DEVELOPMENT AND VALIDATION OF QUALITY OF LIFE SCALE FOR PEOPLE LIVING WITH HIV

Ostras, Rio de Janeiro, Brasil. ABSTRACT Objective: to develop and validate a scale to measure the quality of life of people living with HIV in Brazil. Method: methodological study conducted in a Brazilian care service specialized in sexually transmissible infections/AIDS between 2017 and 2019 addressing people living with HIV. The scale’s development and validation included exploratory factor analysis to describe its factor structure and psychometric properties, Multitrait-Multimethod analysis to verify its validity and Cronbach’s alpha for reliability. Floor and ceiling effects were described according to the responses’ frequency distribution. Results: a total of 460 people living with HIV participated. Most were men 276 (60.0%) aged 43 on average (SD=±12.4). The Exploratory Factor Analysis revealed four factors with 39.9% of explained variance. The total scale presented satisfactory reliability with a Cronbach’s alpha equal to 85.0%. Most items presented satisfactory convergent


INTRODUCTION
By the end of 2019, 38 million people lived with the Human Immunodeficiency Virus (HIV) worldwide 1 . HIV is considered a debilitating, transmissible chronic disease, which under immunosuppressive conditions, allows the emergence of opportunistic infections that worsen an individual's clinical conditions 2 .
Because it is a chronic disease, after the antiretroviral therapy (ART) is implemented, people living with HIV experience changes in their quality of life (QoL), possibly compromising aspects such as life satisfaction and HIV mastery, treatment adherence, in addition to increased concern with financial issues and health 3 . According to the World Health Organization, the concept of QoL is "an individual's perception of their position in life, in the context of the culture and value systems in which they live and in relation to their goals, expectations, standards, and concerns 4:1405 ." QoL is a multidimensional and subjective concept that includes parameters such as wellbeing, self-care with health and diet, and how satisfied people are with their living conditions 4 . The QoL construct among people living with HIV (PLHIV) is strongly linked to their health conditions 5 .
Tools intended to measure QoL are supposed to include efficacy and efficiency indicators and indicators measuring the impact of treatments on patients to support comparisons of procedures and dimension both costs and benefits. Therefore, the results found here will support the implementation of health actions intended to improve this population's QoL [3][4][5][6] .
Given a growing concern with QoL and a search for instruments to measure it, the World Health Organization conducted a multicenter project called The World Health Organization Quality of Life Project to develop a QoL measure composed of 100 items 4 , the WHOQOL-100.
A measure intended to address specific issues concerning HIV emerged from this generic instrument; that is, a specific module was developed to assess the QoL among PLHIV, the so-called "WHOQOL-HIV" 7 . The psychometric properties of this scale's Brazilian version were tested and showed satisfactory reliability and concurrent validity 8 .
Another instrument widely used in Brazil is HAT-QoL. It comprises 34 items distributed into nine domains addressing overall function, life satisfaction, health worries, financial worries, medication concerns, HIV mastery, disclosure worries, provider trust, and sexual function. Its version in Portuguese was also validated in Brazil, presenting satisfactory reliability and validity 9 .
Despite the existence of these scales [3][4] , this study updates this construct considering impacts influencing this epidemic over 50 years. The Quali-HIV considers current needs and the dynamic nature of this epidemic, in addition to new epidemiological and social aspects. Thus, this study's objective was to develop and validate a scale to measure QoL among individuals living with HIV in Brazil.

METHOD
This methodological study based on the Psychometric Theory 10 was conducted between 2017 and 2019 in a Specialized Care Service for Sexually Transmitted Infections (STIs)/AIDS located in a city in the interior of São Paulo, Brazil.
Rigor is required for each of the stages concerning the development of an instrument, in which systematic and standardized methods are used to obtain a reliable measure [10][11] .
The process of establishing the first set of items to compose the Quali-HIV scale was guided by an integrative literature review and interviews held with this population. An initial version composed of 148 items emerged from this process. An expert committee composed of two experts in the topic, two experts in the method, and one individual representing this population, verified its face and content validity. A 76-item instrument resulted from this stage, which was then submitted to semantic validation and a pre-test, from which the 51-item version used in this study emerged [12][13] .
This study proceeded with the scale's validation addressing a population composed of PLHIV in follow-up at a health service. Individuals of both sexes, aged 18 old or older, with an HIV diagnosis for at least six months and taking the antiretroviral therapy for at least three months participated in the study. The guidelines provided by the Ministry of Health concerning the ART protocol were considered to include the participants. According to this protocol, all PLHIV, regardless of their viral load, should take antiretroviral therapy 14 . Imprisoned and institutionalized individuals were excluded.
The sample size was determined according to the literature 15 , and a convenient sample was selected. The individuals were invited to participate in the study while they awaited their scheduled medical appointments. A previously validated instrument in terms of form and content, addressing the individuals' sociodemographic and clinical aspects and lifestyle, was used to interview and characterize the participants.
The participants rated the first set of items composing the self-reported Quali-HIV on a fivepoint Likert scale.
The scale's development and validation included a description of its factor structure, reliability, convergent and divergent construct validity, in addition to floor and ceiling effects.
An Exploratory Factor Analysis was performed to describe the scale's factor structure concerning the number of factors and allocation of items in each factor. The Kaiser-Meyer-Olkin (KMO) test and Bartlett's sphericity test were performed to verify whether the data matrix could be factorized 16 .
The unweighted least squares method and Varimax rotation were used in the extraction of the factors. Factor loadings below 0.30 were the criterion used to exclude items 11 . A scree plot helped to determine the number of factors 15 .
In this study, reliability was verified using Cronbach's alpha coefficient. Therefore, the items' internal consistency and dimensions were verified. Coefficients can range between 0 and 1, whereas 0 indicates lack of internal consistency and 1 represents total internal consistency; 11 values above 0.7 were considered acceptable 17 .
The Multitrait-Multimethod analysis was used to describe convergent and divergent construct validity. Acceptable product-moment correlation coefficients, between an item and the factor to which it belongs, are supposed to be higher than 0.40 to confirm convergent validity. Divergent validity is verified through the percentage of times in which product-moment correlation coefficients, between an item and the factor to which it belongs, are higher or statistically higher than those between the item and the factors it does not belong. The closer to 100, the more discriminant the dimensions are (goodness of fit) 11 .
Floor and ceiling effects occur when the scores are not symmetrically distributed, and more than 15% of the responses are concentrated either on the minimum or maximum scores of a scale, respectively. The occurrence of these effects reduces a scale's responsiveness, preventing or hindering verifying whether the construct changed in situations in which one's health condition worsened or improved 17-18 .

5/14
IBM ® SPSS version 20.0 was used to determine the factor structure and reliability, and the Multitrait Analysis Program (MAP) was used to perform the Multitrait-Multimethod analysis and verify the correlations between the items and their respective factors 11,19 .
This study complied with all the ethical guidelines recommended by Resolution 466/12 and the participants signed free and informed consent forms.

RESULTS
A total of 460 PLHIV participated in this study, and the information concerning 367 participants was used to describe the scale's factor structure and remaining psychometric properties.
The participants were 43 years old on average (SD=±12.4), ranging from 18 to 73. Most were men 276 (60.0%), heterosexual 313 (68.0%), single 226 (49.1%), and completed high school 182 (39.6%). The participants' laboratory characteristics concerning HIV infection and clinical characteristics are presented in Table 1. The Quali-HIV Scale composed of 51 items was used in the factor structure analysis. The correlation matrix showed satisfactory covariance (0.832) for factor extraction through the KMO test, and Bartlett's sphericity test presented a significant result (<0.001).
The Scree Plot showed eigenvalues equal to or higher than 1, showing the scale possibly had four factors. To compose the factors and considering the criterion of explained variance, the distribution and arrangement of items revealed a model composed of 51 items distributed into four factors. This was the most appropriate outcome, with an explained variance of 36.45% (Table 2).   (Table 3).  The Quali-HIV's reliability was verified with a Cronbach's alpha equal to 0.85. The factors' coefficients ranged from 0.68 and 0.89. The four rotated factors were distributed as follow: Factor 1 was composed of 16 items with a Cronbach's alpha equal to 0.89; Factor 2 grouped 11 items, with Cronbach's alpha equal to 0.75; Factor 3 comprised three items with a Cronbach's alpha equal to 0.85, and Factor 4 was composed of 5 items with a Cronbach's alpha equal to 0.68.

6/14
Regarding convergent validity, Table 4 presents the product-moment correlations that resulted from the MAP analysis. Most items (93.3%) presented satisfactory values, above 0.40. According to MAP, convergent validity also presented satisfactory results as the scale as a whole presented goodness of fit equal to 96.3%. When the factors were analyzed separately, Factors 1 and 2 presented 98.7% and 97.0% respectively; Factor 3 presented goodness of fit of 100% and Factor 4 presented goodness of fit of 80% (Table 5).

Goodness of fit 100%
Of 45 items resulting from the exploratory factor analysis, floor and ceiling effects occurred in 29 and 16 items, respectively.

DISCUSSION
Quality of life among PLHIV is a relevant topic in the historical context of the disease, considering that improved knowledge concerning diagnosis and its treatment has impacted how the infection is managed and how people live with it 20 .
The individuals participating in the development and validation of the Quali-HIV Scale were PLHIV, most of whom were men. Even though the infection profile changed in recent years, such as the feminization of HIV, predominantly male individuals become infected. Data presented in this study corroborate the epidemiological bulletin issued in 2018 by the Secretaria de Vigilância em Saúde  [Health Surveillance Department]. The ratio between sexes for HIV diagnosis was 26 men for every ten women diagnosed with the disease, without considering infections among pregnant women 21 . Data reported between 2007 to June 2019 revealed 300,496 HIV infections in Brazil,136,902 (45.6%) of which were identified in the southeast. The rate of detection according to sex may vary according to the region; however, a predominance of cases among men was found in most regions; a finding also reported by other studies 3,[22][23][24] .
This study shows a prevalence of heterosexual men; however, most men exposed to the virus in the southeast and mid-west reported to be gay or bisexual 21 .
Factor analysis was applied to verify the instrument's structural construct, establishing the factors and the relationship between the variables 15 . Factor analysis is a tool used to explore the dimensions of a scale. In this study, this analysis enabled identifying four factors: Factor 1 (26 items), Factor 2 (11 items), Factor 3 (3 items), and Factor 4 (5 items). Therefore, the final instrument was composed of 45 items.
Note that Factor 1, called "Impact of HIV infection on QoL," addresses items concerning daily activities, the treatment, diagnosis, and friendships. Factor 2 "Well being" comprises items concerning leisure, the use of antiretroviral therapy, diet, and religion, among others. Factor 3 "Physical Activity," includes items that concern exercise, and Factor 4 "HIV diagnosis," includes items concerning serological status, confidentiality, fears, and difficulties. These factors represented 40.0% of the construct's explained variance.
The factors included in the Quali-HIV Scale reflect aspects that are essential to assess QoL, considering it is a multidimensional construct 4,11 .
Factor analysis is widely used by studies developing and validating scales to assess construct validity because it enables identifying factors that explain the construct under study. Construct validity is the most important psychometric assessment for a measurement tool 11 .
The results obtained for the Quali-HIV Scale showed satisfactory convergent validity for most items (42 items) and their respective factors, with a satisfactory correlation between them 11 . Corroborating this study's findings, one study recommends similar correlations for this validity 25 . Only three items presented small correlations with their factors. However, after a qualitative analysis, we decided to keep the factors where they were allocated.
Note that a high proportion of common variance in convergent validity is supposed to be contained in the items indicating a specific construct 15 .
Regarding discriminant validity, it shows the degree in which a construct diverge from the remaining 15 . In this study, the instrument presented satisfactory goodness of fit. It can be verified considering the percentage of items that present higher correlations with their respective factors than with the remaining factors 11 . The high percentages of factors and of the scale as a whole confirmed the instrument's validity.
Validating or adapting a specific instrument to measure health-related QoL ensures that the participants access the same construct 26 .
Regarding the Quali-HIV reliability, satisfactory indexes were found for the total scale and each of its factors 17 . A similar result was reported by a study that used this instrument to assess the QoL of PLHIV 27 . Additionally, the investigation developed to assess the psychometric properties of the WHOQOL-HIV Bref obtained a Cronbach's alpha equal to 0.93 28 .
The description of floor and ceiling effects is based on the answers provided to the instrument, considering that the participants may opt for one extreme or the other. This dynamic may compromise the variability of responses. Respondents may use this strategy for finding it difficult to rate their experiences 29 , and these effects may influence the instrument's responsiveness and reliability 17 .
Positive and negative terms included in the items may result in these effects 30 . Other studies validating instruments measuring health-related QoL also identified these effects 29 . Nonetheless, it is worth noting that these effects may be related to the perception of QoL held by PLHIV.
After the development and validation processes, the conclusion is that the Quali-HIV Scale is a valid and reliable instrument to measure QoL among PLHIV.
This study's limitation refers to the fact that the development and validation of this scale were conducted in the population assisted by Specialized Care Units IST/HIV located in a single Brazilian city.

CONCLUSION
This study's relevance lies on the fact that there is a lack of updated tools to measure QoL among PVHIV and this study provides a valid and reliable instrument to be used by nursing workers and other professionals providing individualized care to this population.
The Quali-HIV scale was developed in Brazil, considering the current context of the HIV epidemic. Its importance is linked to the methodological rigor applied during its development. Thus, the Quali-HIV scale is an updated instrument that can be replicated and used to diagnose the QoL of individuals living with HIV to support health care actions implemented to this population.
This scale can be used in various settings and adapted to other contexts. However, its use has to be authorized by the primary author, considering that the studies addressing this scale will contribute to improved knowledge.