Study design and participants
A total of 954 older adults from the Moravian-Silesian Region who live in a home environment participated in the research. The criterion for inclusion in the research group was that the person had to be aged 65 or older, and they had to be cognitively intact (without diagnosed dementia, and, able to sign an informed consent form). The older adults were approached in all districts of the Moravian-Silesian Region through more than 10 organizations (e.g., seniors clubs, community centers), through libraries, and through the Center for Prevention and Support of Healthy Aging of the Faculty of Medicine, University of Ostrava. The questionnaires were distributed to the participants in both printed and electronic form. According to data from the Czech Statistical Office from 2021, approximately 236,000 people over 65 years of age live in the Moravian-Silesian Region. Our sample included 0.4% of these seniors.
Instrument
To evaluate the QoL, we chose the OPQoL-brief questionnaire [6], which is the short version of the OPQoL-35 and which measures the QoL of people over 65 years of age. The OPQOL-35 questionnaire was developed by Ann Bowling of University College London [16]. A shortened version of OPQoL-brief was later developed by Bowling et al. [6]. The OPQoL-brief consisted of 13 statements, with the participants being asked to indicate the extent to which they agree with each statement by selecting one of five possible options (“strongly disagree,” “disagree,” “neither agree nor disagree,” “agree,” and “strongly agree”). The range in the original version is based on the principle of point allocation (1–5). The items are summed to provide a total OPQoL-brief score. The total score of OPQoL-brief ranges from 13 to 65 and higher scores indicate better QoL. The OPQoL-brief questionnaire also includes a preliminary single item on global OoL. This single item is not scored with the OPQOL; it is coded as very good (5) to very bad (1). Bowling et al. [6] found a highly reliable and valid measure of QoL in old age in the OPQoL-brief scale.
Translation and linguistic validation in four phases: (1) translation, (2) reverse translation, (3) cognitive debriefing, and (4) proofreading. The OPQoL-brief was first translated by two local professional translators into Czech. Then, both translators and the local coordinator discussed the translation and created the first Czech version based on these two independently performed translations. Another professional translator then translated the OPQoL-brief back into English, and the local coordinator compared the reverse translation with the original English version. Any discrepancies were discussed between the translators, and a consensus was reached for the second version of the translation. Two translators and two experts from the field corrected the detected deviations. As a preliminary check, 20 Czech-speaking elderly people (mean age 71.2; 60% women) were then asked to read through the questionnaire with a research assistant and to indicate whether the instructions or any of the items were unclear. All items were deemed clear (cognitive debriefing). The proofreading was done by a proof-reader (native speaker). The final version was then created.
The following questionnaires were used to evaluate other parameters:
-
GDS-15 [22]. A Short Form of Geriatric Depression Scale consisting of 15 questions was developed in 1986 (response: yes/no). Scores of 0–4 are considered normal; 5–8 indicate mild depression, 9–11 indicate moderate depression, and 12–15 indicate severe depression. The Czech version was published by Jirák [23].
-
GAI [24]. Geriatric Anxiety Inventory Scale consists of 20 “agree/disagree” items designed to assess common anxiety symptoms. A sum of these ratings composes a measure of general anxiety symptoms (ranging from 0 to 20), with higher scores indicating greater anxiety [24, 25].
-
SOC-13 [26]. The Sense of Coherence Scale is the short form of the SOC scale and consists of 13 items that comprise three components: comprehensibility (5 items), manageability (4 items), and meaningfulness (4 items). The respondents indicate whether they agree or disagree on a 7-category semantic differential scale with two anchoring responses tailored to the content of each item. The total score can range from 13 to 91, and a higher score indicates higher SOC.
-
RSES [27]. Rosenberg Self-Esteem Scale is a 10-item Likert type scale, with items answered on a four-point scale: from strongly agree to strongly disagree.
Data analysis
To test the psychometric properties of the questionnaire created, we tested the validity (construct validity, discriminant validity, convergent validity) and reliability (internal consistency, test–retest reliability). Also, the psychometric properties test sample size (≥ 500) was met, which can be considered very good [28]. The statistical program SPSS, v. 24.0 was used for data analysis.
Structural validity
At first, we evaluated the single-dimensionality of the scale using the confirmatory factor analysis. The confirmatory factor analysis was performed using the robust maximum confidence value method (MLR), which corrects for abnormal distribution of items. The values of the parameters RMSEA (root mean square error of approximation), CFI (comparative fit index), TLI (Tucker-Lewis index), and SRMR (standardized root mean square residual) and GFI (Goodness of Fit Index) are given for individual models. CFI and TFI values should be close to 1.0 or at least exceed 0.90 [29]. The cut-off value for RMSEA is a recommended value of less than 0.06 [30] or with a strict limit of 0.07 [31]. The lower limit of the RMSEA confidence interval should be close to 0; the upper limit should not exceed 0.08 [31]. The SRMR value should be less than 0.05; however, a value under 0.08 is acceptable. The value acceptable for GFI is ≥ 90 [32]. The model showed borderline values, although the level of statistical significance of the chi-quadrate value was unsatisfactory. For this reason, we decided to perform an exploratory factor analysis, a principal component method, with Varimax rotation. It would help us better understand the factor structure of the OPQoL questionnaire. Prior to factor analysis, the suitability of factor analysis was verified using KMO (Kaiser–Meyer–Olkin measure) and Bartlett sphericity test. The model was tested as a single-factor model and subsequently with a value of 1.0 and greater.
Convergent validity
Convergent validity was verified through Spearman’s correlation coefficient between the OPQoL-brief score and selected scales (GDS-15, GAI, SOC, RSES) and social support. We hypothesized that the QoL (OPQoL-brief) correlates negatively with anxiety [33] and depression [33,34,35] and correlates positively with sense of coherence [36, 37] and self-esteem [38]. Hendl [39] distinguishes the strength of the relationship association according to the value of the correlation coefficient “r” as follows: weak dependence (r = 0.1–0.3), medium dependence (r = 0.3–0.7), and strong dependence (r = 0.7–1). We supposed at least the medium correlation among the analyzed parameters. Correlation analysis between the selected parameters was performed because of the abnormal data distribution (Kolmogorov–Smirnov test) through Spearman’s correlation coefficient.
Discriminant validity
Discriminant validity was assessed based on the OPQOL-brief ability to discriminate between healthy older individuals and older adults suffering from mental and physical illnesses. We hypothesized that the QoL of a person with mental/ physical illnesses was significantly different from that of a person without it [40]. The validity of the measure is supported if the mean of the QoL levels is significantly different between two groups. We tested the difference of the total score of QoL between two groups (older individuals with and without the illnesses) using the independent Wilcoxon test.
Reliability
Internal consistency was determined through Cronbach’s alpha coefficient (α). The acceptable minimum value was set at α > 0.70 [41, 42]. Furthermore, we assessed the Cronbach’s alpha of domains without any items and the correlation of the individual items and the given domain (item-total correlation) with the acceptable minimum r > 0.40 [31].
To evaluate test–retest reliability, the ICC coefficient using two-way mixed model along with 95% confidence was computed. The coefficient of more than 0.70 was considered as excellent stability. Over a period of no longer than 5 days, the questionnaire was completed by 95 older adults to assess the test–retest reliability.