Quality-of-life attributes of chiropractic students that are significantly associated with students’ sense of coherence: A cross-sectional study

Abstract

Objective

The objective of this study was to characterize chiropractic students, assess their sense of coherence (SOC) and QOL and determine what items of QOL measures were associated with the subdomains of SOC.

Methods

In this cross-sectional study, we surveyed students on their sociodemographic information, student characteristics, and intended practice characteristics. In addition, we measured their SOC using the SOC-29 and QOL using the PROMIS-29 and PROMIS Global Health instruments. Elastic net regression identified items from the PROMIS-29 and PROMIS Global Health surveys that were associated with the subdomains (ie, comprehensibility, manageability, and meaning) of SOC.

Results

With a response rate of 32.2%, 113 students composed our study population. The mean (SD) SOC-29 score was 148.5 (21.09) with global physical and mental health T-scores of 52.3 (7.24) and 51.0 (8.96), respectively. The mean (SD) T-scores for the QOL domains as measured by the PROMIS-29 were physical function = 54.7 (5.21), anxiety = 54.7 (8.94), depression = 49.0 (8.38), fatigue = 50.0 (9.88), sleep disturbance = 48.4 (8.20), ability to participate in social roles and activities = 54.9 (8.28), pain interference = 47.3 (7.85), and pain intensity = 1.6 (1.74). Items associated with the mental health domains had the strongest association with comprehensibility. Items in physical health and social health were most strongly associated with manageability and meaningfulness, respectively.

Conclusion

The SOC and QOL of chiropractic students were assessed at 1 institution. Items from the PROMIS-29 and PROMIS Global Health strongly associated with SOC and its subdomains were identified.

INTRODUCTION

Chiropractic is a globally recognized health care profession with more than 103,000 practitioners.^1,2 Obtaining a doctor of chiropractic degree presents challenges, particularly due to the rigorous curriculum that impacts students’ mental, physical, and social health.^3–5 Whereas student QOL in other health care fields (ie, medicine, nursing, dentistry, and pharmacy)^6–23 is well-studied, research on chiropractic students’ QOL is relatively limited but suggests that their QOL is compromised.^{3–5,24–40} Assessing QOL in clinical settings helps identify effective interventions,⁴¹ assess patient needs,⁴² and predict outcomes.⁴³ A similar approach in health care education may identify factors to improve students’ QOL, boost their academic performance and clinical skills and enhance their overall educational experience.^44,45 Exposure to chronic and high levels of stress in the academic setting⁴⁶ have been linked to poor performance,⁴⁷ challenges with mental and physical health,^48–50 and overall deteriorating health.⁵¹ It has been demonstrated that there is a direct relationship between QOL and academic success, which correlated with increased energy, mobility, daily activities, and work capacity. Additionally, students with better QOL ranked better in physical safety, financial resources, educational opportunities, home environment, and health care access.⁵² Addressing students’ QOL is, therefore, crucial as mental health challenges can hinder concentration and academic engagement,⁵³ impair motivation,⁵⁴ and increase dropout rates.⁵⁵ Not surprisingly, academic performance has been linked to professional competence.⁵⁶

The Antonovsky salutogenic model has been offered as a framework to manage and cope with stress (ie, physical, mental, and emotional) and maintain health and QOL.^57,58 Central to salutogenesis is sense of coherence (SOC). An SOC reflects an individual’s level of confidence that the individual’s environment is structured, predictable, and explicable and that resources are available to address stressful stimuli worthy of investment and engagement.⁵⁹ Resources are categorized as general resistance resources (GRRs) and specific resistance resources (SRRs). SRRs are tailored to specific stressors,^59,60 whereas GRRs encompass broader factors such as income and social support.^59,61 An SOC has the subdomains of comprehensibility (ie, an individual’s ability to comprehend stress), manageability (ie, the ability to mobilize resources to address the stress), and meaningfulness (ie, addressing the stress is valuable and worthy of investment).^62,63

SOC is measured using the SOC-29 survey instrument. The construct validity of the SOC-29 scale has been confirmed through factor analysis using a 3-factor model representing comprehensibility, manageability, and meaningfulness. Significant standardized factor loadings and covariances indicated a good model fit (χ²_{(374, n=2069)} = 2238.001, root mean square error of approximation [RMSEA] = 0.049, and standardized root mean square [SRMR] = 0.052). A second order factor model also demonstrated good model fit (χ²_{(374, n=2069)} = 2238.002, RMSEA = 0.049, and SRMR = 0.052). Standardized factor loadings of the 3 latent variables on SOC displayed a perfect relationship between SOC and manageability, whereas the other 2 relationships were slightly lower (comprehensibility = 0.88 and meaningfulness = 0.87). These results imply that, under the 3 dimensions, there is 1 theoretical construct (SOC) that accounts for 100% of the variance in manageability, 77.4% of the variance in comprehensibility, and 75.7% of the variance in meaningfulness.⁶⁴

A number of studies demonstrate that a high SOC can lead to improved health outcomes^65–68 and optimized QOL.^62,69 Two mechanisms have been proposed to strengthen SOC.⁷⁰ The first, tension management, involves the mobilization of GRRs and SRRs to address a stressor. This occurs when the individual mobilizes GRRs and SRRs to manage tension. Successful tension management will move the individual toward total health in a continuum anchored by total health and total ill health.⁷¹ The individual’s SOC is further strengthened and further equipped to recognize/identify, access, and effectively utilize new GRRs and SRRs. With increasing SOC, the person recognizes that consistent and socially valuable life experiences are worthy of time and effort. This perspective transforms tension management into coping experiences.^72,73 The second means of strengthening an individual’s SOC is the perspective that a strengthened SOC results in an enhanced QOL,⁶² which, in turn, allows the individual to be more equipped (ie, a greater capacity to utilize resources and adapt positively to stress) to successfully manage tension and contribute to further development of an individual’s level of SOC.⁷⁴ The 2 mechanisms of strengthening SOC are interdependent and interactive. In essence, there is a synergistic relationship between SOC and QOL.

Given this synergistic relationship, this study sought to examine the relationship between QOL and SOC. Our review of the literature revealed that, to date, no study has examined the SOC and QOL of chiropractic students. This study aimed to explore chiropractic students’ sociodemographic and student characteristics (ie, previous education, intended practice), and assess their SOC and QOL. Specifically, we assessed what domains of QOL were strongly associated with the subdomains of SOC (ie, comprehensibility, manageability, and meaningfulness). We hypothesized that mental health factors would have stronger correlations with the SOC subdomains more than physical or social factors, thus addressing an overlooked relationship between SOC and QOL in chiropractic education.

METHODS

Sample

This was a cross-sectional study conducted using an anonymous 78-item paper and pencil survey. The participants were a convenience sample of chiropractic students (n = 350) at 1 chiropractic college. Students were invited to participate in the study by email. Inclusion criteria for study participation were (1) the student was ≥18 years of age, (2) currently enrolled as a chiropractic student, and (3) completed an informed consent form.

Variables

Variables included sociodemographic information, student characteristics, intended practice characteristics, QOL, and SOC.

Sociodemographic variables were sex and age. Student characteristics included the student’s current quarter in the curriculum, educational level achieved prior to chiropractic matriculation, and student professional membership/affiliation. Intended practice characteristics included intended country of practice, geographic locale of practice, practice environment, and intended chiropractic technique and adjunctive therapies.

QOL and SOC variables included the PROMIS-29,⁷⁵ PROMIS Global Health,⁷⁶ and SOC-29 instruments.⁷⁷ The PROMIS-29 V.2 profile instrument assesses 7 QOL domains (ie, physical functioning, fatigue, pain interference, depressive symptoms, anxiety, ability to participate in social roles and activities, and sleep disturbance) using 4 items per domain and 1 item for pain using a numeric rating scale (NRS) (0 = no pain, 10 = maximum pain).⁷⁸ The PROMIS Global Health is a 10-item instrument that uses 4 items to measure global physical health (GPH) and another 4 items to measure global mental health (GMH).⁷⁹ Items 1 and 9 were found to be highly correlated with the other 8 items and not used.⁸⁰ A scoring table was developed to associate the raw scores to a T-score metric (ie, mean of 50 and standard deviation of 10) for each QOL domain, which is referenced to (and centered upon) the US general population.⁸¹ The greater the T-score, the greater the measured QOL domain.

The PROMIS-29 has been demonstrated with good face validity having been published in a number of languages.^82–87 The PROMIS-29 has also been demonstrated to have good construct,^{78,80,83,88–92} criterion validity,^90,93 and discriminant validity^94,95 and is able to detect meaningful changes in patients over time.^96,97 The PROMIS-29 has been shown to have excellent internal consistency with a Cronbach alpha (α) ranging from .77 to .98 in its subscales.^78,80,90,92

Validity of the PROMIS Global Health has also been established through construct validity,⁷⁹ criterion validity,⁷⁶ divergent or discriminant validity,⁷⁹ test–retest reliability, and responsiveness to change.^76,98 The reliability of the PROMIS Global Health has been shown to have a Cronbach α of .81 and .82 for GPH and 0.86 and 0.88 for GMH.^76,79

The SOC-29 is a 29-item questionnaire with a 7-level scale response. The summed score from the 29 items defined the respondents’ SOC score (ie, minimum = 29, maximum = 203), whereas items belonging to a subdomain are also summed for their subdomain scores. The greater the score, the greater its measure.

Reviews on the validity and reliability of the SOC scale have concluded that the scale is valid and reliable.^77,99 In terms of face validity, the SOC-29 has been published in 49 different languages in 48 different countries.⁶⁵ Construct validity,^64,66–71 consensual validity,⁶⁵ and predictive validity^72–75 have also been demonstrated with the SOC-29. The Cronbach α has been found to range from 0.70 to 0.95 in 124 studies.⁴¹

Ethics

This study was approved by the institutional review board of Sherman College of Chiropractic (#SCQLCS30216). Informed consent forms were distributed to students willing to take the survey.

Statistical Analysis

Data were exported to an Excel spreadsheet (Microsoft Corp) and stored digitally in a username- and password-protected computer, and the scantrons were stored under lock and key. To assess the normality of the data, the Shapiro-Wilk test was conducted. Physical function (W = 0.0129, p = 0), anxiety (W = 0.0210, p = 0), depression (W = 0.0226, p = 0), GPH (W = 0.0028, p = 0) and SOC (W = 0.0017, p = 0) were not normally distributed. The Mann-Whitney U test was applied to assess if 2 group distributions were different. The Kruskal-Wallis test was applied to assess if ≥3 or more group distributions were different. Cronbach α and Pearson product-moment correlation (r) were also determined. Only moderate (|r| = .30–.50) to strong correlations (|r| > .50) are provided as per Cohen.¹⁰⁰ Cronbach α is a measure of how well the questions in a survey instrument consistently and reliably measure the same construct (ie, depression or physical function). Expressed between 0 and 1, the closer α is to 1, the more strongly related the questions are in a survey and likely measuring the same construct. Conversely, a low α value may reflect poor interrelatedness between items, that there is a low number of questions, or that the questions are highly heterogeneous. Cronbach α values of .7 to .8 are regarded as satisfactory.¹⁰¹ Ultimately however, the interpretation of Cronbach α depends on the context and the specific research area.¹⁰²

To identify the QOL attributes that have a significant or strong predictive relationship with the subdomains of SOC, we performed elastic net regularization, a regression technique that combines the penalty functions of both least absolute shrinkage and selection operator and ridge regression.¹⁰³ Three elastic net regressions were performed to examine the association of items of the PROMIS-29 and PROMIS Global Health to the subdomains of SOC (ie, comprehensibility, manageability, and meaning). Final models were selected based on an optimized 10-fold cross-validated mena squared error (MSE) as determined by the lambda (λ) shrinkage parameter. Lambda is 1 of 2 tuning parameters (the other is α) in the loss function for elastic net regularization and is referred to as the shrinkage parameter. When the value of λ = 0, no shrinkage of the regression coefficients occurs. As the value of λ increases, the regression coefficients shrink toward 0.

The final elastic net regression model was chosen with an MSE within 1 SE (ie, standard deviation of the sample divided by the square root of the sample size). The λ_1SE is the value of λ that performs within 1 SE of the model with the lowest cross-validated error. Elastic net regression was performed with the statistical platform R.¹⁰⁴ Further description of this statistical approach specific to our study is presented in the Supplementary File, Part 1.

RESULTS

One-hundred thirteen students (60 women, 53 men) completed the survey for a response rate of 32.3%. There were no missing values from the responses and, therefore, no imputation was required. The respondents’ mean scores from the SOC-29, mean T scores from the PROMIS-29 and PROMIS Global Health and respondent characteristics (ie, sociodemographic information and intended practice characteristics) are shown in Table 1. The majority (n = 103, 91.1%) of respondents were aged 18–35 years with 10 (8.9%) aged ≥36 years. A greater proportion of respondents were in quarters 3, 4, 8, or 12 (Table 1).

Table 1.

Findings from the SOC-29, PROMIS-29, and PROMIS Global Health instruments along with findings from the respondents’ sociodemographic information, student characteristics, and intended practice characteristics

Outcome measure	Mean (SD)	Outcome measure	Mean (SD)
SOC	148.5 (21.09)	Fatigue	50.0 (9.88)
Comprehension	49.5 (8.34)	Sleep disturbance	48.4 (8.20)
Manageability	54.0 (8.53)	Ability to participate in social roles and activities	54.9 (8.28)
Meaning	45.0 (7.21)	Pain interference	47.3 (7.85)
Physical function	54.7 (5.21)	Pain NRS	1.6 (1.74)
Anxiety	54.7 (8.94)	GPH	52.3 (7.24)
Depression	49.0 (8.38)	GMH	51.0 (8.96)
Age group of respondents, No. (%)		18–25 years	58 (51.3)
		26–35 years	45 (39.8)
		36+ years	9 (8.0)
Educational level prior to matriculation, No. (%)		Some college	6 (5.3)
		Associate’s degree	10 (8.8)
		Baccalaureate degree	91 (80.5)
		Master’s degree	5 (4.4)
		PhD/doctorate	1 (0.9)
Quarter (Q) of study (No., %)		Q1 (0, 0)	Q8 (13, 11.5)
		Q2 (7, 6.2)	Q9 (3, 2.7)
		Q3 (17, 15.0)	Q10 (2, 1.8)
		Q4 (33, 29.2)	Q11 (0, 0)
		Q5 (4, 3.5)	Q12 (17, 15.0)
		Q6 (6, 5.3)	Q13 (1, 0.9)
		Q7 (9, 8.0)	Q14 (1, 0.9)
Intended country of practice, No. (%)		USA	99 (87.6)
		Canada	3 (2.7)
		Other	11 (9.7)
Intended geographic locale of practice, No. (%)		Outer city suburbs	47 (41.6)
		City (ie, central business district)	27 (23.9)
		Inner-city suburbs	24 (21.2)
		Rural/remote region	15 (13.3)
Practice environment, No. (%)		Solo practice	73 (64.6)
		With partner(s)	26 (23.0)
		Group of alternative medicine professionals	4 (3.5)
		Group of orthodox medicine professionals	4 (3.5)
		With orthodox and CAM professionals	3 (2.7)
		Within an institution	3 (2.7)
Intended primary modality for practice, No. (%)		Spinal adjustment/Manipulation	108 (95.6)
		Nutrition	2 (1.8)
		Homeopathic remedies	1 (0.9)
		Other	2 (1.8)
Intended primary adjunctive therapy, No. (%)		Spinal Adjustment/Manipulation	57 (50.4)
		Nutrition	19 (16.8)
		Massage	17 (15.0)
		Acupuncture	3 (2.7)
		Homeopathic remedies	1 (0.9)
		Naturopathic remedies	1 (0.9)
		Other	13 (11.5)
Student affiliation/membership, No. (%)		ICPA	25 (22.1)
		student ACA	1 (0.9)
		student ICA	5 (4.4)
		None	58 (51.3)
		Other	24 (21.2)

ICA = American Chiropractic Association; ACA = American Chiropractic Association; ICPA = International Chiropractic Association; International Chiropractic Pediatric Association

Overall, the PROMIS-29 domains demonstrated strong internal consistency with α values ranging from .74 to .97. The GMH scale also showed good reliability (α = .86) with the GPH scale exhibiting a lower internal consistency (α = .65). Cronbach α for the SOC-29 subdomains were .76 (comprehensibility), .82 (manageability) and .82 (meaningfulness).

The results of the correlational analysis between SOC and the various QOL domains of the PROMIS-29 and PROMIS Global Health are provided in the Supplementary File, Part 2. With the PROMIS-29 QOL domains, strong intracorrelations were observed within the mental health domains (ie, anxiety and depression: r = .62) and physical health domains (ie, physical function and pain interference: r = .54) and moderate intercorrelations between the mental health domains and physical health domains (ie, physical function and depression: r = −.44). Fatigue, sleep disturbance and ability to participate in social roles and activities (ie, sleep disturbance and fatigue: r = .43) were moderately intracorrelated, whereas the mental health domains (ie, fatigue and anxiety: r = .66) as well as the physical health domains (ie, pain interference with sleep disturbance: r = .44) were moderately to strongly intercorrelated. The PROMIS Global Health and PROMIS-29 were strongly intercorrelated with the mental and physical health domains.

SOC were strongly correlated with depression (r = −.59), anxiety (r = −.50), GPH (r = .61), and GMH (r = .75) and moderately correlated with physical function (r = .32) and pain interference (r = −.33), fatigue, sleep disturbance, and ability to participate in social roles and activities (ie, (|r| = .40–.46). Lastly, the direction of correlation was as expected with measures of compromised versus optimized QOL having a negative correlation (ie, physical function and depression).

The Mann-Whitney U test failed to reject the null hypothesis (ie, p ≥ .05), indicating no difference in the median scores of primary outcome measures between male and female respondents (Supplementary File, Part 3). The Kruskal-Wallis test with a Bonferroni correction of 0.017 also failed to reject the null hypothesis (ie, p ≥ .05) demonstrating no difference in median scores of primary outcome measures based on year of study (Supplementary File, Part 4) or age group (Supplementary File, Part 5).

For the final elastic net regression, Figure 1 shows the iterative assessment of cross-validated model performance with increasing shrinkage parameter for all 3 models.

Figure 1.

The iterative assessment of the cross-validated model performance with increasing shrinkage parameter for all elastic net regression models for the subdomains of SOC.

The results of the elastic net regression are shown Tables 2–4. The larger the absolute value of coefficients, the stronger the influence of the QOL items on the subdomain of SOC. A positive or negative sign of the coefficient indicates the direction of the relationship between the independent variable and the dependent variable. A positive coefficient indicates that an increase in the independent variable leads to increase in outcome and vice versa. Therefore, the relative magnitudes and signs of the coefficients for each item provide insights into their importance and the nature of their relationship with SOC.

Table 2.

Elastic net regression identifying QOL survey items strongly associated with comprehensibility

Variable	Health domain	Estimate
Intercept		37.64
PROMIS global health item 5: In general, how would you rate your satisfaction with your social activities and relationships?	Mental health	1.92
PROMIS global health item 4: In general, how would you rate your mental health, including your mood and your ability to think?	Mental health	1.47
Depression item 2: I felt helpless…	Mental health	−0.78
PROMIS global health item 10: How often have you been bothered by emotional problems such as feeling anxious, depressed or irritable?	Mental health	0.46
Sleep disturbance item 1: My sleep quality was…	Physical	−0.24
Anxiety item 3: My worries overwhelmed me….	Mental health	−0.14
Anxiety item 1: I felt fearful…	Mental health	−0.08
Fatigue item 2: I have trouble starting things because I am tired….	Physical health	−0.01

Note: Coefficients represent the estimated effect of each independent variable on the dependent variable of comprehensibility. A positive coefficient means that as the independent variable increases, the dependent variable tends to increase as well. A negative coefficient means that as the independent variable increases, the dependent variable tends to decrease. As an example, for every 1-unit increase in the independent variable of global item 5, comprehensibility is predicted to increase by 1.92 units. Conversely, for every 1-unit increase in the independent variable of fatigue item 2, comprehensibility is predicted to decrease by 0.01 units.

Table 4.

Elastic net regression identifying QOL survey items strongly associated with meaningfulness

Variable	Health domain	Estimate
Intercept		33.40
PROMIS global health item 9: In general, please rate how well you carry out your usual social activities and roles. (This includes activities at home, at work and in your community, and responsibilities as a parent, child, spouse, employee, friend, etc.)	Social health	0.85
Sleep disturbance item 2: My sleep was refreshing.	Physical health	−0.68
PROMIS global health item 2: In general, would you say your quality of life is:	Mental health	0.57
PROMIS global health item 6: To what extent are you able to carry out your everyday physical activities such as walking, climbing stairs, carrying groceries, or moving a chair?	Physical health	0.56
PROMIS global health item 1: In general, would you say your health is:	Mental health	0.46
Depression item 3: I felt depressed	Mental health	−0.39
PROMIS global health item 4: In general, how would you rate your mental health, including your mood and your ability to think?	Mental health	0.38
Ability to participate in social roles and activities item 3: I have trouble doing all of my usual work (include work at home)	Physical health	0.34
PROMIS global health item 5: In general, how would you rate your satisfaction with your social activities and relationships?	Mental health	0.17
PROMIS global health item 10: How often have you been bothered by emotional problems such as feeling anxious, depressed or irritable?	Mental health	0.17
Ability to participate in social roles and activities item 2: I have trouble doing all of the family activities that I want to do…	Physical health	0.09
Depression item 4: I felt hopeless	Mental health	−0.07

Note: Coefficients represent the estimated effect of each independent variable on the dependent variable of comprehensibility. A positive coefficient means that as the independent variable increases, the dependent variable tends to increase as well. A negative coefficient means that as the independent variable increases, the dependent variable tends to decrease. As an example, for every 1-unit increase in the independent variable of global item 5, meaning is predicted to increase by 0.17 units. Conversely, for every 1-unit increase in the independent variable of depression item 3, meaning is predicted to decrease by 0.39 units.

For comprehensibility, the λ_1SE was at 1.69, ensuring minimal jeopardization to the MSE. Whereas reducing variance and preventing overfitting, this model introduced bias, potentially leading to a higher MSE and affecting predictive accuracy. Eight variables strongly correlated with comprehensibility, originating from PROMIS-29 QOL domains (ie, anxiety [2 items], depression [1 item], fatigue [1 item], and sleep disturbance [1 item]), and the PROMIS Global Health items 4, 5, and 10 (Table 2).

For the SOC subdomain of manageability, the λ_1SE was at 2.62 for minimal jeopardization of the MSE and identified 13 variables. These items belonged to the PROMIS-29 QOL domains of anxiety (1 item), depression (3 items), sleep disturbance (1 item), and PROMIS Global Health items 1–6, 9, 10 (Table 3).

Table 3.

Elastic net regression identifying QOL survey items strongly associated with manageability

Variable	Health domain	Estimate
Intercept		34.44
PROMIS global health item 6: To what extent are you able to carry out your everyday physical activities such as walking, climbing stairs, carrying groceries, or moving a chair?	Physical health	1.93
PROMIS global health item 4: In general, how would you rate your mental health, including your mood and your ability to think?	Mental health	1.11
PROMIS global health item 1: In general, would you say your health is:	Physical health	1.05
Depression item 4: I felt hopeless	Mental health	−0.73
Depression item 2: I felt helpless	Mental health	−0.55
PROMIS global health item 10: How often have you been bothered by emotional problems such as feeling anxious, depressed or irritable?	Mental health	0.50
PROMIS global health item 5: In general, how would you rate your satisfaction with your social activities and relationships?	Social health	0.29
PROMIS global health item 2: In general, would you say your quality of life is:	Mental health	0.33
PROMIS global health item3: In general, how would you rate your physical health?	Physical health	0.13
PROMIS global health item 9: In general, please rate how well you carry out your usual social activities and roles. (This includes activities at home, at work and in your community, and responsibilities as a parent, child, spouse, employee, friend, etc.)	Social health	0.13
Depression item 3: I felt depressed	Mental health	−0.11
Sleep disturbance item 2: My sleep was refreshing.	Physical health	−0.11
Anxiety item 1: I felt fearful	Mental health	−0.08

Note: Coefficients represent the estimated effect of each independent variable on the dependent variable of comprehensibility. A positive coefficient means that as the independent variable increases, the dependent variable tends to increase as well. A negative coefficient means that as the independent variable increases, the dependent variable tends to decrease. As an example, for every 1-unit increase in the independent variable of global item 6, manageability is predicted to increase by 1.93 units. Conversely, for every 1-unit increase in the independent variable of depression item 2, manageability is predicted to decrease by 0.55 units.

For the SOC subdomain of meaning, the λ_1SE was at 2.94 with minimal jeopardization of the MSE and identified 12 variables. These items belonged to the PROMIS-29 QOL domains of depression (2 items), sleep disturbance (1 item), ability to participate in social roles and activities (2 items), and PROMIS Global Health items 1, 2, 4–6, 9, 10 (Table 4).

DISCUSSION

This study characterized chiropractic students at 1 chiropractic college in terms of their sociodemographic information, student characteristics, SOC, and QOL using the SOC-29 and PROMIS instruments, respectively.

The respondents’ sociodemographic information is consistent with those of other chiropractic students in North America¹⁰⁵ in terms of their age group (ie, between 18 and 35 years of age), education attained prior to chiropractic matriculation (ie, a baccalaureate degree), and membership (or lack thereof) in chiropractic organizations. In contrast, this study found the majority of respondents to be female. Consistent with patterns observed among US chiropractors,¹⁰⁶ the majority of our respondents aimed to practice in urban or suburban locations, either independently or in collaboration with other chiropractors.

Our findings indicate the need for a more comprehensive examination of the QOL of chiropractic students and its determinants such as SOC. SOC facilitates adaptive coping strategies^107–109 and contributes to individuals’ resilience^110,111 and self-efficacy.¹¹² A high SOC serves to protect against the negative impact of stress and adversity, enabling individuals to better navigate challenges and maintain a higher QOL.^62,69,113 For chiropractic students, an increased SOC may enhance their resilience, coping skills, and self-efficacy to allow them to navigate through their academic and clinical training and eventually navigate life’s challenges beyond their studies in a successful manner.

The relationship between SOC and QOL has often been described as synergistic. Whereas SOC can influence a person’s QOL⁶² and overall health,⁶⁹ QOL can also affect an individual’s SOC.^70,74 For example, improving social relationships and addressing social withdrawal can help strengthen SOC and alleviate the severity of depression.¹¹⁴

To inform our study, we performed a literature review (from journal inception to June 2025) using the Allied and Complementary Medicine Database, Cumulated Index in Nursing and Allied Health Literature, PsychInfo, Pubmed, and Index to Chiropractic Literature databases with the search terms “chiropractic or chiropractic care or chiropractic manipulation AND students or college students or higher education or further education or university students AND sense of coherence or SOC or salutogenesis or Antonovksy or salutogenic theory.” No studies were found assessing SOC in chiropractic students. Therefore, this is the first study of its kind assessing SOC and QOL among chiropractic students. The mean SOC score for this cohort of chiropractic students was 148.5 (SD = 21.09). As a context, Eriksson and Lindstrom⁶⁹ found the population distribution of the SOC-29 score to range from 100.5 (SD = 28.50) to 164.5 (SD = 17.10).

Whereas a full literature review of SOC measures in health care students is outside this manuscript’s scope, the following narrative samples and provides context to measures of SOC from medical, nursing, and dental students.

Among medical students, 1 study found that a large proportion at that end of their first semester were found to have low-to-moderate resilience scores and distinctly reduced SOC in comparison to the population average.¹¹⁵ Their perceived stress scores were inversely correlated with the SOC (r = −.662, p < .001) such that, with an increase in SOC, the perceived stress level decreased.¹¹⁶ In another study, predictors of student fatigue were identified as comprehensibility, recognizing and using emotions, QOL, satisfaction with studying medicine, and stress connected with medical studies.¹¹⁷ Medical students categorized as clever-satisfied with life had the highest SOC prior to and throughout medical school, had the highest QOL and income after medical school, and were the least overwhelmed by work stress but showed the least satisfaction with their medical career and a low level of professional competence.¹¹⁸

Dental students have been found to experience moderate burnout, SOC, perceived stress levels, and high levels of perceived social support during the COVID-19 pandemic.¹¹⁹ In another study, SOC score and self-directed learning readiness (SDLR) score had a moderate strength-significant correlation, whereas SOC and academic performance showed a weak strength-insignificant correlation. SDLR score and academic performance had a moderate strength-significant positive correlation.¹²⁰ Scholz et al¹²¹ found increasing average scores for cognitive and emotional burnout and decreasing levels of SOC as the students progressed through their semesters. Gambetta-Tessini et al¹²² found that coping strategies correlated positively with dental students’ SOC, whereas maladaptive coping strategies were negatively correlated with the SOC. Silva et al¹²³ found that better perception of the academic environment decreased the likelihood of depression, anxiety, and stress, which, in turn, led to greater SOC.

Among nursing students, Liu et al¹²⁴ found that the mediating effects of academic emotion and self-esteem between SOC and professional identity included 4 paths: (1) path 1: SOC → positive academic emotion → self-esteem → professional identity, (2) path 2: SOC → negative academic emotion → self-esteem → professional identity, (3) path 3: SOC → positive academic emotion → professional identity, (4) path 4: SOC → self-esteem → professional identity. Kleiveland et al¹²⁵ found that SOC was associated with QOL and that SOC significantly interacted the relationship between stress and QOL. Using a prospective, correlational design, Salamonson et al¹²⁶ examined the relationship between SOC, self-regulated learning, and academic performance among first-year nursing students (n = 563). A cluster analysis based on students’ sociodemographic variables (age, gender), learning strategies (elaborative, organization, rehearsal, self-efficacy, and task value), and academic performance (ie, overall grades) found no statistically significant differences between the 3 clusters. However, students with higher SOC were consistently more likely to self-report using all 5 components of learning strategies (ie, elaboration, organization, rehearsal, self-efficacy, and task value). Additionally, students with the highest SOC cluster scores consistently were more likely to achieve higher academic grades. He et al¹²⁷ found no statistical difference in the mean SOC score among first-, second-, and third-year nursing Chinese international students (n = 119). However, significant differences were found in the level of acculturative stress with third-year students having the highest stress scores, followed by the first-year students and the lowest scores for the second-year students. Colomer et al¹²⁸ demonstrated that a high SOC was associated with greater motivation to study and better academic performance among students (n = 921) in a certification of nursing assistant program.

In this study, we assessed the association between specific items of the PROMIS-29 and the PROMIS Global Health with the subdomains of SOC among chiropractic students. Given the synergistic relationship between SOC and QOL, the targeting of specific subdomains of SOC that are associated with specific items of the PROMIS instruments may help clarify how different QOL domains contribute to a chiropractic student’s overall SOC. This, in turn, could inform chiropractic educators to effectively allocate resources to improve QOL and strengthen SOC among chiropractic students.

Items identified as having a strong association with comprehensibility were 6 mental health items and 2 items physical health items. Although GMH item 5 makes inquiries on rating of satisfaction with social activities and relationships, this item is a mental health item because social well-being is a crucial aspect of overall mental health in the PROMIS framework.⁷⁶ Sleep disturbance and fatigue were included in the physical health domain but considered a measure of physical and mental health.¹²⁹ PROMIS Global Health items 4, 5, and 10 and PROMIS-29 depression item 2 had coefficients with the largest absolute magnitude and, therefore, were the 4 most important variables associated with comprehensibility (see Table 2).

For manageability, 13 items were identified as significantly associated with this SOC subdomain. Seven of the items belonged to the mental health domains, 4 belonged to physical health, and 2 belonged to social health. Note that, in the PROMIS framework, questions about QOL (PROMIS global health item 2) are considered a strong indicator of mental health and, therefore, categorized as such.¹³⁰ The PROMIS Global Health items 1, 4, 6, and the PROMIS-29 depression item 4 were the 4 most important variables (see Table 3).

For meaningfulness, 12 items were significantly associated with this SOC subdomain. Seven items belonged to mental health, 3 items with physical health, and 1 item with the social health domain. The ability to participate in social roles and activities was initially categorized under physical health within the PROMIS framework and supported by confirmatory factor analyses.¹³¹ However, a later exploratory factor analysis of the PROMIS-29 found that satisfaction with participation in social roles correlated with a distinct social health factor.¹³² The key difference between these 2 related concepts is that “ability” focuses on the perceived capacity to perform social roles, whereas “satisfaction” relates to one’s contentment with those roles.^133,134 PROMIS Global Health items 1, 4, and 9, and sleep disturbance item 2 were identified as the 4 most important variables (see Table 4).

Overall, items from the mental health domains had the strongest influence on comprehensibility, manageability, and meaning. This confirms the findings of Alcantara et al¹³⁵ that GMH, depression, and anxiety were more important predictors for SOC when compared with fatigue; sleep disturbance; ability to participate in social roles and activities; or pain intensity, pain interference, and physical function.

Our findings also provided some interesting insights on the relationship between SOC and QOL that are not as clear cut as previously described in the literature. Comprehensibility refers to the extent to which one perceives internal and external stimuli as rationally understandable and is orderly, coherent, clear, and structured.⁹⁹ Comprehensibility is primarily associated with mental health items, specifically those involved in scoring GMH in PROMIS Global Health. As observed in Table 2, physical health items such as sleep disturbance and fatigue made only minor contributions to changes in the SOC score (ie, a 1 T-score-unit increase in fatigue results in a 0.01 decrease in SOC). Additionally, sleep disturbance and fatigue have both physical and mental health characteristics.¹³² Overall, comprehensibility is associated predominantly with the mental health domains, fundamentally a cognitive dimension.⁹⁹

Manageability, the instrumental or behavioral dimension, is defined as the degree to which one feels that there are resources at one’s disposal.⁹⁹ The variable selection found that the majority of items associated with this domain were mental health along with physical health and social health. The strongest association with manageability was PROMIS Global Health item 6, a measure of physical function, with a coefficient of 1.93 (see Table 3). This indicates the importance of an individual’s functional capacity to engage in activities of daily living (ie, accessing resources such as health care services, work, and social interactions), all of which contribute to a sense of control and manageability.

According to Antonovsky^136,137 meaningfulness was the most important subdomain of SOC. In Table 4, the majority of items identified as strongly associated with meaningfulness were mental health items and also included social health and physical health item with relatively larger magnitudes of association to meaningfulness. Our findings highlight the important role of social health, improving sleep disturbance and increasing physical functioning in the development of meaningfulness. For example, the extent to which one’s social support or community gives meaning to its members, provides challenges that are worthy of investment and engagement.¹³⁸ The role of social support,¹³⁹ neighborhood, or community cohesion¹⁴⁰ has been demonstrated as constructive in the development of a strong SOC. The results of our findings provide specific strategies inform the chiropractic curriculum to increase SOC in students. For example, educational institutions should ensure/facilitate that students are satisfied with their social activities and relationships, that they maintain optimum physical functioning and mental health, and that they receive adequate and quality sleep.

The findings from this study of moderate-to-strong correlations between the physical health domains of the PROMIS-29 and PROMIS Global Health to SOC contradicts the conclusion by Flensborg-Marsden et al¹⁴¹ that there is only a weak association/correlation between measures of SOC and physical health. Correlational analysis can be misleading when multicollinearity is present as was demonstrated in this study. Multicollinearity makes it difficult to determine which QOL survey items are primarily responsible for the observed relationship with SOC. Furthermore, as Alcantara et al¹³² demonstrated, certain QOL domains of the PROMIS-29 or the PROMIS Global Health were interrelated with mental, physical, and social health. Lastly, this study demonstrated that the relationship between various QOL domains and SOC is more complex. Elastic net regression identified that the majority of items strongly associated with the SOC subdomains were mostly of mental health items of the PROMIS-29 and PROMIS Global Health. This may explain the relatively stronger correlation of mental health domains with SOC. However, the 3 individual elastic net regression analysis revealed nuances in the relationship between QOL and SOC. Items belonging to the mental health domains were strongly associated with comprehensibility, whereas physical and mental health items were also strongly associated with manageability, and social and physical health items were strongly associated with meaningfulness. Although both correlational and regression analysis measure the strength of association and direction of 2 continuous variables,¹⁴² in regression analysis, the 2 variables are not interchangeable and allow for the evaluation of more than 1 independent variable to a dependent variable.¹⁴³ Alcantara et al¹³⁵ demonstrated that the relationship between physical, mental, and social health domains to SOC are intertwined such that interaction may be taking place between physical, mental, and social health in predicting SOC. Random forest partial dependence plots revealed that, despite the presence of low physical health, a high SOC is predicted in the presence of high mental health and, presumably, a high QOL.¹³⁵ This is referred to as the disability paradox and confirmed by Albrecht and Devlieger¹⁴⁴ by demonstrating that, despite moderate-to-serious disability, their study respondents reported excellent or good QOL.

Elastic net regression can model complex relationships involving large numbers of variables with multicollinearity along with variable selection to identify the most relevant variables associated with the dependent variable.¹⁰³ As demonstrated in this study, the association between various QOL domains depends on the subdomain of SOC. Flensborg-Madsen et al¹⁴¹ commented that the SOC scales are unable to explain health that is measured only in physical terms. Our study findings directly contradict this conclusion. We found that physical health was strongly associated with the SOC subdomains of manageability and meaningfulness. Flensborg-Madsen et al¹⁴¹ proposed a new model to better understand the distinct impacts of physical and mental health (ie, “emotional coherence” for physical health and “mental coherence” for psychological health).^141,145 Eriksson and Contu¹⁴⁶ argued that this breaks significantly from Antonovsky’s fundamental notion of SOC as an “orientation to life” and the holistic approach to patient care by reinforcing the physical versus mental health divide in modern health care. We concur with Eriksson and Contu¹⁴⁶ in view of the present findings and the findings of Alcantara et al¹³⁵ on the intertwined relationship between QOL domains and SOC.

In scoring the PROMIS Global Health, item 1 (In general, would you say your health is…) was excluded due to its high correlation (ie, r = 0.80) with item 3 (In general, how would you rate your physical health?). Item 3 was retained to emphasize the physical nature of the construct of GPH. Item 9 (In general, please rate how well you carry out your usual social activities and roles…) was excluded due to its high correlation with GPH.¹⁴⁷ According to Hays et al,¹⁴⁷ items 1 and 9 can be used as single items to assess overall health perception and social health in a questionnaire, respectively.

In our analysis, the relationship between mental health and SOC was significant with a strong effect on all the subdomains of SOC. We also found that physical and social health had a significant influence on manageability and meaningfulness. This is intuitive in the sense that physical capabilities and social relationships have an influence on a person’s belief in their skills, ability, and support and their belief that their experiences in life are interesting, provide satisfaction, and are purposeful. It is of particular interest that mental health attributes were almost exclusively attributed to explaining a chiropractic student’s comprehensibility. This implies that mental health domains greatly influence a chiropractic student’s ability to understand life events and reasonably predict future events, such as success in clinical practice.

In all of the models examined, the PROMIS Global Health items had a tendency to remain in the model, whereas similar PROMIS-29 items were eliminated. This may be due to the high collinearity between the items of the PROMIS Global Health and PROMIS-29, but the PROMIS Global Health items remained more stable and provided more predictive accuracies by remaining in the model. Indeed, Pearson r found GMH correlated highly with the PROMIS-29 anxiety (r = −.65) and depression (r = −.63), whereas GPH correlated highly with the PROMIS-29 physical function (r = .55), fatigue (r = −.68), sleep disturbance (r = −.53), pain interference (r = −.61), and pain NRS (r = −.59). Based on these findings, we echo our previous sentiment that chiropractic educators should integrate chiropractic training focused on the preservation and promotion of the students’ mental, physical, and social health into the chiropractic curriculum.

In the training of health care professionals, challenges to QOL and the ability to cope with physical, mental, and social stressors have been linked with academic performance and future clinical practice.³ As indicated earlier, compromised QOL measures have been assessed among chiropractic students^5,40 with the World Health Organization Quality of Life (WHOQOL)–Brief Version,¹⁴⁸ the RAND SF-36,¹⁴⁹ and the Global Well-being Scale.¹⁵⁰ Baca et al⁵ found that 41% of the variance in the overall Response to Stressful Events Scale score can be predicted by just 3 variables: male gender, the WHOQOL Psychological Health, and the WHOQOL single-item QOL. Shepherd and Kirk⁴⁰ found that students in a spinal hygiene course demonstrated statistically significant improvement in 6 of the 8 RAND SF-36 scales (ie, physical functioning, general health, vitality, social functioning, role emotional and mental health) and the Global Well-Being Scale when compared with a control group. Recently, Cade et al¹⁵¹ compared previously recorded QOL scores of students at the New Zealand College of Chiropractic using the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36)¹⁵² before (n = 276) and after (n = 120) COVID lockdowns. After COVID, most QOL scores dropped substantially with emotional, social, and fatigue-related domains showing the greatest decline.

The use of the T-score metric (ie, mean = 50, SD = 10) provides a valuable tool for research and clinical practice when comparing PROMIS scores across studies.¹⁵³ Our study revealed that the QOL of chiropractic students is more compromised when compared with chiropractors. Comparing data obtained by Alcantara et al¹⁵⁴ on the QOL and SOC of chiropractors, the mean T score for anxiety and depression were 7.61 and 5.47 T score units larger (respectively) among chiropractic students and clinically significantly different.¹⁵⁵ A consensus among the PROMIS leadership pointed to a difference of 3 T-score points as minimally clinical important difference.¹⁵⁶ Therefore, our student respondents were also compromised in a clinically significant manner with fatigue (difference = −4.58), sleep disturbance (difference = −4.46), and pain interference (difference = −3.48). The GPH and GMH T-scores among students were 3.37 and 5.56 units lower, respectively, when compared with chiropractors.

Future Studies

The current cross-sectional study provides a snapshot of the SOC and QOL of chiropractic students at a single institution at 1 point in time. Future research should involve a longitudinal study tracking SOC and QOL across the chiropractic education continuum to determine whether SOC and QOL improve, decline, or remain stable across the chiropractic academic program. Additionally, a multi-institutional study across several chiropractic programs globally would strengthen generalizability.

Strengths

Strengths of this study include the simultaneous application of 3 validated instruments (eg, SOC-29, PROMIS-29, and PROMIS Global Health) providing a comprehensive and psychometrically robust assessment of SOC and QOL. The use of elastic net regression to identify item-level associations with the SOC subdomains of comprehensibility, manageability, and meaningfulness represents a methodological strength, particularly given its ability to handle multicollinearity among survey items. The collection of student sociodemographic information and intended practice characteristics further enriched the findings. Finally, the study contributes baseline SOC and QOL data for an understudied health professions student population, laying the groundwork for future research.

Limitations

A number of limitations must be acknowledged in our study. Surveys have inherent biases (ie, recall bias and intentional misreporting of responses). Our respondents were from 1 chiropractic college, and given inherent differences from 1 chiropractic college to another,^157,158 selection bias may occur and place into question the generalizability of the results of our findings.

Our study involved hierarchical data with potential unobserved individual-specific effects correlated with independent variables, for example, variations in QOL measures across different student groups. However, the results of our Mann-Whitney U test and Kruskal-Wallis test found no statistically significant differences in QOL measures among various groups. However, our analysis lacked consideration of random effects to enhance model validity and interpretability across student groups. Additionally, various factors such as baseline measures, academic performance, financial stress, and life experiences could influence QOL and SOC, potentially confounding the observed relationships in our analysis.

A small sample size can lead to imprecise parameter estimates and the risk of overfitting may occur, impacting model performance with new data and limiting its generalizability of the model.^159,160 Challenges in conducting a power analysis hindered determining the appropriate sample size for detecting significant effects accurately. Incorrect assumptions on effect sizes, variance, covariance, and the appropriate λ can lead to several issues (ie, inaccurate coefficient estimates, inaccurate standard errors) that ultimately affects model reliability and validity. This study refrains from making causal claims or broad inferences beyond its scope, serving as groundwork for future research to explore associations between QOL and SOC. If the rule of thumb for sample size requirements for regression analysis¹⁶¹ is followed, our study is underpowered and challenges the detection of meaningful effects. Future studies should consider larger, multi-institutional, and multicohort approaches to enhance the depth and relevance of findings in chiropractic education and psychometric analysis.

As our participants were chosen based on accessibility and the willingness to participate rather than randomly chosen, this convenience sample of students introduced the element of selection bias. We reiterate the importance of addressing sample size considerations in future studies to bolster the external validity of the results.

Given the exploratory nature of our study and the established reliability and validity of the PROMIS-29, PROMIS Global Health, and SOC-29 with other populations, our use of these instruments with our target population was justified. Furthermore, our assessment of Cronbach α showed good internal consistency for the 3 instruments, indirectly contributing to the construct validity of the instruments to our target population. However, our failure to validate the instruments to our target population opens the possibility of biased results, making meaningful inferences or generalizability challenging.

Whereas cross-sectional studies are valuable for gathering preliminary data, understanding prevalence, and identifying associations, they have limitations in assessing incidence and establishing causal relationships. Our statistical method of elastic net regression with cross-sectional data primarily provided associations between variables. (ie, the association between QOL and SOC variables). Whereas elastic net regression identified various QOL domains (ie, anxiety, depression, GMH) as potentially important predictors for the subdomains of SOC (ie, comprehensibility, manageability, meaning), no cause-and-effect inferences can be made from our findings.

Additional limitations to our study were the poor response rate and our sampling method. Whereas typical response rates for the health professions range between 3% and 50%,¹⁶² our response rates of 32.29% is relatively poor. It has been recommended that a response rate of approximately 60% for most research should be the goal of researchers.¹⁶³ A poor response rate can lead to nonresponse bias, making the results unrepresentative of the target population and potentially invalidating the findings. Regarding our convenience sampling method, the inherent nonrandom nature of this approach can introduce sampling bias, potentially compromising the external validity of our study and restricting the generalizability of our findings.¹⁶⁴

CONCLUSION

The mental health domains of QOL had a significant association with a students’ SOC and its subdomains. Physical health and social health domains of QOL had a notable association with the SOC subdomains of manageability and meaning, whereas depression and sleep disturbance were associated with all the SOC subdomains. Whereas the 3 SOC subdomains are interdependent, they are differentially influenced by distinct domains of QOL, informing the design of targeted student wellness interventions aimed at strengthening specific components of SOC in chiropractic students.

SUPPLEMENTARY FILE

Supplementary material associated with this article can be found in the online version at doi:10.7899/JCE-25-8.

DECLARATIONS

Funding

This study was funded by the International Chiropractic Pediatric Association and Sherman College of Chiropractic.

Competing Interests

This study received funding from the International Chiropractic Pediatric Association (ICPA). Joel Alcantara serves as a member of the board of directors for the ICPA, which is a funding source for this research. Andrew Whetten and Brian Dooley declare no competing interests.

Data Availability

The data set of this study is available from the corresponding author upon reasonable request subject to ICPA institutional approval and/or data-sharing agreements.

Artificial Intelligence

The authors used elastic net regression, a foundational algorithm in artificial intelligence and machine learning, for data analysis and predictive modeling. Beyond this, AI was not used to prepare this paper.

Contributorship

Concept development: BD, JA. Design: BD, JA. Supervision: BD. Data collection/processing: BD. Analysis/interpretation: BD, AW, JA. Literature search: BD, JA. Writing: BD, JA, AW. Critical review: BD, JA, AW.