Study design and context
This is a cross-sectional study recruited from health-care professionals working with pediatric and adult populations in public, private, and philanthropic hospital institutions in the states of Espírito Santo, Ceará, Pernambuco, Alagoas, Piauí, Bahia, Acre, Minas Gerais, Rio de Janeiro, São Paulo, and the Rio Grande do Sul from October 2018 to April 2019.
Health professionals participated in the study (physicians, nurses, and physiotherapists), working in direct care to patients and who have at least 6-months experience in the service. The inclusion criterion of the sample was those signing the free and informed consent form for participation in the research by sending the invitation or in person. The professionals answered the questionnaires to collect the sociodemographic variables and later to the other instruments of this research.
Sociodemographic variables were obtained through structured interviews and included age (years), sex (male or female), and professional designations (title, number of works, work experience time, typical work shift, and state of origin).
Regret Coping Scale for Healthcare Professionals (RCS-HCP)
The daily strategies for coping with repentance related to the care used by health professionals were measured using the RCS-HCP scale. The scale consists of 15 items that measure strategies for coping with regret that can be focused on the problem or adaptive or maladaptive emotions. The questions assess the change in patient care practices, usually performed after events of regret. Answer options range from 1 (never or almost never) to 4 (always or almost always). The original scale features three domains. The first domain focused on the problem, and the other focused on adaptive emotion and maladaptive emotion(12). Although the best strategy is situation-specific, and no single strategy can be described as generally better than others, some strategies in domains focused on maladaptive emotion are more associated with negative results(14). The estimate of the latent trait ‘dealing with regret’ is obtained through the Average Score; the total score of the instrument is not calculated since it is believed that there are three types of coping strategies and not a global strategy.
Self-Reporting Questionnaire – SRQ-20
The SRQ-20 scale was validated in Brazil(15,16) and assessed the prevalence of Common Mental Disorders (depressive, anxious, and somatic complaints). This instrument has 20 questions, and the final score can range from zero (null probability) to 20 (high probability) of common mental disorders(15,16).
Life Satisfaction Scale
The Life Satisfaction Scale comprises five items answered using a 7-point Likert scale, with 1 = totally disagree, 2 = disagree, 3 = disagree slightly, 4 = neither agree nor disagree, 5 = agree slightly, 6 = agree, and 7 = totally agree(17),(18). The total score can range from 5 points (extremely dissatisfied) to 35 points (extremely satisfied)(18).
The validation process was composed of two phases, and the methodology adopted for the translation of the scale followed international standards(19,20).
Translation of the RIS-10 encompassed the following steps: (i) translation by two German-Brazilian Portuguese translators; (ii) harmonization between both Portuguese versions, resulting in a single version in Portuguese; (iii) back-translation of the harmonized version by two Brazilian Portuguese-German translators; (iv) harmonization between both translators, resulting in a single German version; and (v) general harmonization, where the versions resulting from the first and second harmonization were discussed by the four translators to obtain a consensus version(20).
We also translated the RIS-10 from French into Portuguese by two translators and harmonized these translations to assess the differences between the translated versions of German and French. Given that no differences were found between these translations, we adopted the German-to-Portuguese translation as the official translation.
Evaluation of psychometric properties.
Phase II comprised the evaluation of the psychometric properties of validity (content, construct, and criterion) and reliability through field testing.
After the scale was translated, the process of cultural adaptation began. For this, this version of the scale was evaluated in relation to content by judges with clinical experience in the studied latent trait. Six judges who have been working in the health care area for more than 5 years participated from each of the following areas: two physicians, two nurses, one psychologist, and one physiotherapist.
First, the evaluation was done qualitatively to obtain the possible suggestions for a better cultural adaptation of the translated terms. The level of agreement among the judges regarding the relevance and representativeness of the items was evaluated by the Content Validity Index (CVI). A 4-point Likert scale was used, where: 1 = not relevant; 2 = item needs a large revision to be representative (not relevant); 3 = quite clear, but needs a small review (very relevant); and 4 = quite clear and representative (highly relevant)(19).
This index is calculated by the sum of the 3- and 4-point answers divided by the total number of judges, yielding a proportion of judges who deemed the item valid. However, 1- and 2-point answers required revision or elimination. To calculate the general CVI of the instrument, the sum of all CVI calculated separately was performed, divided by the number of items(19). A CVI exceeding 0.78 is considered an acceptable agreement rate when six judges participate, which was the case in our study(19,21). The scale’s content was evaluated through a pilot study of ten professionals, six nurses, three physicians, and one physiotherapist.
Construct validity testing was performed with exploratory and confirmatory factor analysis. Exploratory factor analysis was performed with the Promax rotation method and used the Kaiser measure to assess the adequacy of the sample to a latent factorial structure. The evaluation of the adequacy of a latent factorial structure to the data was measured using the Kaiser-Meyer-Olkin (KMO) test(22). KMO values exceeding 0.5 were considered adequate (22).
Confirmatory factor analysis (CFA) verified the factorial structure suggested in the original scale with three factors using the structural equation model(12). The adjustment and quality of the sample of this study to the factorial structure were examined using the following: χ2 (chi-square model), the goodness of fit index (GFI), root mean square error of approximation (RMSEA), standardized root mean squared residual (SRMR), normed fit index (NFI), comparative fit index (CFI), Tucker-Lewis index (TLI), and Bollen’s incremental fit index (IFI). The cut-off points considered acceptable for scale adjustment were: χ2: p > 0.05, GFI > 0.90; RMSEA < 0.08, SRMSR < 0.10, NFI ≥ 0.90, CFI > 0.90, TLI > 0.95, and IFI > 0.90(23–26).
For criterion validity, the total score of the RCS-HCP scale was correlated with the questionnaires validated in Brazil, namely, the SRQ-20 (15,16) and the Life Satisfaction Scale(17). We hoped that some emotion-centered strategies would be more often associated with negative outcomes, such as a higher prevalence of the common mental disorder, and secondly, that problem-centered, emotion-centered strategies would be associated with greater satisfaction with life. Correlations were evaluated using Spearman’s rho (ρ), and values of r ˃ 0.3 were considered acceptable(27).
The reliability measures of internal consistency, floor and ceiling effects, test-retest, and Spearman-Brown coefficient were used(27). Cronbach’s α was used for internal consistency (28). The floor and ceiling effects were evaluated by determining the lowest and highest percentage of the population in the application of the scale(29,30). The Spearman-Brown coefficient(27) was analyzed by the split method, as detailed in the following strategies. First, the items were randomly divided into two equal halves. A scale mean was computed for each half, and then the two sets of scale means were correlated to estimate a split-half correlation. The split-half correlation was adjusted by the Spearman-Brown formula to create a split-half reliability(31,32). Test-retest reliability was analyzed using the intraclass correlation and Bland-Altman plots (33). Data collection for test-retest analysis was performed within a maximum period of 30 days.
Interpretations of the reliability test items were as follows: Cronbach’s α was ≥ 0.7, as recommended33, the criterion considered to floor and ceiling effect was > 20%(29,34), the intraclass correlation (CIC) was considered acceptable when ≥ 0.7(35) and the Spearman-Brown coefficient was > 0.3(27). The data were analyzed using the statistical software SAS v.9.4 and the Lavann package v.0.6-5 of R. This study uses a p of 0.05 as the statistical threshold of significance.
Calculation of the sample size was based on the psychometric properties evaluated and aimed for a ratio of 15:1 (15 respondents for 1 item of the instrument)(36). Since the scale contains a total of 15 items, 150 participants would be needed. A total of 341 professionals participated in this study.
This study was approved by the ethics committee of the Pontifícia Universidade Católica do Rio Grande do Sul – PUC/RS (CAAE: 2.462.827/2018). All participants signed an informed consent form prior to the study.