Abstracts

Introduction

Sense of coherence (SOC) is a concept introduced by Aaron Antonovsky’s salutogenesis framework. It aims to understand the contributing factors in the development of health and to explain how individuals can manage their lives despite adverse conditions ¹1. Antonovsky A. Health, stress and coping. London: Jossey-Bass; 1979.^,22. Antonovsky A . Unraveling mystery of health: how people manage stress and stay well. San Francisco: Jossey-Bass; 1987.. SOC comprises three components: comprehensibility (ability to understand an event), manageability (perceived potential to manipulate or resolve the event), and meaningfulness (significance attributed to this event) ²2. Antonovsky A . Unraveling mystery of health: how people manage stress and stay well. San Francisco: Jossey-Bass; 1987..

A strong SOC empowers individuals to mobilize internal and external resources to effectively cope with stressors and manage tension, thereby promoting and maintaining their health ¹1. Antonovsky A. Health, stress and coping. London: Jossey-Bass; 1979.. Individuals with a strong SOC would be more efficient in creating coping mechanisms and strategies to maintain health in unfavorable situations ²2. Antonovsky A . Unraveling mystery of health: how people manage stress and stay well. San Francisco: Jossey-Bass; 1987.. Urakawa & Yokoyama ³3. Urakawa K, Yokoyama K. Sense of coherence (SOC) may reduce the effects of occupational stress on mental health status among Japanese factory workers. Ind Health 2009; 47:503-8. observed that SOC is negatively associated with stress levels and positively correlated with the ability to cope with stress. This reduction in stress would positively influence health-related behaviors, contributing to maintain a positive health status.

Evidence suggests that SOC is associated with various health-related behaviors like tobacco smoking ⁴4. Silarova B, Nagyova I, Rosenberger J, Studencan M, Ondusova D, Reijneveld SA, et al. Sense of coherence as a predictor of health-related behaviours among patients with coronary heart disease. Eur J Cardiovasc Nurs 2014; 13:345-56., alcohol intake ⁵5. Morita Y, Ohta M, Inoue T, Honda T, Konno Y, Eguchi Y, et al. Sense of coherence is significantly associated with both metabolic syndrome and lifestyle in Japanese computer software office workers. Int J Occup Med Environ Health 2014; 27:967-79., and illicit drug use ⁶6. Ristkari T, Sourander A, Helenius H, Nikolakaros G, Salanterä S, Multimäki P, et al. Sense of coherence among Finnish young men - a cross-sectional study at military call-up. Nord J Psychiatry 2005; 59:473-80.. These behaviors are associated with the development of noncommunicable diseases (NCDs) such as cardiovascular diseases, cancer, diabetes, and chronic respiratory conditions, which account for two-thirds of the overall burden of disease in low- and middle-income countries ⁷7. GBD 2019 Risk Factors Collaborators. Global burden of 87 risk factors in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet 2020; 396:1223-49.^,88. Institute for Health Metrics and Evaluation. GBD results. https://vizhub.healthdata.org/gbd-results/ (accessed on 13/Jun/2023).
https://vizhub.healthdata.org/gbd-result... . In 2019, smoking resulted in 8.71 million attributable deaths (15.4% of all deaths), alcohol use led to 2.07 million attributable deaths among men and 0.40 million among women, and drug use contributed to 0.45 million attributable deaths ⁸8. Institute for Health Metrics and Evaluation. GBD results. https://vizhub.healthdata.org/gbd-results/ (accessed on 13/Jun/2023).
https://vizhub.healthdata.org/gbd-result... . SOC would also have an impact in the development of NCDs through health-related behaviors.

It is crucial to recognize and critically assess the methodological limitations within research. Many studies assessed the association between SOC and health-related behaviors without controlling for potential confounding factors. Failure to address known confounders can overestimate the association. For example, low socioeconomic status is associated with lower SOC scores and less favorable behavioral habits. Consequently, analyses that fail to control for confounding by socioeconomic status are susceptible to residual confounding which would overestimate the magnitude of the association.

Currently, only the study by da-Silva-Domingues et al. ⁹9. da-Silva-Domingues H, Del-Pino-Casado R, Palomino-Moral PÁ, López Martínez C, Moreno-Cámara S, Frías-Osuna A. Relationship between sense of coherence and health-related behaviours in adolescents and young adults: a systematic review. BMC Public Health 2022; 22:477. has reviewed the relation between SOC and substance use, but as part of a broader analysis that focused on health behaviors such as eating habits, time spent on computers, rest periods, as well as smoking, alcohol consumption, and oral health care. However, it specifically evaluates the association between SOC and substance use (tobacco, alcohol, and illicit drugs). It is crucial to recognize how SOC influences both general health behaviors and specifically risky behaviors like substance use, given their significant impact on public health and the development of chronic diseases. While da-Silva-Domingues et al.’s study ⁹9. da-Silva-Domingues H, Del-Pino-Casado R, Palomino-Moral PÁ, López Martínez C, Moreno-Cámara S, Frías-Osuna A. Relationship between sense of coherence and health-related behaviours in adolescents and young adults: a systematic review. BMC Public Health 2022; 22:477. assessed individuals aged 12 to 30 years, our research seeks to bridge a gap by concentrating on the adult population as defined by the World Health Organization (WHO). By identifying heterogeneity sources and conducting a meta-analysis, the present study intends to offer insights beyond the existing literature which can guide future research and interventions aimed at reducing harmful substance use among adults.

In short, this study reviewed the literature on the association between substance use (tobacco, alcohol, and illicit drugs) and SOC, while exploring sources of heterogeneity.

Methods

This systematic review was registered on PROSPERO (protocol n. CRD42023402776) and conducted according to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and AMSTAR2 (A MeaSurement Tool to Assess systematic Reviews) guidelines (Supplementary Material - Boxes 1 and 2; https://cadernos.ensp.fiocruz.br/static//arquivo/suppl-e00141323_3035.pdf). We formulated the following research question: “What is the association between sense of coherence and substance use in adults?” (P = adults; I = high SOC; C = low SOC; O = substance use outcomes).

Search protocol and selection criteria

Bibliographic search was conducted in March 2023 on the PubMed, LILACS, Web of Science, and PsycINFO databases. Search strategies combined the terms for SOC (“sense of coherence” OR salutogen* OR “general resistance resources”) with the following terms for each of the studied outcomes:

Supplementary Material - Table S1 (https://cadernos.ensp.fiocruz.br/static//arquivo/suppl-e00141323_3035.pdf) shows the search strategy and the number of studies identified.

Inclusion criteria consisted of original articles that evaluated the association of SOC with at least one measure of substance use among adult participants (age > 19 years). We set no restrictions on language, publication date, or the SOC measurement scale used. Papers involving animals, research protocols, editorials, comments, and those with insufficient data were excluded.

Additionally, we searched for grey literature in the Google Scholar, CAPES Portal of Theses and Dissertations, and ProQuest databases.

Data extraction (selection and coding)

Study selection was performed in two phases. First, two independent reviewers evaluated the title and abstract of each identified study. Articles considered as possibly eligible for inclusion in the review were retrieved and read in full. Discrepancies between the reviewers regarding the inclusion or exclusion of a paper were resolved by a third reviewer.

Last name of the first author, year of publication, country where the study was conducted, sample size, age and gender of the studied population, study design, scale used for substance use measurement, categorization of substance use, scale used for SOC assessment, SOC categorization, control for confounding factors, and effect measurement along with its 95% confidence interval (95%CI) were extracted. Two independent reviewers extracted the data using a Microsoft Excel spreadsheet (https://products.office.com/). Any discrepancies between reviewers were resolved through consensus or consultation with a third reviewer. Subsequently, the extracted data were transferred to Stata software (https://www.stata.com).

Methodological quality of the included studies was evaluated using the Risk of Bias in Non-randomized Studies of Exposures scale (ROBINS-e) ¹⁰10. Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, et al., editors. Cochrane handbook for systematic reviews of interventions. 2nd Ed. Chichester: John Wiley & Sons; 2019., recommended by the Cochrane Collaboration for assessing effectiveness and safety in nonrandomized intervention trials. This instrument has seven domains of bias categorized by the timing of occurrence: pre-intervention (confounding and selection bias in participant enrollment), at intervention (bias in classifying interventions), and post-intervention (deviations from intended interventions, missing data bias, outcome measurement bias, and bias in selection of reported results). Assessment classified items as low, moderate, severe, or critical risk of bias, or as having no information, following the Cochrane Handbook for Systematic Reviews of Interventions¹¹11. Higgins JPT , Thomas J , Chandler J , Cumpston M , Li T , Page MJ , et al., editors. Cochrane handbook for systematic reviews of interventions, version 6.4 (updated August 2023). http://www.training.cochrane.org/handbook (accessed on 18/Apr/2024).
http://www.training.cochrane.org/handboo... .

Data synthesis and analysis plan

For inclusion in the meta-analysis, the studies had to report a crude or adjusted measure of the association between SOC and substance use. We considered the following reported measures: mean values of SOC for both users and non-users, linear regression coefficients, odds ratios (OR), and prevalence ratios of substance use in different SOC categories along with their corresponding 95%CI or standard errors.

Pooled measure of association was calculated using a random effects model. We conducted two distinct meta-analyses - one focused on substance use outcomes; the other focused on the mean SOC - using Stata. The analyses were stratified according to the type of substance, follow-up rate, sample size, age group, confounding variables (socioeconomic status), study setting (America, Asia, Europe), SOC categorization, smoking and alcohol.

Heterogeneity between studies was assessed using the Q-test and I-square. Estimates were pooled using a random-effects model. Publication bias was assessed using a funnel plot and Egger’s test. We also stratified the analysis according to sample size to further evaluate the impact of publication bias on the estimates.

Certainty of evidence was evaluated using the GRADEpro software (https://www.gradepro.org/). The GRADE (Grading of Recommendations Assessment, Development, and Evaluation) system categorizes evidence quality into four levels - high, moderate, low, and very low - based on considerations such as study design limitations, indirect evidence, inconsistency of results, imprecision of results, and probability of publication bias.

Results

Characteristics of the included studies

Database search identified a total of 566 records (Figure 1). After removing duplicates (n = 206 articles), 360 titles and abstracts were read resulting in 54 studies selected for full text review. In the end, the review sample included 21 articles. Exclusion criteria consisted of articles with outcomes and exposure that were not of interest (n = 14), user sample (n = 6), and lack of data on the association between SOC and substance use (n = 8). Supplementary Material - Box S3 (https://cadernos.ensp.fiocruz.br/static//arquivo/suppl-e00141323_3035.pdf) shows the references excluded after full text review and the respective reason. An additional manual search was performed on the references of the 21 selected articles, but we identified no additional study.

Characterization and qualitative synthesis of the selected studies

Table 1 presents the main characteristics of the included studies. Most were cross-sectional (n = 14) and longitudinal (n = 6), conducted in European countries (n = 17) and published from 1987 to 2022. Sample size varied considerably, ranging from 120 to 40,674 participants. Participant age also exhibited a wide range, from 15 to 88 years. SOC was assessed by different instruments, namely: SOC-L9 (n = 2), SOC-7 (n = 1), SOC-3 (n = 3), SOC-29 (n = 7), and SOC-12 (n = 2).

Risk of bias assessment

Quality assessment of the risk of bias was predominantly moderate (n = 10), eight studies presented critical risk and four low risk of bias (Box 1). Missing data was the main criteria contributing to moderate or critical risk of bias.

Table 2 summarizes the certainty of the evidence for the outcomes included in the meta-analysis.

OR for substance use

Figure 2 summarizes the results from the 21 studies evaluating the association between SOC and substance use. Individuals with high SOC had a 22% lower odds of using any substance compared with individuals with low SOC (pooled effect: OR = 0.78, 95%CI: 0.68; 0.88, very low degree of certainty). When stratified by substance type, we observed that a strong SOC slightly reduced the odds of smoking (n = 11), but the confidence interval included the reference (pooled effect: OR = 0.92, 95%CI: 0.82; 1.01). Regarding alcohol use (n = 8), individuals with a strong SOC had 30% lower odds of using it (pooled effect: OR = 0.70, 95%CI: 0.50; 0.90, very low degree of certainty). Only two studies analyzed the association with use of illicit drugs, showing a pooled OR of 0.31 (95%CI: 0.04; 0.59).

Table 3 presents the subgroup analyses according to characteristics of the analyzed studies. Regarding smoking, studies that evaluated young adults showed a similar association (pooled effect: OR = 0.91, 95%CI: 0.74; 1.09) to those that evaluated older individuals (pooled effect: OR = 0.99, 95%CI: 0.75; 1.23). As for alcohol use, studies that adjusted their estimates for confounding variables reported a weaker association (pooled OR = 0.89, 95%CI: 0.80; 0.98) than those reporting crude estimates (pooled OR = 0.51, 95%CI: 0.22; 0.80). The pooled OR remained similar across studies that evaluated different age groups and continents.

Egger’s test results indicated a trend of small-study effects or publication bias, albeit not statistically significant (p = 0.07). However, we must consider that this analysis is underpowered.

Mean SOC

Seven studies assessed the mean SOC among substance users and non-users assessed using the SOC-13 scale. The pooled mean difference was -3.50 points on the SOC scale (95%CI: -5.47; -1.53) (total score can range from 13 to 65 points) for substance users compared with non-users (Figure 3).

Other measures - regression coefficients

One study estimated the beta coefficient, indicating that a 0.1-point decrease on the SOC scale corresponds to a 1-point increase on the Alcohol Problem Index (a Stern index scale).

Discussion

This systematic review identified 21 studies on the association between SOC and substance use in adults. Our findings indicate a negative association between SOC and substance use, suggesting that individuals with a strong SOC are less likely to use alcohol and illicit drugs.

Regarding confounding control, half of the studies included in the meta-analyses did not adjust their estimates for confounding by socioeconomic status or other variables. As for alcohol use, studies that reported crude estimates showed a stronger association compared with those that controlled for confounding variables, thereby suggesting that confounding overestimated the association between SOC and substance use. Socioeconomic and demographic factors play a crucial role in shaping meaningful experiences that contribute to developing a strong SOC in adulthood ¹²12. Bernabé E, Watt RG, Sheiham A, Suominen-Taipale AL, Nordblad A, Savolainen J, et al. The influence of sense of coherence on the relationship between childhood socioeconomic status and adult oral health-related behaviours. Community Dent Oral Epidemiol 2009; 37:357-65.. Socioeconomic status has been positively associated with SOC ¹³13. Togari T, Inoue Y, Oshima G, Abe S, Hosokawa R, Takaku Y. Socioeconomic status and the sense of coherence among Japanese people living with HIV. Int J Environ Res Public Health 2022; 19:7673. but negatively associated with substance use ¹⁴14. Hiscock R, Bauld L, Amos A, Fidler JA, Munafò M. Socioeconomic status and smoking: a review. Ann N Y Acad Sci 2012; 1248:107-23.^,1515. Probst C, Kilian C, Sanchez S, Lange S, Rehm J. The role of alcohol use and drinking patterns in socioeconomic inequalities in mortality: a systematic review. Lancet Public Health 2020; 5:e324-e32.. Consequently, socioeconomic status would overestimate the magnitude of the association between SOC and substance use. As previously described, the magnitude of the association between SOC and alcohol use was weak in those studies that controlled for confounding variables. Thus, further studies evaluating the association of SOC with substance should adjust their estimates to socioeconomic variables.

This study has several strengths, such as the independent literature search conducted by two authors. All studies included collected self-reported data on both substance use and SOC, thus minimizing the occurrence of information bias. Additionally, only studies using validated and standardized instruments to assess SOC were included, thus reducing the possibility of misclassification. One limitation concerns the small number of identified studies measuring issues such as illicit drug use.

Analysis revealed a more consistent association of SOC with alcohol use than with smoking; however, precise measurement of this relation for illicit drugs was hindered by the limited number of studies available. Scarcity of evidence prevents definitive conclusions about the association between SOC and illicit substance use. This research gap suggests that SOC may be more strongly linked to behaviors related to more severe substances like illicit drugs and hazardous alcohol consumption, but confirming this hypothesis would require more comprehensive and specific investigation. Acknowledging this limitation highlights the pressing need for future research focused on understanding the correlation between SOC and use of different types of substances, particularly those considered severe.

Certainty of evidence as assessed by the GRADE system was very low, thus more robust study designs such as randomized clinical trials and longitudinal studies are needed to better understand the impact of a strong SOC on substance use (alcohol, smoking, and illicit drugs). Future research could examine how specific socioeconomic aspects interact with SOC to influence substance use, looking at factors like income, education, and occupation individually. Increasing the certainty of evidence through rigorous study designs is crucial for elaborating effective interventions and public health policies aimed at reducing harmful substance use and promoting overall well-being. Including diverse populations in research, considering ethnicity, culture, and geography, would enhance result generalizability and provide a more comprehensive understanding of how SOC operates across various demographic groups. Moreover, the consistency of our findings with another recent review involving adolescent and young adult populations highlights SOC as a potential protective factor against harmful substance use, reinforcing the clinical and public health relevance of these findings.

Overall, this systematic review and meta-analysis suggest that a strong SOC protects against substance use (alcohol, smoking, and illicit drugs) among adults, regardless of age. As practical implications, these findings suggest that early identification of individuals with low SOC may indicate the need for preventive interventions related to substance use. Recognizing and addressing a diminished SOC early could guide tailored interventions aimed at fortifying resilience and mitigating the risk of substance-related issues, especially in high-risk individuals.

References

Publication Dates

History