Predictors of quality of life of patients with chronic obstructive pulmonary disease

Background/Aim. Chronic obstructive pulmonary disease (COPD) has a significant impact on quality of life of patients. We investigated which demographic and social characteristics can predict the global quality of life (QoL) of COPD patients. Methods. The patients (n = 288) were divided into three groups according to the stage of disease: Group I = stage 0 ? at risk; Group II = Stages I and II; Group III = stages III and IV. The patients fulfilled a questionnaire related to the demographic and social characteristics and the validated multidimensional questionnaire ? Serbian version of the St. George?s Respiratory Questionnaire (SGRQ). The Student?s t test, ?? test, ANOVA, univariate and multivariate logistic regression tests were used for statistical analyses. Results. In the group I, prevailed the men, employed persons, with a moderate financial status and no family history of COPD. In the group II dominated women, pensioners, with a moderate financial status, duration of illness up to five years, and no family history of COPD. In the group III prevailed women, unemployed persons, a moderate financial status, COPD duration up to 5 years and no family history of COPD. The predictors of the Symptoms score were grades of COPD and duration of the disease, and the predictors of Activity grades of COPD, sex, age and financial status. All variables were found to have a statistically significant relationship in the Impact score in the pre-analyses, were also significant in the univariate regression model. They were age, employement status, financial status and COPD duration. The same predictors that significantly contributed to the explanation of the Impact score, contributed to the explanation of the Total score on SGRQ. In the multivariate regression model, the predictors of the Activity score, Impacts score and Total score were the COPD grade and financial status; only the COPD grade contributed to the explanation of the Symptoms score. Conclusion. Financial status is the most important social factor, and the grade of COPD is the best disease-related predictor of QoL of COPD patients.


Introduction
The World Health Organization (WHO) recognizes that chronic obstructive pulmonary disease (COPD) is of a major public health importance, causing huge economic burden not only to developed countries but even more to low and middle income countries, and in particular, to the vulnerable population 1 . Since COPD is progressive disorder, it has a significant negative impact on quality of life (QoL) of patients. Today, QoL is very important outcome measure in any chronic disease including COPD. After the importance of QoL in COPD patients has been increasingly recognized, several research groups started to study QoL of these patients [2][3][4] in more detail.
It is now known that COPD affects QoL by causing numerous physical, functional, psychological and social stigmata 5 . In order to measure health-related QoL in this chronic disease, several instruments were developed 6 and compared 7 . The most commonly used is the St George's Respiratory Questionnaire (SGRQ) 8 , which was translated and validated into several languages 9 including Serbian 10 .
We investigated the COPD patients-generated data of their social structure and which of these have the greatest impact on their QoL.

Patients
This investigation was performed from July to December 2016. A total of 288 outpatients suffering from COPD in a stable phase of the disease COPD entered this ethically approved study (Decision of Ethics Committee of Primary Health Center "Zemun" N o 03-887/2). The responders were all the patients of Primary Health center "Zemun". The eligible criteria included confirmed diagnosis of COPD according to the Global Initiative for Obstructive Lung Disease (GOLD) criteria 11, 12 . All participants signed the written informed consent. The patients younger than 18 years, those with bronchial asthma, lung cancer or any other respiratory disease that might induce chronic airflow limitation, were excluded. The patients were divided into three groups according to the severity of their disease: Group I -stage 0 (risk group); Group II included stages 1 (mild) and 2 (moderate); and Group III included stages III (severe) and IV (very severe) airflow limitation.

Method
The respondents were asked to fill out a questionnaire including demographic (sex, age) and social characteristics (employment, self-estimated financial status). The patients also fulfilled the validated multidimensional questionnaire -Serbian version of SGRQ 10 , which is designed to measure and quantify health-related status in the patients with chronic airflow limitation. The first part of this questionnaire ("Symptoms") evaluates symptoms (frequency of cough, sputum production, wheeze, breathlessness and the duration and frequency of attacks of breathlessness or wheeze). The second part has two components: "Activity" and "Impacts". The "Activity" section addresses activities that cause breathlessness or are limited because of breathlessness. The "Impacts" section covers a range of factors including influence on employment, being in control of health, panic, stigmatization, the need for medication, side effects of prescribed therapies, expectations for health and disturbances of daily life.

Statistical analysis
All calculations were performed using the Statistical Package for the Social Sciences (SPSS) statistical package, version 21. The baseline quantitative characteristics of patients were expressed as mean, median (M), standard deviations (SD) and rank, while categorical variables were expressed as frequencies and percentages. Statistical significance of differences between the groups was determined using the Student's t test, ANOVA and χ² test for qualitative variables. The univariate logistic regression test was used for variables found significant in pre-analyses, and those that gave statistically significant contribution to the explanation of dependent variable were tested by the multivariate model. Therefore, both univariate and multivariate models were used for prediction.
All statistical tests were considered significant with probability of 0.05.

Results
A majority of patients were females (54.5%). The women also dominated in the group I and group II of patients. The median age for all patients was 62 years, raising steadily from the group I to the group III (48, 64 and 68, respectively). A half of the patients (50.0%) were retirees, while the employed and un- employed contributed almost equally to the second part (23.6% and 26. 4%, respectively). More than one third (36.8%) of patients considered their financial conditions to be moderate. There were also the patients who thought that it was bad (14.9%) or very bad (24.3). Only one quarter (24.0%) of patients considered their financial status to be good or very good. The COPD duration was up to 4 years in a great majority of patients in stage 0, which is (47.4% of all patients), while the duration up to 10 years and more was observed as dominant in patients from the groups II and III. Family history of COPD was denied by 63.6% of all patients. Statistically significant differences were found for all characteristics of patients in relation to the stage of the disease.
Taking all together, in the group I (COPD stage 0), men and employed persons dominated, being of moderate financial status and illness duration up to 5 years with no family history of COPD. In the group II (stage I & II) there were dominatly presented women, pensioners, of moderate financial status with the illness duration up to 5 years and no family history of COPD. In the group III (stage III & IV), prevailed women, unemployed persons, of moderate financial status with COPD duration up to 5 years and no family history of COPD (Table 1).
The average values that the responders achieved at the Symptoms score, Activity score, Impacts score and Total score obtained by using the SGRQ instrument, are shown in Figure 1. The highest value was calculated for the Symptoms score (57.18), followed by the Activity score and Impact score (56.06 and 40.26, respectively). The Total score, which measured global quality of life, was 47.86, thus indicating that our patients had moderate QoL ( Figure 1). All scores were expressed on the scale ranging from 0-100 (0 = the best; 100 = the worst).
In relation to the severity of disease, the groups differed significantly (p < 0.001) regarding all four scores of the questionnaire. The responders in the stadium III and IV had the highest values of all scores (Symptoms score: 58.1 ± 4.0; Activity score: 74.2 ± 29.5; Impacts score: 55.1 ± 22.5; Total score: 61.4 ± 19.5). The women had higher Activity score than men (59.6 ± 31.2 versus 51.8 ± 32.4). The eldest category (aged 71-95 years) had the highest Activity, Impact and Total scores (67.1 ± 32.4, 48.6 ± 25 and 55.7 ± 21.6 respectively). This difference among the categories was statistically significant (p < 0.001). On the contrary, no significant difference among the different categories regarding Symptoms score was found.
The responders having different employment status differed in the Impacts score and Total score (p < 0.05). The retired people had the highest values in both dimensions (Impacts score: 43 ± 25.3; Total score: 50 ± 22.8). The responders of the worst financial status had the highest values of the Activity score (75.9 ± 30.70), Impacts score (58. 8 ± 22.4) and Total score (63.9 ± 20.3), the differences in these scores were statistically significant (p < 0.001). The patients with the longest history of disease had the highest values of the Symptoms score: (58.8 ± 4.2), Impacts score (46.3 ± 27.1) and Total score (52.3 ± 24.2), and these differences were statistically significant (p < 0.001). There was no difference between the patients with and without family history of disease ( Table 2).

OR -odds ratio; CI -confidence interval.
After testing of differences, the variables found to be statistically significant were tested by univariate regression model, and those that were significant were tested by the multivariate regression model. All variables found to have statistically significant relationship in the Impacts score in pre-analyses were also significant in the univariate regression model. The In the multivariate regression model, the predictors of the Activity score, Impacts score and Total score were COPD grade and financial status; these predictors explained 17% of the variance of dependent variable Activity score, 24% of the variance of Impacts score and 23% of the variance of the Total score. In the multivariate regression model, only the COPD grade contributed to the explanation of the Symptoms score (Table 3).

Discussion
Current approach to the investigation of QoL comprises the combination of objective indicators on different life domains with the subjective evaluation given by the individuals, using data on subjective well-being 13 . The subjective evaluation can be the limitation of this approach although such an approach is also recommended for investigations of QoL of patients suffering from COPD 14,15 . According to these recommendations, we used this approach in our investigation of social factors that can predict QoL of COPD patients.
Several multi-country surveys presented at the 2011's European Respiratory Society's Annual Congress 16 , revealed that COPD had the harmful impact on many aspects of quality of life. Our results showed that the financial factor as a social factor emerged in three dimensions that measured quality of life of the COPD patients, while the grade of COPD emerged as a statistically significant predictor of all four dimensions of this questionanaire. Similarly to our results, it was reported that the decrease of financial status was one of the main reasons for patients' feelings of being unable to fulfill their life goals 17,18 . It was reported that the lower social class in terms of the financial status had lower levels of QoL 17,[19][20][21] . A high percentage (39.2%) of our patients described their financial status as bad (14.9%) or very bad (24.3%), and therefore the poverty marked their financial status. Bad financial situation is probably influenced by their employment status since a great majority of our patients were unemployed or retired.
The disease severity (advanced stage of the disease) was also found to have a negative predictive effect on QoL of COPD patients 22 . In our study, the stage of COPD was in a positive correlation with the patients' age -our oldest patients had the most advanced stages of disease. It was shown that older age can also be a predictor of lower QoL of COPD patients 23,24 , although severe COPD may affect negatively QoL of even younger people 2, 7 . In very old people, any chronic disease is the main cause of lower QoL; the gender differences were insignificant 25 .
However, it should be noted that the experience of aging may be influenced by some social and cultural factors that characterize different nations 26 .
As far as gender is concerned, in our study, the women dominated in medium and advanced stages of disease, and their activity score was higher than that of the men. Several studies revealed that the female gender had poorer QoL than the male 27 , especially in relation with a psychological wellbeing 28 . This was true even if women were younger and in the earlier stages of disease 29 .
We conducted this study according to the recommendation that, in order to gain insight into QoL of COPD patients, both demographic and disease-specific impact and general impact of the disease should be used 30 . We found that the most important social factor that predict the QoL of our patients best was their financial status. The other disease-related predictor was the grade of COPD. The knowledge of not only the demographic but also the social characteristics of patient migh help the carers to predict quality of life of their patients. Better QoL of patients could be achieved by higher levels of positive social support, which perhaps may be influenced by efforts of health care providers in this sense. Since the main goal of medical care is to improve and maintain QoL of the patients 31, 32 , we believe that our results might contribute to this ultimate goal achievement.

Conclusion
Financial status is the most important social factor that can best predict quality of life (QoL), and the grade of COPD is the best disease-related predictor of QoL of COPD patients.