The adoption and maintenance of a healthy lifestyle based on five domains, namely nutrition, physical activity, sleep, stress management, and substance abuse (WHO, 1999), has been recognized as a priority by the World Health Organization (WHO, 2024b). With regard to the physical activity (PA) behaviors in Europe, the last Eurobarometer on Sport and Physical Activity (European Union, 2022) identified that Europeans are more sedentary; it also acknowledges that the overweight, obesity, stress, and substance abuse (alcohol and tobacco) rates have risen. The survey identified that approximately 46% of Europeans never exercise or engage in PA and that this number has slightly increased from previous years. Additionally, sedentary behaviors like sitting for long periods of time (over 8.5 h per day) have become more common, leading to public health concerns related to chronic conditions like cardiovascular diseases and type 2 diabetes (T2D) that are associated with unhealthy lifestyles. Effectively, low levels of PA and poor nutritional habits characterized by high calorie intake are highly associated with T2D (Amanat et al., 2020). Diabetes is known to negatively impact the quality of life and subjective wellbeing (SWB) of individuals (Fisher et al., 2010; Hajos et al., 2012). Given the physical limitations and emotional burden associated with diabetes (Lloyd et al., 2001), the satisfaction with life (SWL), positive and negative affect, as well as cognitive and affective variables of SWB (Diener et al., 1985) are influenced by this disease. According to available research, individuals affected by T2D have lower perceptions of SWL, lower levels of positive affect (Anderson et al., 2001; Lloyd et al., 2001), and higher levels of negative affect (Fisher et al., 2010) compared to individuals without the disease.

Lately, there has been an increase in T2D prevalence worldwide; nearly 537 million adults have been diagnosed with diabetes, and it is expected that this number will increase to 783 million by 2045 (IDF, 2021). In Portugal, diabetes reportedly affect 13% of the adult population. The IDF (2021) also projects a significant increase in the number of diabetes cases in this country given its large elderly population as well as high prevalence of obesity and sedentary behaviors. Hence, physical exercise (PE) is an essential tool in the management of T2D subjects as it can improve glucose control, potentiate insulin sensitivity, and promote cardiovascular and overall health (Colberg et al., 2016). According to Umpierre et al. (2011), exercise helps lower blood sugar levels by increasing its usage; thus, exercise increases the body’s demand for energy, which the muscles obtain from glucose. PE practice along with dietary changes can improve weight management, which in turn improves blood glucose regulation and insulin sensitivity (Knowler et al., 2002). With regard to exercise typology, research has shown that combining different exercise forms (aerobic and resistance aerobic) could improve some of the glucose metabolism markers and inflammatory parameters in sedentary adults without diabetes (Ferreira et al., 2022; Silva et al., 2024); combination exercises may also be more effective for management of T2D than aerobic or resistance exercises alone (Schwingshackl et al., 2014). Moreover, according to Colberg et al. (2010), exercise programs designed specifically for T2D subjects can improve mobility and aerobic resistance, which could facilitate the daily PAs in these individuals’ routines (e.g., climbing stairs, walking more, or doing household tasks). Adherence to an exercise program can build a virtuous cycle, where the perceived benefits reinforce a physically active lifestyle. Exercise can also contribute to behavioral changes in other areas essential to T2D management. In addition, individuals involved in exercise programs are more aware of the importance of dietary control for optimizing glycemic control. According to Wing and Phelan (2005), exercise practice is often associated with a concern for the diet since people who adopt active lifestyles are also looking for complementary improvements to their nutritional lifestyles.

The benefits of a healthy lifestyle are well-defined and linked to a higher average life expectancy as well as a lower likelihood of contracting serious chronic diseases (Oftedal et al., 2019; Silva et al., 2024); thus, there is emerging interest for researchers to be able to evaluate the lifestyle of the population, which in turn indicates the need to aggregate different lifestyle parameters in a single instrument. To address this need, the Fantastic Lifestyle (Wilson et al., 1984) was developed to assess the lifestyle of subjects; this questionnaire comprises nine factors and 25 items that explore the habits and behaviors concerning population lifestyle. This instrument analyzes the physical, psychological, and social behaviors associated with the FANTASTIC acronym, namely, family and friends (F), activity (A; physical activity), nutrition (N), tobacco/toxins (T; tobacco and drugs), alcohol (A), sleep/stress (S), type of personality (T), insight (I), and career (C). The original version of this questionnaire (Wilson et al., 1984) has nine factors and 25 items, where some factors only have two items; this instrument has already been tested in Mexico in people with high blood pressure (López-Carmona et al., 2000) and T2D (Moctezuma et al., 2003). Additionally, Añez et al. (2008) and Ramírez-Vélez Agredo (2012) confirmed that the initial structure of this scale was reliable for accessing the lifestyles of Brazilian and Colombian adults, respectively. In Portugal, Silva et al. (2024) identified that this instrument can be used to assess the lifestyles of Portuguese young adults through a factorial structure comprising nine factors and 30 items; even with some factors having less than three items, this factorial structure is different from the original one proposed by Wilson et al. (1984), which reveals that the questionnaire can have different factorial compositions.

Independently of the above, the presence of factors with only two items is a drawback of the Fantastic Lifestyle questionnaire since it is paramount to have factors with at least three items (Kline, 2016). According to Nunnally and Bernstein (1994), factors with only two items may have issues regarding stability, and the reliability may be affected considering that two items are not sufficient for robust measurements with the model. Moreover, factors with few items (i.e., <3) may not be sufficiently developed to capture the full scope of the constructs in question (Worthington and Whittaker, 2006). Beyond these concerns, to the best of our knowledge, previous studies on the psychometric qualities of the Fantastic Lifestyle questionnaire were based on exploratory factor analysis (EFA). To validate the internal structures of the scales and confirm the theoretical constructs, it is essential to perform confirmatory factor analysis (CFA) (Cid et al., 2022); this methodology helps ensure that the relationships between the observed variables and their underlying latent factors are consistent with theoretical expectations. Considering that the assessment instruments developed specifically for this area are not abundant or inclusive of some limitations, such as the Fantastic Lifestyle questionnaire, researchers have two options for action (Fonseca and Paula Brito, 2005) as follows: (1) develop new instruments; (2) make adaptations of instruments already existing in other languages into Portuguese.

Given the intention of developing a questionnaire with more robust psychometric qualities to assess the lifestyles of Portuguese adults, including those with T2D, and considering the abovementioned concerns, we aimed to develop and analyze the factorial structure of the new Portuguese version of the Fantastic Lifestyle questionnaire (nine factors and 28 items) through EFA and CFA in study I. This version is derived from the original and also includes the full Brazilian version (Añez et al., 2008) as well as a part of the Portuguese version (Silva et al., 2014), thereby departing from the original by having one more factor and five more items. This new version includes three items per factor to address the reliability issues in the previous version. Furthermore, considering the issues related to the population’s lifestyle; prevalence of T2D in Portugal and globally; positive impacts of combined exercises on disease management, quality of life, and SWL; as well as promotion of daily PA behaviors and nutritional alterations, we aimed to implement and assess the effects of the Diabetes em Movimento ^® on the SWB, daily physical activity, and nutritional habits in Portuguese subjects with T2D through study II.

The effects of combining PE programs on the health and SWB of T2D subjects are well defined (Chudyk and Petrella, 2011; Hayashino et al., 2012; Mendes et al., 2016; Oliveira et al., 2012). However, many of these programs require expensive equipment and are difficult to access by subjects with pathologies and older people, which represents a high burden on healthcare and unavailability to most people with T2D (Mendes et al., 2016). Beyond these community PE programs for specific population groups, the WHO has suggested recommended interventions to promote lifestyle changes (Mendes et al., 2017; WHO, 2009); these interventions appear to be cost-effective and more applicable than individual interventions and have also been linked to better adherence and long-term results in the T2D population (Plotnikoff et al., 2013; Sazlina et al., 2013). The main goal of this work was to test the effectiveness of a PE program - Diabetes em Movimento (Mendes et al., 2016, 2017) designed specifically for people with T2D on their SWB (i.e., satisfaction with life and different affect) as well as lifestyle changes based on the main dimensions of the PANSAS model that is based on the WHO’s global recommendations for physical activity and health. Furthermore, we propose to validate (in study I) the Portuguese version of the FANTASTIC questionnaire (Wilson et al., 1984) that allows assessment of the various dimensions of lifestyle.

EFA: From the data presented in Table 1, we observe that nine factors were extracted from EFA with eigenvalue ≥1, justifying overall 57.75% total variance of the results; one item (item 12) that showed cross-loading between the factors tobacco and nutrition was not considered acceptable and was eliminated. Thus, the final analysis included nine factors with 27 items, and the results of the factors were quite acceptable (≥0.50), except for item 12. The values of the factorial weights for the respective factors varied between 0.51 and 0.17, and all items consequently explained at least 25% of the variance of the latent factor, while the Cronbach’s alpha of factors varied between 0.74 and 0.82. From Table 2, we observe that family was the factor with the highest value (3.54 ± 0.51) while sleep/seatbelts/stress was less valued by the subjects (0.72 ± 0.53). The descriptive analysis revealed that the subjects valued factors related to family, nutrition, personality traits, and career more than the others. However, most of the correlations were significant both directly and inversely, with a negative correlation between family and alcohol/addictive substance consumption as well as a positive correlation between introspection and career.

CFA: A preliminary analysis of the data was carried out on the missing values as well as univariate and multivariate distributions, as suggested by Hair et al. (2019). The results showed that there were no missing values or violations of the univariate normal distribution since the asymmetry was between −2 and +2 and kurtosis was between −7 and +7 (Byrne, 2016). However, the assumption of multivariate distribution was found to be violated as the multivariate Mardia kurtosis coefficient exceeded the recommended cutoff value (>5) (Byrne, 2016). In this sense, the bootstrap procedure (2,000 samples) was used for subsequent analyses, as suggested by Nevitt and Hancock (2001). The analysis of the Fantastic Lifestyle measurement model (nine factors/27 items) fitted the data based on the cutoff values adopted: (χ2 = 556.40; SRMR = 0.035; bootstrapped p = <0.001; RMSEA = 0.040 [90% CI = 0.035, 0.045]; TLI = 0.946; CFI = 0.956). We also identified that AVE was ≥0.50 and composite reliability was >0.70 for all the factors, as indicated in Table 3, which consequently explained at least 25% of the latent factor variance. However, some discriminant validity problems were identified involving the factors family and alcohol consumption/addictive substances; physical activity and sleep/stress; nutrition and sleep/stress; and introspection and career, where the squared values of the correlations between the factors were higher than the AVE values of the individual factors. There were no observed problems with the discriminant validity for the remaining factors.

The EFA showed that the subjects valued the factors under analysis and that they all had adequate internal consistency since the Cronbach’s alpha value always exceeded 0.70, which is considered the minimum value for internal validation of an instrument (Hill and Hill, 2008). In the study by Silva et al. (2014), only the overall Cronbach’s alpha was calculated but not the other parameters. Even so, in our analysis, the lowest alpha value is 0.71 and highest is 0.82, which are higher than the values obtained for the preliminary validation (which were below the minimum recommended values). Thus, we observe that all the factors in the final instrument are applicable for their intended measurements. The factor weights were also satisfactory since they were significant and higher than 0.50, thus verifying that all the items explained at least 25% of the variance of the latent factor (Hair et al., 2019). During this analysis, a cross-loading was found, where item 12 was associated with both nutrition and tobacco factors. Considering that the contents of items 11 and 12 are integrated into a single item in the original version and are separated in the Brazilian version of the questionnaire as well as considering that the factor tobacco had four items, we decided to remove item 12 from the final factorial structure as in the study by Dinzeo et al. (2013), as suggested by Kline (2016). However, the total variance of our results was 57.75%, which is higher than the 55.02% reported by Dong et al. (2012) and the 33.33% reported by Dinzeo et al. (2013). The present version (i.e., with 27 items) was built on the basis of theoretical and psychometric criteria by considering that three items were sufficient to represent the latent constructs and warrant model identification in the CFA process, as suggested in previous studies (Worthington and Whittaker, 2006). Based on this initial model, a CFA was conducted as we identified the need to carry out more robust analyses.

In terms of the descriptive analysis, as shown in Table 2, we found that the subjects valued the factors family, nutrition, personality traits, and career more highly, whereas the subjects surveyed by Silva et al. (2014) valued the factors alcohol, personality, introspection, and sleep/stress more highly. This difference can be explained by the fact that the population in the present study is made up of both students and older people while the preliminary validation sample consisted of only university students. During the analysis, most of the correlations between the factors were significant, with the strongest correlations observed being between family and alcohol (−0.91) as well as introspection and career (0.85). Such correlations were not calculated in the preliminary validation (Silva et al., 2014), and only correlations between the factors and overall lifestyle score were analyzed, which showed the strongest correlation between the introspection and personality factors. In the study by Darviri et al. (2014), all the correlations were positive and significant, suggesting that individuals who adopted healthier eating habits and followed an organized routine including PA also sought social and mental support. We also found that there were no convergent validity problems since the AVE from the factors was equal to or greater than 0.50 (Hair et al., 2019), indicating that the items converged strongly for the corresponding factors. The main objective of this work was achieved with the adjustment of the measurement model. However, we are unable to compare these results because the previous versions of the Fantastic Lifestyle questionnaire did not report analyses of this nature, as mentioned above. Thus, the present study contributes to the dissemination of knowledge regarding lifestyle and manages to empirically demonstrate the factor structure of this questionnaire, thus addressing an existing gap in literature. In short, the present study shows that this modified questionnaire can be used to generally assess the lifestyle of the Portuguese population. Although the preliminary validation showed some gaps, the more robust analyses that were carried out later showed that the instrument in its current final version reveals good psychometric qualities in terms of internal consistency, convergent and discriminant validities, as well as adequate fit of the measurement model to the data.

Although this study adds to knowledge in this area and helps to fill the gaps identified previously, it has some limitations. First, the sample cohort is composed of only students and people with T2D, so the results cannot be generalized to other populations; this sample was obtained by convenience and comprises two different populations, which can influence the identified factorial structure when considering that different samples may interpret the items differently. Therefore, future studies should use the current instrument and validate it for other segments of the population. Second, considering the sample heterogeneity, it was not possible to carry out an analysis of invariance owing to the different characteristics of the samples under study; in this sense, future studies should try and analyze the invariance of the measurement model based on gender, age, and other sociodemographic characteristics. Third, considering the cross-sectional nature of this study, future studies would need to carry out experimental and/or longitudinal analyses, e.g., through longitudinal invariance.

Based on the results obtained through EFA and CFA, we show that the present version of the Fantastic Lifestyle questionnaire (nine factor and 27 items) is a valid and reliable tool for assessing the lifestyle factors among the Portuguese population.

A preliminary analysis of the data revealed that there were no missing values. The sample size was then calculated using the G*Power 3.1 software (Faul et al., 2009), following the recommendations of Denis (2019), and included the following parameters: anticipated effect size of f2 = 0.6, α = 0.05, statistical power = 0.95, and a minimum of 31 subjects. All of these criteria were respected in this study. The exercise program was paused for 1 week each during the Christmas and Easter seasons, and adherence was monitored using a presence register. There was no participant dropout, and the average adherence in each session was 70%. Based on the results shown in Table 4, from a descriptive point of view, the mean values increased in the case of the positive affect, SWL, PA, and nutrition while decreasing for negative affect. From an inferential point of view, the results showed significant differences in all factors from the initial to final sessions (p < 0.05) with large effects, revealing that the PE program was important for both SWB and certain lifestyle factors like PA and nutrition.

In study II, we aimed to implement and assess the effects of the Diabetes em Movimento ^® program on SWB, daily PA, and nutritional habits in Portuguese subjects with T2D. The results indicate an increase in the average values from the initial to final phases of the program, with significant differences in all parameters except the values of the negative affect, which decreased considerably. This means that the people involved in the Diabetes em Movimento ^® program reported improved positive indicators of SWB, PA, and nutritional habits as well as diminished negative affect. These results are very encouraging and highlight PE as a mechanism for improving healthy habits and SWB in this population. Effectively, the combination of PEs developed herein were associated with physical benefits such as fat loss (especially visceral fat), improved lean muscle mass, improved metabolic efficiency, better physical fitness, and reduced risk of cardiovascular diseases (Church et al., 2010; Schwingshackl et al., 2014). The program can also contribute to social support and interaction development since such exercise programs are influenced by social elements (e.g., group class, coach, and teacher) that often impact loneliness and isolation (Ahn et al., 2024). The combined exercise program also has great influences on improving anxiety and depression, social dysfunction, loss of confidence, and general psychological wellbeing among patients with T2D (Gebeyehu et al., 2022).

According to Diener et al. (1985), global life satisfaction is a cognitive judgment of the quality of one’s own life. Hence, the benefits of PE can also extend to improved quality of life and consequently SWL. Effectively, exercise and SWL are positively linked since physically active people have higher levels of SWL than physically inactive people (Iwon et al., 2021). Otherwise, the affective dimension of SWB reflects the basic experience within events occurring in one’s life. There are two dimensions to the affective components of SWB, namely positive and negative affect. The positive affect or pleasant emotions are part of SWB because these emotions reflect individual reactions to events that show that life goes as desired (Diener et al., 1985), and more daily PA is associated with higher positive affect (Ludwig and Rauch, 2018). Exercise is also linked to the release of endorphins that act as painkillers and mood elevators to strengthen the feelings of wellbeing (Boecker et al., 2008). Exercise is known to help improve insulin sensitivity and glucose control, which can reduce the psychological stressors associated with T2D (Colberg et al., 2016). Moreover, exercise programs that include social interactions and peer support have been shown to result in higher levels of emotional wellbeing and positive affect in T2D subjects compared to those exercising alone (Paul et al., 2007). Since the T2D subjects reported more PA and healthy nutritional habits after intervention, the findings corroborate the results of previous studies (Colberg et al., 2016; Wing and Phelan, 2005). Beyond improving physical fitness that would allow people with T2D to be more physically active during their daily lives, the structured exercise programs provide an environment that encourages regular PA, making it easier to develop and maintain healthy habits (WHO, 2024a). Furthermore, being a part of a group can offer social support and motivation, which are essential for sustaining long-term behavioral changes (CDC, 2024). People with T2D who were involved in a positive feedback loop with nutrition emphasized that the combination of regular exercise and improved dietary habits created a synergistic effect, leading to sustained improvements in both PA levels and nutrition, especially in individuals with severe metabolic conditions (Dalle Grave et al., 2010).

Despite the encouraging results, the generalization of the results should be carefully considered given the main limitations of this study. The sample size limits generalizability and also makes it difficult to extend the results to larger populations with different sociodemographic characteristics. The present study design does not allow us to assess the results through a follow-up period, which is essential for evaluating the long-term effects of the observed results. A comparative analysis with a control group that is not involved in the same intervention is therefore essential to guarantee that the results are exclusively linked to the Diabetes em Movimento ^® program. This methodological limitation also restricts internal validity and compromises the capacity to securely affirm that the observed effects are derived exclusively from the intervention. Therefore, regarding the future directions of this work, we consider it essential that the study be developed with access to a larger sample group along with a follow-up period as well as the presence of a control group through a randomized process to strengthen the evidence.

The main findings of study II suggest that people involved in the Diabetes em Movimento ^® program reported improved SWB and key lifestyle factors through significant increases in the positive affect, SWL, PA, and nutrition habits along with a notable decrease in the negative affect between the baseline and post-intervention period. The present results support evidence confirming the positive effects of PE through the Diabetes em Movimento ^® program to foster SWB and promote healthier lifestyle behaviors among T2D subjects.

Based on the main findings of study I, we identified that the factorial structure of the Fantastic Lifestyle questionnaire (i.e., nine factors, 27 items) is a valid and reliable tool for assessing the lifestyle variables of the Portuguese population. Furthermore, we identified in study II that the implementation of the PE program Diabetes em Movimento ^® was beneficial for improving the positive affect, SWL, PA, sleep, stress, and nutritional behaviors significantly while notably diminishing the negative affect in our sample of people with T2D. As noted before, these results should be interpreted with some caution as our study did not have a control group and the sample was one of convenience. Although the present work can be classified as a pretest/posttest study without a control group, which would allow us to compare the measurements before and after the intervention, the lack of a control group and absence of sample randomization necessitates a somewhat cautious claim that the changes observed herein are exclusively attributable to the intervention. Nevertheless, these results can provide insights into the effectiveness of an intervention even though they are limited in terms of internal validity. We should therefore conduct further studies in the future to consolidate the results achieved in the present study, particularly by including control groups or applying sample randomization methods.