Psychometric Properties of the “Lasher and Faulkender Anxiety about Aging Scale” among Iranian older people

Background Assessing anxiety in the elderly and the factors affecting this phenomenon will help the health care providers to provide appropriate and effective support and health care services for older adults. The aim of the present study was to assess the psychometric properties of the Aging Scale (AAS) among Persian speaking older adults. A sample of 703 community-dwelling older adults was recruited for the study. A 'forward-backward' translation procedure was conducted to develop the Iranian version of the AAS. Conrmatory factor analysis (CFA) and Rasch model were then used for construct validity, and GHQ-12 and MSPSS were utilized for assessing concurrent validity of the AAS.


Background
With the growing years, the senile section of society is exposed to multiple stressors encompassing; decease of dear ones, dysfunctional mobility, nancial dependence, dependence for survival, numerous and chronic illnesses conditions [1][2][3] Additionally, the perceived lack of ability, fear of gradual degradation of the physical self, and the feeling of loneliness induce fear and anxiety. The negative perception of aging leads to enhanced levels of anxiety, which is termed as anxiety of aging [4,5].
The anxiety of aging refers to worrying about the capacity to maintain sound health [6,7] expressed that aging anxiety has fundamental dimensions such as psychological, social, physical, and transpersonal or spirituality. As a result, understanding how the anxiety of aging can in uence behavior and interaction would be a critical research question.
Several factors can in uence the anxiety of aging. The anxiety of aging denotes anxiety about growing old, probably in uenced by culture, gender, age, and socioeconomic status [8]. Studies have demonstrated that individuals have varied perceptions of levels of aging anxiety, such as low negative anxiety and very high levels [9]. Individuals suffering from anxiety of aging have little tendency to socialize with other older adults [9]. Some studies have shown that the anxiety of aging is more re ective in middle age and younger people rather than older adults [10]. Results might be accurate in general.
However, physical, mental, and social alterations would make older people more dependent, consequently leading to higher levels of anxiety of aging among older adults.
Related psychological issues of older adults might lead to numerous psychopathologies such as depression, cognitive impairment and/or death [11], lower life satisfaction and marital satisfaction [12].
Since anxiety is a critical issue, it is necessary to focus on designing relevant interventions to reduce anxiety to foster the quality of life. Therefore, valid and reliable tools are required to investigate the anxiety of aging and potential factors escalating it, especially among the aging population. Most of the scales evaluating anxiety of aging are uni-dimensional, and the item factor loadings are generally small [13,14]. However, concrete diagnoses of the anxiety of aging require several components to be considered, and the scale must be multi-dimensional. The uni-dimensional scales suffer from the limitations of [7]. It seems, one of the common multi-dimensional scales that is used to assess anxiety of aging is Lasher's Anxiety about Aging Scale [AAS] was developed by Lasher and Faulkender in1993 [7]. Since Lasher's AAS or similar tools have not been psychometrically adapted for people who are Persian or talk in Farsi language; Therefore psychometric evaluation of this tools can be classi ed as ve reasons for conducting: 1) psychometric an appropriate tool for assessing the anxiety of the older adults 2) Designing interventions to reduce the anxiety of aging after evaluating the anxiety about aging in the diagnostic evaluation phase 3) Investigating Factors affecting on aging anxiety of older adults 4) Investigating the effect of the gender and age pattern of aging anxiety across the life among the older adults, and nally, 5) Assessing the concurrent validity of AAS.

Participants and procedures
In this cross-sectional design, a sample of 703 community-dwelling older adults was recruited by using a multistate randomized sampling method in the cities of Bukan and Qazvin, Iran, between February and May of 2018. The general population covered by the regional health care centers in urban areas formed the sampling frame. In the rst stage, out of fteen urban regions/clusters, six regions/clusters were selected by using proportionate random sampling. Then, the sample size in each cluster was calculated using the quota distribution for the older adults who had registered in health care centers. The inclusion criteria were being at least 60 years old, lack of cognitive disorders, and having a health record or pro le in the relevant health centers. Additionally, eligibility criteria included being mentally and cognitively able to be interviewed and complete questionnaires. Eligibility screening was administered on the day of enrollment.
The study was approved by the Tabriz University of Medical Sciences Ethics Committee [Ethics Code: IR.TBZMED.REC.1397.304) and the written informed consent form was obtained from all the participants before the study.

Translation procedure
The translation procedure was conducted based on international guidelines to con rm the accuracy of the translation. Several steps were taken as (1)  AAS has four subscales determine overall aspects of the anxiety of aging. The AAS consists of a 20-item measure that produces a summative score based on the answers to all 20 items, including 1. Fear of old people ("I feel very comfortable when I'm around an old person") 2. Psychological concerns ("I fear it will be very hard for me to nd contentment in old age") 3. Physical appearance ("When I look in the mirror, it bothers me to see how my looks have changed with age") and 4. Fear of losses ("I fear that when I am old all my friends will be gone)." Lasher anxiety of aging has high internal consistency, and the four factors explain 50.6 percent of the total variance. Lasher's anxiety of aging scale measures on a 5-point Likert-scale, and items are scored from 1 (Strongly disagree) to 5 (Strongly agree). The higher score indicates a higher level of aging anxiety [7]. General health questionnaire (GHQ_12] The Persian version of GHQ-12 [15] is used to measure mental health. The response was captured on a Likert scale where the scores between 0 to 3 indicating ''less than usual'' to higher scores re ecting "'much more than usual" with a score range of 0 to 36. The higher score indicates lower psychological well-being.
The reliability coe cient reported for GHQ-12 was 0.82, which indicates an acceptable internal consistency.

The Multidimensional Scale of Perceived Social Support [MSPSS]
For measuring perceived social support, we used the Persian version of the MSPSS[16], which had been developed by [17]. It included twelve items in three domains-family, friends, and signi cant others. The response rate was a seven-point Likert-type scale, ranging from 0 to 6-a higher score indicating greater perceived social support. In our study, an estimated reliability coe cient internal consistency of 0.94 was reported.

Statistical analysis
The reliability of the Lasher Anxiety of aging scale was evaluated using internal consistency and item-tototal correlation [corrected for overlap]. Cronbach's α values of ≥ 0.7 and item-to-total correlation of < 0.4 are considered acceptable.
The construct validity of the AAS was assessed by performing con rmatory factor analysis (CFA) and the Rasch model. Diagonally weighted least squares (DWLS) estimator was used as an appropriate estimator for the ordinal data (i.e. Likert-type scale) in the CFA. Further, to assess the model t, the following t indexes were considered: comparative t index (CFI) and Tucker-Lewis index (TLI); root mean square error of approximation (RMSEA) and standardized root mean square residual (SRMR) and chi-square (χ 2 ) test. The acceptable model t was de ned as CFI and TLI > 0.9, RMSEA and SRMR < 0.08 and a nonsigni cant chi-square.
Average Variance Extracted (AVE) and Composite Reliability (CR) were computed using the factor loadings extracted from the CFA. Values > 0.5 and > 0.7 for AVE and CR indicate he evidence of the convergent validity of the AAS.
To further assessment of the AAS, Rash analyses were performed using the partial credit model (18-21). Item t in the Rasch model was evaluated using information-weighted t statistic (in t) mean square (MnSq), and outlier-sensitive t statistic (out t) MnSq. The values between 0.5 and 1.5 were considered as an acceptable t. Additionally, person and item separation indices and reliability were computed for each subscale of the AAS. Permissible values for Person and item separation reliability were > 0.7, while that for the person and item separation indices were > 2. To evaluate measurement invariance across different subsample of older adults with gender and living status, in each item of the AAS, the differential item functioning (DIF) was carried out. A DIF contrast < 0.5 logits indicates a small absence of DIF.
In other to assess the concurrent validity of the AAS, Pearson correlations were computed between the AAS, the GHQ-12 and MSPSS.

Result
In total, 703 older adults have participated in this study. The participant's characteristics are shown in Table 1. The mean age of older adults was 69.4 ± 8.1 years. The majority of the participants were male (59.2%), illiterate (79.7%), married (66.6%) and housekeeper by occupation (40.0%). The corrected item-total correlations exceeded the recommended threshold of 0.40 value, ranging from 0.40-0.80 (Table 2). e DIF contrast across living status = Di culty for participants who were living Alone-Di culty for participants who were living with family MnSq = mean square error; DIF = differential item functioning.
Item-t statistics are shown in Table 2. The In t and Out t MNSQ lied within the acceptable range of 0.5-1.5. For the older adults, the most challenging item was ' I get nervous when I think about someone else making decisions for me when I am old', whereas the most comfortable item was 'I will have plenty to occupy my time when I am old'. Additionally, all items were invariant across gender and living status groups.
The subscale level analysis of the items yielded item separation indices ranged between 3.41 and 10.42, having reliability range between 0.87 and 0.99, and person separation indices ranged between 2.30 and 2.86 with person reliability ranged 0.73 and 0.81 (Table 3). As Table 3 shows, all values of Cronbach's alpha, CR, and AVE were above the acceptable thresholds in four subscales.

Discussion
In the present study, the psychometric properties of the Persian version of the AAS were examined using classical test theory (CTT), item response theory (IRT) approaches. CTT is based on the fact that a person's score is a set of true scores and error scores. However, CTT does not provide a clear picture of the relationship between items and attribute questions with the subject's ability to be measured by these questions. In this CTT, for all abilities (subjects with different abilities), a constant value of measurement standard error is obtained, so that the reliability index remains constant for all ability levels. Of course, the accuracy of measuring each test varies at different levels of ability. This problem is not evident in the case of CTT. Unlike CTT, in IRT, the index of accuracy of parameter estimation is called information function, and its error rate is equal to the standard error of parameter estimation. The results of the CTT indicated that the Persian version of the AAS was internally consistent and had an acceptable four-factor structure using CFA, and acceptable AVE and CR across older adults. In addition to this, the Rasch model further con rmed that the Persian version of the AAS had a four-factor structure and was invariant across gender and living status groups.
Previous studies also supported the four-factor structure of the AAS [7,[22][23][24]. However, our study was the rst study to apply it particularly on the elderly adults.
A fundamental principle in examining the measurement invariance is to compare the individuals from diverse subgroups to detect bias at the item or scale level. The results of the study revealed that the AAS items were invariant across gender and living status among subgroups of the participants. Consistent with previous studies [24,25], the AAS can be conveniently applied for gender groups as well as the living conditions of the elderly in Iran.

Conclusion
In general, the AAS has robust psychometric properties to measure levels of anxiety among the elderly population of Iran. The validity and reliability of the questionnaire and the factor structure of the items have been approved. Therefore, the results of the present study can be applied in diverse circumstances, including mental health research of the elderly with high performance.

Abbreviations
Anxiety About Aging Scale (aas)

Full term Abbreviation
Anxiety about aging scale AAS General health questionnaire-12 items GHQ-12 The Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.