Abstract
Background
The use of Bayesian Structural Equation Model (BSEM) to evaluate the impact of TB on self-reported health related quality of life (HRQoL) of TB patients has been not studied.
Objective
To identify the factors that contribute to the HRQoL of TB patients using BSEM.
Methods
This is a latent variable modeling with Bayesian approach using secondary data. HRQoL data collected after one year from newly diagnosed 436 TB patients who were registered and successfully completed treatment at Government health facilities in Tiruvallur district, south India under the National TB Elimination Programme (NTEP) were used for this analysis. In this study, the four independent latent variables such as physical well–being (PW = PW1-7), mental well-being (MW = MW1-7), social well-being (SW = SW1-4) and habits were considered. The BSEM was constructed using Markov Chain Monte Carlo algorithm for identifying the factors that contribute to the HRQoL of TB patients who completed treatment.
Results
Bayesian estimates were obtained using 46,300 observations after convergence and the standardized structural regression estimate of PW, MW, SW on HRQoL were 0.377 (p<0.001), 0.543 (p<0.001) and 0.208 (p<0.001) respectively. The latent variables PW, MW and SW were significantly associated with HRQoL of TB patients. The age was found to be significantly negatively associated with HRQoL of TB patients.
Conclusions
The current study demonstrated the application of BSEM in evaluating HRQoL. This methodology may be used to study precise estimates of HRQoL of TB patients in different time points.
Introduction
Tuberculosis (TB) disease causes significant negative impact on Health Related Quality of Life (HRQoL) of the patient [1]. The need for evaluation of TB therapeutic interventions is no longer restricted to clinical outcomes but also includes psycho-social well-being and HRQoL. The measurement of HRQoL are being used in direct patient care processes, clinical trials, program evaluations, and for monitoring health status in populations [2]. Quality of life has become an instrumental outcome measure in clinical research, and advances have been made in assessing the impact of many diseases on HRQoL.
HRQoL is a multidimensional concept which is evaluated by a number of different latent constructs such as physical function, general health, mental health and social relationships [3]. As these latent constructs often cannot be measured objectively and directly, they are treated as latent variables in HRQoL analysis. In medical statistics, traditional regression models are limited for modeling latent variables and may result in biased estimates.
The Bayesian SEM approach utilises prior information for updating the current information on the parameter and to estimate the parameters based on posterior distribution defined for latent variables [4–8]. There is paucity of studies which applied BSEM for modelling socio-psychological variables which are latent in nature [9–12]. With this background, we applied BSEM for understanding the structural relationship of the factors contributing to HRQoL of TB patients for the first time.
Methodology
Study design
This is a latent variable modeling with Bayesian approach using secondary data.
Source of data
HRQoL data collected after one year from newly diagnosed 436 TB patients who were registered and successfully completed treatment (Table 1) at Government health facilities in Tiruvallur district, south India under the National TB Elimination Programme (NTEP) were used for this analysis [13]. The study was approved by the Institutional Ethics Committee of the National Institute for Research in Tuberculosis, Indian Council of Medical Research, Chennai.
Table 1. Distribution of TB patients’ well-being in different domains.
| HRQoL domains | No. | % |
|---|---|---|
| Physical well-being | ||
| Free from TB symptoms | 263 | 60 |
| Other health problem | 96 | 22 |
| Current general health—Absolutely normal | 391 | 90 |
| No body pain | 300 | 69 |
| Body pain interfere activity- Not at all | 323 | 74 |
| Rate your health | ||
| Good | 161 | 37 |
| Very good | 140 | 32 |
| Health status compared to pre-illness | ||
| Somewhat better | 94 | 22 |
| Much better | 215 | 49 |
| Mental well-being | ||
| Happy most of the time | 249 | 57 |
| Never frustrate or impatient | 234 | 54 |
| Never feeling of fear or panic | 220 | 51 |
| Never depressed | 124 | 28 |
| Always free of tension | 94 | 22 |
| Always energetic | 161 | 37 |
| Feel low energy | 192 | 44 |
| Social well-being | ||
| Visiting friends/relatives/neighbours | 363 | 83 |
| Feel free to attend social function | 365 | 84 |
| Discussing illness with family members | 401 | 92 |
| Feeling inhibited to discuss illness with friends | 344 | 79 |
Tool used to measure HRQoL
The 36 item Short Form Health Survey (SF-36) questionnaire was used to measure HRQoL of the TB patients [14]. All the components under each domain were given equal weight for calculating of quality of life scoring based on SF-36. All scales and the component scores were positively scored and higher scores represented better HRQoL.
Model specification
In this study, 21 independent variables were used. The physical well-being variables were: Free from TB symptoms, other health problem, general health status, body pain, pain interfering normal activity, self-health rate and compared to pre- illness (PW1-PW7). The mental well-being variables were: happy person, frustrate/impatient, feeling fear/panic, depressed, free of tension, energetic and feeling low energy (MW1-MW-7). The social well-being variables were: visiting friends/relatives/neighbours, feel free to attend social functions, discussing illness with family members and feeling inhibited to discuss illness with friend (SW1-SW4). The life style related variables were: drinking alcohol, and smoking tobacco, substance abuse. The four independent latent variables which were grouped from these 21 independent variables are physical well–being (PW = PW1-7), mental well-being (MW = MW1-7) and social well-being (SW = SW1-4). The life style variables such as drinking alcohol, smoking tobacco, substance abuse were formed the latent variable habits. In addition, age and occupation variables were also considered for this model. The path diagram of different health domain influencing HRQoL of TB patients is given in Fig 1.
Fig 1. Path diagram for HRQoL of TB patients.
Legend: pw- Physical well- being, mw-Mental well-being, sw-Social well-being, occu–Occupation, subsabus–substance abuse, smoke–smoking habits, tpqol–Health Related Quality of Life of tuberculosis patients.
BSEM analysis
The BSEM was constructed using Markov Chain Monte Carlo (MCMC) algorithm for identifying the factors that contribute to the HRQoL of TB patients who completed treatment. The MPlus software version 7.1 was used for analysis. The non-informative prior which is default in MPlus was used for BSEM. For BSEM, the model convergence was assessed by the trace plots and estimated potential scale reduction (EPSR). Convergence of the sequence has been achieved where the ESPR values are less than 1.2. A Posterior predictive p-value is used to test the goodness of fit of the posited model. The codes used for BSEM analysis are incorporated in S1 Appendix in S1 File.
Results
Estimates of Bayesian SEM
The Bayesian estimates for measurement model (Table 2) and structural model (Table 3) were obtained using 46,300 observations after discarding the first 46,300 burn-in iterations. It was observed that EPSR values are less than 1.2 after 39,000 iterations indicating model convergence. The trace plots for selected parameters demonstrates that a tight horizontal band which indicates the parameters are converged (Fig 2). The posterior probability density plots that indicate the posterior densities for these parameters are approximating normal density (Fig 3).
Table 2. Parameter estimates of measurement model for HRQoL of TB patients.
| Factors | Estimated mean score | Posterior SD | 95% credible interval | p-value |
|---|---|---|---|---|
| Physical well-being by | ||||
| Free from TB symptoms | 0.586 | 0.038 | (0.511, 0.657) | <0.001 |
| Other health problem | 0.408 | 0.058 | (0.289, 0.515) | <0.001 |
| General health status | 0.778 | 0.039 | (0.692, 0.844) | <0.001 |
| Body pain | 0.830 | 0.025 | (0.777, 0.875) | <0.001 |
| Pain interfering normal activity | 0.864 | 0.024 | (0.813, 0.906) | <0.001 |
| Rate your health | 0.838 | 0.020 | (0.795, 0.873) | <0.001 |
| Compared to pre-illness | 0.661 | 0.034 | (0.590, 0.723) | <0.001 |
| Mental well-being by | ||||
| Happy person | 0.786 | 0.023 | (0.738, 0.828) | <0.001 |
| Frustrate/impatient | 0.714 | 0.028 | (0.654, 0.765) | <0.001 |
| Feeling fear/panic | 0.717 | 0.029 | (0.656, 0.769) | <0.001 |
| Depressed | 0.736 | 0.027 | (0.679, 0.785) | <0.001 |
| Free of tension | 0.540 | 0.039 | (0.461, 0.612) | <0.001 |
| Energetic | 0.814 | 0.021 | (0.768, 0.851) | <0.001 |
| Feeling low energy | 0.763 | 0.025 | (0.711, 0.808) | <0.001 |
| Social well-being by | ||||
| Visiting friends/relatives/neighbours | 0.961 | 0.019 | (0.913, 0.984) | <0.001 |
| Feel free to attend social functions | 0.977 | 0.017 | (0.934, 0.996) | <0.001 |
| Discussing illness with family members | 0.732 | 0.053 | (0.617, 0.824) | <0.001 |
| Feeling inhibited to discuss illness with friends | 0.182 | 0.077 | (0.028, 0.327) | <0.05 |
| Habits by | ||||
| Smoking | 0.866 | 0.041 | (0.777, 0.931) | <0.001 |
| Drinking | 0.886 | 0.040 | (0.796, 0.950) | <0.001 |
| Substance abuse | 0.096 | 0.103 | (-0.108, 0.293) | 0.179 |
| Physical well-being on | ||||
| Age | -0.205 | 0.046 | (-0.295, -0.112) | <0.001 |
| Occupation | 0.119 | 0.046 | (0.025, 0.206) | <0.01 |
| Mental well-being on | ||||
| Age | -0.161 | 0.044 | (-0.244, -0.072) | <0.001 |
| Occupation | 0.115 | 0.043 | (0.028, 0.196) | <0.01 |
| Social well-being on | ||||
| Age | -0.141 | 0.064 | (-0.268, -0.019) | <0.05 |
| Occupation | 0.249 | 0.058 | (0.130, 0.354) | <0.001 |
| Habits on | ||||
| Age | 0.355 | 0.055 | (0.243, 0.457) | <0.001 |
| Occupation | 0.244 | 0.059 | (0.126, 0.356) | < 0.001 |
Table 3. Parameter estimates of structural model for HRQoL of TB patients.
| Factors | Estimate | Posterior SD | 95% credible interval | p-value |
|---|---|---|---|---|
| HRQoL on | ||||
| Physical Well-being | 0.377 | 0.033 | (0.312, 0.440) | <0.001 |
| Mental Well-being | 0.543 | 0.034 | (0.478, 0.611) | <0.001 |
| Social Well-being | 0.208 | 0.026 | (0.155, 0.259) | <0.001 |
| Habits | -0.006 | 0.025 | (-0.056, 0.040) | 0.405 |
| HRQoL on | ||||
| Age | 0.040 | 0.019 | (0.003, 0.076) | <0.05 |
| Occupation | 0.005 | 0.018 | (-0.030, 0.040) | 0.383 |
| Physical Well-being with | ||||
| MW | 0.752 | 0.032 | (0.684, 0.810) | <0.001 |
| SW | 0.468 | 0.061 | (0.341, 0.581) | <0.001 |
| Habits | -0.121 | 0.073 | (-0.263, 0.022) | <0.05 |
| Mental Well-being with | ||||
| SW | 0.622 | 0.050 | (0.516, 0.711) | <0.001 |
| Habits | 0.088 | 0.071 | (-0.051, 0.226) | 0.108 |
| Social Well-being with Habits | 0.202 | 0.095 | (0.009, 0.381) | <0.05 |
Fig 2. Trace plots of value of parameter for different iterations.
Fig 3. Posterior distribution of kernel density.
Estimates of measurement model
It was observed that the estimate of all the variables except substance abuse were found to be significant with their respective latent variables (Table 2). The variables age and occupation were also found to be significant with the latent variables physical, mental, social well-being and habits.
Estimates of structural model
The standardized estimate of physical, mental and social well-being on the variable percentage of total quality of life were 0.377 (p<0.001), 0.543 (p<0.001) and 0.208 (p<0.001) respectively. Hence mental health had the most important effect on HRQoL, followed in turn by physical health and social relationship. The physical, mental and social well-being latent variables were significantly associated with HRQoL, while bad habits was not significantly affecting the HRQoL negatively. The variables age was found to be significantly associated with HRQoL of TB patients. The covariance between social and physical, physical and mental, and social and mental well-being latent variables were found to be significant. Posterior predictive p-value for the model fit assessment was 0.045 which was reasonably close to the nominal 5% level.
Discussion
Our study is first of its kind to use BSEM models to explore HRQoL data of newly diagnosed TB patients. Measurement of the HRQoL adds a new dimension to the evaluation of psychosocial variables of TB patients. We also examined the influence of covariates such as age and occupation on HRQoL scores for four physical, mental, social and life style latent variables. Of this, the mental well-being latent variable had most significant effect on HRQoL of TB patients followed by physical and social well-being latent variables.
The study findings highlights the significant association between mental well-being and chronic physical conditions of patients that significantly impact their HRQoL. Findings from a SEM modeling to detect HRQoL changes among cancer patients after invasive surgery found deteriorated physical well-being and improvement in mental well-being [15]. Other previous studies also reported almost similar findings [16–18].
In addition, we found that the covariate age was found to be significantly negatively associated with HRQoL of TB patients. This is logical as age increases, older persons have more health-related problems than the younger ages. This finding was corroborated with the findings from other studies [19–21]. To avoid this multiple effect, there is a need for an early diagnosis to improve the health of the elderly population in the later stages of life.
This study has few limitations; the first is we used the non-informative prior, which is default in the software MPlus that may influence our estimates. The second is that we used HRQoL of TB patients after completion of treatment and we didn’t get data during treatment.
Conclusion
The current study demonstrated the application of BSEM in evaluating HRQoL. The physical, mental and social well-being latent variables were significantly associated with HRQoL of TB patients. This is a first attempt to develop methodology to apply BSEM to study HRQoL of TB patients. This methodology may be used to study precise estimates of HRQoL of TB patients in different time points and also to be applied to study the HRQoL of patients with other diseases.
Supporting information
(DOCX)
Acknowledgments
The authors acknowledge the ICMR-National Institute for Research in Tuberculosis, Chennai for permitting us to use the data for modeling.
Data Availability
All relevant data are within the paper and its Supporting Information files.
Funding Statement
We didn’t receive any funding for this study.
References
- 1.Bauer M, Leavens A, Schwartzman. A systematic review and meta-analysis of the impact of tuberculosis on health-related quality of life. Quality of Life Research. 2013; 22: 2213–2235. 10.1007/s11136-012-0329-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Dujaili JA, Sulaiman SA, Hassali MA, Awaisu A, Blebil AQ, Bredle JM. Health-related quality of life as a predictor of tuberculosis treatment outcomes in Iraq. International Journal of Infectious Diseases. 2015; 31: 4–8. 10.1016/j.ijid.2014.12.004 [DOI] [PubMed] [Google Scholar]
- 3.Staquet MJ, Hayes RD, Fayes PM. Quality of Life Assessment in Clinical Trials. 1998; Oxford University Press, New York. [Google Scholar]
- 4.Scheines R, Hoijtink H, Boomsma A. Bayesian estimation and testing of structural equation models. Psychometrika. 1999; 64: 37–52. [Google Scholar]
- 5.Lee SY, Shi JQ. Bayesian analysis of structural equation model with fixed covariates. Structural Equation Modeling. 2000; 7: 411–430. [Google Scholar]
- 6.Ansari A, Jedidi K, Dube LL. Heterogeneous factor analysis models: A Bayesian approach. Psychometrika. 2002; 67: 49–78. [Google Scholar]
- 7.Lee S, Song X. Bayesian model comparison of nonlinear structuralequation models with missing continuous and ordinal categorical data. BritishJournal of Mathematical and Statistical Psychology. 2004; 57: 131–150. [DOI] [PubMed] [Google Scholar]
- 8.Lee SY. Structural Equation Modeling: A Bayesian Approach. 2007; John Wiley & Sons Ltd, England. [Google Scholar]
- 9.Van de Schoot R, Kaplan D, Denissen J, Asendorpf JB, Neyer FJ, van Aken MAG. A Gentle Introduction to Bayesian Analysis: Applications to Developmental Research. Child Development. 2014; 85: 842–860. 10.1111/cdev.12169 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Jenatabadi SH, Moghavvemi S, Wan Mohamed Radzi CWJB, Babashamsi P, Arashi M. Testing students’ e-learning via Facebook through Bayesian structural equation modeling. PLoS One. 2017; 12: e0182311. 10.1371/journal.pone.0182311 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wan Mohamed Radzi CWJB, Hui H, Jenatabadi SH. Comparing Bayesian and Maximum Likelihood Predictors in Structural Equation Modeling of Children’s Lifestyle Index. Symmertry. 2016; 8:104. 10.3390/sym8120141 [DOI] [Google Scholar]
- 12.Stenling A, Ivarsson A, Johnson U, Lindwall M. Bayesian Structural Equation Modeling in Sport and Exercise Psychology. Journal of Sport and Exercise Psychology. 2015; 10.1123/jsep.2014-0330 [DOI] [PubMed] [Google Scholar]
- 13.Muniyandi M, Rajeswari R, Balasubramanian R, Nirupa C, Gopi PG, Jaggarajamma K, et al. Evaluation of post-treatment health-related quality of life (HRQoL) among tuberculosis patients. International Journal of Lung and Tuberculosis Disease. 2007; 11: 887–892. [PubMed] [Google Scholar]
- 14.Ware JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992; 6: 473–483. [PubMed] [Google Scholar]
- 15.Oort FJ, Visser MRM, Spranjers MAG. An application of structural equation modeling to detect response shifts and true change in quality of life data from cancer patients undergoing invasive surgery. Quality of Life Research. 2005; 14: 599–609. 10.1007/s11136-004-0831-x [DOI] [PubMed] [Google Scholar]
- 16.Atif M, Sulaiman SAS, Shafie AA, Asif M, Sarfraz MK, Low HC, et al. Impact of tuberculosis treatment on health-related quality of life of pulmonary tuberculosis patients: a follow-up study. Health and Quality of Life Outcomes. 2014; 12: 19. 10.1186/1477-7525-12-19 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Rajeswari R, Muniyandi M, Balasubramanian R, Narayanan PR. Perceptions of tuberculosis patients about their physical, mental and social well-being: a field report from south India. Social Science and Medicine. 2005; 60: 1845–1853. 10.1016/j.socscimed.2004.08.024 [DOI] [PubMed] [Google Scholar]
- 18.Guo N, Marra F, Marra CA. Measuring health-related quality of life in tuberculosis: a systematic review. Health and Quality of Life Outcomes. 2009; 7: 1–10. 10.1186/1477-7525-7-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kastien-Hilka TK, RosenKranz B, Sinanovic E, Bennett B, Schwenkgens M. Health-related quality of life in South African patients with pulmonary tuberculosis. PlosONE. 2017; 12: e0174605. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Li CT, Chu KH, Reiher B, Kienene T, Chien LY. Evaluation of health-related quality of life in patients with tuberculosis who completed treatment in Kiribati. Journal of international medical research. 2017; 45: 610–620. 10.1177/0300060517694491 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Oort FJ, Visser MRM, Spranjers MAG. An application of structural equation modeling to detect response shifts and true change in quality of life data from cancer patients undergoing invasive surgery. Quality of Life Research. 2005; 14: 599–609. 10.1007/s11136-004-0831-x [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
(DOCX)
Data Availability Statement
All relevant data are within the paper and its Supporting Information files.