Validity and reliability of the Brazilian version of the Patient Dignity Inventory (PDI – Br)*

Objective: to perform the psychometric validation of the Brazilian version of the Patient Dignity Inventory (PDI – Br) in patients with advanced diseases in palliative care. Method: a methodological study to verify the psychometric properties of the Patient Dignity Inventory (PDI – Br) instrument, through validity and reliability tests. Results: the exploratory factor analysis showed a factorial solution with three factors, responsible for 40.9% of the explained variance, with adequate internal consistency for the Presence of Symptoms (α=0.859), Dependence (α=0.871), and Existential Suffering (α=0.759) domains. The test-retest was performed and indicated moderate to strong correlations. Convergent validity demonstrated a positive correlation between the Presence of Symptoms and the sadness (r=0.443) and anxiety (r=0.464) variables. Weak negative correlations were observed between the PDI – Br domains and functionality, spiritual well-being and quality of life. Conclusion: composed of three domains and 25 items, the PDI – Br instrument presented satisfactory psychometric properties for its use in our environment, through the evidence of validity and reliability.


Introduction
The progressive aging of the population shows that people are living longer, but with a high prevalence of chronic diseases, neoplasms or dementia (1)(2) ; consequently, their quality of life decreases.
According to the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística, IBGE, in Portuguese), in 2050 the population over 60 will be 2 billion people and 25% of them will be over 65 (3) . Neoplasms also represent a public health problem and Brazil is expected to have 625,000 new cases per year from 2020 to 2022 (4) .
However, the increase in life expectancy has not been accompanied by a better quality of life in the phenomenon of aging and illness. Technological advances and the variety of therapies available lead to a constant search for the cure of diseases, leaving interventions that focus on the dignified end of life in a second place (5) .
Differentiating itself from curative medicine, the palliative care approach is a type of assistance that proposes multifaceted care, with the management of physical, social, emotional, and spiritual symptoms for patients facing advanced life-threatening diseases (5) .
In the context of palliative care, patients can experience situations that affect their perception or sense of dignity, which can be defined as a value, from which the person is perceived by the world and by himself/herself as a valuable and respectful human being who keeps its essence intact, even when facing the physical degradation caused by different circumstances (6) .
The literature shows that the sensation of loss of dignity has been associated with an absence of the will to live, indicating a strong connection with depression, lack of hope, and expression of an interest in anticipating death (7) . Thus, the investigation of the sense or perception of dignity in the scope of palliative care has been gaining prominence in the health area.
A Brazilian study investigated the concept of dignity of patients in palliative care and showed that both health professionals and caregivers can influence self-perception of dignity. Being a "correct" person, maintaining autonomy and being cared for with respect were the elements that positively influenced the perception of dignity; on the other hand, urban violence and the lack of public accessibility policies had negative influences (8) .
Physician Harvey Max Chochinov from Canada is one of the main researchers in this area and proposed the Dignity Model, with the purpose of establishing the association between dignity and psychosocial factors in patients in an advanced stage of an incurable disease (9)(10) . For Chochinov, maintaining dignity allows patients in palliative care to continue performing their usual roles (9)(10) .
Based on the Theoretical Dignity Model, Chochinov proposed an instrument composed of 25 items distributed in five domains, called the Patient Dignity Inventory (PDI), in order to identify the problems associated with the loss of dignity (9)(10) .
The Dignity Inventory proposed by Chochinov was translated, adapted and validated in Germany, Spain, Italy, Portugal and Greece (11)(12)(13)(14)(15) . So far, in Brazil there is still no single questionnaire that identifies the problems related to dignity for patients in palliative care, but there is a growing demand for this type of care.
Investigating the concept of dignity in patients in palliative care can contribute to direct the focus of care, in addition to allowing for the evaluation of interventions with the potential to improve the sense of dignity of these patients (2) .   (18) .
The Brazilian version of the Functional Assessment of Chronic Illness Therapy-Spiritual Well-Being (FACITsp 12) scale was used to assess spiritual well-being by means of two subscales: "meaning/peace" and "faith", in which the higher score, the higher the patients› spiritual well-being (19) .
The Karnofsky Performance Scale (KPS) was used to assess the functional capacity of the patients, with a score of 100 to 0, with 100 being preserved functional capacity and 0 representing a patient in the process of death (20) , in addition to the translated and adapted version of the Patient Dignity Inventory (PDI -Br).
Construct validity was performed by exploratory factor analysis, which verifies correlations between several variables, grouping them into a set of common latent dimensions, the factors, domains or dimensions (21)(22)(23)(24) . For the adequacy of the data in the factor analysis, the Kaiser-Meyer-Olkin (KMO) coefficient was used, for which the literature indicates By means of concurrent validity, the accuracy of an instrument is verified, contrasting it with a gold standard or an external criterion. This type of verification is divided into two types: convergent (when there is a correlation with the criterion) and divergent (when there is no correlation with the criterion) (25) . Concurrent validity was analyzed using Pearson's correlation coefficient, a numerical measure that verifies the relationship between two variables. It ranges from 0 to 1, either positive or negative and the closer to 1, the stronger the correlation (24) .
In Brazil, until now, there is no model of instrument that quantifies self-perception of dignity.
Based on the PDI validation carried out in Germany, Italy, Canada, and Spain (11)(12)(13)(14) and according to the literature recommendation, other scales were applied to the patients in association with the Patient Dignity Inventory (PDI), for the analysis of the correspondence of the phenomenon of dignity with external criteria.
For the test of association between dignity and physical symptoms, the Edmonton Symptom Assessment System (ESAS-Br) (17) and the Hospital Anxiety and Depression Scale (HADS) (16) were used.
The established assumption was that the loss of dignity has a positive correlation with depression, anxiety, and physical symptoms.
The assumption was also made that the loss of dignity is negatively correlated with functional capacity, spirituality, and quality of life. In order to assess the association between loss of dignity and decreased functional capacity, the Karnofsky Performance Scale (20) , the FACIT-sp 12 scale (19) and the two items on quality of life of the QLQ-C30 were used (18) .
indicating whether the items on the scale measure the same characteristic and, when they do, they are inclined to demonstrate a good correlation with each other (24) . The analysis of the α value also contributes to the conformation of the items, to the resolution of sustaining their permanence or the removal (14) .
The conjecture of reliability by the test-retest is based on establishing the association of the scores achieved on the same instrument by the same people at two different times and the expectation is that the scores achieved show an association (23)(24) . In the study, the PDI -Br was used for patients in the consultation following the first approach.

Results
In total, 135 patients were analyzed, followed-up at the ICESP (61.5%) and at the HC-FMUSP (38.5%) and were approached on average in the 4 th consultation; their mean age was 65 years old, with a mean schooling of 5.9 years, and most of them were retired (68.2%).
Neoplasms were predominant, accounting for 68.2% of the presentations, followed by diseases of the respiratory system (11.8%), cardiovascular diseases (6.7%), and neurological diseases (4.4%), as shown in Table 1.  (Table 2).  in the three domains detected in the exploratory factor analysis were identified for the reliability of the instrument, described in Table 3.   (Table 3).
The correlation analysis of the retest for items 2, 5,  (Table 5).

Suffering (SS) domains, weak and very weak correlations
were found with the scales that measure symptoms.
The evaluation of the hypothesis of negative correlation between the PDI -Br domains and the KPS, FACIT-sp 12 and EORTC-QLQ-C30 instruments showed negative and weak associations, as detailed in Table 5.

Discussion
For psychometric validation studies, the literature suggests samplings over 50 individuals, recommending at least 100 people; these recommendations are necessary to guarantee more solid conclusions, from the mean of 5 or more observations per item (24) . 135 patients were interviewed in this study, which guaranteed a mean of 5.4 observations for each item of the PDI -Br.
Evidence from the literature also shows that values greater than 0.30 are admissible for factor loads (24) .
The correlation matrix of the exploratory factor analysis resulted in a three-factor solution for the PDI -Br; in this analysis, items 11, 16, and 17 showed factor loads lower than the recommended ones. However, the 25 items of the instrument were organized only once in each of the three domains.
An Italian study that examined the factorial structure of the PDI applied exploratory factor analysis and also found three factors, with an explained variance of 64.4% (26) , a value which is higher than the one observed in this study (40.9%). In the Italian research the three factors were named as follows: existential suffering, psychological suffering, and physical suffering (26) . In the present study, the three factors were named presence of symptoms, dependence and existential suffering.
The definition for the final organization of an instrument should not be based only on one criterion (24) .
According to the recommendations in the literature, the   (11) ), Germany (n=112 and α=0.96 (12) ), Spain (n=124 and α=0.89 (13) ), Italy (n=266 and α=0.96 (14) ) and Greece (n=120 and α=0.71 to 0.9 (15) ). were kept in the instrument, because they were highly related to a loss of sense of dignity, according to the Chochinov's Theory (11) . In the PDI -Br all the items were loaded into the data correlation matrix after exploratory factor analysis.
Evidence from the literature demonstrates that the reliability of the test-retest assessment can change due to approaches taken over very distant periods of time (22,24) . This factor may explain the lower correlation of this study, given the 31-day interval. and weak with depression (r=0.374, p<0.001) (11) .
A validation study developed in Greece also found moderate and strong positive correlations between some domains of the PDI and the symptoms of anxiety (r=0.44 to 0.71; p<0.005), and weak to moderate correlations with the symptoms of depression (r=0.31 to 0.57; p<0.005) (15) . An Italian study found weak to moderate positive correlations (r=0.33 to 0.55; p<0.001) between the PDI and symptoms of depression (26) .
Evidence from the literature corroborates to explain that the loss of dignity is directly related to worse levels of anxiety and depression. A recent systematic review study found that, by raising patients' sense of dignity, the anxiety and depression levels also improved significantly (28) . and Germany (11)(12) regarding the pain variable, for which the correlation was positive, weak, and significant.  (11,(13)(14) .
For the items on the EORT-QLQ-C30 scale, weak and significant correlations were found between the domains of the PDI -Br and the questions of general health and general health quality. In Germany, the same instrument to measure quality of life was used and the correlation found was significantly moderate and negative (r=-0.42, p<0.001) (12) .
A research study carried out in 2015 in Spain showed that the sense of dignity was heightened by the application of the dignity therapy, which also brought significant beneficial effects on spiritual wellbeing (p<0.001) and on quality of life (p=0.011) (13) . A previous study conducted in Canada in 2002 highlighted a directly proportional relationship between better quality of life rates and a greater sense of dignity (29) . This study has strengths and limitations, which can be transposed in new investigations. The great aspect highlighted as a strong point is the provision of a valid and reliable instrument to estimate the dignity of patients, especially those in palliative care. On the other hand, its limitation was the fact that the sample includes patients with low schooling and low socioeconomic status, which represents a portion of the population, but the results may be different in other population stratum.

Conclusion
The tests performed demonstrate evidence of validity and reliability of the PDI -Br instrument, composed of three domains and 25 items, confirming its psychometric properties for its use in our country. This instrument offers the health professionals the possibility to assess the perception of dignity of patients in palliative care, contributing to the study of this phenomenon in the national context.