Introduction

Health disparities among communities have long been a concern in public health studies, with considerable differences in health outcomes revealed [1]. The concept of Social Determinants of Health (SDOH) is a critical lens through which these disparities can be examined. The term "social determinants of health" refers to the multifaceted factors encompassing an individual's birth, growth, life experiences, occupational situations, and aging processes, collectively contributing to their overall well-being [2]. SDOH comprises several health-related societal, economic, and environmental factors that can be categorized into groups of healthcare resources and access, education quality and access, economic stability, social and community context, and neighborhood and built environments [2-4]. Social determinants of health (SDOH) significantly impact health outcomes [3], [5], [6].

Multiple studies have explored the relationship between social determinants of health (SDOH), cancer screening habits, and cancer stage upon detection.  Current research indicates that SDOH substantially impacts an individual's propensity towards cancer screening and overall trajectory throughout the diagnostic process [7-9]. Studies have shown that a person's income, education, and other socioeconomic factors can affect their ability to access cancer screening programs and follow-up care [10-11]. Additionally, the availability of healthcare facilities and access options in their community can play a significant role in timely screening [10], [12], [13]. Furthermore, research has emphasized the importance of social determinants of health (SDOH) in affecting the stage at which cancer is diagnosed [7]. Differences in exposure to risk factors, access to healthcare, and health awareness can contribute to disparities in cancer diagnosis across different communities [7], [12], [13].

This systematic review builds on reports of racial disparities in cancer diagnosis and outcomes, especially within the Black community [7], [9], [14]. Despite advancements in medicine and healthcare infrastructure, these disparities persist, highlighting the need for a comprehensive investigation of the underlying reasons [7], [15]. Reports show that black people experience higher rates of late-stage cancer diagnosis and lower survival outcomes than other racial groups [9], [16], [17]. In a study by Haque et al. [16], black men and women reported the highest incidences of cancer deaths at 298.2 per 100,000 for men and 206.5 per 100,000 for women. This systematic review aimed to analyze the existing literature on the relationship between social determinants of health (SDOH), cancer screening behavior, and cancer stage at diagnosis, specifically focusing on Black/African-American individuals. By examining the role of social determinants in cancer screening behavior and stage at diagnosis, this review intends to shed light on the nature of this relationship.

Methods

This systematic review followed the guidelines outlined in the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) statement [18].

Information Sources and Study Selection

A systematic literature search was conducted for articles published up to 1st July 2023. PubMed, Scopus, ScienceDirect, and Google Scholar index databases were used. Topic keywords were used to generate search strings that contained text words and mesh terms. The search strings used are listed in Table 1 in the Appendix. Only the first ten pages of Google Scholar results were assessed. The identified studies were subjected to the study selection process.

Inclusion and exclusion criteria

Articles considered eligible for inclusion had to be original research articles written in English. The race of the population of the study had to be Black. Apart from race, no other population attributes were considered eligible. This study examined the relationship between SDOH, cancer screening behavior, and cancer stage at diagnosis.

The exclusion criteria included non-original research articles, such as systematic reviews, meta-analyses, editorials, article comments, and literature reviews. Studies that did not meet the required study objectives were also excluded.

Review of methodological quality

The included studies were assessed using the Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies, developed by the National Heart, Lung, and Blood Institute [19]. SDOH variables were used as the exposure variables for the assessment, and cancer screening was used as the outcome variable.

Data Extraction

Each article included in the review is summarized in a table, including basic study characteristics. The extracted attributes were the study author, publication year, study design, participants’ demographics, type of cancer under study, and SDOH variables under study.

Results

Search results

The initial search identified 3,897 articles from the databases. One 3,380 articles were obtained from Scopus, 142 from ScienceDirect, 2,275 from PubMed, and 100 from Google Scholar.  Eighty-seven duplicates were excluded from the analysis. During title and abstract screening, 3,734 articles were excluded following the eligibility criteria, and the remaining 76 articles were subjected to full-text review. Fifty-seven articles were excluded because they did not fully satisfy the inclusion criteria. Only 19 articles were included in this systematic review. The reasons for exclusion are shown in the PRISMA flowchart in Figure 1.

Figure 1: PRISMA flow diagram showing the study selection process.

Data extraction results

Check Table 2 in the Appendix.

Quality assessment results

All the cross-sectional studies scored ‘Fair,’ and the cohort study scored ‘Good’ for overall quality. For Item 6: ‘For the analyses in this study, were the exposure(s) of interest measured prior to the outcome(s) being measured? and Item 7: Was the timeframe sufficient so that one could reasonably expect to see an association between exposure and outcome if it existed?’ all cross-sectional studies scored ‘N/A’ since exposure and outcome data in all cross-sectional studies were collected simultaneously.

Characteristics of included studies: a summary

This systematic review included 18 cross-sectional studies and 1 cohort study. The total population was 13,813. Colorectal cancer (CRC) is the most studied type of cancer. This review aimed to examine how SDOH affects cancer screening and stage at diagnosis; however, only studies on cancer screening have been found. This limits the scope of this study. SDOH variables differed broadly across studies. The results of this review are presented thematically. SDOH variables were defined according to groups developed by the U.S. Department of Health and Human Services [2].

Results of included studies

Healthcare access and quality

Buehler et al. [21] found that cancer screening percentages were significantly higher among insured than uninsured. Uninsured people had an adjusted Risk Ratio (aRR) of 1.30 (95%CI, 1.22-1.39) for being unscreened for CRC. Buehler et al. [21] reported that uninsured women are less likely to be screened for cervical cancer. Being uninsured had an aRR of 1.39 (95%CI, 1.22-1.59) for being unscreened. In Palmer et al. [32], being insured led to a likelihood of CRC screening, with an OR of 2.816 (95%CI, 1.53-5.17). Lehto et al. [29], Patel et al. [34], and Patel et al. [35] found that health insurance was a predictor of screening. According to Mitchell et al. [31], health insurance (correlational value = 0.334, p < 0.001) was a positive predictor of CRC screening. Earl et al. [23], Ford et al. [24], Patel et al. [35], Patel et al. [36], and Palmer et al. [33] identified significant determining factors for not undergoing screening. However, Halbert et al. [26] and Kim and Alhassan [28] found that being insured did not significantly affect screening. Griffith [25] reported that insured people were 3.25 (95%CI, 1.14-9.31) times more likely to screen than uninsured.

The roles of doctors and physicians have also been examined in several studies. In Halbert et al. [26], the likelihood of screening increased with good healthcare provider communication (OR = 10.78, 95 %CI = 4.85, 29.94, p = 0.0001). In Katz et al. [27], having good patient-provider communication meant a 1.95 times (95%CI, 1.29-2.94) likelihood of being screened. Brittain and Murphy [20] reported a significant relationship between having a primary care provider and a colonoscopy (CRC screening). Brittain and Murphy [20], Christie et al. [22], Mitchell et al. [31], and Patel et al. [35] reported that having a regular health provider is a determinant for undergoing screening. Patel et al. [35] and Kim and Alhassan [28] showed that annual health checkups were significantly related to prostate cancer screening behaviors. In Patel et al. [35], participants who did not make at least one medical visit in the previous 12 months were 0.27 times as likely to have been screened compared to those who did. Receiving a doctor’s recommendation for screening was also found to increase the likelihood of screening by 3.90 times (95%CI, 1.37-7.93) [32]. Palmer et al. [33] also reported that access to CRC screening services is crucial for participants undergoing screening.

Some studies have examined how participants and patients interact with healthcare facilities. These studies have examined patient trust and knowledge about cancer screening. Palmer et al. [33], Rogers et al. [37], and Sanchez et al. [38] reported medical mistrust as a significant screening barrier. Brittain and Murphy [20], Palmer et al. [33], and Sanchez et al. [38] found that participants had a positive belief in screening procedures, indicating a greater likelihood of undergoing the procedure. Brittain and Murphy [20] found a small, negative, and significant relationship between beliefs about CRC and colonoscopy (r = −.25, p < .01). Palmer et al. [33] reported that participants who believed they were at an increased risk of CRC were adherent.

Christie et al. [22], Earl et al. [23], Patel et al. [36], Palmer et al. [33], and Sanchez et al. [38] found that patients who underwent screening were more knowledgeable about CRC screening, including the purpose of screening, types of CRC screening tests, and appropriate testing intervals. In Katz et al. [27], having adequate CRC Knowledge resulted in an OR of 1.82 (1.14, 2.89) for getting screened. Adequate knowledge was, in turn, associated with a higher level of education, a higher level of income, and health insurance. In Ford et al. [24], a lack of knowledge regarding cancer and confusion between prostate cancer screening and prostate cancer diagnostic tests were barriers to screening. Patel et al. [34] and Patel et al. [35] reported that not knowing where to go for screenings and not having enough information about screenings was an obstacle to screening.

Education access and quality

The highest level of education was a predictor of the intention to get screened for cancer [26], [32], [35]. In Palmer et al. [32], education, i.e., college graduate or postgraduate, was a predictive factor for CRC screening with an odds ratio of 1.31 (95%CI, 0.56-3.07). In Halbert et al. [26], being educated, i.e., above high school level, meant a higher likelihood of screening OR= 1.26 (95%CI, 0.65, 2.45). Patel et al. [35] found that participants with more than a high school education were 2.47 times as likely to have been screened than those with less than a high school education.

Buehler et al. [21] found that colorectal cancer screening percentages were highest among patients living where a greater proportion of residents had a bachelor’s degree.

Economic stability

Kim and Alhassan [28] reported that annual household income is significantly related to prostate cancer screening behaviors. In Earl et al. [23] and other studies, insufficient finances were mentioned as a reason for not undergoing CRC screening. The income amount was different across studies, but it was clear that the more income a person earned, the higher the chance of undergoing screening. In Halbert et al. [26], having an annual income level of >$20,000 meant an OR of 2.09 (95%CI, 1.07- 4.06) for being screened. In Mitchell et al. [31], having an annual household income below $20,000 was negatively associated with CRC screening (correlation value = − 0.186, p < 0.001).  Palmer et al. [32] reported that an annual household income of >$65,000 was a predicting factor for CRC screening. According to Patel et al. [34], having a higher income was a predictor for screening. In Patel et al. [36], participants with annual household incomes of ≥$15,000 were 2.15 times more likely to have been screened for breast cancer in the past two years compared to participants who had annual household incomes <$15,000.

Employment status also plays a role in screening behavior. In some studies, being employed was a predictive factor for screening [26], [30]. Lozano et al. [30] reported that unemployment was a significant factor for not being screened, and Halbert et al. [26] reported that being employed meant a higher likelihood of screening OR= 1.40 (95%CI, 0.68-2.92). In some studies, however, being employed became an obstacle [31], [32], [34]-[36] under some conditions. According to Mitchell et al. [31], employment is a negative predictor of screening. However, the association was positive (encouraging screening) if employment offered paid sick leave. Palmer et al. [32] found that employment type affected the intention to undergo screening, as the study showed that caregivers were less likely to undergo CRC screening. In Patel et al. [36], difficulty getting time off from work was cited as an obstacle to screening by 52% of the participants. In Patel et al. [34] and Patel et al. [35], participants who were not employed were 3.46 times and 2.09 times as likely to have screened compared to those who were employed.

Neighborhood and built environment

Buehler et al. [21] reported that participants living in the most racially segregated neighborhoods were 10% more likely than others to not be screened for colorectal cancer. The adjusted OR for screening was 0.91 (95%CI, 0.84-0.99) for those living in the most racially segregated neighborhoods. Additionally, women living in areas with higher racial segregation were less likely to be screened for cervical cancer [21]. Having a racial segregation index, z score of ≤1.96, an aRR of 0.97 (0.86-1.10) for being screened, and a racial segregation index, z score of >1.96 used as a reference.

Brittain and Murphy [20] reported that women living in census tracts with the highest levels of perceived neighborhood safety had lower cervical cancer screening percentages than those residing in tracts perceived as least safe.

Lozano et al. [30] reported that higher levels of community disadvantage, such as neighborhood insecurity, housing cost burden, household poverty, violent crime, limited food access, and vacant housing, were associated with a lower likelihood of CRC screening, even after controlling for individual-level factors. In Halbert et al. [26], the likelihood of screening increased with greater neighborhood satisfaction (OR = 1.38, 95%CI, 1.01-1.90).

Social and Community contexts

In Brittain and Murphy [20], being religious or having a future orientation toward a particular religion positively correlated with colorectal cancer screening. According to Mitchell et al. [31], being religious, that is, church membership (correlation value = 0.136, p < 0.01), was a positive predictor of CRC screening.

Ford et al. [24] and Rogers et al. [37] found that support from family members or social networks was the most consistent facilitator of screening decisions. Brittain and Murphy [20] reported a positive correlation between family and family influence and screening for colorectal cancer. Family support was a predictive factor for colorectal cancer screening among men in a study by Earl et al. [23]. In the study, 91.6% of respondents found the statement “Get screened for your family” to be persuasive [23]. Lehto et al. [29] found that men who had undergone screening were more likely to be married than those who did not undergo prostate cancer screening. According to Mitchell et al. [31] and Patel et al. [35], being married is a positive predictor of CRC screening. Patel et al. [35] found that participants who were single and had never been married were 0.37 times as likely to have been screened compared to those who were married or living with a partner.

Discussion

Healthcare access and quality impact on cancer screening behavior cannot be overstated. Several studies have consistently shown that health insurance is linked to higher screening rates [21], [23], [24], [29], [31], [32], [34], [35], [36]. These findings emphasize the importance of having health insurance and developing policies to resolve insurance problems in black communities. Furthermore, the role and communication of healthcare providers is crucial [26], [27], [35]. It is worth noting that access to healthcare professionals may vary by location, which could affect the observed relationships.

Educational attainment is a crucial factor that affects cancer screening practices. Based on studies conducted by Palmer et al. [32] and Patel et al. [35], individuals with higher education levels (higher than high school) are more likely to undergo cancer screening. This may support the belief that higher education enhances health literacy and awareness, emphasizing the need for educational access and ensuring that tertiary education is easily attainable in Black communities.

The impact of economic stability on screening has been a recurring topic in many studies. Several studies have reported that higher income is a predictor of increased screening rates [28], [32], [36]. However, the relationship between screening and employment status is complex. While some studies have suggested that being employed leads to better screening rates [26], other studies have highlighted work-related obstacles, such as the lack of paid sick leave [31] and difficulty getting time off from work [36]. The effect of employment status on screening behavior can be seen as individual- and situation-based since it is difficult to generalize it. Nevertheless, it can be suggested that a good working environment coupled with a deserving income level leads to an improved likelihood of screening.

The influence of neighborhood and built environment on cancer screening behavior is significant. Buehler et al. [21] reported that racial segregation presents a substantial obstacle to screening, as individuals residing in such neighborhoods exhibit lower screening rates. This underscores the far-reaching impact of structural inequality. Moreover, perceptions of neighborhood safety and community satisfaction also contributed to screening decisions [20], [26]. Mitigating disparities within neighborhoods and creating safe and supportive environments could potentially increase screening rates, thereby enhancing overall healthcare outcomes.

Various social and community contexts influence cancer screening behaviors, including family support and religious beliefs. Ford et al. [24] and Brittain and Murphy [20] found that family support and social networks can encourage screening decisions, whereas religious affiliation and beliefs are positively correlated with CRC screening [20], [31]. These study findings highlight the socioeconomic and cultural factors that can be leveraged to improve cancer screening practices and, in turn, reduce negative cancer outcomes among Black individuals. This review also highlights the importance of addressing disparities within neighborhoods and creating safe and supportive environments to increase screening rates and improve overall healthcare outcomes.

Study limitations

It is important to note that the findings may be inconsistent due to the heterogeneity arising from different methodologies, sample sizes, and demographics across studies. Additionally, all except one of the studies had a cross-sectional design, which limited the ability to establish causal relationships.

Conclusion

This systematic review aimed to examine how social determinants of health (SDOH) are linked to cancer screening behaviors in Black individuals. The results indicate that having access to healthcare services, receiving good-quality care, and having insurance coverage are all critical factors that affect screening rates. Studies suggest that insured individuals are more likely to undergo screenings [21], [33], [35]. It is essential to consider geographic disparities in access to healthcare services, but healthcare providers' effective communication and advice also exert a significant influence [26], [27], [35]. Increasing access to education and developing health literacy is crucial in promoting cancer screening behaviors, as identified by studies conducted by Palmer et al. [32], Halbert et al. [26], and Patel et al. [35]. Additionally, higher income and employment levels are associated with higher screening rates, as highlighted by Kim and Alhassan [28], Halbert et al. [26], Palmer et al. [32], and Patel et al. [36]. However, the relationship between job status and screening is complex and influenced by factors such as workplace circumstances and leave regulations [31], [36]. Neighborhood and built environments play a significant role in determining cancer screening rates. Factors such as racial segregation, perceived safety, and community satisfaction can affect screening decisions [20], [21], [26]. Creating supportive environments and addressing structural disparities may help increase screening rates among Black individuals. Social and community circumstances also impact screening practices, with family support, religious views, and social networks identified as favorable drivers of cancer screening by Ford et al. [24], Brittain and Murphy [20], and Mitchell et al. [31]. To promote cancer screening and reduce cancer-related inequities in Black communities, the identified social, economic, and cultural factors can be leveraged to improve screening rates. This systematic review emphasizes the crucial role of healthcare access, socioeconomic factors, and community dynamics in shaping cancer screening behaviors among Black individuals. The insights gained from this review can be used to create targeted interventions and policies to improve screening rates and healthcare equity and ultimately reduce the burden of cancer in the Black community.

Declarations

Conflict of Interest declaration:

The authors declare that they have NO affiliations with or involvement in any organization or entity with any financial interest in the subject matter or materials discussed in this manuscript.

Author Contributions:

SM solely contributed to the design and implementation of the research, analysis of the results, and the writing of the manuscript.

References

1. M. Islam, “Social determinants of health and related inequalities: Confusion and implications,” Frontiers in Public Health 7 (2019): 1-4.
2. S. Department of Health and Human Services, “Social determinants of health,” Healthy People 2030, (2023).
3. World Health Organization, “Social determinants of health,” www.who.int, (2023).
4. Frank, T. Abel, S. Campostrini, et al, “The Social Determinants of Health: Time to Re-Think?,” International Journal of Environmental Research and Public Health 17 (2020): p. 5856.
5. Braveman and L. Gottlieb, “The Social Determinants of Health: It’s Time to Consider the Causes of the Causes,” Public Health Reports 129 (2018): 19-31.
6. Centers for Disease Control and Prevention, “Social Determinants of Health,” Centers for Disease Control and Prevention (2022).
7. I. Alcaraz, T. L. Wiedt, E. C. Daniels, et al, “Understanding and addressing social determinants to advance cancer health equity in the United States: A blueprint for practice, research, and policy,” CA: A Cancer Journal for Clinicians 70 (2019): pp. 31-46.
8. E. Gaubatz et al., “Racial and socioeconomic disparities associated with 90-day mortality among patients with head and neck cancer in the United States,” Oral Oncology 89 (2019): 95-101.
9. M. Roche, X. Niu, A. M. Stroup, and K. A. Henry, “Disparities in Female Breast Cancer Stage at Diagnosis in New Jersey: A Spatial-Temporal Analysis,” Journal of Public Health Management and Practice, vol. 23 (2017): 477-486.
10. S. Coughlin, “Social determinants of breast cancer risk, stage, and survival,” Breast Cancer Research and Treatment 177 (2019): 537-548.
11. S. Coughlin, “Social Determinants of Health and Cancer Survivorship,” Journal of environment and Health Sciences 7 (2021): 11-15.
12. S. Kurani et al., “Association of Neighborhood Measures of Social Determinants of Health With Breast, Cervical, and Colorectal Cancer Screening Rates in the US Midwest,” JAMA Network Open 3 (2020): e200618
13. E. Zettler, B. A. Feinberg, Y. Jeune-Smith, et al, “Impact of social determinants of health on cancer care: a survey of community oncologists,” BMJ Open 11 (2021): e049259.
14. A. Zavala et al., “Cancer health disparities in racial/ethnic minorities in the United States,” British Journal of Cancer 124 (2020): 1-18.
15. R. Lawrence et al., “Trends in Cancer Mortality Among Black Individuals in the US From 1999 to 2019,” JAMA Oncology 8 (2022): 1184-1189.
16. T. Haque et al., “Cancer mortality rates by racial and ethnic groups in the United States, 2018-2020,” 115 (2023): 822-830.
17. S. Pinheiro et al., “Cancer mortality among US blacks: Variability between African Americans, Afro-Caribbeans, and Africans,” Cancer Epidemiology 66 (2020): p.101709.
18. J. Page et al., “The PRISMA 2020 statement: An updated guideline for reporting systematic reviews,” British Medical Journal 372 (2021): n71.
19. National Heart, Lung, and Blood Institute (NHLBI), “Study Quality Assessment Tools,” Nih.gov (2009).
20. Brittain and V. P. Murphy, “Sociocultural and Health Correlates Related to Colorectal Cancer Screening Adherence Among Urban African Americans,” Cancer Nursing 38 (2015): 118-124.
21. W. Buehler, J. C. Castro, S. Cohen, Y et al, “Personal and Neighborhood Attributes Associated with Cervical and Colorectal Cancer Screening in an Urban African American Population,” Preventing Chronic Disease 16 (2019): E118.
22. Christie et al., “Sociodemographic correlates of stage of adoption for colorectal cancer screening in African Americans,” Ethnicity & disease 19 (2009): 323-329.
23. Earl et al., “Barriers and Facilitators to Colorectal Cancer Screening in African-American Men,” Digestive Diseases and Sciences 67 (2021): 463-472.
24. E. Ford, S. W. Vernon, S. L. Havstad, et al, “Factors influencing behavioral intention regarding prostate cancer screening among older African-American men.,” Journal of the National Medical Association 98 (2006): 505-514.
25. A. Griffith, “Biological, Psychological and Behavioral, and Social Variables Influencing Colorectal Cancer Screening in African Americans,” Nursing Research 58 (2009): 312-320.
26. C. H. Halbert et al., “Neighborhood Satisfaction and Colorectal Cancer Screening in a Community Sample of African Americans,” Journal of community health 41 (2016): 38-45.
27. L. Katz et al., “Colorectal cancer screening among African American church members: A qualitative and quantitative study of patient-provider communication,” BMC Public Health 4 (2004).
28. Kim and M. Alhassan, “Analyzing Factors Enabling Prostate Cancer Screening Behaviors Among African American Males in the South Region Using the Andersen’s Behavioral Model of Healthcare Services Utilization,” Journal of Prevention 44 (2023): 253-266.
29. H. Lehto, L. Song, K. F. Stein, and P. Coleman-Burns, “Factors Influencing Prostate Cancer Screening in African American Men,” Western Journal of Nursing Research 32 (2010): 779-793.
30. Lozano et al., “The impact of neighborhood disadvantage on colorectal cancer screening among African Americans in Chicago,” 34 (2023): 102235-102235.
31. A. Mitchell, D. C. Watkins, and C. S. Modlin, “Social Determinants Associated with Colorectal Cancer Screening in an Urban Community Sample of African-American Men,” Journal of Men’s Health 10 (2013): 14-21.
32. C. Palmer, D. Chhabra, and S. McKinney, “Colorectal Cancer Screening Adherence in African-American Men and Women 50 Years of Age and Older Living in Maryland,” Journal of Community Health 36 (2010): 517-524.
33. C. Palmer, L. A. Midgette, and I. Dankwa, “Colorectal Cancer Screening and African Americans: Findings from a Qualitative Study,” Cancer Control 15 (2008): 72-79.
34. Patel et al., “Factors Influencing Recommended Cancer Screening in Low-Income African American Women in Tennessee,” Journal of Racial and Ethnic Health Disparities 7 (2019): 129-136.
35. Patel et al., “Factors Influencing Colorectal Cancer Screening in Low-Income African Americans in Tennessee,” Journal of Community Health 37 (2011): 673-679.
36. Patel et al., “Factors Influencing Breast Cancer Screening in Low-Income African Americans in Tennessee,” Journal of Community Health 39 (2014): 943-950.
37. R. Rogers et al., “Psychosocial determinants of colorectal Cancer screening uptake among African-American men: understanding the role of masculine role norms, medical mistrust, and normative support,” Ethnicity & Health, 27 (2020): 1-20.
38. A. Sanchez, D. J. Bowen, A. Hart, and C. Spigner, “Factors Influencing Prostate Cancer Screening Decisions Among African American Men,” Ethnicity & Disease 17 (2007): 374-380.

Appendix

Search in ScienceDirect was limited to title, abstract, and keywords.

Table 1: Search strings

CRC- Colorectal cancer; CCS-cross-sectional study; CS-cohort study

Table 2: Study Descriptor table

Item 1: Was the research question or objective stated clearly? Item 2: Was the study population clearly defined? Item 3: Was the participation rate of eligible individuals at least 50%? Item 4: Were all subjects selected or recruited from the same or similar populations (including the same time period)? Were the inclusion and exclusion criteria in the study pre-specified and applied uniformly to all participants? Item 5: Was sample size justification, power description, or variance and effect estimates provided? Item 6: For the analyses in this study, were the exposure(s) of interest measured before the outcome(s) being measured? Item 7: Was the timeframe sufficient so that one could reasonably expect an association between exposure and outcome if it existed? Item 8: For exposures that can vary in amount or level, did the study examine different levels of exposure as related to the outcome (e.g., categories of exposure or exposure measured as continuous variables)? Item 9: Were the exposure measures (independent variables) clearly defined, valid, reliable, and consistently implemented across all study participants? Item 10: Was the exposure(s) assessed more than once over time? Item 11: Were the outcome measures (dependent variables) clearly defined, valid, reliable, and consistently implemented across all study participants? Item 12: Were the outcome assessors blinded to the participants’ exposure status? Item 13: Was lost to follow-up after baseline 20% or less? Item 14: Were key potential confounding variables measured and statistically adjusted for their impact on the relationship between exposure(s) and outcome(s)?

NA = Not applicable, NR = not reported, CD =cannot determine

Table 3: Quality Appraisal results