Death anxiety measurement using Templer Death Anxiety Scale: a systematic review and validation study

Abstract

Background

Death anxiety, defined as the fear of mortality of oneself and others, is a significant psychological concern among healthcare professionals, particularly nurses, who frequently encounter end-of-life situations. Despite its global relevance, in the Indian context, cultural and religious attitudes toward death, often marked by avoidance of open discussion combined with high patient loads, limited resources, and limited validation studies, underscore the necessity of studying death anxiety in this context.

Aims

To address this gap, two studies were conducted. Study 1 systematically reviewed previous validation studies on Templer’s death anxiety scale; the findings highlighted limitations in the current literature. Informed by these gaps, Study 2 aimed to empirically assess the psychometric properties of a widely used Templer’s death anxiety scale among Indian nurses.

Method

For study 1, a systematic review was conducted following PRISMA guidelines, covering studies published between 1970 and 2024 across five databases. After applying inclusion and exclusion criteria, 19 studies were retained for final synthesis following the STROBE quality checklist. For study 2, a cross-sectional design using cluster sampling was employed to collect data from nurses (N = 1360) working in both public and private hospitals in India. Data were collected using the 15-item TDAS.

Results

Study 1 found that existing validation studies often reported multidimensional structures of the TDAS, contrary to Templer’s original unidimensional model, but were limited by small sample sizes, cultural biases, and reliance on exploratory methods without confirmatory validation. For Study 2, Exploratory and confirmatory factor analyses identified a three-factor structure comprising (1) Fear of Personal Mortality, (2) Philosophical and Temporal Fear/Anxiety, and (3) Fear of Societal Violence and Uncertainty. The 9-item model demonstrated acceptable reliability (Cronbach’s α = 0.71), with CFA indices indicating good model fit (RMSEA = 0.06, SRMR = 0.07, GFI = 0.96, CFI = 0.97, NFI = 0.96). Both of these analyses identified and confirmed a three-factor structure for Templer’s death anxiety scale.

Conclusion

This study strengthens the psychometric foundation of death anxiety assessments in Indian healthcare settings and underscores the importance of culturally relevant validation processes to enhance clinical and research applications.

Introduction

Death anxiety (DA), also known as the fear of death, is a universal and complex human experience. It has been a topic of interest historically, with philosophers, psychologists, scholars, and healthcare professionals exploring its diverse aspects across various disciplines [1]. DA represents a primary existential concern that shapes humans’ thoughts, actions, and choices [2]. Individuals’ views on death are shaped by personal, cultural, social, and philosophical beliefs that influence their behaviors related to death [3]. DA is described as a “negative emotional reaction provoked by the anticipation of a state in which the self does not exist” [4], accompanied by feelings of fear or dread [5].

Theoretical background

Death anxiety is a psychological phenomenon that has been studied extensively through various theoretical frameworks. The Two-Factor Model of DA [6, 7] distinguishes between general anxiety and depression as one factor, and specific life experiences or the fear of death as another. This model suggests that the breakdown of defense mechanisms, such as denial, can lead to both DA and other psychological issues [8]. Terror Management Theory [9] argues that cultural systems buffer DA by offering symbolic immortality and reinforcing self-esteem. Search-for-meaning theories [10, 11] suggest that finding purpose, especially in adversity, helps individuals cope with DA by altering perceptions of death or adopting an “extended self” that helps to reduce anxiety [12]. Studies show those with high DA view death as far off, while those with low DA focus on the present [13, 14], and individuals with an external locus of control tend to experience more DA [15]. Recent approaches also deepen this understanding. Meaning Management Theory [16] emphasizes the role of meaning-centered coping in reducing DA. Existential Positive Psychology [17] reframes DA as an opportunity for spiritual and personal growth, suggesting that confronting mortality fosters meaning-making and authentic living. These advancements highlight DA’s relevance across various domains, such as mental health, decision-making, and cultural dynamics. Together, these theories provide a multifaceted understanding of DA, underscoring its complexity and the need for nuanced interventions.

Measures of death anxiety

DA can be measured through various methods such as qualitative interviews, projective techniques, and standardized self-report scales. While interviews and projective techniques provide rich and individualized insights, they are often limited by subjectivity and lack of standardization, making them less feasible for large-scale research [18].

As a result, self-report questionnaires have become the most widely used tools for measuring DA. These tools are easy to administer, cost-effective, and suitable for large samples. There are various psychometric tools, such as the Revised Death Anxiety Scale (RDAS), which captures cognitive and emotional dimensions of fear [19], the Multidimensional Fear of Death Scale [20], and the Death Concern Scale [21]. These tools have aimed to assess different components of DA, such as fear of the dying process, fear of the unknown, or fear of being forgotten [22]. Thakur and Thakur [23] created a 16-item Likert-type scale specifically for the Indian population, aimed at capturing cultural nuances related to death. Similarly, Dhar, Mehta, and Dhar [24] developed a 10-item dichotomous scale that focuses on environmental and illness-related aspects of DA and is standardized for Indian adults aged 25–55 years. Among these tools, the most widely used scale is the Templer Death Anxiety Scale (TDAS) developed by Donald Templer in 1970 [25]. TDAS consists of 15 true-false items aimed at assessing conscious anxiety related to death, which reflects a wider range of life experiences [26]. The scale was designed as unidimensional, positing that DA is a single, coherent construct. Sample items include statements like “I am very much afraid to die” or “The sight of a dead body is horrifying to me.”

Despite the availability of Indian-developed tools, the TDAS was selected for the current study for several reasons. It has a strong foundation in the literature and has been used in cross-cultural comparisons, making it suitable for studies seeking broader generalizability [25]. Its widespread use allows for greater comparability with existing global findings and meta-analyses [27]. While Indian-developed tools offer cultural specificity, they have not been as extensively validated or psychometrically evaluated across varied professional groups, such as nurses. In contrast, the TDAS has demonstrated reliability and validity in diverse healthcare settings and lends itself well to psychometric analyses such as exploratory and confirmatory factor analysis [28].

Death anxiety among nurses

Exposure to the processes accompanying the death of others makes individuals conscious of their mortality, giving rise to anxiety [29]. Among nurses who are exposed to death as a part of their occupation, DA is a significant challenge [30]. DA among nurses can often negatively impact on the delivery of high-quality care to terminally ill patients [31]. This anxiety often arises from exposure to illness, trauma, and violence, leading to adverse attitudes toward the care of dying patients [31]. Factors such as personal anxiety, occupational stress, and burnout have been linked to heightened levels of DA among nurses [32]. Consequently, it is imperative for nursing policymakers to address these factors and offer support mechanisms to assist nurses in managing their DA.

Nurses’ understanding and experience of death have been studied in different cultural contexts. Research has consistently highlighted the role of these cultural factors in shaping nurses’ attitudes toward mortality. In England [33], regular exposure of hospice nurses to terminally ill patients allowed them to become more familiar with the dying process, resulting in lower levels of DA compared to emergency nurses who may not have the same opportunities to acclimatize, leading to heightened anxiety. In contrast, in Israel [34], oncology nurses observed moderate DA, despite their extensive experience with dying patients. This suggests that exposure to death alone may not be enough to reduce anxiety and that other factors, such as personal, cultural, or religious beliefs, could play a significant role. Similarly, in the USA [35] oncology nurses, with prolonged exposure to dying patients, exhibited more favorable attitudes toward death, suggesting that professional experience may contribute to desensitization to death-related fears. Moreover, Greek renal nurses who had received specialized palliative care education experienced reduced DA and were more comfortable discussing death [36]. Despite global literature on DA among nurses, studies within the Indian context remain limited. Given that cultural factors significantly influence death-related attitudes, as evidenced above, there is a need for context-specific validation of assessment tools. Although TDAS is widely used across diverse populations, its psychometric properties have not been examined among Indian nurses. The current study also addresses this gap by evaluating the validity and reliability of the TDAS in this population.

Rationale

The healthcare sector is recognized as one of the most stressful domains of work [37]. Within this sector, nurses constitute the largest segment of the healthcare workforce and maintain the closest and most sustained proximity to patients [38]. Importantly, India is currently experiencing a shortage of approximately 2 million nurses, resulting in a ratio of one nurse per 483 patients [39]. The World Health Organization (WHO) recommended the norm of 3 nurses per 1,000 people [40]. While the situation in India is exacerbated by the fact that there are only 1.7 nurses per 1,000 population, according to data from the Union Health Ministry, which is markedly below the prescribed recommendation by WHO. Specifically, in Bihar, the situation is even more critical, with the nurse-to-population ratio among the lowest in the country [41]. The shortage of staff nurses against sanctioned strength ranged from 18% (Patna) to 72% (Purnea) and is compounded by the fact that more than 40% of sanctioned nursing positions in government hospitals remain vacant, as highlighted in a recent report by the Comptroller and Auditor General (CAG) of India [41]. The strain on the existing nursing workforce in Bihar is further amplified by inadequate infrastructure [42] and the stigma attached to the nursing profession [43, 44], which compromises healthcare delivery. In addition, the practices in Bihar are heavily influenced by the Indian philosophical tradition, where it is believed that the aim of death is to achieve moksha, which means getting free from suffering and illness. The smoke from the funeral pyre that rises up into the sky is believed to take the soul of the deceased to integrate with god [45].

As a consequence, the burden on existing nurses is intensified since they have to provide care to more patients than can reasonably be expected at a single time, with no rest intervals. This increased workload often leads to poor-quality of patient care. For instance, reports have indicated issues such as missed care, incomplete tasks, and inadequate documentation of patient care [46]; this also places immense physical and emotional pressure on nurses, contributing to a higher risk of mental health challenges. These challenges not only widen the gap in achieving the WHO-recommended ratio but also hinder progress toward meeting Sustainable Development Goals (SDGs) related to health and well-being. This underscores the importance of studying nurses in Bihar to develop evidence-based interventions aimed at enhancing mental health among nurses. Given the work-related stress and frequent exposure to patient deaths, addressing their mental health is a growing concern. DA is one of the indicators of poor mental health and is increasingly recognized as a significant issue among healthcare professionals, particularly nurses, who routinely encounter death and dying in their line of work [47–49]. Despite its importance, DA is less studied in the Indian healthcare workforce, leaving a gap in its understanding in this context. Therefore, the current study aims to study the factor structure of DA among Indian nurses to ascertain if the existing widely used instrument for measuring DA, the TDAS [26], is a reliable and valid tool for this population. This aim gains prominence in the light of literature detailed above which elucidates that this scale has a multi-factor structure, which is contrary to the originally proposed unidimensional structure by Templer [26]. Considering nurses’ frequent encounters with death, the lack of studies on DA among nurses in India is a gap to be addressed.

Hence, this paper has 2 major objectives and its associated hypotheses divided across two studies.

Study 1: Objective 1

To systematically review the factor structure and validation studies of the TDAS across different populations and identify gaps in the existing literature.

Study 2: Objective 1

To explore the underlying factor structure of the TDAS among Indian nurses using Exploratory Factor Analysis (EFA).

Hypothesis 1

TDAS will demonstrate good construct validity among Indian Nurses.

Study 2: Objective 2

To validate the factor structure identified through EFA by using Confirmatory Factor Analysis (CFA) and test its goodness-of-fit in the Indian nursing population.

Hypothesis 2

TDAS will demonstrate a good model fit in CFA.

Method

Study 1

Procedure

The present systematic review of empirical quantitative studies examining the validity and factor structure of the TDAS followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. The inclusion criteria were: (a) studies published from the year 1970, since the TDAS was published in 1970 by Templer; (b) studies examining the factor structure of TDAS. The exclusion criteria were: (a) qualitative research, (b) grey literature, and (c) studies that did not provide adequate information regarding the methods employed. The search was conducted in the following databases: Web of Science, ScienceDirect, PubMed, EBSCO, and Google Scholar. The search strategy combined key terms using Boolean operators (“Templer Death Anxiety Scale” OR “TDAS”) AND (“factor structure” OR “exploratory factor analysis” OR “confirmatory factor analysis”), (“Death Anxiety Scale” OR “Templer DAS”) AND (“validity” OR “validation” OR “psychometric properties”), (“Templer DAS” OR “Death Anxiety Scale”) AND (“reliability” OR “internal consistency” OR “Cronbach’s alpha”), (“Death Anxiety Scale” AND “nurses”) OR (“TDAS” AND “healthcare professionals”), (“Death Anxiety Scale” OR “Templer DAS”) AND (“cross-cultural validation”). Filters were applied to restrict results to peer-reviewed journal-published articles. Key terms used were “factor structure of Templer death anxiety scale”, “Templer Death Anxiety Scale”, “Death Anxiety Scale validation”, “Factor structure of TDAS”, “Psychometric properties of death anxiety scale”, “Cross-cultural validation of TDAS”, “Exploratory Factor Analysis of TDAS”, “Confirmatory Factor Analysis of TDAS”, “Reliability and validity of TDAS”, from 1970 to November 2024. Among the identified studies, the authors manually went through the references to determine if there were any relevant studies meeting the above criteria, to be added for analysis. The quality of the included papers was checked using the STROBE checklist, most papers did not include any explanation about efforts to address any sources of potential bias. Therefore, this criterion was not included because most papers did not explicitly address this concern. Figure 1 (PRISMA flowchart) shows the selection of the included studies. The final number of studies retained for analysis were N = 19.

Fig. 1.

PRISMA flow diagram of systematic review

Results

From the 19 selected studies, it was gathered that previous research has extensively examined the validity and factor structure of the TDAS across diverse populations and multiple countries (see Table 1). Despite spanning from the 1980s to 2018, there are intermittent gaps in studies of factor structure analysis of TDAS, for example, very few studies appeared between 1996 and 2002, and again between 2002 and 2011.TDAS’s factor structure has been studied most in Iran. Most studied samples were students (N = 7), nurses (N = 2), patients (N = 7), and war veterans (N = 2). Studying nurses, patients and war veterans can be understood in the light of the fact that they have seen illness/death closely. Students, on the other hand, are usually a population that is easily accessible and hence are likely to be studied more. The minimum factor identified was 1 (N = 1), and the maximum was 6 (N = 1). Most studies found a 4-factor structure (N = 8). Most studies used Principal component analysis (PCA) with varimax rotation (N = 15). Notably, they predominantly relied on conventional statistical methods for factor analysis, overlooking newer techniques that may yield more robust results. Moreover, CFA, a crucial analytical approach, was often omitted in prior studies. Collectively, the existing findings suggest that the TDAS is a multidimensional instrument. This is different from the single-factor structure proposed by the original author. Importantly, the systematic review revealed that only one study by Royal and Elahi [50] identified a single factor. These results paved the way for study 2, as no investigations specific to the Indian population were found.

Table 1.

Studies that have validated Templer Death Anxiety Scale (TDAS)

Author (year)	Sample	Factors extracted	Total variance	Reliability	Country	Extraction method	Rotation method	Software used	TDAS version	CFA conducted
[8] Lonetto et al., (1979)	5 different student samples	4	64%	-	Canada and North Ireland	PCA	Varimax Rotation	-	15 Items; T/F	no
[79] Warren and Chopra (1979)	Students and Faculty	3	38%	0.69	Australia	PCA	Varimax Rotation	-	15 Items; T/F	no
[86] Martin (1982-83)	Canadian Nurses	5	41.2%	-	Canada	Principal Axes Analysis	Varimax Rotation	-	15 Items; T/F	no
[6] Gilliland and Templer (1985-86)	general population and psychiatric patients	4	63%	-	US	PCA	Varimax Rotation	-	15 Items; T/F	no
[91] Levin (1989)	Male and female (Age: 17–46)	4	35%	0.72	Canada	-	Oblique Rotation	Big.Iif& factor analysis computer program	15 Items; T/F	no
[80] Abdel-khalek et al., (1993)	Students	5	53%	0.70–0.73	Egypt	PCA	Varimax Rotation	-	Arabic version	no
[92] Lester and Castromayor (1993)	Students	6	-	-	Philippines	PCA	Varimax Rotation	SPSS-X	15 Items; T/F	no
[83] Saggino & Kline (1996)	Students	3	40.30%	0.75	Italy	PCA	Direct Oblimin	-	15 item Italina version	no
[93] Tomás-Sábado and Gómez-Benito (2002)	Students	4	53.90%	0.73	Spain	PCA	Varimax Rotation	NA	15 Items ; T/F (Spanish version)	no
[50] Royal and Elahi (2011)	Cancer patients	1	22.60%	0.70–0.74	US	Rasch measurement–based PCA	-	Winsteps measurement software	15 items; 5-point Likert scale	no
[94] Ali and Behrooz (2011)	Students	5	51.40%	0.75	Iran	PCA	Varimax Rotation	SPSS	15 Items; T/F	no
[95] Nia et al. (2014)	warfare veterans (male)	4	40.60%	0.89	Iran	PCA	Varimax Rotation	SPSS 16	Extended Persian version (51 item); Likert scale	no
[49] Nia et al., (2016)	warfare veterans	4	37.51%	0.73–0.89	Iran	PCA	Varimax Rotation	SPSS 16	Extended Persian version (51 item); Likert scale	yes
[84] Soleimani et al. (2016a)	Family care givers of cancer patients	3	60.38%	0.78–0.81	Iran	PCA	Varimax Rotation	SPSS 22	15 items; Likert scale (Persian version)	yes
[85] Soleimani (2016b)	cancer patients	3	63.93%	0.81	Iran	PCA	Varimax Rotation	SPSS 22	15 items; Likert scale (Persian version)	yes
[52] Yang et al. (2017)	cancer patients	4	63.78%	0.81	China	PCA	Varimax Rotation	SPSS	15 items; Likert scale (Chinese version)	no
[51] Soleimani et al. (2017)	Heart patients	3	68.11%	0.49–0.86	Iran	PCA	Varimax Rotation	SPSS 22	15 item; Likert scale (Persian version)	yes
[82] Nia et al., (2017)	end stage renal patients	3	46.30%	0.83	Iran	Maximum Likelihood	Oblique Rotation	SPSS 22	Extended Persian version (51 item); Likert scale	yes
[27] Dadfar et al. (2018)	Iranian nurses	4	51.76%	0.60	Iran	PCA	Varimax Rotation	SPSS 23	15 Items; T/F (Farzi version)	no

Note. “-” denotes information not available. PCA = Principal Component Analysis

Study 2

Design and sample

A cross-sectional study employing a cluster sampling design was conducted in Bihar, India. Patna district of Bihar was selected for data collection as it has a healthy mix of both private and public hospitals and all patients from other parts of the state are referred here. A list was made by the authors of all the hospitals in the Patna district which was not publicly available through any government body, from which private and public hospitals were randomly selected using the fishbowl method, wherein each hospital was considered as one cluster. Data were collected from all employed nurses within each selected hospital. The inclusion criteria for participant selection included the following: (a) Nurses currently employed in Patna district (b) be able to read and write in English, while the exclusion criteria were: (a) any other health care staff, e.g., paramedical specialties and nursing students (b) Nurses with a current or past diagnosis of any psychiatric disorder. After removal of incomplete responses and outlier data, the sample size for the current study was N = 1360.

Tool

Templer’s Death Anxiety Scale- This scale was developed by Templer [26]. It consists of 15 items. Six items were keyed “false,” and nine were keyed “true.” Answers conforming to the key equals a score of one. The score ranges from 0 to 15. Its test-retest reliability is 0.83, and multiple studies have reported Cronbach’s alpha coefficients ranging from 0.7 to 0.9. Items include: “I am very much afraid to die,” “I fear dying a painful death,” and “Sight of a dead body is horrifying to me.” Higher scores indicate higher DA. TDAS is one of the most widely used questionnaires to assess DA and has been adapted to different cultures [see 51, 52].

Procedure

After ethical approval for the study was provided by the Institute’s ethics committee (Letter No: 2065/IEC-AIIMSRPR/2021), permission for data collection was sought from all the hospitals’ management. Every nurse in each hospital was then approached in person, and rapport was established. Participants were provided with a brief description of the study and were informed about their rights to refuse to participate or withdraw at any time. Informed consent was obtained from all participants, after which the questionnaire was provided to them. Data were collected in person using the traditional paper and pencil method between January 2022 and July 2023. A total of (N = 1360) responses were considered for analysis after removal of incomplete responses and outliers. This study was conducted in accordance with the principles outlined in the Declaration of Helsinki. All the ethical guidelines set by the American Psychological Association [53] were followed.

Statistical analysis

To meet the objectives of the Study 2, the internal consistency of the TDAS scale was calculated using Cronbach’s (α) to assess the reliability. To establish the construct validity of the DA scale, a combination of Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) was performed to investigate and confirm the underlying factor structure. EFA was performed using FACTOR software version 12.04.02 [54, 55]. Factor software is robust and comprises all the best practices for conducting EFA [56, 57]. CFA was conducted through the R programming language and its “laavan” library [58]. The Solomon method [59] was applied to split the 1360 samples into two equivalent subgroups of 680 each. To assess the equivalence of the subgroups, the ratio communality Index (RCI) was found to be 0.98, which is closer to 1, representing an equivalent sample subgroup. The minimum sample size was calculated using a Monte Carlo simulation for each analysis, for EFA, CFA. Monte Carlo is the most robust and rigorous way to determine the power of specific sample sizes or to decide the appropriateness of a sample size for a given power level [60, 61]. The minimum sample size for EFA and CFA was found to be 300, it also passed the recommended minimum sample criteria for EFA, i.e. 10 participants per item [62].

Results

Sample characteristics

The final sample comprised 1360 working nurses. The average age of nurses was 35 years (SD = 8.5 yrs), and average experience was 15 years (SD = 6.8 yrs). The mean of DA was 7. The sample was predominantly female, with only 13 male participants, reflecting the skewed gender distribution in this profession. This is since, historically, nursing has been seen as an extension of domestic roles, aligning with traditional gender expectations that associate caregiving and nurturing responsibilities with women. Such perceptions reinforce the idea that women are more naturally suited for the emotional and physical demands of caregiving, which may contribute to their higher representation in this field [63]. This is similar to global trends, wherein the nursing profession is predominantly a female profession. These participants were not removed from the sample, as the aim was to study DA as it is present in the population. In the current sample, 41% of nurses work in private hospitals and 59% work in public hospitals. 57% have sometimes faced language problems while dealing with patients.

Exploratory factor analysis

Templer proposed a 15-item, unidimensional scale of DA with two response categories (true and false). However, various studies present contradictory findings that this scale has multiple (3,4,5) subfactors [64]. Therefore, EFA was carried out for this population to measure the construct validity of the scale in this particular population and also to check the unidimensionality of the scale. To estimate whether the common variance required a factor analysis, the Kaiser-Meyer-Olkin index (KMO) and Bartlett’s statistic were used. Values over 0.50 were considered suitable for KMO, while p values 0.01 were considered statistically significant for Bartlett’s statistic. An adequacy analysis of the items was carried out according to the values of the Measure of Sampling Adequacy (MSA), where factors lower than 0.50 indicated that the corresponding item did not measure the same construct as the rest, and should therefore be removed from the factorial solution [65]. The adequacy of the factorial solution was evaluated through the different fit indexes enumerated further. An expected mean value of RMSR for an acceptable model is 0.0383 (Kelley’s criterion [66], page 13); RMSEA values between 0.05 and 0.08, are considered to be a good fit. CFI values of 0.95 or higher and GFI values higher than 0.90 [67] and AGFI values higher than 0.8 were considered indicators of a good fit [68].

An initial EFA was performed for all the items without specifying a specific number of subscales. A matrix of polychoric correlations was used; factors were extracted by Robust Diagonally Weighted Least Squares (RDWLS) and oblique rotation direct oblimin [57, 68]. Horn’s parallel analysis (PA) has emerged as one of the most accurate and recommended dimensionality assessment techniques for continuous data [69, 70], and is considered better than principal component analysis [71]. Hence used to establish the number of factors to be retained, and the consistency of the retained factors was calculated. EFA results showed good adequacy for the data, with KMO = 0.717, and a statistically significant Bartlett’s value (p < 0.001). However, the fit analysis and the factor loadings of the items suggested that a few items should be removed. Each item was removed one by one until it gave statistically good results. Total number of items removed was six (items 3,5,6,7,11,12). Finally, parallel analysis recommended a three-factor solution. Final EFA was carried out with the remaining 9 items and three factors-subscales. In this EFA, the KMO values and Bartlett’s statistic indicated a better fit (KMO = 0.641); Bartlett (p ≤ 0.001), and the MSA analysis did not suggest the removal of any item. Table 2 shows the three-factor scale solution, which showed that the first factor accounts for 40.14% variance, the second factor accounts for 22.04% of the variance, and the third factor accounts for 17.86% of the variance, according to the parallel analysis. Using the method of mean of random variance extraction, three factors were retained as the real-data percentage of common variance was found higher than the mean of the PA-MRFA’s random dataset. Table 3 shows the factor loadings (after oblique rotation) of this model. All items received loadings over 0.70, except item 1, which received the lowest factorial loading (0.577). According to this analysis, items 1, 2, and 4 comprised the first factor, which corresponded to a dimension-subscale we named as “ Fear of Personal Mortality”; items 8, 9, and 10 were included in a second factor-subscale “ Philosophical and temporal fear/anxiety”; items 13,14 and 15 were included in the third factor “Fear of societal violence and uncertainty”. Unidimensionality was assessed through the indexes Unidimensional Congruence (UniCo), Explained Common Variance (ECV), and Mean of Item Residual Absolute Loadings (MIREAL) [72]. Data can be suggested as essentially unidimensional for UniCo values higher than 0.95, ECV values higher than 0.85, or MIREAL values lower than 0.30 [72]. Values displaying closeness to unidimensionality were obtained from UniCo = 0.849, ECV = 0.654, and MIREAL = 0.359. These values show that there is no unidimensionality. The fit values for this model were: RMSEA = 0.07, CFI = 0.974, GFI = 1.0, and AGFI = 1.0, which indicated a good fit for the three-factor model. The RMSR was 0.044, which is close to Kelley’s criteria of acceptable fit. Thus, hypothesis 1 was accepted.

Table 2.

Parallel Analysis (PA) based on minimum rank factor analysis (Polychoric correlation)

Variable	Real data % of variance	Mean of random% of variance	95 percentiles of random % of variance
1	40.14*	22.62	26.94
2	22.04*	19.39	22.52
3	17.86*	16.39	18.85
4	7.30	13.58	15.48

Note. Only the first 4 factors are shown. The factors with stars are retained as the real-data percentage of common variance is higher than the mean of the PA-MRFA’s random dataset

Table 3.

Factor loadings and their reliability score (N = 680)

Items	Rotated factor loadings
	Factor 1	Factor 2	Factor 3
1. I am very much afraid to die.	0.57
2. The thought of death seldom enters my mind.	− 0.73
3. I dread to think about having to have an operation.	0.71
4. I am often distressed by the way time flies so very rapidly.		0.88
5. I fear dying a painful death.		0.81
6. The subject of life after death troubles me greatly.		0.69
7. I shudder when I hear people talking about World War 3.			0.77
8. The sight of a dead body is horrifying to me.			0.73
9. I feel that the future holds nothing for me to fear.			− 0.79
Orion reliability	0.74	0.83	0.86
Total Omega coefficient of the scale	0.76
Greatest Lower Bound to Reliability (GLB)	1.03
Standardized Cronbach’s alpha coefficient	0.76

Note. Extraction Method: Robust Diagonally Weighted Least Squares (RDWLS); Rotation Method: Direct Oblimin

Serial numbers are for display only, not the original TDAS item numbers

Table 2 shows factor consistency, which was evaluated by using the ORION coefficients (Overall Reliability of fully Informative prior Oblique N-EAP scores) [73]. It shows ORION coefficient values were: 0.745 for factor-subscale 1, 0.837 for factor-subscale 2, and 0.868 for factor 3, showing adequate consistency, with values higher than 0.80. Additionally, the reliability of the questionnaire was analyzed by calculating the Omega coefficient (ω) and Greatest Lower Bound (GLB) [74]. The total Omega coefficient was 0.764, while the GLB was 1.03; the standardized Cronbach’s alpha coefficient was 0.763 for the total scale. These values indicate that Templer’s DA scale is a reliable measure of DA.

Confirmatory factor analysis (CFA)

CFA was performed on the remaining 680 participants to confirm the EFA, which uncovered a three-factor structure. Following the previous studies, multiple fit indices were used to assess model fitness. Table 4 shows the result of CFA for the 3-factor structure (nine-item scale) as well as the reference levels used by various indices. As shown in Table 4, each of these indices confirmed the model fitness except the chi-square value, which was significant. This may be attributed to a large sample size, as it can lead to erroneous results of chi-square [62]. Thus, the CFA result cross-validated the 3-factor structure found in EFA. Thus, hypothesis 2 was accepted.

Table 4.

Model fit indices of 3-factor model of Templer’s Death Anxiety Scale after confirmatory factor analysis (N = 680)

Number of factors	3 factor model	Reference level
Absolute Terms
X2 Chi-Square	133.3
(p-value)	0.01	> 0.05
Root Mean Square Error of Approximation (RMSEA)	0.06	< 0.07
Standardized Root Mean Square Residual (SRMSR)	0.07	< 0.08
Goodness of fit Index (GFI)	0.96	> 0.95
Weighted by the number of estimated parameters
Adjusted Goodness of Fit Index (AGFI)	0.84	> 0.8
Parsimony Goodness of Fit Index (PGFI)	0.81	> 0.8
Comparison to the baseline model
Comparative Fit Index (CFI)	0.97	> 0.95
Normed Fit Index (NFI)	0.96	> 0.95

Note. Reference Level by Schumaker & Lomax (2004)

Reliability and validity of the Templer’s DA scale

The Cronbach’s alpha coefficient for TDAS as a whole for the 15-item scale was 0.76, indicating its good reliability. Instruments with a Cronbach’s α value of 0.60 or greater are considered to have satisfactory internal consistency [75, 76]. Cronbach’s alpha was also calculated for the 3-factor TDAS for the 9-item scale, which was 0.71. It was found that the reliability of the three dimensions was acceptable (⍺ = 0.68 for factor 1, ⍺ = 0.65 for factor 2, ⍺ = 0.69 for factor 3). Lower reliability of individual factors compared to the total scale (α = 0.71) is expected in multi-factor scales, as subscale alphas are often lower due to fewer items, and this does not indicate poor reliability [77]. The convergent validity of this three-factor model was assessed through Average Variance Extracted (AVE) of 0.58 (factor 1 = 0.57, factor 2 = 0.61, factor 3 = 0.58), which is above the recommended value of 0.50 by [78]. Hence, convergent validity for TDAS was established.

Combined discussion and implications

Study 1 aimed to systematically review and summarize previous validation studies of TDAS across different populations. While the objective of study 2 was to establish the validity of the TDAS among Indian nurses. The systematic review identified 19 studies investigating TDAS factor structure. It also highlighted that no research has verified the factor structure of the scale in the Indian population. An important finding from the review is the inconsistency in the factor structures reported across different cultural contexts. Most supported a multidimensional structure, identifying between two and five factors. Notably, Royal and Elahi [50] was the only study to report a single-factor model. Overall, this suggests that while TDAS items are interrelated, they cluster into distinct subdimensions rather than forming a single latent construct. The review also revealed that most previous studies primarily relied on EFA to explore the factor structure of TDAS, with limited use of CFA to confirm these structures. Only a few studies applied CFA or other advanced psychometric techniques, leaving uncertainty regarding the stability and replicability of the factor models. It was found that the selected studies converge on identifying specific factors that encapsulate DA. Based on this review, these factors were represented as cognitive and affective responses to death, physiological changes associated with illness and dying, awareness of temporal progression, and experiences of pain and stress. These conclusions are also echoed by Lonetto and Templer [8]. The review also found significant cultural differences in the reliability and validity of TDAS. While most studies reported high internal consistency (Cronbach’s alpha: 0.76–0.92), some studies noted that certain TDAS items exhibited cultural biases, leading to cross-loadings or weak item fit [79, 80].

These inconsistencies raise questions about measurement invariance across populations, suggesting that cultural adaptation may be necessary when using TDAS in different contexts. The gaps identified in Study 1, including the lack of validation studies in the Indian population, the limited use of CFA in prior research, and inconsistent multidimensional factor structures across different cultural contexts, informed the design of Study 2. To address these gaps, Study 2 aimed to validate the TDAS specifically among Indian nurses, employing both EFA and CFA to establish a robust three-factor structure that reflects culturally relevant dimensions of death anxiety. This approach ensured that the scale could capture both personal and societal aspects of death anxiety, providing a reliable tool for research and clinical use in this population. Having a valid and reliable tool is an important step in carrying out studies that can be used to understand DA in the affected Indian population, particularly nurses, because they are directly exposed to death. Study 2 results demonstrated that the TDAS is a reliable and valid instrument for assessing DA within the Indian nursing population. Given the increasing emphasis on rigorous validation methods in psychometric research [81], this study addressed this gap by employing both EFA and CFA to test the robustness of TDAS in this population. EFA revealed a three-factor structure, confirming that the dimensions of DA are consistent with previous findings [51, 79, 82–85]. Based on the factor loadings, the first factor was named “fear of personal mortality”(consisting of items 1, 2, and 4), reflecting anxiety related to the individual’s own death. The second factor was called “philosophical and temporal anxiety” (consisting of items 8, 9, and 10), capturing concerns about the existential and temporal aspects of death. Finally, the third factor, named “fear of societal violence and uncertainty” (consisting of items 13,14,15), is related to anxiety emerging from broader societal threats and the unpredictability of death. The identification of these distinct dimensions will provide insights into how DA is interpreted across different aspects of nurses’ personal, philosophical, and societal experiences. Among different nursing populations, both four and five-factor structures have been found in Iran [27] and Canada [86], respectively. The findings of the present study are similar to Dadfar’s [27] investigation that identified four factors of TDAS to be “death anxiety”, “fear of the future”, “time passing”, and “thoughts of the death”; where the same items as identified in the present study contribute to a large portion of the variance in both investigations. This is different from Martin’s [86] five factors of “death anxiety denial”, “general death anxiety”, “fearful anticipation of death”, “physical death fear”, and “fear of catastrophic death” wherein a large portion of variance was explained by death anxiety denial, not found in the present study. These differences in factor structures may be due to cultural and contextual differences within nursing populations: Dadfar’s study was conducted with Iranian nurses, Martin’s with Canadian nurses, and our study with Indian nurses, who experience different societal norms, healthcare system challenges, patient mortality exposure, and workplace environments. Methodological differences across studies, such as variations in sample size, EFA extraction and rotation methods, and the use of CFA, may also contribute to the observed differences in factor structures. From The three factors identified in the current study enumerated above, societal violence and uncertainty” align with Langs’s [87] typology of death anxiety. Factor 1 “fear of personal mortality” corresponds to “existential death anxiety”, reflecting concerns about personal death and the end of existence. Similarly, factor 2 “philosophical and temporal anxiety” is also linked to “existential death anxiety”, as it reflects anxiety related to the finite nature of life and existential questions surrounding the afterlife. Factor 3 “fear of societal violence and uncertainty” aligns with “Predatory death anxiety”, which is triggered by external threats, such as societal violence or instability that challenge one’s survival. This dimension is particularly relevant to Indian nurses due to their exposure to high-intensity hospital settings, large patient loads, and occasional workplace hazards, which may heighten anxiety about societal threats beyond personal mortality. These findings highlight the multidimensional nature of death anxiety, encompassing both personal and societal dimensions of existential concern. In the current investigation, CFA further supported the scale’s validity, confirming that the three-factor model identified through EFA adequately fits the observed data. Previous studies, though performed on different populations, have revealed a three-factor structure for TDAS. These, however included dissimilar items forming the factors. The closest of these factors were Factor 2 and Factor 1 from investigations on heart [51] and cancer patients [85]. Other studies did not reveal any clear similarities or common factors. It can hence be gathered that DA has different expressions in different populations.

The current findings have implications for researchers who can conduct comparative cross-cultural investigations of the three-factor structure for TDAS. It can also be used within different samples in the Indian context to verify the three-factor structure. This three-factor TDAS model provides greater insight into the complex nature of DA among Indian nurses, offering a deeper understanding of how personal mortality fears, philosophical/existential concerns, and anxiety about others’ deaths uniquely manifest in this cultural and occupational context. By distinguishing these dimensions, the model enhances theoretical knowledge for designing culturally sensitive interventions, support programs, and training modules, thereby allowing hospital administrators and mental health professionals to address specific aspects of DA rather than treating it as a single construct. Interventions can be tailored to each factor that has been identified in this study specifically for the Indian Population. For example, for Factor 1 (“Fear of Personal Mortality”), cognitive-behavioral therapy (CBT) techniques such as cognitive restructuring and exposure tasks can help reduce anxiety. For Factor 2 (“Philosophical and Temporal Anxiety”), Acceptance and Commitment Therapy (ACT), including mindfulness and values clarification, can help nurses accept existential concerns and commit to meaningful actions. For Factor 3 (“Fear of Societal Violence and Uncertainty”), resilience-building strategies and coping skills interventions can address anxieties related to external threats and occupational stress. These are supported by [88, 89]. Among practical implications, these targeted interventions will enable lowering specific dimensions of DA that have been identified above. Moreover, since the scale has been found to be valid on Indian nurses, mental health professionals can also use the scale for early screening to identify individuals at risk of moderate to high DA, enabling timely interventions to improve their coping styles and enhancing well-being. Hospital management can use this information for staff wellness programs, monitoring, and addressing DA among healthcare workers. By identifying and managing DA early, institutions can protect nurses’ mental health, prevent burnout, and enable high-quality patient care [90].

Limitations and future directions

We have tried to address the objectives adequately; however, there are some limitations to this investigation. An important consideration is the need for broader validation across diverse populations. Given that death anxiety may manifest differently based on cultural, professional, and personal experiences, future research could extend these findings to other healthcare professionals, caregivers, and non-medical populations. In study 2, the factor-specific Cronbach’s alphas were less, yet still fall within the acceptable range for early-stage scale validation, particularly given the limited number of items within each factor [77] as six items were removed during the factor analysis, in order to meetmodel fit criteria. Future researchers could also employ longitudinal designs to evaluate whether these factors remain consistent over time or are sensitive to other factors, thereby strengthening the predictive validity of the scale. The gender distribution of the sample may appear skewed, with only 13 male participants; however, this is consistent with the demographic reality of the nursing profession in India, which is predominantly female. Future research could examine gender-related differences in death anxiety by including more balanced samples or extending validation to healthcare professions with higher male representation.Further, longitudinal studies among healthcare professionals could provide valuable insights into how death anxiety evolves over time and how it interacts with occupational stress, resilience, and mental health outcomes. In Study 1, grey literature was not considered based on the objective and exclusion criteria, however, future studies could incorporate unpublished sources also if possible to minimize potential publication bias. In Study 2, DA was measured through self-report questionnaires which may have led to potential response biases. Another important direction for future research is ensuring measurement consistency across different demographic subgroups. Given the potential variations in age, gender, professional experience, and religious beliefs, future studies could employ multi-group Confirmatory Factor Analysis (MGCFA) to assess the measurement invariance of TDAS across different populations. Additionally, considering the rich spiritual and cultural perspectives on death in India, further exploration of how these influences shape the expression and experience of death anxiety would be beneficial. Future studies could also examine culturally specific factors such as Hindu beliefs in karma and rebirth, or ritual practices surrounding death, using qualitative or mixed-method approaches to understand how these shape the expression of death anxiety in Indian contexts. Adapting TDAS linguistically and conceptually for different regional and cultural contexts could enhance its applicability and relevance in cross-cultural settings.

Conclusion

This study combined a systematic review and empirical validation to examine the factor structure and psychometric properties of the TDAS among Indian nurses. The systematic review highlighted the multidimensional nature of TDAS across different populations and the frequent omission of CFA in prior studies. Addressing this gap, the present study employed both EFA and CFA, confirming a robust multidimensional structure and acceptable reliability and validity in the Indian nursing context. Findings revealed moderate reliability and a three-factor structure of Templer’s DA scale, validated through both exploratory and confirmatory factor analysis, demonstrating a strong model fit. This confirmed the multidimensional structure of the TDAS in this population, contrary to Templer’s original unidimensional proposal. Thus, the 9-item TDAS can serve as a brief screening tool to identify nurses who may benefit from targeted support, although its reduced length may limit direct comparability with prior studies. However, multiple studies have shown different factor structures of TDAS across populations, indicating that death anxiety may manifest differently among various groups. Future research should focus on longitudinal studies, cross-cultural adaptations, and specific intervention-based approaches, such as Acceptance and Commitment Therapy and cognitive-behavioral strategies, tailored to the distinct factors of death anxiety identified in this study, to further enhance its applicability.

Abbreviations

DA

Death anxiety

RDAS

Revised Death Anxiety Scale

TDAS

Templer Death Anxiety Scale

WHO

World Health Organization

CAG

Comptroller and Auditor General

SDGs

Sustainable Development Goals

EFA

Exploratory Factor Analysis

CFA

Confirmatory Factor Analysis

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

PCA

Principal component analysis

RCI

Ratio communality Index

RDWLS

Robust Diagonally Weighted Least Squares

PA

Parallel analysis

UniCo

Unidimensional Congruence

ECV

Explained Common Variance

MIREAL

Mean of Item Residual Absolute Loadings

MGCFA

Multi-group Confirmatory Factor Analysis

Author contributions

Both authors have made substantial contributions to the design of the study, data collection, and interpretation of data. Both authors read, revised, and approved the final manuscript.

Funding

The author(s) reported there is no funding associated with this work featured in this article.

Data availability

The data will be made available to the reader upon reasonable request to the first author.

Declarations

Ethics approval and consent to participate

The ethical approval was provided by the institute ethics committee, from All India Institute of Medical Sciences, Raipur (Chhattisgarh). The letter number is 2065/IEC-AIIMSRPR/2021.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.