001, 95% CI [−6.68, −3.28]).

The results of the bivariate and the post-hoc pairwise analyses are summarized in Table 2.

Pearson’s correlation analysis

ASE had a significantly weak negative association with anxiety (−0.205, p < 0.001), stress (−0.148, p < 0.001), and depression (−0.310, p < 0.001). No significant correlation was found between ASE and sedentary behavior. Indices of mental health showed a significantly strong positive correlation to each other and a weak positive correlation to sedentary behavior. The results of the correlation analysis are depicted in Fig. 3.

Pearson Correlation Matrix: An illustration of the correlation between the different continuous study variables and themselves

Full size image

Multiple Linear regression analysis

Having enough money with savings (β = 3.21, p = 0.004) and the stress score (β = 0.57, p = 0.018) were predictors of improved ASE, while smoking (β = −5.84, p = 0.004), sleeping ≥ 9 h (β = −2.35, p = 0.015), and depression score (β = −1.36, p < 0.001) were negative predictors. The detailed results of the regression analysis are presented in Table 3.

In this nationwide cross-sectional study on 1,446 medical students among 30 medical colleges in Egypt, we examined for the first time the impact of multiple sociodemographic, lifestyle, physical, and psychological factors on ASE after the recent major educational reform in Egypt.

Association of ASE with the sociodemographic characteristics

In alignment with the recent economic difficulties in Egypt [10], 8% of participants deemed their income inadequate, and 54.6% regarded it as sufficient without the ability to save, which is notably worse than what was reported back in 2009 when 79.9% of respondents expressed financial satisfaction [27]. Notably, a significant portion of the current population lives below the poverty line, particularly in Upper Egypt [28], aligning with the more ASE reduction we observed among Upper Egypt students. Our results show that higher income was correlated to and a predictor of higher ASE. Students from higher-income families may have access to better educational resources and alleviated financial stress, which could potentially influence their ASE. This is consistent with a previous systematic review demonstrating that medical student debt negatively impacts both academic performance and mental health [29]. However, multiple confounders, such as regional disparities, may have influenced the observed associations. For instance, students from Upper Egypt may face limited educational opportunities, healthcare infrastructure, or mental health support as most mental health providers are in the capital, even if their family income is adequate compared to the other more privileged regions. Further research is needed to take these confounding factors into consideration [28, 30]. Notably, we found no significant differences in ASE between public and private universities, which may potentially reflect two opposing balancing factors. Public universities include the most prestigious institutions in Egypt which necessities higher academic scores at baseline by the high-school students for enrollment. On the other hand, private university students, despite having relatively lower academic scores at baseline, benefit from better educational opportunities and possibly a better economic and financial status at baseline, resulting in comparable overall ASE to public universities. Nevertheless, future prospective studies can provide more reliable insight on this area.

Impact of daily lifestyle behaviors on ASE

Consistent with previous studies in Egypt [31], the prevalence of smoking in this study was 2%, which is lower than in neighboring countries like Saudi Arabia (12.4%) and Sudan (48.8%) [32, 33]. This relatively lower prevalence may be attributed to a relatively higher level of awareness among Egyptian students [34], in addition to cultural restrictions [35], and a more extensive representation (68.4%) of the earlier, less stressful years in our sample [36]. Nevertheless, smoking, which is known to be associated with other mental health issues and reduced overall performance [37, 38], was directly associated with and a strong predictor of worse ASE in our study.

About 22.2% of the participants slept less than 6 h daily, an increase from 17.9% before the reform [31]. However, in our analysis, sleep deficiency was not statistically inferior to normal sleep. Although insufficient sleep is known to worsen cognitive outcomes [39], it is not uncommon for medical students to sacrifice sleep to study more [40], potentially improving productivity. Although this sounds more productive in the short term, it could be counterproductive in the long term [41], which we could not assess in our study. Interestingly, oversleeping (> 9 h) was associated with worse ASE than both normal sleep (7–8 h) and sleep deprivation (< 6 h), which was similar to the results of a previous meta-analysis [42]c. Moreover, oversleeping was the only sleep pattern significantly predicting a worse ASE. Oversleeping may reflect poor time management or underlying health conditions, such as narcolepsy or depression, which could explain its association with worse ASE. However, future studies, with longitudinal designs, are required to further explore oversleeping effects [43, 44].

Consistent with the observed positive influence of healthy diet habits on academic performance, we found a positive significant association between regular breakfast consumption, daily fluid intake, and more nutritious diets with ASE [45,46,47]. However, none of these factors demonstrated a predictive capability. Notably, the most significant improvement in ASE was observed when frequent breakfast consumption (at least five times weekly) was compared to avoiding breakfast exclusively, suggesting maintaining consistent breakfast throughout the week to maximize benefits. Moreover, we found that drinking less than 1 L of water had lower ASE than 1–4 L daily. Interestingly, overhydration (> 4 L per day) was also associated with poorer ASE outcomes compared to adequate hydration (1–4 L). Overhydration can reflect underlying health issues, such as diabetes or psychogenic polydipsia that cause excessive thirst and can concomitantly impair ASE [48], however, similar to oversleeping, more longitudinal studies are needed.

Although several previous studies suggested that a sedentary lifestyle is associated with lower self-efficacy [49]. our study did not find such an association despite a high prevalence (54.9%) of moderate-high sedentary lifestyles, which may be attributed to the lower likelihood of medical students engaging in extracurricular physical activities compared to their non-medical peers, at the expense of dedicating most of their time to studying, mitigating the adverse effects of a sedentary lifestyle [50]It is important to note that a pronounced lack of physical activity and disconnection from other aspects of life remains a concern that can impact mental health [51]. which was supported by the significant positive correlation in our results between sedentary behavior and all the examined psychological domains, that were eventually associated with a lower ASE.

Influence of mental health on ASE

We observed a high prevalence of moderate to severe depression (65.8%), anxiety (57.9%) and stress (75.2%) among the participants, higher than the previously reported results across a large medical college in Upper Egypt in 2015 (47.3%, 56.9% and 46.1% respectively) [29]. Although anxiety was broadly similar, a marked elevation in depression and stress levels is evident, which may be attributed not only to the worsening economic situation but also to the increased content load after the reform, which might happen if there was a condensed curriculum duration without enough content tailoring [15, 31]. Compared to other neighboring countries, we had worse results than Saudi Arabia, a wealthier and more stable country (43.3%, 37.23% and 29.88%, respectively), and similar results to Sudan, another developing country with recent economic difficulties except for stress where we still had a higher prevalence (60.8%, 43.8% and 37.6%, respectively) [52, 53].

Although all the investigated mental health conditions were associated with lower ASE, confirming their detrimental effects on cognitive functions [54], depression demonstrated the strongest correlation and highest predictive capability for poor ASE, in contrast to anxiety, which was associated with but not predictive of worsening ASE. However, our findings on the relationship between stress and ASE were more complex. Despite the significant association and correlation between stress and worse ASE, stress emerged as a weak positive predictor of ASE. Previous research has suggested that stress can be partially associated with increased effort and motivation [55]. Therefore, it is possible that high-performing students report more subjective stress, especially during high-stakes exams or deadlines. However, chronic and escalating stress can eventually lead to burnout, resulting in worse performance over time [56]. Therefore, although stress may temporarily enhance academic performance, it is counterproductive in the long term, an observation that requires further investigation since we could not distinguish between short- and long-term stresses. Finally, a positive correlation was consistently observed between all the examined psychological domains and each other, suggesting the presence of additive or synergistic effects that further diminish ASE. Previous studies have indicated that chronic stress may elevate anxiety levels, with anxiety persisting even after the initial stressor is removed, underscoring the compounded effects of mental health issues on one another [55, 57].

Implications

The study results provide insight into multiple potential areas of interventions based on the observed trends that can be then assessed through pre-post intervention well-controlled longitudinal studies to establish their practical implications [58]. First, the adoption of academic support strategies in the form of mentorship programs implemented by the faculty members should be encouraged and supervised to help students get through their curricula, manage their time, and relieve their stress [59]. Mentorship programs can be implemented by recruiting faculty members or senior students and pairing them with junior students based on shared interests or specialties. One-on-one or group sessions should be scheduled to provide academic support, such as exam preparation, career guidance (e.g., specialty selection), and stress relief. Additionally, the new medical curricula should be under continuous evaluation and revision by the faculty quality assurance and student affairs units, removing unnecessary content while obtaining constructive regular feedback from the students on their mental health and academic concerns [15]. A dedicated committee comprising faculty members, education experts, and student representatives should be established to discuss curricular changes, conduct regular surveys, and analyze student performance metrics, including results and dropout rates, to identify potential barriers. Scheduled breaks are recommended between lectures to maintain students’ concentration. In case of any evidence of physical or psychological concerns, students should have the choice to postpone some of the curricular content or take a gap year, if needed, without losing grades [60]. Moreover, given that many medical students usually encounter financial barriers, more attention should be devoted to releasing financial strains by providing financial aid, affordable educational materials, and in-campus educational activities that engage students from different social backgrounds, future research can explore how financial aid programs impact ASE over time particularly in low-income countries. Additionally, the study revealed that only 55.3% of the population had a normal BMI, and only 18.5% had recently adopted healthier dietary changes. Lifestyle modification programs promoting healthy habits, such as regular physical activity and a balanced diet, should be considered in future research, as targeted interventions may have long-term positive effects on students’ ASE. Therefore, medical colleges should implement in-campus nutrition education programs and extracurricular physical activities, with regular re-evaluation [61].

Strengths and limitations

This is the first study among Egyptian medical students to investigate the effects of multiple sociodemographic, lifestyle, and psychological factors on ASE and the first to examine the impact of the recently reformed medical education system on various aspects among medical students. The study was done in 30 Egyptian universities representing variable regions and sociodemographic backgrounds along the nation, with a sample size exceeding three times the estimated sample size to maximize generalizability. We used validated Arabic indices to formulate a comprehensive questionnaire that future studies in other nations can use. In addition, we adhered to the STROBE guidelines while reporting the study results and implemented a robust statistical analysis plan to provide meaningful results. Therefore, the implications of our study can apply to Egypt and other developing countries, especially those undergoing similar financial challenges or planning to implement significant changes or reforms in the medical education system.

Given the nature of cross-sectional study, we faced some inevitable limitations. For instance, we could not differentiate short- from long-term behaviors and mental health conditions, in addition, although significant associations were observed between the study variables, the cross-sectional nature restricts establishing definitive conclusions about temporality and causality. All these concerns can be better assessed in future longitudinal designs. Furthermore, a self-reported survey-based questionnaire is liable to response, non-response, and recall biases. Moreover, we only included medical students without comparing them directly to non-medical students. Additionally, we could not control for certain critical confounders such as regional or socioeconomic differences. Finally, using the convenient and snowball sampling could have introduced selection bias.

Conclusion

A nationwide cross-sectional study was conducted to examine the impact of sociodemographic, lifestyle, and psychological factors on ASE among 1,446 Egyptian medical students after the recent reform in the education system. Increased levels of anxiety, stress, and depression were significantly associated with and correlated to worse ASE. The main predictors for worse ASE were smoking, oversleeping, and depression, while financial stability was a positive predictor. Although the study results highlight potential areas for intervention, the cross-sectional design is observational with limited causal insight. More longitudinal studies are required to validate the causal relationship of the studied variables. The suggested interventions should be further assessed through pre-post interventional studies.

All relevant data are represented in the manuscript or the Supplemental Appendix. Any further data is available upon a reasonable request to the corresponding author.

ASE:

Academic self-efficacy

STROBE:

Strengthening The Reporting of Observational Studies in Epidemiology

SBQ:

Sedentary behavior questionnaire

DASS:

Depression, Anxiety and stress scale

IQR:

Interquartile range

We thank the ASE-Egypt Collaborative Group for their outstanding efforts during the data collection. We also thank the participants for their patience and time participating in our study.ASE-Egypt Collaborative Group: Ahmed Hatem Ibrahim, Abdelwahab Waleed Salah, Rana Mohamed Attallah, Mariam Mohamed El-Saeed, Esraa Mohammed Marei, Ahmed Lotfy, Ahmed Khaled Fathy, Omar Osama, Ibrahim Elsayed Abdelazem, Saara Yahya Ibrahim, Mohamed Alaa El-Tonbary.

ASE-Egypt Collaborative Group

Ahmed Hatem Ibrahim, Abdelwahab Waleed Salah, Rana Mohamed Attallah, Mariam Mohamed El-Saeed, Esraa Mohammed Marei, Ahmed Lotfy, Ahmed Khaled Fathy, Omar Osama, Ibrahim Elsayed Abdelazem, Saara Yahya Ibrahim, Mohamed Alaa El-Tonbary.

No funds were received by the authors for this manuscript.

Authors and Affiliations

1. Faculty of Medicine, Al-Azhar University, Nasr City, Cario, Egypt

   Ahmed Amir Samir

2. Faculty of Medicine, Mansoura University, Dakahlia, Egypt

   Ahmed Hussein Elamir

3. Faculty of Medicine, Menoufia University, Menoufia, Egypt

   Mohamed Basyouni Helal

4. Faculty of Medicine, South Valley University, Qena, Egypt

   Yomna Goudy

5. Faculty of Medicine, Mansoura University, Dakahlia, Egypt

   Khaled Elbarbary

6. Faculty of Medicine, Alexandria University, Alexandria, Egypt

   Mohamed El-Mezayen

7. Faculty of Pharmacy, Alexandria University, Alexandria, Egypt

   Wesam Abo-Elenien

8. Faculty of Medicine, Zagazig University, El-Sharkia, Egypt

   Ebtehal Hashem Abdelazim

9. Faculty of Medicine, Fayoum University, Fayoum, Egypt

   Manar Alaa Mabrouk

10. Minneapolis Heart Institute Foundation, Minneapolis, MN, USA

    Kerollos Abdelsayed

11. Aswan Heart Centre, Magdi Yacoub Foundation, Aswan, Egypt

    Kerollos Abdelsayed

Consortia

Contributions

AAS: Conceptualization, Writing-original Draft, Granting IRB Approval, and data collection supervision. AHE, MBH: carried out the analysis and interpreted the results. YG: Writing original draft and Review. KE, MYA, WAE, EHA, MAM: Data Collection, Writing original Draft. KA: Conceptualization, Writing-Review, editing, and project supervision. ASE-Egypt Collaborative Group: Data Collection. All the authors gave their consent.

Corresponding author

Correspondence to Ahmed Amir Samir.

Ethics approval and consent to participate

Contribution to this study was voluntary and followed the principles of the Declaration of Helsinki (1964, last revised in 2013). Every participant provided their informed consent electronically before participation. The online questionnaire link did not collect any personal information. The information was accessible only to the research team. Ethical approval was obtained from the institutional review board committee at Al Azhar University, Faculty of Medicine (approval code: 0000003/2/24).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions