Validity and reliability of the Mexican resilience measurement scale in families of children with chronic conditions

The resilience to face disease is a process of positive adaptation despite the loss of health. It involves developing vitality and skills to overcome the negative effects of adversity, risks, and vulnerability caused by the disease. Resilience has been defined in multiple ways [1‐5], although one aspect common to all of them is the ability to adapt and achieve optimal functioning in the face of unfavorable conditions that pose risks and threaten one’s integrity. Several studies have shown that overcoming a significant loss or a potentially traumatic event results in increased resilience and therefore greater adaptability to adverse situations [6, 7]. Given the implications for the prevention of mental health problems [8, 9] and promoting human development [10], this construct has been an object of theoretical and empirical research in recent decades [11‐13], particularly in positive psychology [14] and psychometry [15].

In the field of psychometry, several instruments have been developed to measure resilience [16‐18]. One of these is the 25-item Connor-Davidson Resilience Scale (CD-RISC) [19], which is among the instruments with the most robust psychometric properties [12, 20]. This scale in its original version or in its short version with 10 items (CD-RISC10) [21] has been validated for use in populations of eastern [9, 14, 22] and western non-English speaking societies [23‐25]. For the Mexican population, the CD-RISC [19], in addition to the Resilience Scale for Adults (RSA) [17], was the basis for the development of the Mexican Scale of Resilience Measurement (RESI-M) [26]. The RESI-M was validated for use in a general population of students and employees, of both sexes and with an age range of 18 to 25 years. It consists of a Likert scale instrument composed of five factors: strength and self-confidence, social competence, family support, social support, and structure, each with high internal consistency values (from .79 to .92). The overall internal consistency was also high (alpha = .93), and the five factors extracted accounted for 43.60% of the total variance. Sociodemographic differences were reported in the social support factor, with women and single participants scoring higher, and for strength and self-confidence, with married men scoring highest. Other research has confirmed the psychometric properties of RESI-M and the stability of its factors in a subpopulation determined by the geographic area of the country [27]. In subsequent studies, RESI-M was used to evaluate resilience in primary caregivers with normal or complicated grief, the latter characterizing those showing less possession of this ability [28], and to demonstrate resilience as an important variable in the grieving process of women with cancer [29].

Although there is evidence with respect to the adequate psychometric properties of RESI-M [26, 27], it must be considered that this scale was originally validated for use in the general population, so its use in other populations compromises the validity of its results. There is evidence that a validated scale with a general population presents a different factor structure when validated with a population of caregivers [30], results that undermine those obtained by Ornelas-Tavares [28]. In the same way, a scale validated with a general population may produce inaccurate data when applied to a clinical population [15], as the study of Miaja and Moral [29] shows. This situation can be resolved by making validations ex professo for the population to be investigated. For example, the CD-RISC10, validated in Spanish with university students [31], was later validated with patients with fibromyalgia [32] to guarantee the validity of the results in each population.

In Mexico, to date, no studies have validated the RESI-M or some other scale used to measure resilience [33, 34] in the caregiver population. This is despite the fact that the activities they carry out make them a vulnerable population [35], prone to presenting symptoms of stress, anxiety and depression [36‐38] as well as experiencing a decrease in their quality of life [39, 40].

Pediatric chronic diseases represent one of the greatest challenges for family environments, with physical, psychological, socioeconomic, and behavioral effects on patients and their caregivers [41]. In addition, in recent years, the prevalence in Mexico of chronic diseases, particularly cancer [42], in children under 18 years of age has reportedly increased, which in turn has led to an increase in the number of caregivers for this population. These caregivers, usually the parents of the child patient [35], are prone to high levels of stress [43].

Given the importance of resilience for coping, focusing on problems rather than emotions [44] and on the adaptation of family members to the exigencies of chronic disease enables such individuals to face it and transcend it [45, 46]. The present study aimed to validate the five-factor model for the RESI-M in a population of family caregivers of hospitalized children with chronic diseases; calculate the internal consistency of the scale and its five factors; and compare mean scores between both sexes. The hypothesis was that the model of five correlated factors proposed by Palomar and Gomez [26] showed a good or adequate fit with the data in this population of caregivers. The following results were expected: high internal consistency value for the scale; high values for its factors, except an acceptable value for the structure factor [26]; and statistically equivalent averages in the RESI-M total score and its five factors between both sexes, except in the social support factor, in which women could have a higher average than men [26].

Table 1 describes the sociodemographic characteristics of the family caregivers and children with chronic diseases in the initial sample.

The robust maximum likelihood estimation method was used due to the lack of multivariate normality of the data (Mardia’s coefficient = 802.48). All parameters were significant, with measurement weights greater than .50 (Table 2). The arithmetic mean of the squared measure weights or average variance extracted (AVE) of the strength and self-confidence factor (items 1 to 19) was 47.2%, and its internal consistency was very high (α = .93). The AVE of the social competence factor (items 20 to 27) was 46.8%, and its internal consistency was high (α = .87). The AVE of the family support factor (items 28 to 33) was 55.7% and showed a high internal consistency (α = .89). The AVE of the social support factor (items 34 to 38) was 63.6%, and its internal consistency was high (α = .90). The AVE of the structure factor (items 39 to 43) was 40.5% and had an acceptable internal consistency (α = .76). The overall internal consistency was very high (alpha = .95).

The correlations between the five factors were significant and ranged from moderate (r = .41, p < .001) to very high (r = .74, p < .001) (Table 3). The mean shared variance between pairs of factors was 31.5%. Although goodness of fit was rejected by a chi-square test (_S-Bχ² _{[840, N = 446]} = 1397.91, p < .001), the other three indexes showed a good fit to the data: χ²/df = 1.66, CFI = .95, and .03 RMSEA = (IC 90%: .02, .04).

Finally, t-tests did not reveal gender differences in the overall scale score or in the factors (p > .05), as shown in Table 4.

The general objective of this study was to validate the RESI-M in a population of family caregivers of children hospitalized with chronic diseases. The initial hypothesis was that the factorial structure in this population would be identical to that detected in the general population in the original validation of that instrument [26]. The data obtained confirmed this hypothesis. The present factor analysis, using the RESI-M, indicates a structure composed of five factors: Strength and Self-Confidence, Social Competence, Family Support, Social Support, and Structure. Overall, the factors showed good internal consistency, with Strength and Self-Confidence being very high; Social Competence, Family Support and Social Support high; and Structure acceptable. These findings are similar to those of Palomar and Gómez [26] and Camacho-Valadez [27] with respect to the dimensions used to assess the construct of resilience in the Mexican general population.

These data indicate that general psychometric properties of the RESI-M found in the general population are also found in a specific subpopulation facing conditions of vulnerability and in which greater resilience must be developed—specifically, a population of caregivers of children suffering from chronic illness. In addition, our study indicates the relevance of using RESI-M over a broader age range (from 18 to 50 years) relative to that employed in the original validation that ranged from 18 to 25 years. Unlike the data obtained by Palomar and Gómez [26], in whose study the women’s social support factor average was higher than the men’s one, our data do not indicate differences between the sexes in the average of any resilience factor. This may be due to the particular conditions to which family caregivers are exposed.

A limitation that warrants attention is the use of non-probability sampling. Thus, the results do not represent estimates of population parameters, an issue that must be addressed in future studies. The sample size was 34 participants below the minimum number of participants, based on the rule of including at least five participants per parameter to be estimated [51]; nevertheless, the number was greater than 400. This number is considered a good sample size when the number of indicators per factor is high (6 or more), measurement weights are high (greater than .50), and few factors exist (five or fewer) [51, 52]. These criteria were satisfied in the case of the five-factor model for the RESI-M tested in the present study. Therefore, the sample size should be considered sufficient for conducting a one-group analysis confirmatory analysis. Among the strengths of this study is the use of robust methods in the confirmatory factorial analysis due to non-compliance with the assumption of multivariate normality.

This study demonstrated that the model of five correlated factors (strength and self-confidence, social competence, family support, social support, and structure), originally developed for the 43 items constituting the RESI among young adults from the general population, is valid among families of hospitalized children suffering from chronic diseases. In addition, the internal consistency values of the RESI-M and its five factors in this population varied from acceptable to very high and do not require the removal of any item for their improvement. Given that no difference in the averages between women and men was observed, the scale and its factors do not require a different standardization for each sex. Thus, the RESI-M can be considered a valid and reliable instrument for measuring and assessing resilience in family caregivers of children with chronic diseases. These results may contribute to research in the field of family resilience in the context of pediatric disease and to the development and implementation of intervention programs for improving the quality of life and well-being of family caregivers.