The Effect of a Web-Based Deep Breathing App on Stress of Direct Care Workers: Uncontrolled Intervention Study

Abstract

Objectives:

Studies have demonstrated success in reducing stress levels in complex interventions including deep breathing components. Breathing exercise interventions, however, tend not to be studied in isolation. The aim of the study was to examine the impact of a breathing exercise using a web-based app on the stress levels of direct care workers (DCWs) who serve people with intellectual and developmental disabilities.

Design:

Uncontrolled one-group intervention.

Settings/Location:

DCWs were recruited from intellectual and developmental disability service providers in a US Midwestern state.

Subjects:

Sixty-four DCWs who used the breathing exercise app at least 2 times.

Interventions:

Breathing exercise using a web-based app for a month.

Outcome measures:

Five measures were obtained from the data recorded in the app: initial stress level before exercise, exercise duration in minutes, change in stress level between before and after each breathing exercise, and frequency and number of days the app was used during a month. Participants' self-report of the number of days of app use was collected in a 1-month follow-up survey.

Results:

The participants appear to have a moderate stress level indicated by the initial stress level 5 out of 10. After the breathing exercise, the stress level decreased by 1.2 points from 5.0 (standard deviation [SD] = 1.8) to 3.7 (SD = 1.6) on average (paired t-test, p < 0.00005). Cohen's d 0.72 indicates a large effect size. Among within-individual factors, a higher initial stress level and longer app use per occasion were significantly associated with stress reduction. Among between-individual factors, only race was associated with stress reduction. Although there was no effect of being an African American alone on stress level change (coefficient = 0.44, 95% confidence interval = −0.29 to 1.18, p > 0.05), there was an additional reduction among African Americans in relation to the initial stress level, controlling for exercise duration. The number of breathing exercise days recorded in the app was not correlated with that of self-report (Pearson's correlation r = 0.12, p > 0.05). Although the app was used for 4.4 (SD = 4.2) days, participants reported using it for 10.7 (SD = 8.2) days on average in the follow-up survey.

Conclusions:

The findings suggest the benefit of breathing exercises using an app for reducing DCWs' stress levels. Regular use of such apps may assist with stress management and bolster overall health and well-being among DCWs.

Introduction

Background

Direct care workers (DCWs) play important roles to support people with intellectual and developmental disabilities (IDDs) throughout their lives for health, community integration, and independence.¹ DCWs is a general term that refers to workers who provide hands-on assistance. Working as a DCW is a stressful job, including heavy workloads and limited job autonomy.^2
–4 Many DCWs experience poverty status, as well as racial discrimination,^5
–7 which compounds their experience of work-related stress.¹ DCWs' experience of work stress can lead to a myriad of negative consequences, including poorer physical and mental health.^8
–10

Social capital can be a resource that mitigates the negative impact of stress.^11,12 Living or working in communities that provide social support and engagement can serve as social capital to lessen mental health challenges or stress.^11,13,14 For disadvantaged racial minorities, social capital can be limited or does not function positively because people around them may be already overburdened and struggle with similar stressors.^11,12

Regular self-care can help DCWs cope with work-related stress.¹⁵ Stress management programs that address stress from the cognitive, behavioral, and psychological standpoints include cognitive behavioral therapy, meditation, mindfulness practices, progressive muscle relaxation, biofeedback, guided imagery, and breathing exercises.¹⁶

An increasing body of evidence is available regarding the efficacy of digital health interventions.^17

–20 Such digital formats have allowed for greater participant accessibility. More data on longitudinal outcomes are still needed, however, to determine efficacy and whether such formats are beneficial for various conditions or behaviors.^17,18,20

A variety of breathing exercises (e.g., abdominal breathing, yogic breathing, and diaphragmatic breathing) have been shown to assist with stress management.^{16,21

–26} It is suggested that deep breathing operates to functionally reset the autonomic nervous system to regain homeostasis. This, in turn, can help decrease fatigue, anxiety, and stress,¹⁶ and contribute to healthier lifestyle choices.^21,27,28

Although a wide range of intervention studies have demonstrated success in reducing health care professionals' stress levels through relaxation-based techniques,^29,30 this commonly has been done in the context of more comprehensive interventions, such as those employing mindfulness, meditation, and gentle movement,^{9,10,31
–33} and relaxation exercises.^22,34 Deep breathing interventions generally are not studied in isolation.^{22,26,31,35
–37} To the authors' knowledge, no deep breathing interventions have been tested exclusively with DCWs.

Deep breathing exercises have been found to be effective with patients, family caregivers, and community members in a few studies. Beng et al.²⁴ conducted a nonblinded randomized controlled pilot study with 20 palliative care patients and family caregivers to test the efficacy of a 5-min mindful breathing distress reduction exercise. Participants in the mindful breathing group reported significantly lowered distress scores compared with the control group.²⁴ Rawat³⁸ studied the effects of a deep breathing exercise in reducing stress levels with regular 5-min breathing exercises among 60 community-based participants in a one-group pre–post-test design. Stress levels significantly decreased after 4 weeks of data collection.³⁸ In another deep breathing intervention with community-based participants,²³ heart rate variability was measured in a control group and groups engaged in deep breathing for 5, 7, and 9 min. The deep breathing groups demonstrated significantly lower heart rate variability (related to reduced stress levels).²³

Research aims and hypotheses

This study aimed to assess the impact of breathing exercise using a web-based app on DCWs' stress levels and examine within-individual and between-individual factors that affect stress levels using a multilevel modeling approach for repeated measures data. We hypothesized that (1) within-individual-level factors, such as a higher initial stress level and doing the breathing exercise longer, are associated with a larger decrease in stress; (2) between-individual-level factors such as app use frequency, gender, age, race, and tenure in the field are associated with change in stress level, controlling for within-individual factors; and (3) between-individual-level factors moderate the effect of within-individual-level factors on the change in stress levels.

Methods

Study design

The effect of breathing exercises on stress was examined by measuring stress levels before and after each breathing exercise in an uncontrolled intervention study. Due to the study design, findings should be considered explorative.

Human participants

DCWs who completed an online palliative care training voluntarily participated in this study (n = 126) at four community-based service organizations serving people with IDDs in a U.S. midwestern state. DCWs were recruited with flyers posted in the participating organizations. Informed consent was obtained before study participation.

Procedure

Participants completed a baseline survey before the palliative care training and a follow-up survey 1 month after the training. The deep breathing app was included as part of the training since the provision of palliative care can be stressful for a multitude of reasons, including bereavement. Participants received a link to the app through email, which was available for 1 month. Initially, users were asked to type their name, indicate their stress level, perform a breathing exercise after a simple gif-format graphic cue, and indicate their concluding stress level. The stress level was measured with a 10-point scale (i.e., 1 [lowest]–10 [highest] stress levels). The gif displayed 3-sec inhale and 4-sec exhale cycles repeatedly. Participants could do as many breathing cycles as desired.

Measures

Several measures were obtained from the data recorded in the breathing app: (1) the initial stress level before exercise (called before-stress); (2) exercise duration in minutes (called duration), calculated based on timestamps (the start and end time of breathing exercise) recorded in the app; (3) change in stress level between before and after each breathing exercise (called stress change); (4) frequency of app use (called frequency), obtained by counting breathing exercise occasions of each user during a 1-month period; (5) the number of days the app was used (called app use days), obtained by counting the days the app was used in the 1-month period. The frequency and app use days were not the same for those who used the app more than once per day. Participant demographic and work-related information (i.e., gender, age, race, education level, and tenure in the field) was collected in a baseline survey. In a 1-month follow-up survey, participants were asked the number of days they used the breathing exercise app in the past month. This allowed comparison of participant self-reports and objective use data recorded in the app (app use days).

Analysis

When measurements are taken multiple times from individuals, the hierarchical or nested structure of the data should be considered in the analyses.³⁹ Since measurements within individuals are not likely to be independent (e.g., stress levels of the same individual are likely to be correlated), using traditional single-level analysis, which assumes independence of observations, is problematic for hierarchical data.³⁹

Data from participants who used the app at least two times over a month were analyzed to examine the change in stress levels using a two-level multilevel modeling. The analytical data included 279 app use occasions by 64 participants. All analyses were performed using Stata version 14.2.

Sample characteristics were examined using percentages and means. Differences in stress change by demographic or work-related characteristics were examined using an independent t-test or analysis of variance (ANOVA). The correlation between the recorded app use days and self-reported app use days was examined using Pearson's correlation.

A series of random intercept models were estimated using the xtreg command in Stata in order of increasing complexity. First, an unconditional intercept-only model was fitted without any independent variable. The purpose of the unconditional intercept-only model is to divide the total variance in the dependent variable into a variance that occurs between individuals and a variance that occurs within individuals.⁴⁰ Using the estimates from the unconditional intercept-only model, intraclass correlation (ICC) was calculated to assess the proportion of the dependent variable's total variance that is due to individual differences (e.g., stable individual traits).⁴⁰ To examine the change in the dependent variable over time (e.g., whether stress change increases with repeated breathing exercises), dummy variables for app use occasions (occasion 1, occasion 2, etc.) were added to the unconditional intercept-only model and models with independent variables.

To test Hypothesis 1, a model with each variable, before-stress and duration, was fitted. To test Hypothesis 2, a model included significant within-individual-level variables and each between-individual-level variable (frequency, gender [male or female], race indicator variables [White, African American, and other], age in years, and tenure in the field in years). To test Hypothesis 3, a model included significant within-individual and between-individual-level covariates and a cross-level interaction term. Akaike's information criterion was obtained for each model to compare model fits. Akaike's information criterion is used to identify the most adequate model among alternative models by considering both the goodness-of-fit and parsimony of models.⁴¹

Ethics

The study was approved by the university institutional review board. Corresponding review committees of the participating organizations (i.e., IDD agencies) approved the study protocol as well.

Results

Participant characteristics

Table 1 summarizes participant characteristics. The majority of the participants were female (85.9%), White (65.6%), or Black/African American (24.4%), and less than a 4-year college education (high school, some college, or an associate's degree) (51.6%). On average, participants were 38.7 (standard deviation [SD] = 11) years old and had worked for 11.5 (SD = 8.3) years in the field.

Table 1.

Participant Characteristics (n = 64)

	Mean (SD)	Range
Age in years	38.7 (11.0)	20–62
Working in IDD field in years	11.5 (8.3)	1–32
	N	%
Gender
Female	55	85.9
Male	9	14.1
Race
White	42	65.6
Black/African American	15	24.4
Other	7	10.9
Education
High school	9	14.1
Some college	18	28.1
Associate's degree	6	9.4
College	25	39.1
Graduate/Professional	6	9.4

IDD, intellectual and developmental disability; SD, standard deviation.

Breathing exercise and stress levels

Table 2 summarizes information on breathing exercise and stress levels during a 1-month period. On average, participants did the breathing exercise 4.4 (SD = 4.1) times (range = 2–29) with ∼41% performing the breathing exercise twice and 20% more than five times during that time (Fig. 1). Eight participants performed the breathing exercise more than once per day (range = 1.1–2 times). On average, participants spent 1.9 (SD = 1.2) min per breathing exercise. Among 64 participants, 58 reported the number of days of app use in the 1-month follow-up survey. The number of days performing the breathing exercise recorded in the app (mean = 4.4, SD = 4.2) was smaller than the self-reported in the 1-month follow-up survey (mean = 10.7, SD = 8.2). The discrepancy between the reported and recorded use days (mean = 6.3, SD = 8.8, ranging −9 to 28), however, was not associated with other variables, including the outcome variable.

FIG. 1.

Frequency of breathing exercise in the past month (n = 64).

Table 2.

Breathing Exercise and Stress Levels in the Past Month (n = 64)

	Mean (SD)	Range
Frequency of breathing exercise	4.4 (4.1)	2 to 29
Duration of breathing exercise in minutes^a	1.9 (1.2)	0.5 to 7.4
No. of days, recorded in app (n = 58)	4.4 (4.2)	1 to 29
No. of days, self-reported (n = 58)	10.7 (8.2)	1 to 30
Stress level before breathing exercise^b
All (n = 64)	5.0 (1.8)	1 to 9.5
White (n = 42)	5.2 (1.6)	1.1 to 9.5
African American (n = 15)	4.5 (2.3)	1 to 7.3
Other (n = 7)	3.0 (1.8)	2.2 to 7.7
Stress level after breathing exercise^b
All (n = 64)	3.7 (1.6)	1 to 7
White (n = 42)	4.0 (1.5)	1 to 7
African American (n = 15)	3.0 (1.7)	1 to 5.7
Other (n = 7)	3.7 (1.7)	1.5 to 6.2
Change in stress from before to after^b
All (n = 64)	−1.2 (0.9)	−4.3 to 0.5
White (n = 42)	−1.2 (0.8)	−3 to 0.5
African American (n = 15)	−1.4 (1.2)	−4.3 to 0
Other (n = 7)	−1.1 (0.8)	−2 to 0

Per breathing exercise occasion.

No significant difference in means of the three race groups from analysis of variance (p > 0.05).

SD, standard deviation.

Our sample appears to have a moderate stress level indicated by the initial stress level 5.0 out of 10. A study of DCWs in the United Kingdom⁴² reported a mean stress level of 14.4 out of 40 using the perceived stress scale. Though different metrics, our sample seems to report a little higher stress level than the U.K. DCW sample. After the breathing exercise, the stress level decreased by 1.2 points from 5.0 (SD = 1.8) to 3.7 (SD = 1.6) on average (p < 0.0005 in a paired t-test). Cohen's d 0.72 indicates a large effect size.^43,44 Figure 2 shows patterns of stress change over time by connecting the measures of stress change at breathing exercise occasions. Upon visual examination, no clear pattern was observed. Stress reduction appeared to be larger among African Americans than among other groups although the difference was not statistically significant (Fig. 3). There was no difference in the stress change by age, gender, education, or tenure in the field in an independent t-test or ANOVA (results not shown).

FIG. 2.

Patterns of change in stress level over time, overlaid (n = 64). Note: The measures of change in stress level at breathing exercise occasions are connected with lines. Individuals are identified in different lines.

FIG. 3.

Mean stress level before and after breathing exercise by race (n = 64).

Results from multilevel modeling

According to the unconditional random intercept-only model (with no independent variables, results not shown), ICC was 0.36, meaning that between-individual differences explained 36% of the total variance of the dependent variable. Between-individual factors could be a person's personality, environment, and general long-term health status, etc.⁴⁵ Within-individual factors could be stressful experience or emotional status of the time when a measurement is taken.⁴⁵ According to a likelihood ratio test, the random intercept variance is significantly greater than zero (p < 0.0005), indicating the appropriateness of using a random effects model over a traditional single-level analysis. The regression model with dummy variables for breathing exercise occasions indicated no discernable pattern over time as suggested in Figure 2.

Table 3 presents the results from conditional random intercept models (with independent variables). Model 1 and Model 2 are relevant to Hypothesis 1. The results indicate that stress decreases after breathing exercises as each before-stress and duration increases. Model 3 indicates that both before-stress and duration are still significant when they are included in a model together. Model 4 was run to test Hypothesis 2. The result indicates that African American race is associated with a decrease in stress level (coefficient = −0.66, confidence interval [95% CI] = −1.11 to −0.21, p < 0.01), controlling for before-stress and duration. Other race variables (White and Other) were not associated with the dependent variable (results not shown). Model 5 is relevant to Hypothesis 3. The significant interaction term indicates that there is an additional 0.23-point decrease in stress level with a 1-point increase in before-stress for African Americans (coefficient = −0.23, 95% CI = −0.33 to −0.11, p < 0.0005). Although there is no effect of being an African American alone on stress change (coefficient = 0.44, 95% CI = −0.29 to 1.18, p > 0.05), there is an additional reduction among African Americans in relation to the initial stress level, controlling for duration.

Table 3.

Coefficients and 95% Confidence Intervals of Random Intercept Models for Change in Stress Level Through Breathing Exercise (n = 64)

Variable	Model 1	Model 2	Model 3	Model 4	Model 5
Before-stress^a	−0.325^** (−0.388 to −0.263)		−0.316^** (−0.377 to −0.255)	−0.318^** (−0.378 to −0.258)	−0.230^** (−0.305 to −0.155)
Duration^a		−0.123^* (−0.195 to −0.050)	−0.095^* (−0.157 to −0.032)	−0.104^* (−0.168 to −0.041)	−0.100^* (−0.161 to −0.039)
African American^b				−0.661^* (−1.11 to −0.212)	0.443 (−0.293 to 1.179)
Interaction^c					−0.229^** (−0.350 to −0.108)
Random intercept variance	0.496	0.579	0.438	0.337	0.354
Residual variance	0.780	1.061	0.769	0.778	0.729
ICC	0.389	0.354	0.363	0.303	0.326
AIC	807	886	800	795	783

Within-invididual-level time-variant variable.

Between-invididual-level time-invariant variable.

Interaction between African American race and before-stress.

^*p < 0.005, ^**p < 0.0005.

Before-stress, stress level before breathing exercise; duration, minutes of breathing exercise; ICC, intraclass correlation; AIC, Akaike's information criterion.

AIC indicates improvements in model fit (i.e., smaller AIC) as the model progressed from Model 1 through Model 5 (Table 3).⁴⁶ From the unconditional model to Model 5, ∼43% of the unexplained variance in stress change was reduced, as the proportion of reduction in variance is: (variance of residuals within groups from a reference model − variance of residuals within groups from a conditional model)/variance of residuals within groups from a reference model.

Discussion

The results indicate the effect of a simple deep breathing exercise in reducing stress among DCWs who serve people with IDDs. In regression models, higher initial stress and longer exercise duration were associated with stress reduction. Controlling for time-variant factors, among individual-level characteristics, only race had a significant impact on stress reduction. African Americans seemed to experience greater stress reduction through the breathing exercise in relation to initial stress. There was additional stress reduction if they had a higher initial stress level, controlling for exercise duration.

Greater stress reduction postapp use among participants with a higher initial stress level and a longer exercise duration is as expected. A higher stress level is likely to result in a lower stress level after a breathing exercise, while reducing stress from a lower stress level would be difficult. Given the short breathing exercise duration (mean = 1.9 min), taking a longer time for the breathing exercise may have helped reduce stress additionally.

The reasons for African American participants experiencing greater benefit of the deep breathing exercise are not clear in our data, and the literature does not provide clear empirical evidence for or against this finding. Stress levels either before or after the breathing exercise were not significantly different among racial groups (Table 2). In bivariate analysis, being African American was not associated with either frequency of app use or duration of app use. Since a low percentage of African American participants in our study used the app more than once, they may have been particularly motivated to use it, and received benefit from it. Among studies about the effectiveness of mindfulness-based stress reduction interventions,^29,30 only a few identified participant racial composition without reporting variation in outcomes by race.^47
–49 It is possible that being African American is correlated with unmeasured variables or there are unobserved associations. Greater intervention benefits for African Americans could be related to the level of social capital acquired, due to persistent racial segregation and discrimination.¹¹ Although African Americans tend to report strong social capital from ties to members of their community,^11,12 their social networks tend to be small and composed of individuals who are similarly struggling with their own stressors and lack of resources^12,50,51 Other types of interventions have demonstrated better intervention responses among African American participants than White participants.^52,53 After a 10-week cognitive behavioral stress management intervention for men with advanced prostate cancer, Black men reported greater reduction in prostate cancer-related anxiety than White participants.⁵² In a one-group pre–post-test design study, Black older adults reported larger increase in social connectedness than White counterparts, controlling for gender, after 12-week computer and internet training intervention.⁵³

Study limitations include the use of convenience sampling in a small geographic area that limits the generalizability of results. Life circumstance and working conditions of DCWs, which are likely to be closely related to stress, may vary across different types of organizations and areas in the United States. Information that may directly influence stress levels, such as stressors at work at the time of the app use, was not obtained. Collecting such information might have negative effects, such as deterring app use. It is possible that an improper use, such as leaving the app open without doing breathing exercise, has been recorded as a valid use. The reasonable range of duration (0.5–7.4 min), however, suggests that participants actually used the app for the duration.

Despite the limitations, this study improved upon existing studies in several ways. Objective information on app use behavior allowed comparison of the app data with a self-report. The discrepancy in the number of app use days in the app and survey underscores the importance of having verifications of self-report data. Multiple breathing exercise occasions per individual were recorded for 1 month. This allowed examination of the patterns based on multiple data points. Time-variant independent variables, concurrently measured with the dependent variable, were included in analysis. By using multilevel analysis methods, all cases could be included in analysis despite unequal spacing of occasions and unequal number of occasions across individuals.⁴⁵ Using multilevel analysis, unbiased estimates were obtained of the effect of each time-variant and time-invariant covariate on the dependent variable while considering the nested structure of the variables.

Conclusions

The findings suggest the benefit of a deep breathing app for reducing DCW stress levels. To identify the reasons for larger stress reduction among African American participants, further research is needed with more information on between-individual and within-individual factors using larger samples. In addition to organizational policies and training that prevent and reduce worker stress, regular deep breathing using apps can help support stress management and bolster health and well-being among DCWs.

Footnotes

Acknowledgments

We are thankful for the support of Northern Illinois University and, specifically, the College of Health and Human Sciences.

Authors' Contributions

J.K. formulated the focus of the study, analyzed the data, cowrote the article, and assisted with editing. J.A.G. cowrote the article and assisted with editing. H.J. assisted with preliminary data analysis, the literature review, and general article editing.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

PHI National. Direct care workers in the United States: Key facts. Online document at: https://phinational.org/resource/direct-care-workers-in-the-united-states-key-facts, accessed October 1, 2020 .

Gray-Stanley

, Muramatsu

. Work stress, burnout, and social and personal resources among direct care workers. Res Dev Disabil, 2011; 32:1065–1074.

Vassos

, Nankervis

, Skerry

, Lante

. Can the job demand-control-(support) model predict disability support worker burnout and work engagement?. J Intellect Dev Disabil, 2017; 44:1–11.

Vassos

, Nankervis

. Investigating the importance of various individual, interpersonal, organisational and demographic variables when predicting job burnout in disability support workers. Res Dev Disabil, 2012; 33:1780–1791.

Truitt

, Snyder

. Racialized experiences of black nursing professionals and certified nursing assistants in long-term care settings. J Transcult Nurs, 2019; 31:312–318.

Hurtado

, Sabbath

, Ertel

, et al. Racial disparities in job strain among American and immigrant long-term care workers. Int Nurs Rev, 2012; 59: 237–244.

McCord

, Joseph

, Dhanani

, Beus

. A meta-analysis of sex and race differences in perceived workplace mistreatment. J Appl Psychol, 2018; 103:137–163.

Koinis

, Giannou

, Drantaki

, et al. The impact of healthcare workers job environment on their mental-emotional health. Coping strategies: The case of a local general hospital. Health Psychol Res, 2015; 3:1984.

Griffith

, Hasley

, Liu

, et al. Qigong stress reduction in hospital staff. J Altern Complement Med, 2008; 14:939–945.

10.

Poulin

, Mackenzie

, Soloway

, Karayolas

. Mindfulness training as an evidence-based approach to reducing stress and promoting well-being among human service professionals. Int J Health Promot Educ, 2008; 46:72–80.

11.

Dean

, Subramanian

, Williams

, et al. The role of social capital in African-American women's use of mammography. Soc Sci Med, 2014; 104:148–156.

12.

Mitchell

, LaGory

. Social capital and mental distress in an impoverished community. City Commun, 2002; 1:199–222.

13.

Clausen

, Meng

, Borg

. Does social capital in the workplace predict job performance, work engagement, and psychological well-being? A prospective analysis. J Occup Environ Med, 2019; 61:800–805.

14.

Snowden

. Racial, cultural and ethnic disparities in health and mental health: Toward theory and research at community levels. Am J Community Psychol, 2005; 35:1–8.

15.

Keesler

, Troxel

. They care for others, but what about themselves? Understanding self-care among DSPs' and its relationship to professional quality of life. Intellect Dev Disabil, 2020; 58:221–240.

16.

Vargoli

, Darviri

. Stress management techniques: Evidence-based procedures that reduce stress and promote health. Health Sci J, 2001; 5:74–89.

17.

White

, McConnell

, Clipp

, et al. A randomized controlled trial of the psychosocial impact of providing internet training and access to older adults. Aging Ment Health, 2020; 6:213–221.

18.

Mikolasek

, Berg

, Witt

, Barth

. Effectiveness of mindfulness-and relaxation-based eHealth interventions for patients with medical conditions: A systematic review and synthesis. Int J Behav Med, 2018; 25:1–16.

19.

Newby

, Teah

, Cooke

, et al. Do automated digital health behaviour change interventions have a positive effect on self-efficacy? A systematic review and meta-analysis. Health Psychol Rev, 2019:1–19.

20.

Rose

, Barker

, Maria Jacob

, et al. A systematic review of digital interventions for improving the diet and physical activity behaviors of adolescents. J Adolesc Health, 2017; 61:669–677.

21.

Russo

, Santarelli

, O'Rourke

. The physiological effects of slow breathing in the healthy human. Breathe, 2017; 13:298–309.

22.

Perciavalle

, Blandini

, Fecarotta

, et al. The role of deep breathing on stress. Neurol Sci, 2017; 38:451–458.

23.

Cheng

, Croarkin

, Lee

. Heart rate variability of various video-aided mindful deep breathing durations and its impact on depression, anxiety, and stress symptom severity. Mindfulness, 2019; 10:2082–2094.

24.

Beng

, Ahmad

, Loong

, et al. Distress reduction for palliative care patients and families with 5-minute mindful breathing: A pilot study. Am J Hosp Palliat Care, 2016; 33:555–560.

25.

Saoji

, Raghavendra

, Manjunath

. Effects of yogic breath regulation: A narrative review of scientific evidence. J Ayurveda Integr Med, 2019; 10:50–58.

26.

Brown

, Gerbarg

. Sudarshan Kriya Yogic breathing in the treatment of stress, anxiety, and depression: Part II - clinical applications and guidelines. J Altern Complement Med, 2005; 11:711–717.

27.

Xenaki

, Bacopoulou

, Kokkinos

, et al. Impact of a stress management program on weight loss, mental health and lifestyle in adults with obesity: A randomized controlled trial. J Mol Biochem, 2018; 7:78–84.

28.

Jerath

, Crawford

, Barnes

, Harden

. Self-regulation of breathing as a primary treatment for anxiety. Appl Psychophysiol Biofeedback, 2020; 40:107–115.

29.

Burton

, Burgess

, Dean

, et al. How effective are mindfulness-based interventions for reducing stress among healthcare professionals? A systematic review and meta-analysis. Stress Health, 2017; 33:3–13.

30.

Kriakous

, Elliott

, Lamers

, et al. The effectiveness of mindfulness-based stress reduction on the psychological functioning of healthcare professionals: A systematic review. Mindfulness (N Y), 2020; 1–28.

31.

Yang

, Tang

, Zhou

. Effect of mindfulness-based stress reduction therapy on work stress and mental health of psychiatric nurses. Psychiatr Danub, 2018; 30:189–196.

32.

van der Riet

, Rossiter

, Kirby

, et al. Piloting a stress management and mindfulness program for undergraduate nursing students: Student feedback and lessons learned. Nurs Educ Today, 2015; 35:44–49.

33.

Jones

. A self-care plan for hospice workers. Am J Hosp Palliat Care, 2005; 22:125–128.

34.

Manpreet

, Anil

, Kumar

. Effectiveness of deep breathing exercises and progressive muscle relaxation technique on stress among amputated patients. Baba Farid Univ, 2016; 11:23–27.

35.

Chan

, Kohab

, Teo

, et al. Biochemical and psychometric evaluation of Self-Healing Qigong as a stress reduction tool among first year nursing and midwifery students. Complement Ther Clin Pract, 2013; 19:179–183.

36.

Pizutti

, Carissimi

, Valdivia

, et al. Evaluation of Breathworks' Mindfulness for Stress 8-week course: Effects on depressive symptoms, psychiatric symptoms, affects, self-compassion, and mindfulness facets in Brazilian health professionals. J Clin Psych, 2019; 75:970–984.

37.

Peterson

, Bauer

, Chopra

, et al. Effects of Shambhavi Mahamudra Kriya, a multicomponent breath-based yogic practice (pranayama), on perceived stress and general well-being. J Evid Based Complementary Altern Med, 2017; 22:788–797.

38.

Rawat

. A study to assess the effectiveness of deep breathing exercise to reduce the level of stress among housewives in a selected rural area of Udaipur. Asian J Pharm Res, 2018; 6:93–96.

39.

Raudenbush

, Bryk

. Hierarchical Linear Models: Applications and Data Analysis Methods. 2nd ed. Newbury Park, CA: Sage, 2002.

40.

Merlo

, Chaix

, Yang

, et al. A brief conceptual tutorial of multilevel analysis in social epidemiology: Linking the statistical concept of clustering to the idea of contextual phenomenon. J Epidemiol Community Health, 2005; 59:443–449.

41.

Sawa

. Information criteria for discriminating among alternative regression models. Econometrica, 1978; 46:1273–1291.

42.

Smyth

, Healy

, Lydon

. An analysis of stress, burnout, and work commitment among disability support staff in the UK. Res Dev Disabil, 2015; 47:297–305.

43.

Ialongo

. Understanding the effect size and its measures. Biochemia Medica, 2016; 26:150–163.

44.

Maher

, Markey

, Ebert-May

. The other half of the story: Effect size analysis in quantitative research. CBE Life Sci Educ, 2013; 12:345–351.

45.

Hoffman

, Stawski

. Persons as contexts: Evaluating between-person and within-person effects in longitudinal analysis. Res Hum Dev, 2009; 6:97–100.

46.

Kwok

, Underhill

, Berry

, et al. Analyzing longitudinal data with multilevel models: An example with individuals living with lower extremity intra-articular fractures. Rehab Psych, 2008; 53:370–386.

47.

Crowder

, Sears

. Building resilience in social workers: An exploratory study on the impacts of a mindfulness-based intervention. Aust Soc Work, 2017; 70:17–29.

48.

Fortney

, Luchterhand

, Zakletskaia

, et al. Abbreviated mindfulness intervention for job satisfaction, quality of life, and com passion in primary care clinicians: A pilot study. Annal Fam Med, 2013; 11:412–420.

49.

Geary

, Rosenthal

. Sustained impact of MBSR on stress, well-being, and daily spiritual experiences for one year in academic healthcare employees. J Altern Complement Med, 2011; 17:939–944.

50.

Barnes

, Mendes de Leon

, Bienias

, et al. A longitudinal study of black–white differences in social resources. J Gerontol B Psychol Sci Soc Sci, 2004; 59:146-S153.

51.

Ajrouch

, Antonucci

, Janevic

. Social networks among blacks and whites: The interaction between race and age. J Gerontol B Psychol Sci Soc Sci, 2001; 6:S112–S118.

52.

Bouchard

, Yanez

, Dahn

, et al. Brief report of a tablet-delivered psychosocial intervention for men with advanced prostate cancer: Acceptability and efficacy by race. Transl Behav Med, 2019; 9:629–637.

53.

Kim

, Gray

, Ciesla

. Impacts of a computer and Internet skills training program on communication and social connectedness among low-income older adults. In: Oral presentation at: American Public Health Association Annual Meeting, November 15–19, New Orleans, LA, 2014.