Development and validation of Ferrans & Powers Quality of Life Index: wound version

Abstracts

O artigo visa a apresentar o processo de construção e validação do Índice de Qualidade de Vida de Ferrans & Powers - Versão Feridas (IQVFP-VF) para emprego em pessoas com feridas agudas e crônicas de diferentes etiologias. O estudo metodológico desenvolveu-se por meio de procedimentos teóricos, empíricos e analíticos. Os resultados indicaram que o instrumento possui consistência interna e estabilidade satisfatórias; confirmaram a validade de conteúdo, concorrente (com o item sua satisfação) e convergente (com o WHOQOL-breve), bem como sua capacidade de discriminar os indivíduos conforme o número e duração das feridas, intensidade de dor e idade. A análise fatorial confirmatória mostrou que o instrumento manteve-se razoavelmente ajustado ao modelo original. Em conclusão, pode-se considerar que o IQVFP-VF é válido e tem confiabilidade atestada nos aspectos mais importantes da Qualidade de Vida para a população com feridas - geral, saúde e aspectos psicológicos e espirituais.

Qualidade de vida; Ferimentos e lesões; Reprodutibilidade dos testes; Estudos de validação

ORIGINAL ARTICLE

Development and validation of Ferrans & Powers Quality of Life Index - wound version^*

Desarrollo y validación del Índice de Calidad de Vida de Ferrans & Powers - versión heridas

Beatriz Farias Alves Yamada^I; Vera Lucia Conceição de Gouveia Santos^II

^IEnfermeira estomaterapeuta (TiSOBEST). PhD. in Nursing. Sócia proprietária da Enfmedic Saúde. Coordenadora do Curso de Especialização em Enfermagem em Estomaterapia da Universidade Estadual do Amazonas. Manaus, AM, Brazil. [e-mail]

^IIEnfermeira estomaterapeuta (TiSOBEST). Associated Professor of Medical-Surgical Nursing Department of Nursing School of the University of São Paulo. Coordenadora dos Cursos de Especialização em Enfermagem em Estomaterapia da Nursing School of the University of São Paulo.São Paulo, SP, Brazil. [e-mail]

Correspondence addressed to

ABSTRACT

The article shows the development and validation process of the Ferrans & Powers Quality of Life Index - Wound Version (FPQLI -WV) to be applied in people with different types of acute and chronic wounds. This methodological study was developed using theoretical, empirical and analytical procedures. Results showed that the instrument obtained satisfactory internal consistency and stability; and also confirmed content, concurrent (your satisfaction item) and convergent validities (WHOQOL-Bref), besides its discriminatory validity according to wound number and duration, pain intensity and age. Confirmatory factor analysis suggested FPQLI-WV showed reasonably fit to the original instrument model. In conclusion, FPQLI-WV showed to be reliable and valid for the most important wound patients' quality of life aspects such as health and the psychological and spiritual domains.

Key words: Quality of life. Wounds and injuries. Reproducibility of results. Validation studies.

RESUMEN

El articulo presenta el proceso de desarrollo y validación del Índice de Calidad de Vida de Ferrans & Powers–Versión Heridas (ICVFP-VH) para utilización en pacientes con distintos tipos de heridas agudas y crónicas. El estudio metodológico se ha desarrollado por medio de los procedimientos teóricos, empíricos y analíticos. Los resultados indicaron que el instrumento tiene consistencia interna y estabilidad; han confirmado la validad de contenido, concurrente (con el ítem su satisfacción), convergente (con el WHOQOL-breve) y discriminante de acuerdo con el numero y tiempo con la herida, intensidad del dolor y edad. La análisis factorial confirmatoria ha apuntado que el instrumento se ajusta de manera razonable al modelo original. Es posible concluir que el ICVFP-VH es fiable y valido en los aspectos más importantes para la calidad de vida de las personas con heridas - salud y dimensiones psicológica y espiritual.

Descriptores: Calidad de vida. Heridas y traumatismos. Reproducibilidad de resultados. Estudios de validación.

INTRODUCTION

Cutaneous wounds may affect people in all phases of the vital life cycle. In order to repair eventually damaged tissues, the body applies an intrinsic, dynamic, organized and extremely complex biological process that can be quickly developed whenever the clinical situation is favorable and the extent and degree of tissue loss is not considerable. Nonetheless, countless wounds evolve to chronic statuses, triggering a series of problems that can affect the subject's whole life structure and generate strong negative impacts to the person's quality of life (QOL).

QOL can be assessed by either qualitative or quantitative approaches. The quantitative assessment is carried out by measurement instruments whose advantage is the production of scores that make possible easier comparison processes among populational groups. For this reason, this type of assessment is largely adopted in all fields of human knowledge, especially in healthcare.

There is not a specific and all-embracing instrument to assess peoples' QOL in the domain of wounds. As such, a model to be applied in patients with cutaneous acute or chronic wounds stemming from any etiology was developed. For this purpose, a specific version of a general QOL translated and validated instrument was developed, namely Ferrans & Powers Quality of Life Index (FPQLI)

DEVELOPMENT OF THE FERRANS & POWERS QUALITY OF LIFE INDEX - WOUND VERSION (FPQLI-WV)

The present study is originally a PhD thesis⁽¹⁾ on the Adult Healthcare Nursing Program of the Nursing School of the University of Sao Paulo. The research was approved by the School's Ethics Committee in Research under number 464/2005/CEP-EEUSP.

The development of the FPQLI wound version is based upon the building criteria of measurement instruments. The background model was proposed by Pasquali⁽²⁾ and divided into three broad poles named: theoretical, empirical (experimental), and analytical (statistical) procedures.

The theoretical procedure regards the theoretical foundation laid on the investigation construct; the empirical procedure refers to the pilot instrument's phases and application techniques, as well as the information collection process aimed to assess the psychometric properties of the final instrument; the analytical procedure presents data statistical analyses aimed to the validation of the developed instrument. The author⁽²⁾ considers the two last procedures as part of an instrument validation process.

Theoretical procedures

The theoretical procedures are the most puzzling part of the process of developing an instrument. However, as the study was focused on the development of a new version, some phases had already been fulfilled by the authors of the original instrument in the process of developing the generic version. This phase is subdivided into certain steps: establishment of the construct's dimensionality; constitutive and operational definition of the construct; operationalization of the construct in behavioral tasks; and theoretical analysis of items⁽²⁾.

In this study, quality of life was the construct and the constitutive definition was similar to the one proposed by the generic instrument, where QOL is understood as a person's sense of well-being that stems from satisfaction or dissatisfaction with the areas of life that are important to him/her⁽³⁾.

As per the operational definition (insertion of items), a list of items that represent the investigated construct must be drafted. In order to complete this list, the study triggered the analysis of sixty nine articles found in literature, including other versions of the FPQLI. The researchers' clinical experience and their informal discussions with colleagues and patients were also taken into account.

The process of reviewing literature showed that several items that could be related to people's QOL were already inserted into the generic FPQLI. In such a way, all questions of the generic version were kept and new items of interest to measure the QOL in wounded populations were distributed in their original domains (health/ functioning, socio-economic, psychological/spiritual, and family). They were listed as follows: wound care process; intensity of pain in the wound; pain relieving treatment; be wounded; time for healing; wound conditions (drainage, smell, appearance); changes in daily life due to the wound (e.g. wound dressing, shower, clothes, shoes, medicine, work, movement); sleep; not having any children; not having any husband/wife, boyfriend/girlfriend, or spouse.

At the end of this phase, we proceeded to the item construction process following the FPQLI original structure and inserting the items into the respective original domains. Pilot version 1 was then created. It was compatible with the next to last stage of the theoretical procedures, that is, the analysis of the items.

The analysis of the items was performed in two distinct phases. The first aimed to assess the content validation and the semantic analysis. This procedure was carried out by a committee of judges composed of five knowledgeable professionals who were responsible for assessing the conceptual equivalence of the items and their relationship with the domains, besides the semantics of the items. The second phase was carried out after the committee of professional judges concluded the analysis process and was aimed to develop a semantic assessment only. For that purpose, a heterogeneous focal group, composed of six people with wounds, was created.

Following the conclusion of the focal group phase and in order to finalize the theoretical procedures, a pre-test using the pilot-III version (resulting from the analysis of the items) was applied to a sample composed of 12 patients from three medical institutions located in the Brazilian states of Sao Paulo and Minas Gerais.

The pre-test showed that some people were still unaware of the meaning of some terms, even after the implementation of two semantic assessments. Two items needed to be reshaped, which generated the pilot-IV version that was used in the empirical procedure presented next.

Empirical procedures

The empirical procedures regard the application of the pilot instrument obtained in the phase of the theoretical procedures towards the target population, aimed to generate data to be submitted to the analysis of the instrument's psychometric properties.

In the present study the FPQLI-WV was applied two times in distinct situations. The first application aimed to select a population sample (test sample), while the second aimed to reapply the test to a part of the previously interviewed sample (re-test), fifteen days later. A part of the test sample was selected to test the construct validity by the application of another QOL instrument in parallel with the application of the test.

A questionnaire idealized by the researchers was adopted in order to collect socio-demographic and clinical data; besides, the pilot-IV FPQLI-wound version and the brief World Health Organization's Quality of Life instrument - WHOQOL-bref - were also applied.

An interview-based technique was used for the data collection process and was carried out by the researchers and their instructed assistants.

A quite relevant aspect refers to the establishment of the sample sizes, not clearly defined by literature. For the test sample, there is a variation between five and ten persons for each of the instrument's items^(4,5). This study defined a minimum of eight people for each of the 42 items of the questionnaire (n=336). However, this amount was exceeded and reached 364 people belonging to 16 centers located in the states of Sao Paulo, Minas Gerais and Ceará. As literature presents no standardization, the re-test established a minimum of 50 reassessments and thus the final sample was composed of 63 voluntary people who returned to take the test 15 days later. Finally, the construct validity test determined a minimum of 40% of the test sample, and at the end of the process a total amount of 180 people had been assessed.

After the accomplishment of the empirical phase, the next step was the analytical procedure, which consists of statistical analyses aiming to assess the reliability and validity of the applied instrument.

Analytical procedures

Prior to the beginning of the statistical analyses, some phases were made necessary to organize the data. In this research, content validity analyses were carried out in the theoretical procedures phase. The agreement analysis among the members of the committee of judges considered the level of 80% as adequate. The items that were included in the instrument and did not reach the settled percentage were either excluded or adjusted, following the judges' suggestions. Notwithstanding, there was no exclusion of original items. This decision was made in further statistical analyses.

Following the final collection, socio-demographic and clinical data, as well as the QOL instrument scores, were directly stored in the Statistical Package for the Social Sciences (SPSS) version 12.0 software. For the confirmatory factorial analysis, the data located at the SPSS were transferred to the Lisrel 8.7 program.

For the reliability analysis, the missing data was treated and excluded items that presented over 20% of missing responses. The other missing items recorded their own averages. Such proceeding resulted in the loss of only two interviews, thus using 362 cases for the analyses. Besides, it was necessary to combine mutually excluding items of the questionnaire; these items were turned into a single one.

Items that did not fit in the instrument were excluded by means of the item total correlation (ITC), Cronbach's alpha, and type of correlation. According to literature parameters^(4-8), a TIC minimum value of 0.30 was standardized for the maintenance of an item and a Cronbach's alpha coefficient lower than 0.70 was adopted for the exclusion of new items. Moreover, any original or new items that presented negative correlations in the correlation matrix among items would also be excluded.

Statistical tests

For the description of the socio-demographic and clinical characteristics of subjects, as well as the QOL instrument scores, this research applied the measurement of central trend (simple frequency, relative frequency, mean, median, maximum and minimum) and dispersion measurements (standard deviation). Inferential statistics was used to test the instrument's psychometric properties (reliability, factorial analysis and validity). Tests that achieved a descriptive level lower than 5% were considered as significant.

For the internal consistency (IC) the Cronbach's alpha coefficient was calculated for all items and domains of the FPQLI-WV. The intraclass correlation coefficient was used to measure the stability. The correlation level was standardized as follows: weak £ 0.40; moderate = 0.41 to 0.60; good = 0.61 to 0.80; excellent = 0.81 a 1(9-10).

For the analysis of the instrument's validity, two different analyses were carried out:

• Concurrent criterion validity: assessed by the correlation among the scores of your satisfaction with life item, as well as the FPQLI-WV domains and general quality of life, by means of the Pearson's or Spearman's correlation tests, depending on the distribution of items. In order to analyze the magnitude of correlations, the reference values adopted were: weak < 0.30; moderate = 0.30 to 0.60; strong = 0.60 to 0.99; and perfect = 1.00⁽¹¹⁾.

• Construct validity: convergent validity was assessed through the analyses of the correlation between the FPQLI-WV domains and the WHOQOL-Brief.

• The discriminant based construct validity was tested in the domains by comparing the values among subjects from different age groups, amount and duration of wounds, as well as intensity of pain at the moment of the test and the most intense pain in the previous week. The statistical tests applied were the Kolmogorov-Smirnov, the Mann-Whitney, the t-Student, the ANOVA and the Bonferroni.

At the end of the internal consistency analysis, after the definition of the items that would compose the domains, a confirmatory factorial analysis (CFA) was carried out aiming at assessing the adjustment of the wound version. The model was assessed by the item correlation degree with its respective domain. In order to analyze the magnitude of the correlation, the reference values were similar to those adopted by other analyses. The adjustment of the model was assessed by the following methods⁽¹²⁾: Goodness of Fit Index (GFI), with acceptance level lower than 0.8; Root Mean Square Error Approximation (RMSEA), with desirable interval of 0.08 or lower, being 0.10 the acceptable maximum reference value; the Adjusted Goodness of Fit Index (AGFI) and the Normed Fit Index (NFI) are incremental fit measurements of the model whose desirable reference values should reach 0.90; the Normed Chi-Square Measurement is a general fit quality measurement of the model, being calculated by dividing the chi-square by the degrees of freedom (some authors consider values lower than 2.103 as being within the acceptability limits; however, this measurement exceeds the values pointed out by other researchers)⁽⁴⁾; the Comparative Fit Index (CFI) is a more parcimonious fit measurement of the model. Values close to zero suggest a lack of over adjustment in the model in comparison with other models, that could be possible if the items were allocated in other domains. Values close to 1 suggest a substantial improvement of the current model in comparison with all others⁽¹²⁾.

Final instrument

The final version of the FPQLI-WV - a specific instrument that assesses the QOL of people with wounds from any etiology - is composed of 34 items to be filled-in (Appendix)⁽¹⁾.

The socio-demographic characteristics of the people who composed the study's sample were as follows: 55.2% women, mean age of 59.2 + 15.8; 2.9 + 2.6 children; 47.4% married; 60.4% with educational level below the elementary level (12.4% illiterate) and 40.7% retired. The clinical characteristics were: 90.6% of wounds were located on lower members; average of 1.5 ± 1.0 wound per patient; 50.1% of wounds had a venous origin and showed average duration of 51.1 ± 90.7 months. The average intensity of pain at the moment of the test, as well as the worst and lightest pain of the previous week were 2.1 ± 2.8; 5.0 ± 3.5; and 1.6 ± 2.1, respectively.

The reliability and validity values are presented in Tables 1 through 5, and also in Figure 1.

The internal consistency and stability ratified the FPQLI-WV reliability. The instrument also showed concurrent validity in all the domains, as well as convergent construct validity for all the domains, except the Family domain after the correlation with the WHOQOL-Bref physical domain. For the discriminant validity, the amount of wounds, the duration of the current wound, the intensity of pain, and the age were variables that distinguished the QOL of the groups.

The results of the Confirmatory Factor Analysis (Figure 1) indicate that the FPQLI-WV model is at least peripherally acceptable and accurately adjusted to the original model.

CONCLUSIONS AND RECOMMENDATIONS

The FPQLI-WV showed reliability and validity to be applied in patients with acute and chronic wounds, at least in those wounds encompassed by the inclusion criteria employed here, except for patients with wounds that stemmed from severe burns, cancer and wounds originated from the Acquired Immunodeficiency Syndrome (AIDS).

New projects must be developed in order to test the instrument's application in patients with wounds originated from third degree burns, cancer and fungating wounds, and wounds in people with AIDS in order to analyze the instrument's reliability and validity in those samples.

REFERENCES

• 1. Yamada BFA. Índice de Qualidade de Vida de Ferrans & Powers: construção e validação da versão feridas [tese]. São Paulo: Escola de Enfermagem, Universidade de São Paulo; 2006.
• 2. Pasquali L. Princípios de elaboração de escalas psicológicas. Rev Psiq Clin. 1998;25 (5):206-13.
• 3. Ferrans CE, Powers MJ. Psychometric assessment of the quality of life index. Res Nurs Health. 1992;15(1):29-38.
• 4. Hair JF, Anderson RE, Tatham RL, Black WC. Análise multivariada de dados. Porto Alegre: Bookman; 2005.
• 5. Nunnaly JC, Bernstein HP. Psychometric theory. 3^rd ed. New York: McGraw-Hill; 1994.
• 6. LoBiondo-Wood G, Haber J. Nursing research: methods, critical appraisal, and utilization. 4^th ed. St. Louis: Mosby; 1998.
• 7. McHorney CA, Ware JE Jr, Lu JF, Sherbourne CD. The MOS 36-Item Short-form Health Survey (SF-36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups. Med Care. 1994;32(1):40-66.
• 8. Streiner DL, Norman GF. Health measurement scales: a practical guide to their development and use. 2^nd ed. New York: Oxford Press; 1995.
• 9. Bartko JJ. The intraclass correlation coefficient as a measure of reliability. Psychol Rep. 1966;19(1):3-11.
• 10. Wilson KA, Dowling AJ, Abdolell M, Tannock IF. Perception of quality of life by patients, partners and treating physicians. Qual Life Res. 2000;9(9):1041-52.
• 11. Levin J. Fox JA. Estatística para ciências humanas. 9Ş ed. São Paulo: Prentice Hall; 2004.
• 12. Novy DM, Nelson DV, Berry LA, Averill PM. What does the Beck Depression Inventory measure in chronic pain: a reappraisal. Pain. 1995;61(2):261-70.

⁽

^ª

⁾. This instrument has been operating for 27 years now and was one of the first pioneers in the QOL field. The instrument is grounded on a humanistic proposal. It is distinct from other instruments and was aimed to meet the philosophical presuppositions of nursing, and ours as researchers at the same time.

Publication Dates

• Publication in this collection
  23 Mar 2010
• Date of issue
  Dec 2009

History

• Received
  01 July 2009
• Accepted
  24 Aug 2009