Development and validation of a quantitative tool for assessment of sense of grip on disease among PL-HIV: a confirmatory factor analysis

INTRODUCTION

World Health Organization defines chronic disease as ‘an illness for which there is no definite treatment are there but a range of various medical treatments and health behavior is there for its management’.¹ All chronic diseases have their specific characteristics but they all share a common ground that they stay for the lifelong, demanding addressal of disease-specific medical, social, and psychological needs such that people diagnosed with chronic diseases which is a critical experience in their lifetime have to engage themselves in the management of their disease, has to understand the diagnosis and its implications, has to cope with the new challenges that come with the diagnosis.^2-4

This critical experience can be described as an amalgamation of affective, visceral, and bodily experiences that cannot be expressed by words. These inexpressible phenomena are translated into comprehensible words by a process known as ‘sense-meaning-making’ or in short ‘sense’. On the other hand development of psychological competencies required for disease management as well as personal and contextual change post-diagnosis is known as ‘grip on disease’.⁵ Maiello et al gave the sense of grip on disease (SoGoD) model developed based on narrative, semiotic, and socio-constructive perspectives, explores the critical experiences a person goes through post-diagnosis of a disease (sense) and its effect on mental competencies required for comprehensive disease management.^5-7 It has the following narrative domains with their dimensions (Table 1) with two ultimate grip profiles of ‘dynamic grip’ and ‘compliant grip’.

*As per Maiello et al⁵

So, the SoGoD status of an individual living with chronic disease can be explored using semi-structured face-to-face interviews but often these can be lengthy, and difficult to analyze, and nonetheless, subjective biases can’t be excluded. It showed that there is a lack of an objective quantitative tool to assess the SoGoD status. Thus, our study aimed to develop a quantitative tool to assess the SoGoD status.

METHODS

An institution-based tool development study was conducted among 120 adult patients diagnosed with HIV visiting the ART center of the School of Tropical Medicine, Kolkata, from September to December 2023.

Sample size and sampling

As the final revised version of the tool contained 24 items, as per respondent item ratio of (5:1) total of 120 participants were included in the study.⁸

There were on average 70 people diagnosed with HIV visiting the study setting on a regular working day. Considering the selection of a maximum of 10 participants each day the sampling interval came to seven.

Thus, systematic random sampling was applied using the daily register to select every 7^th visitor until the daily sample size was met.

Data collection

Upon selection, the individual was approached and they were explained about the research and its objectives. Those who were sick, and/or refused/unable to give written informed consent were excluded on ethical grounds. Upon agreement they were interviewed, using a close-ended semi-structured interviewer-administered questionnaire containing the socio-demographic details (age, gender) and the revised SoGoD tool.

Sense of grip on disease (SoGoD) tool

The quantitative tool to assess coping was formulated based on the SoGoD model given by Maiello et al.⁵ As per the model constructs wise items were framed through deductive reasoning, after a thorough review of the literature. The initial pool of 36 items was generated and then they were reviewed by three independent domain experts i.e. public health specialist, physiatrist, and infectious disease specialist. Based on their feedback, the items were revised. The revised versions were translated into local languages i.e. Bengali and Hindi and then again back-translated into English by two independent people to maintain compatibility.⁹ The revised tool was pre-tested among 20 adult PLHIV visiting another ART center for acceptability, clarity, and readability. Based on their suggestions further revision was done.

This final version containing 24 items was applied among 120 study participants.

Responses to the questions were coded on a five-point Likert scale of strongly disagree (1) to strongly agree (5), in such a direction that higher values reflected higher SoGoD status. Item numbers two, six, nine, ten, eleven, and thirteen were reverse scored.

Statistical analysis

The SoGoD tool was developed with a congeneric, and unidimensional five-factor confirmatory model containing 24 items. Initially, the dataset was screened for the assumptions needed to meet for conducting CFA i.e. Adequate sample size, Multivariate normality, etc.^10,11

The validity of the model was tested based on factor loading of the items with an acceptable cut-off >0.7, average variance extract (AVE) >0.5, and composite reliability (CR) >0.5.^10,11

The model fit was assessed and determined by: 1) statistical fit (χ²); 2) absolute fit using the root mean square error of approximation (RMSEA) with values ≤0.08 indicating acceptable model fit and values ≤0.05 indicating close fit and the goodness of fit index (GFI) with values ≥0.90 indicating acceptable model fit and values ≥0.95 indicating close fit; 3) relative fit using the Tucker-Lewis index (TLI) and comparative fit index (CFI) with values ≥0.95 indicating close model fit and values ≥0.90 indicating acceptable fit.¹²

The internal consistency of the tool was assessed by Cronbach’s alpha (α) statistic and the cut-off was taken as ≥0.7.^13,14 At first item-rest correlation was calculated with a cut of 0.20.¹⁵ Then average inter-item correlation within a construct was also assessed, with a desired value between 0.15 to 0.50.¹⁶ For each item, changes in the alpha of the tool if the item had been removed were also calculated. To determine the cut-off, item analysis was applied.^17,18 Data were analyzed using MS Excel 2021 and SPSS software version 16.

Ethical consideration

Ethical permission was taken from the institutional ethics committee from STM, Kolkata (ref no. 2023-AS38), AIIH and PH, Kolkata, and NACO.

RESULTS

The mean age of the study participants was 46.8±12.2 years. Out of 120 participants, 85% of them were literate. Table 2 shows the responses of the participants to the questions intended to formulate the SoGoD tool.

SA: Strongly Agree (5), A: Agree (4), N: Neutral (3), D: Disagree (2), SD: Strongly Disagree (1). *Items number two, six, nine, ten, eleven, and thirteen were reverse-scored.

Cut-off point = Ʃ [(Mean raw score – 2* Standard error of mean) * Correction factor] Where, Correction factor = (Sum of total weighted score/ Sum of total raw score) =918.45/4676 =0.19 Standard error of mean = SD/√n

Assumptions of CFA

Five theoretical constructs along with the communality of each item were more than 0.6 indicating a sample of 120 was adequate to conduct CFA.¹⁹ There was multi-variate normality. There were no missing values. We arrived at three items per construct. Thus, it was a Just-identified model.

Model validity

The model was valid with the factor loading (λ) of each item on their respective construct >0.7, average variance extract (AVE) >0.5, and composite reliability (CR) >0.5. As for nomological validity, the observed findings coincide with theoretical constructs (Table 3). The five-factor model is shown in Figure 1.

Model fit statistics

The chi-square value was 101.62, with χ²/df value of 1.27 (p value 0.152), suggesting a good fit to the data. The RMSEA (0.048), CFI (0.938), and GFI (0.043) all suggested an acceptable fit, while the TLI (0.978) suggested a close fit. The standardized residual of all items was <2.5, thus no item has been dropped.

Reliability statistics

The tool was internally consistent with the alpha of 0.883 (Table 4).

Weightage and cut-off determination

Considering the percentage of total variance explained by each construct (Table 4) we asserted weight, to arrive at a global SoGoD score = [(Item 1 + Item 2 + Item 3) * 0.5] + [(Item 4 + Item 5 + Item 6) * 0.2] + [(Item 7 + Item 8 + Item 9) * 0.1] + [(Item 10 + Item 11 + Item 12) * 0.15] + [(Item 13 + Item 14 + Item 15) * 0.05]. The cut-off was found to be 6.8, rounding off to seven (Table 5).

DISCUSSION

This study confirmed the SoGoD model, based on the data collected from 120 PL-HIV visiting a tertiary care hospital in Kolkata. The items included were based on the theoretical framework given by Maiello et al.⁵ 15 items were mounted on five factors, organization of temporality, integration of disease, orientation to action, social sharing, and expression of emotion, found to be well fitting.

Considering the imposed weight, the global SoGoD score ranges from 3 to 15. A higher global SoGoD score indicates higher SoGoD status. As for the cut-off value of seven, any SoGoD global score of seven or more will be considered to be a dynamic sense of grip, otherwise compliant sense of grip.

This study has some limitations. This tool is developed based on the response from PL-HIV only, who face a harsher environment in comparison to persons living with other chronic diseases. Thus, their responses can be extreme which renders the tool with a stringent cut-off value, which can be contextualized based on setting and population.

CONCLUSION

This study revealed a simple and concise, quantitative tool to assess SoGoD status, that can be applied easily, easy to analyze, and interpret. Thus, this tool can be used further to measure coping behavior in PLHIV and also to find out the factors associated with it, which could be helpful to innovate new strategies for the improvement of quality of life for PLHIV.

Funding: No funding sources

Conflict of interest: None declared

Ethical approval: The study was approved by the Institutional Ethics Committee

REFERENCES

1. 1. World Health Organization. Noncommunicable Diseases Global Monitoring Framework: Indicator Definitions and Specifications. World Health Organisation. Available from: https://www.who.int/publications/m/item/noncommunicable-diseases-global-monitoring-framework-indicator-definitions-and-specifications. Accessed on 7 July 2023.
2. 2. NCD Countdown 2030 collaborators. NCD Countdown 2030: worldwide trends in noncommunicable disease mortality and progress towards Sustainable Development Goal target 3.4. Lancet Lond Engl. 2018;392(10152):107-88.
3. 3. Aujoulat I, Marcolongo R, Bonadiman L, Deccache A. Reconsidering patient empowerment in chronic illness: a critique of models of self-efficacy and bodily control. Soc Sci Med 1982. 2008;66(5):1228-39.
4. 4. Asbring P. Chronic illness- a disruption in life: identity-transformation among women with chronic fatigue syndrome and fibromyalgia. J Adv Nurs. 2001;34(3):312-9.
5. 5. Maiello A, Auriemma E, De Luca Picione R, Pacella D, Freda MF. Giving meaning to noncommunicable illness: mixed-method research on sense of grip on disease (SoGoD). Healthcare Basel Switz. 2022;10(7):1309.
6. 6. De Luca Picione R, Freda MF. Borders and modal articulations. semiotic constructs of sensemaking processes enabling a fecund dialogue between cultural psychology and clinical psychology. Integr Psychol Behav Sci. 2016;50(1):29-43.
7. 7. Salvatore S, Freda M. Affect, unconscious and sensemaking. A psychodynamic, semiotic and dialogic model. New Ideas Psychol. 2011;29(2):119-35.
8. 8. Gorsuch RL. Factor Analysis. 2nd edn. New York: Psychology Press; 1983:448.
9. 9. DeVellis RF. Scale Development: Theory and Application. 4th edn. New Delhi: SAGE Publications Inc.; 2017.
10. 10. Brown TA. Confirmatory Factor Analysis for Applied Research. 2nd edn. New York: Guilford Publications; 2015:482.
11. 11. Roos JM, Bauldry S. Confirmatory Factor Analysis; vol. 189 (Quantitative applications in the social sciences). USA: SAGE Publications Inc.; 2021.
12. 12. University of California, Los Angeles (UCLA). A Practical Introduction to Confirmatory Factor Analysis (CFA). Available from: https://stats.oarc.ucla.edu/r/seminars/rcfa/. Accessed on 31 March 2024.
13. 13. Cronbach LJ. Coefficient alpha and the internal structure of tests. Psychometrika. 1951;16(3):297-334.
14. 14. Tavakol M, Dennick R. Making sense of Cronbach’s alpha. Int J Med Educ. 2011;2:53-5.
15. 15. Zijlmans EAO, Tijmstra J, van der Ark LA, Sijtsma K. Item-score reliability in empirical-data sets and its relationship with other item indices. Educ Psychol Meas. 2018;78(6):998-1020.
16. 16. Piedmont RL, Hyland ME. Inter-item correlation frequency distribution analysis: a method for evaluating scale dimensionality. Educ Psychol Meas. 1993;53(2):369-78.
17. 17. Arumugam B, Suganya E, Nagalingam S. Derivation of cut-off value for a 10 item opinion based ordinal survey questionnaire. Int J Community Med Public Health. 2018;5(1030):10-8203.
18. 18. Barua A, Kademane K, Das B, Kumar S, Verma R, Al-Dubai S. A tool for decision-making in norm-referenced survey questionnaires with items of ordinal variables. Int J Collab Res Intern Med Public Health. 2014;6(3).
19. 19. MacCallum R, Widaman K, Zhang S, Hong S. Sample size in factor analysis. Psychol Methods. 1999;4:84-99.