According to data from the Secretary of Health of Maringá, PR, 42,258 older people attended the Basic Health Units (UBS) of the city in 2016. Considering the population found, the initial sample to be considered was 595 older people, adding 10% of possible losses. The final sample consisted of 654 older people of both gender, considering a 95% confidence level and 4% of margin of error. The software used to obtain the calculations was the StatDisk version 8.4.

The UBSs that older people attended to were subdivided into four regions: East (7 UBS) which covers 21.8% of the population, northern region (8 UBS) with 34.5% of the population, western region (8 UBS) with 23.2% of the population, and south region (8 UBS) that covers 20.4% of the total older population of the city. Knowing the regions composition, three UBSs were selected by lot to be evaluated in each of the regions.

After defining the size of the sample in each region and selecting the UBSs, it was important to maintain the proportion of older people population in the sample, so the calculations to obtain the final sample by UBS according to gender were proportional to the population.

The study included older people of both gender, age 60 years and older, with preserved speech and hearing ability, which allowed the questionnaires to be applied. The Mini Mental State Examination (MMSE) was used to exclude older people with significant cognitive deficits. The MMSE consists of questions grouped into seven categories: temporal orientation, spatial orientation, three-word register, attention and calculation, memorization of the three words, language and visual-constructive capability.12 The cut-off score used by the MMSE for exclusion were: 17 for the illiterate; 22 for the older people with education between 1 and 4 years; 24 for those with schooling between 5 and 8 years; and 26 for those with 9 years or more of schooling. These cut-off scores were based on the criteria of Brucki et al.13 They correspond to the average obtained by these authors for each schooling range, minus one standard deviation. Older people were excluded if classified below the cut-off score specific to their schooling.

To characterize older people's sociodemographic profile, a semi-structured questionnaire was used consisting of information regarding age (60–69 years, 70–79 years, 80–90 years), gender (male, female), marital status (married or living with a partner, single, divorced, widower), race (white, black, other), the occupational situation (working or not working for own income), monthly income with minimum wage (MW) in 2016 Census of Brazilian Institute of Geography and Statistics (IGBE) as reference (1–2MW, 2.1–3MW, more than 3MW), retirement (yes, no), schooling (did not study, incomplete elementary school, complete elementary education, complete high school, complete higher education).

The functionality has been evaluated by the WHODAS 2.0 instrument. This 12-item instrument covers six domains of disability within 30 days (last month). These domains are: Understanding and Communication (UC), Mobility (MB), Self-Care (SC), Interpersonal Relations (IR), Activity of Daily Living (ADL) and Social Participation (SP). The items are answered on a 5-point Likert scale, which varies from no difficulty (0 points) to the extreme self-reported difficulty (4 points).14

The older people's physical activity level was evaluated using the short version of the International Physical Activity Questionnaire (IPAQ), composed of seven open questions with information that allow to estimate the time spent per week in different dimensions of physical activity (moderate and vigorous physical and walking efforts) and physical inactivity (sitting position). Sedentary behavior was assessed by the last two IPAQ questions: (1) How much time, in total, do you spend sitting on a weekday? (2) How much time, in total, do you spend sitting on a weekend day? The answer is given in hours and minutes. The physical activity data were processed and cleaned according to the protocol indicated by the IPAQ.15,16

This is an analytical, cross-sectional and observational study approved by the Ethics Committee of Human by University Center of Maringá, through opinion number 1.777.797/2016. The data were collected in 12 UBS, of the 33 UBS of Maringá, divided in four regions (north, south, east and west), selected by lot, after authorization of the Permanent Committee for Formation and Training of Health Workers (CECAPS).

Before the beginning of the data collection, a team of 10 researchers was properly trained, and a pilot test was conducted with 30 older people. The volunteers were approached by the researcher in charge or by the research team. They were informed about the justification, objectives and procedures to be carried out, according to guidelines for research with human beings included in Resolution 196/96 of the National Health Council. After these procedures, those who agreed to participate in the research signed the Informed Consent Form (TCLE). The collection was carried out in different days, shifts and schedules, according to the availability of the researchers.

The direct interview was chosen in the application of the questionnaires, due to the possible difficulty of reading, visual problems and comprehension.

From the 654 patients evaluated, prevailed women (56.0%), married (61.3%), aged between 60 and 69 years (59.2%), monthly income between one and two minimum wages (70.0%), Caucasian (81.0%) and retired (75.0%). It was also observed that most of them had incomplete primary education (43.0%).

Table 1 shows the descriptive values of each study variable, as well as the correlation values between the variables. Sedentary behavior had the following correlations with the functionality: sitting time during the week with functionality in cognition (r=0.15), mobility (r=0.17), self-care (r=0.17), interpersonal relations (r=0.21), and activities of daily living (r=0.29); sitting time at the weekend with minutes of moderate activity a day (r=−0.14), and a week (r=−0.14), with days of vigorous activity (r=−0.22), minutes of vigorous activity a day (r=−0.24) and a week (r=−0.23), with the functionality in cognition (r=0.16), mobility (r=0.16), self-care (r=0.18), interpersonal relations (r=0.21), activities of daily living (r=0.31), and social participation (r=0.14).

The minutes of walking a day are related with self-care functionality (r=−0.10), interpersonal relations (r=−0.10), activities of daily living (r=−0.10), and social participation. Minutes of walk per week with mobility functionality (r=−0.13) and self-care (r=−0.10). Days of moderate activity per week with functionality in interpersonal relations (r=0.10) and in social participation (r=0.10). Minutes of moderate activity per day with functionality in interpersonal relations (r=0.10), activities of in daily living (r=0.10) and in social participation (r=0.10). Minutes of moderate activity per week with functionality in mobility (r=−0.10), in interpersonal relations (r=0.10) and in social participation (r=−0.13). Finally, the days of vigorous activity per week, minutes of vigorous activity per day and per week related with functionality in mobility (r=−0.14/r=−0.13/r=−0.13, respectively).

Initially, the measurement models of the five latent variables (Step 1 of the SEM) were tested, verifying the relationships of the observed items in their respective factors by CFA. All sub-models presented adequate adjustment indices and the quality of the local adjustment. The internal reliability of the items were also confirmed, since all trajectories presented significant factorial loads (p<0.05), greater than 0.50. These results confirm the latent models and allow the test of the structural equations model (Step 2).

A model was tested for trainers without SB mediation, with direct trajectories between practice of physical activity and functionality (M1). This model did not present adequate adjustment indicators [X2(87)=1161.66; p=0.001; X2/gl=13.352; CFI=0.084; GFI=0.83; AGFI=0.77; NFI=0.83; TLI=0.80; RMSEA=0.14; P (rmsea<0.05)=0.0.001, explaining only 2% of the functionality variance (Fig. 1). In addition, the trajectories of light (β=−0.03) and moderate (β=0.03) activities for the functionality were not significant (p>0.05). Therefore, these two latent variables were deleted, and the model was tested again.

Figure 1.

Structural model (M1) of the effect of physical activity practice on the functionally of older people.

(0.19MB).

The second model (M2) without the latent variables of light and moderate activities presented an acceptable fit [X2(26)=118.46; p=0.001; X2/gl=3.556; CFI=0.97; GFI=0.96; AGFI=0.94; NFI=0.96; TLI=0.95; RMSEA=0.07; P (rmsea<0.05)=0.002] but explaining only 1% of the functionality variance (Fig. 2). The trajectory of the vigorous activities had a significant (p=0.028) and weak (β=−0.11) effect, indicating that each increase of 1 standard deviation in the vigorous activity unit had a reduction of 0.11 standard deviation in the unit of functionality. This finding indicates that the practice of vigorous activities is associated with a poor improvement in functionality.

Figure 2.

Structural model (M1) of the effect of vigorous physical activities practice on the functionary of older people.

(0.11MB).

In the third model (M3), SB was inserted as a mediator variable of the effect of vigorous physical activities practice on functionality. The M3 represents several hypothetical relations of association between the variables, and these magnitudes are described by the direct and indirect effects. The model (M3) provided acceptable adjustment values [X2(38)=165.54; p=0.001; X2/gl=3.36; CFI=0.96; GFI=0.96; AGFI=0.92; NFI=0.96; TLI=0.95; RMSEA=0.07; P (rmsea<0.05)=0.001] (Fig. 3).

Figure 3.

Structural model (M3) of the effect of vigorous physical activities practice on the functionally mediated by sedentary behavior.

(0.15MB).

The latent functionality variable is explained in only 3% by the structural model (Vigorous Activities+Sedentary Behavior), while RTA is explained in only 1% by Vigorous Activities (Fig. 3). In the direct relationship established between Vigorous Activities and Functionality, the effect is significantly positive and weak (β=−0.10, p<0.05) (for each increase of 1 standard deviation of the Vigorous Activities Practice, Functionality decreases by 0.10 standard deviations), that is, the greater the practice of Vigorous Activities, the better the functionality. Regarding sedentary behavior, the direct effect on the Functionality was weak and significant (β=0.10, p<0.05) (for each increase of 1 standard deviation of the sedentary behavior, the Functionality increases by 0.10 standard deviations), that is, the longer the sedentary behavior, the worse the functionality. In addition, vigorous activities had a significant (p<0.05) and weak (β=−0.11) direct effect on sedentary behavior (Table 2).

When mediated by sedentary behavior, the relationship between Vigorous Activities and Functionality showed a weak reduction in the total effect (−0.10 to −0.09), with an indirect effect of β=0.01 (Table 2). Thus, it is evident that although sedentary behavior acts to reduce the effect of vigorous activities on functionality, this moderating/damaging effect is very weak, indicating that if older people practice vigorous activities their functionality will have a slight improvement, regardless of the time of sedentary behavior.