Journal of Religion and Health

, Volume 51, Issue 4, pp 1188–1201

Spiritual Integration Predicts Self-Reported Mental and Physical Health

Authors

    • Department of Psychology and AnthropologyUniversity of Texas - Pan American
  • Susan Troncoso Skidmore
    • Department of Educational Leadership and CounselingSam Houston State University
  • Gary T. Montgomery
    • Department of Psychology and AnthropologyUniversity of Texas - Pan American
  • Mary Ann Reidhead
    • Department of HumanitiesFlorida Institute of Technology
  • Van A. Reidhead
    • Office of Academic AffairsEast Stroudsburg University
Original Paper

DOI: 10.1007/s10943-010-9425-9

Cite this article as:
Rogers, D.L., Skidmore, S.T., Montgomery, G.T. et al. J Relig Health (2012) 51: 1188. doi:10.1007/s10943-010-9425-9

Abstract

Data from 167 participants were used to establish the psychometric properties of the Reidhead spiritual integration scale, 31-item version (SI-31). Structural equation modeling was used to empirically evaluate influences on perceived health functioning, while accounting for possible confounds. The analyses showed that SI-31 predicted perceived mental and physical health while controlling for life satisfaction, religious variables, mood patterns, depression symptoms, and demographics. The importance of SI as a predictor of health-related outcomes is supported, as is the usefulness of the SI-31 in predicting these outcomes.

Keywords

SpiritualitySpiritual integrationHealthMental healthPhysical health

Introduction

Religious practice, belief, and spirituality have been important elements of Western Civilization (Koenig et al. 2001, Chap. 3) and continue to be important elements of the US culture (Princeton Research Center 1998). With the rise of empiricism and the success of the scientific method during the early modern and enlightenment periods in Europe, a supporting philosophy of science arose, which separated mental and physical events; mental events and the soul were the purview of the church, while the body and physical world were seen as appropriate for scientific study (Plante and Sherman 2001b). As secularism came to dominate Western academic culture, opening previously proscribed areas of study to scientific inquiry, this dualism remained, manifest as a perceived division between the processes of the body and those of the mind or—as materialism took firmer hold in science—of the body and the brain (Dolan 2007). Toward the end of the 20th century, however, the zeitgeist in investigations of the human condition in the USA began to revolve around a biopsychosocial model in which “…health and illness are viewed as a reflection of reciprocal interactions among biological, psychological and social influences” (Plante and Sherman 2001b, p. 2). The scientific study of religiosity and spirituality, and their relationship with outcome measures such as health, has burgeoned since the 1960s and gained momentum in recent years (Thoresen et al. 2001).

Religiosity presents a knotty measurement problem, partly because it is a complex and multi-faceted concept. Scores of measures of religiosity/spirituality have been proposed. For example, Hill and Hood (1999) compiled an encyclopedia of 125 measures classified into 17 different categories. Rabin and Koenig (2002) identified 10 major religious dimensions that have been measured, including, for example, religious coping and religious maturity. Hackney and Sanders (2003) also pointed out the great variety in conceptualization and measurement of religiosity. In concert with a greater interest in spirituality at the end of the last century (Hall et al. 2008; Roof 1993; Weaver et al. 2006), Rabin and Koenig’s (2002) review noted the recent trend to measure spirituality as a domain in its own right, distinct from general religiosity. A problem in defining spirituality, however, is that it has been framed in such general terms that it has little core meaning as it is usually expressed (Koenig et al. 2001).

Rabin and Koenig (2002) opined that spirituality is a distinctive term only if it involves “…a relationship with or search for the transcendent” (p. 222). Following a similar rationale, Reidhead and Reidhead (2001) and Hurwicz (Reidhead et al. 1999) developed the concept of spiritual integration (SI) from the reference point of the lives and beliefs of nuns and monks of the Catholic Benedictine Tradition. While the term SI has been used in diverse contexts (e.g., Dunne 1985; Morris et al. 2003; Richo 1991), and assessed by psychometric instruments such as the Principles of Living Survey of Brady et al. (1999), Reidhead and Reidhead (2001) defined SI as “a way of understanding, behaving, and being that operates on a principle of integrated wholeness, in which the parts of one’s life are unified into a common field of spiritual understanding and practice” (p. 3). This conceptualization of SI represents a defined perspective on specific aspects of the previously diffuse and varied concepts of “spirituality;” it bears resemblance to Allport’s concept of “intrinsic religiousness,” where one “lives his religion” (Allport and Ross 1967, p. 434); and is germane to Allport’s discussion of religious maturity, where one’s beliefs and faith are integrated within one’s lifestyle (Batson and Ventis 1982). The Reidheads’ SI concept is also consistent with Ellison’s (1983) statement that spirituality represents “an integrative force in the individual’s life, providing meaning, and core values and principles for organizing one’s life.”

SI Scale Development

Reidhead and Reidhead (2001; Reidhead et al. 1999) developed their conceptualization of SI into a psychometric scale through a four-phase procedure. In the first phase, 29 senior contemplative nuns and monks were asked to create “free lists” that produced several hundred phrases that described beliefs and behaviors consistent with SI. The free lists were refined through consensus analysis into a list of 36 core areas. In the second phase, 10 senior (50+ years of age) Catholic oblates (i.e., lay people affiliated with a monastery for spiritual development) and a senior monk participated in developing 109 statements consistent with the 36 core SI areas (Reidhead 2004). The investigators converted each statement into a 5-point Likert-type item and ensured that each of the 36 core areas of SI was represented by at least three items. In the third phase, the 109 Likert-type items were administered to a group of 15 seniors of mixed religious affiliation in a Midwestern Lutheran-affiliated non-residential community center and refined down to 73 items to which two more were added. In the fourth phase of scale development, the 75-Likert-item questionnaires were administered to 30 seniors of mixed religious affiliation living independently in a Midwestern Protestant-affiliated senior residential center and refined to 41 items. We refer to these 41 SI items as the Reidhead spiritual integration scale—41-item version or SI-41.

The conceptualization of the SI-41 bears initial resemblance to Allport’s concept of Intrinsic Religiousness; Donahue (1985), however, would suggest that the resemblance is largely superficial. In his review of intrinsic religiousness studies conducted with Allport’s religious orientations scale he concluded that, by itself, intrinsic religiousness seemed to “correlate with little besides other measures of religiousness” (p. 415). The 36 core areas of spiritual integration from which the SI-41 was developed consists of behaviors and attitudes that mature monks and nuns agreed were necessary to develop a spiritually integrated life.

SI: Convergent and Criterion Validity

The Fetzer Brief Multidimensional Measure of Religiousness/Spirituality (BMMRS; Fetzer Institute and National Institute on Aging 1999/2003) was chosen as a reference for measuring the convergent validity of the newly developed SI scale. The BMMRS and its parent scale, the full-length Fetzer Institute Spirituality Measure (MMRS), are widely cited in research and theory relating to the study of spirituality [e.g., Neff (2006) reported 140 citations], and it has been recommended as a comprehensive scale to measure religious and spiritual beliefs and practices (Rabin and Koenig 2002). Like the SI, the MMRS was developed initially through a form of focus group analysis that bears some resemblance to the statistical method called consensus analysis used by the Reidheads et al. However, whereas the Reidheads’ initial experts were monks and nuns steeped in a contemplative career/lifestyle, the Fetzer Institute’s experts were academic scholars from diverse disciplines. In addition, the Reidheads’ concept of spiritual integration does not map perfectly onto any one of the Fetzer subscales. Thus, to the extent that the Reidheads’ SI measure is a valid measure of their construct of SI, correlations between the SI and conceptually relevant subscales of the BMMRS should be expected to be moderate in strength.

To evaluate the proposition that those scoring high in SI should “…live productive, healthy, satisfying lives…” (Reidhead and Reidhead 2001, p. 4), self-report criterion measures of affect (i.e., mood) patterns, overall life satisfaction, and depressive symptoms were included in the current study. Individuals with high SI scores would be expected to demonstrate a pattern of positive affect, high life satisfaction, and few depressive symptoms.

SI: Relationships with Mental and Physical Health

Since antiquity, humans have assumed a relationship between religiosity/spirituality and health. In Western Civilization, for example, some of the earliest physicians were licensed by the Christian church (Kuhn 1988). With the recent burgeoning interest in measuring the relationship between religiosity/spirituality and other variables, relationships between physical and mental health have been much investigated (e.g., Hill and Pargament 2008; Koenig et al. 2001). In 2001, for instance, Koenig et al. reviewed more than 1,200 investigations of religiosity/spirituality in health care delivery and outcome. SI is arguably an important and useful conceptualization of spirituality for this area of research; Reidhead and Reidhead (2001, p. 4) posited that SI might be associated with mental and physical health and, if so, that those scoring high in SI should “…consume less for mental and physical health care than their neighbors…” In order to determine relationships between SI and self-reported physical and mental health, the Rand 36-item Health Survey 1.0 (SF-36), developed by Ware (2000; Hays et al. 1992, pp. 4–6) was administered. The SF-36 is “…a set of generic, coherent, and easily administered quality-of-life measures” focused on self-perceptions of physical and mental/mental health (Rand 2010a). Versions of this health survey have been widely used; the Rand Institute website includes a bibliography of over 90 citations in which these have been used with a Rand Institute-funded multi-state medical outcomes study (Rand 2010b). Additionally, Ware (2010) cites an extensive bibliography of the SF-36’s use in tracking the course of medical treatment outcomes, and the SF-36 is frequently used by managed care organizations and Medicare for routine monitoring and assessment of care outcomes in adult patients (Rand 2010a).

Ware (2010), in discussing the psychometric properties of the SF-36, presented a measurement model relevant to our evaluation of the usefulness of SI for predicting health-related outcomes. The measurement model casts physical and mental health as higher-order latent variables (LVs), each indicated by its four conceptually affiliated SF-36 subscales as endogenous manifest variables (MVs). The present study expanded Ware’s measurement model of the SF-36 by evaluating SI and other exogenous variables as predictors of physical and mental health, testing the plausibility of a causal model in which SI and other variables determine variations in physical and mental/mental health status as measured by the SF-36.

Method

Participants

The sample was composed of 167 senior individuals, with full physical mobility and mental acuity, residing independently or in a senior living community (SLC) in the greater St. Louis, MO area. Participants were recruited from seven different organizations: an Ethical Society men’s group, seniors from a Unitarian congregation, and five SLCs. Participants’ ages ranged from 56 to 96 years (M = 78.4, SD = 8.7), with 64.7% of the sample self-identified as female and 35.3% as male. Self-reported religious affiliation was fairly evenly distributed between Protestants (55.8%) and non-Protestants (44.2%). The majority of the respondents (79.8%) were either married or widowed. The average amount of time spent per week on religious activities was 6.4 h (SD = 10.2). However, as Table 1 indicates, the majority of the respondents (67.7%) spent 5 h or less on religious activities weekly.
Table 1

Sample demographics

Characteristic

n

%

M

SD

Religious preference

 Protestant

87

55.8

  

 Non-protestant

69

44.2

  

Life status

 Widow/widower

63

39.9

  

 Married

63

39.9

  

 Professed member of a catholic religious order

15

9.5

  

 Divorced or separated

10

6.3

  

 Other

7

4.4

  

Age at time of survey

 Female

99

64.7

78.7

8.2

 Male

54

35.3

77.8

9.5

Time spent on religious activities weekly

 0–2 h

46

36.2

0.6

0.7

 3–5 h

40

31.5

3.8

0.8

 6–8 h

17

13.4

6.9

0.8

 9–15 h

15

11.8

12.3

2.3

 >15

9

7.1

36.9

17.2

Total sample size = 167. Deviations from this total indicate a missing response to that item

Measures

In addition to the SI-41, BMMRS, and SF-36, which have previously been described, three additional measures were included in the survey: a single-item Life Satisfaction measure (LS; Inglehot and Rabier 1986), the Affect-Balance Test (Bradburn 1969; NORC 2010), and the widely used Center for Epidemiologic Studies Depression Scale (CES-D; Radloff 1977). The single-item LS prompted respondents to indicate their satisfaction “On the whole… with the life you lead;” response options included very satisfied, fairly satisfied, not very satisfied, or not at all satisfied. The affect-balance test consists of ten questions answered in yes/no format. The first five measure positive feelings (e.g., “during the past few weeks, did you ever feel pleased about having accomplished something?”) and the last five tap negative feelings (e.g., “during the past few weeks, did you ever feel bored?”). See Table 2 for reliability estimates of all measures as observed in our sample and Table 3 for intercorrelation among the measures.
Table 2

Reliability of measures

Measure

Cronbach’s α

Spiritual integration (SI-31)

0.84

CES-D

0.81

Positive affect (affect balance)

0.52

SF physical functioning

0.94

SF physical role limitations

0.86

SF pain

0.88

SF general health

0.72

SF social functioning

0.80

SF emotional role limitations

0.76

SF energy

0.83

SF emotional well-being

0.72

Table 3

Correlation matrix of all manifest variables in the structural model

  

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

1

SFPHYSFUNC

1

               

2

SFPHYSROLE

0.67

1

              

3

SFPAINFUNC

0.65

0.58

1

             

4

SFGENHLTH

0.62

0.50

0.51

1

            

5

SFSOCFUNC

0.61

0.60

0.68

0.60

1

           

6

SFEMOROLE

0.41

0.47

0.24

0.39

0.49

1

          

7

SFENERGY

0.61

0.53

0.54

0.72

0.68

0.48

1

         

8

SFEMOWELL

0.43

0.35

0.39

0.61

0.49

0.42

0.63

1

        

9

SI-31

0.19

0.19

0.19

0.35

0.38

0.28

0.43

0.51

1

       

10

LS

0.02

0.12

0.17

0.08

0.20

0.18

0.08

0.06

0.12

1

      

11

RELIGHRS

0.19

0.16

0.07

0.23

0.13

0.20

0.21

0.19

0.21

0.14

1

     

12

AGE

−0.05

−0.03

−0.01

0.15

−0.03

0.07

0.04

0.23

0.31

−0.06

0.03

1

    

13

MALE

0.25

0.17

0.23

0.19

0.01

0.13

0.12

0.09

−0.15

0.01

−0.14

0.15

1

   

14

PROTESTANT

−0.10

−0.06

0.04

0.04

0.03

−0.08

−0.02

−0.04

0.15

0.17

−0.12

0.13

−0.14

1

  

15

POSAFFECT

0.25

0.22

0.13

0.31

0.21

0.27

0.26

0.35

0.28

0.04

0.28

0.27

0.17

0.05

1

 

16

CES-D

−0.04

−0.09

−0.15

−0.14

−0.25

−0.20

−0.10

−0.09

−0.09

−0.30

−0.04

0.11

−0.05

−0.06

0.01

1

SF PHYS FUNC SF-36 physical functioning subscale, SF PHY ROLE SF-36 role limitations due to physical functioning subscale, SF PAIN FUNC SF-36 pain functioning subscale, SF GEN HLTH SF-36 general health subscale, SF SOC FUNC SF-36 social functioning subscale, SF EMO ROLE SF-36 role limitations due to emotional health subscale, SF ENERGY SF-36 energy/vitality subscale, SF EMO WELL SF-36 emotional well-being subscale, SI-31 Reidhead spiritual integration scale, 31-item version, LS single-item life satisfaction, RELIG HRS number of hours per week spent in religious activities, MALE sex (binary), PROTESTANT religious denomination (binary), POS AFFECT affect balance (positive values as more positive affect/mood), CES-D CES depression scale

Procedure

Each organization was contacted and provided with flyers, which were used for recruitment in addition to existing organizational communications (e.g., newsletters). Paper-and-pencil questionnaires were administered to groups of participants, in a single sitting for each individual, in locations provided by the recruiting organizations. Participants were instructed to carefully read and mark their answer to each question, to take as much time as they wished, and to ask clarifying questions as needed. The scales in the questionnaire packet were administered in the following order: SI-41, SF-36 Health Survey, single-item LS, 31-item BMMRS, seven demographic questions, 10-item Affect-Balance Test, and 20-item CES-D. Ethical procedures were followed at all times, including provision of informed consent to all participants and protection of the anonymity of their responses. This study was approved and monitored by the Institutional Review Board of the University of Missouri—St. Louis.

Data Analyses

The data were reviewed for evidence of non-normality and potential outliers. The shape of all variables was well within the recommended cutoff values (i.e., skewness <2.0 and kurtosis <7.0; Curran et al. 1996), with the exception of the number of hours spent on religious activities (RELIG HRS). Examination of Q-Q plots for signs of non-normality also indicated that all variables except for RELIG HRS can be considered to be normally distributed. All models were evaluated with Mplus (V5.21; Muthén and Muthén 1998–2007), using the MLR option (maximum likelihood estimation with robust standard errors). Robust statistics provide security against violation assumptions (e.g., non-normal data) especially when the deviations are near or below the limits of what can be detected (Huber and Ronchetti 2009). Another advantage of MLR is its ability to account for missingness in the data. Unlike imputation methods, where missing values are replaced before analysis with new values generated by algorithms based on the non-missing data, in MLR “information is borrowed from the complete data during the estimation of parameters that involve missing values” (Enders 2001, p. 140). Maximum likelihood estimators—sometimes called full-information maximum likelihood (FIML)—are known to be less biased and more efficient than other missing data approaches (Enders and Bandalos 2001).

The composition of Reidhead’s (2004) SI-41 was subjected to exploratory factor analysis to investigate the number of underlying factors, resulting in the final version of this scale (SI-31). Next, two separate confirmatory factor analysis (CFA) models (one each for the latent variables physical health [PHYS HEALTH] and mental health [MENT HEALTH]) were estimated. After confirming the SF-36 measurement model, we examined the impact of the newly specified SI-31 on PHYS HEALTH and MENT HEALTH via the fitting of a structural model. Powell et al. (2003) highlighted the importance of controlling for confounds. We included several potential confounding variables in the structural model to simultaneously control for their effects on the relationships between SI and physical/mental health and to evaluate any predictive power they might have for the two health-related LVs. These predictors/potential confounds were participant age (AGE), single-item life satisfaction (LS), RELIG HRS, sex (male vs. female; MALE), religious affiliation (Protestant vs. non-Protestant; PROTESTANT), Affect-Balance Test score (POS AFFECT), and level of depression (CES-D).

Results

SI Factor Analysis and Convergent Validity

Analysis of scree plots and eigenvalues suggested a single-factor solution as the most sensible composition of the SI. Ten items were removed, as they negatively impacted the internal consistency of the scale. This resulted in a more parsimonious 31-item Spiritual Integration Scale (SI-31; α = 0.84). The collection of items in this unidimensional scale exhibits face validity for the conceptualization of SI that guided this work; items address general spirituality in one’s daily life (e.g., “I think about God in the course of my daily activities,” “most days I read, view, or listen to something that is spiritually uplifting, like scripture or a religion show on TV”), one’s intellectual and emotional approach to living (e.g., “I don’t lose hope when things get difficult,” “I like listening to the ideas of people whose lives and ideas are different from mine,”), one’s approach to interpersonal relationships (“I am on guard to make sure people don’t take advantage of me,” “I try to stay out of sight when there is work to be done”), holistic integrations of spirituality with existential issues (e.g., “Knowing that I’m going to die troubles me,” “I am unhappy with the way my life has turned out,”), and other aspects of integrating spirituality into one’s way of living.

The SI-31 composite score showed a pattern of associations with subscales of the FBMMRS that strongly supports the former’s construct validity. The SI-31 was correlated at r = 0.56 with the BMMRS “daily spiritual experiences” scale, r = 0.38 with “values and beliefs,” r = 0.46 with “forgiveness,” r = 0.48 with “private religious practice,” r = 0.53 with “positive religious coping,” r = −0.34 with “negative religious coping,” r = 0.32 with “religious support,” r = 0.50 with “religious commitment,” r = 0.15 with “organizational religiousness,” r = 0.47 with self-ranking of religiousness, and r = 0.43 with “religious/spiritual history.”

Measurement and Structural Models of Perceived Physical and Mental Health

Ware’s (2000) original model specified that each SF-36 latent variable be indicated by four manifest variables. In our measurement model, however, better fit was achieved by adding two additional indicator paths, one from each LV (see Fig. 1). In our measurement model, PHYS HEALTH is indicated by the SF-36 summary measures of (1) physical functioning, (2) role limitations due to physical health, (3) pain-related functioning, (4) general health, and (5) social functioning. MENT HEALTH is measured by the SF-36 summary scales of (1) general health, (2) social functioning, (3) role limitations due to mental health, (4) energy/fatigue, and (5) general emotional well-being. Despite the wording of some labels, all SF-36 subscales were scored so that higher values described greater physical or emotional functioning.
https://static-content.springer.com/image/art%3A10.1007%2Fs10943-010-9425-9/MediaObjects/10943_2010_9425_Fig1_HTML.gif
Fig. 1

Measurement Model. \( \chi_{15}^{2} \) = 19.71, P = 0.184; CFI = 0.991; RMSEA = 0.043, 90% CI [0.000, 0.090]; SRMR = 0.027. *P < 0.05. PHYS HEALTH: perceived physical health functioning; MENT HEALTH: perceived mental health functioning. Other abbreviations as in Table 3

The fit of the models was assessed using multiple fit indices; the final results of these analyses are presented in Fig. 1. The global fit index, χ2, is interpreted by examining the probability of obtaining the given χ2 value and its corresponding degrees of freedom. Larger probabilities are associated with better overall model fit (Loehlin 2004). The comparative fit index (CFI) measures the improvement of the model fit over the independence or null model. CFI values greater than 0.95 indicate good model fit (Hu and Bentler 1998). For the standardized root mean residual (SRMR) and the root mean square error of approximation (RMSEA) indices, values closer to zero indicate better fit; higher values indicate poorer fit. An RMSEA value less than 0.06 is considered good (Hu and Bentler 1999; Loehlin 2004). Browne and Cudeck (1993) have indicated that RMSEA values between 0.05 and 0.08 can be considered to indicate adequate model fit. The 90% CIs are also reported (Steiger 1990). For the SRMR, values less than 0.08 are considered good (Hu and Bentler 1999; Loehlin 2004).

The fit of the measurement model for the SF-36 presented in Fig. 1 was not statistically significant (\( \chi_{15}^{2} \) = 19.71, P = 0.18). Other fit indices similarly demonstrated the good fit of our proposed model to the data (CFI = 0.991; RMSEA = 0.043, 90% CI [0.000, 0.090]; SRMR = 0.027).

Although the test for the overall structural model fit presented in Fig. 2 was statistically significant (\( \chi_{63}^{2} \) = 103.52, P < 0.05), the other fit indices (i.e., CFI = 0.944, RMSEA = 0.62, 90% CI [0.040, 0.083] and SRMR = 0.042) indicated an adequate to good fit of our hypothesized structural model. All of the factor loadings were statistically significant (P < 0.01), and each observed health subscale (MV) was positively related to its corresponding latent variable (i.e., PHYS HEALTH, MENT HEALTH). The two LVs were also positively correlated with each other within the structural model. The proportion of variance of the individual observed SF-36 health outcome variables explained by the structural model ranged from 0.32 to 0.77.
https://static-content.springer.com/image/art%3A10.1007%2Fs10943-010-9425-9/MediaObjects/10943_2010_9425_Fig2_HTML.gif
Fig. 2

Structural Model. \( \chi_{63}^{2} \) = 103.52, P < 0.05; CFI = 0.944; RMSEA = 0.062, 90% CI [0.040, 0.083]; SRMR = 0.042. *P < 0.05. Abbreviations as in Table 3 and Fig. 1

SI and Perceived Health

The structural model expanded the measurement model by positing the effect of the exogenous variables (SI-31, LS, RELIG HRS, AGE, MALE, PROTESTANT, POS AFFECT, and CES-D) on the two LVs already defined (Fig. 2). These exogenous variables accounted for 22.9% of the variance in participants’ perceived physical health (PHYS HEALTH). Physical health was statistically significantly impacted by SI-31 (γ = 0.27, P < 0.01) and RELIG HRS (γ = 0.14, P < 0.01). Being male also had a statistically significant positive relationship with physical health (γ = 0.34, P < 0.01). In other words, men tended to score higher on perceived physical health than women. Age, on the other hand, had a statistically significant negative effect on physical health (γ = −0.21), P < 0.05). In other words, on average, the older the participant was, the lower their physical health score was. The rest of the variables did not have a statistically significant impact on physical health. All path coefficients presented are standardized.

The exogenous variables mentioned above collectively accounted for 39.3% of the variance in participants’ perceived mental health (MENT HEALTH). SI-31 had a statistically significantly positive effect on mental health (γ = 0.51, P < 0.01). In other words, there was a direct relationship between high scores on the SI-31 and high scores on perceived mental health. Similarly, three other variables had a statistically significantly positive relationship with mental health: RELIG HRS (γ = 0.12, P < 0.05), MALE (γ = 0.19, P < 0.01), and POS AFFECT (γ = 0.17, P < 0.05). The rest of the variables did not have a statistically significant impact on mental health. Again, all path coefficients are standardized.

Discussion

Spiritual Integration

“Spiritual integration” and related terms have been used in the theoretical and research literature of several fields for decades, though without a clear consensus as to the meaning or application of the term. The initial indications of reliability and validity for the SI-31 reported here may be a useful starting point for clarifying the validity not only of this measure but also of the construct of spiritual integration.

The SI-31’s relationship—and plausible predictive ability—with perceived mental and physical health suggests that spiritual integration may play an important role in the way individuals respond to or interpret their physiological and psychological health. The link between general spirituality and health is not a new research avenue (Plante and Sherman 2001a; Young and Koopsen 2005), but the success of a measure of SI in predicting health indices represents an important refinement in our understanding of the relationship between spirituality and individuals’ perceptions of their physical and mental health. This study further raises the possibility that spiritual integration may affect actual mental and physical health functioning; this conclusion is beyond the scope of our results, but these findings are suggestive and provide sufficient “probable cause” to justify further investigation using more labor- and cost-intensive research methods, such as longitudinal, experimental, or quasi-experimental studies of the role of SI in this domain.

In their critical review of the diverse approaches that have been taken to measuring religiosity and spirituality, Hall et al. (2008) pointed out that scales that measure “spirituality in general are vulnerable to over-generalization” (p. 155). Hall et al. (2008) further asserted that a difficulty with spirituality scales is that they are free of context, persuasively arguing that context is crucial for interpreting the meaning of spirituality. The SI-31 scale addresses this problem through its development within the context of the monastic tradition of Benedictine and Cistercian monks and nuns (all within the overarching Benedictine Tradition). The 109 pre-test questions from which the SI-41 was derived were developed in collaboration with Benedictine oblates—lay Catholics engaged in developing their spirituality in partnership with monks or nuns. We agree that it is important to avoid overgeneralization of findings—in the strictest sense, our results merely show that the SI-31 can account for a large portion of the variability in self-reported mental and physical health of a sample of elderly people in St. Louis, Missouri—however, this sample seems generally representative of many North Americans, giving confidence that these results might be replicated in other samples.

The relationship between mental and physical health and religiosity/spirituality has been widely debated and increasingly investigated (e.g., Koenig and Cohen 2002; Koenig and Larson 2001; Koenig et al. 2001; George et al. 2008; Krause 2010; Plante and Sherman 2001a). While our findings are suggestive, the generality of the relationship between SI-31 and mental and physical health will hopefully be evaluated in subsequent investigations. We believe the SI-31 will be applicable to “a wide cross-section of religious and non-religious people alike” (Reidhead 2004, p. 18) and that the SI-31 will hold up well under critical scrutiny (e.g., Meezenbroek et al. 2010).

Participant Demographics and Mood

In line with some previous research, we found that participant sex was associated with better perceived physical (Markides 1990) and mental health (Boughton and Street 2007), as was our finding that increasing age was associated with decreasing perceptions of physical health (Andersen et al. 2007).

It is notable that our measures of depression, life satisfaction, and religious affiliation failed to have any significant impact on perceived physical or mental health functioning in the structural model. This may mean that, in other research, any apparent relationship between individuals’ level of depression symptoms, perceived life satisfaction, and perceived health functioning would actually be better accounted for by the individuals’ level of spiritual integration, overall mood pattern, and basic demographics. The fact that participants’ general mood self-ratings were predictive of self-reported health status, where depression symptoms and life satisfaction were not, is especially interesting, suggesting that mood patterns—possibly at the trait level, though that cannot be fully evaluated in this study—may be a more sensitive predictor of perceived health status than acute, short-term mood fluctuations are.

Conclusion

The SI-31 appears to be a viable measure of spiritual integration (SI), as this concept was defined by Reidhead and Reidhead (2001). It shows very good internal consistency reliability, and its validity is strongly supported by the pattern of correlations it displays with demographic variables and the established Fetzer Institutes Brief Multidimensional Measure of Religiousness/Spirituality (BMMRS). Even more importantly, the model fit in this study demonstrated a plausible, important, and theoretically sound place for spiritual integration (in the form of the SI-31) as a predictor of self-reported physical and mental health functioning. In fact, SI outperformed basic predictors such as participant age, sex, mood self-rating, depression symptoms, and religious involvement in predicting self-reported physical and/or mental health functioning. These results strongly suggest that spiritual integration may be an important predictor of health-related outcomes, and that the SI-31 is a useful measure for clarifying and assessing spiritual integration.

Acknowledgments

This work has been supported in part by grants from the John E. Fetzer Institute (projects 1204.1, 1204.2), the University of Texas–Pan American Graduate School, and the UTPA College of Social and Behavioral Sciences Summer Writing Institute Grants (SWIG) program. The authors wish to thank Margo-Lea Hurwicz for her role in designing the free lists and consensus analysis that underly the SI scale development; Michelle Varón for her assistance with data management and the literature review; and the nuns and monks of Mt. St. Scholastica Monastery, St. Vincent Archabbey, and Gethsemani Abbey—especially Frances Watson, OSB (PhD), Judith Sutera, OSB, Demetrius Dumm, OSB (SD), Donald Raila, OSB, Vernon Holtz, OSB (PhD), Joseph Martinez, OCSO (MD), and the late Rafael Prendergast, OCSO.

Copyright information

© Springer Science+Business Media, LLC 2010