The relationship between educational attainment and health o
The relationship between educational attainment and health outcomes is complex . The consistently strong association with low educational attainment and registration of CUP in our study, independent of overall health and smoking history, suggests low health literacy may play a role in the late diagnosis of cancer. Low educational attainment and thus low literacy is a barrier to health information seeking, comprehension and self-management , it can negatively influence the way doctors communicate with patients , and it is associated with poorer health outcomes . Our data do not appear to support an independent association with person-level economic and social relationship factors, although we had no data on the strength of social connectedness.
We utilised a large, contemporary prospective cohort study with comprehensive data on demographic, social and lifestyle risk factors, and incident cancers and deaths ascertained by high-quality population-based registries. We used two sets of controls to generate a complete risk profile in relation to individuals diagnosed with metastatic cancer of known primary site and to unselected cohort participants. Further, given the consistent prior evidence of an increased risk of a CUP diagnosis in association with comorbid disease [5,20,21], we minimised confounding by adjusting for relevant comorbid conditions and self-reported overall health at baseline . Nevertheless, we cannot exclude residual confounding for individuals whose health status changed between the cohort baseline and their cancer diagnosis, or for individuals whose self-reported overall health poorly correlated with performance status. We also cannot discount residual confounding due to unmeasured demographic, social, occupational, environmental and lifestyle factors, for example, waist circumference, which may be a more accurate measure of body fatness than BMI .
Whilst the 45 and Up cohort study was designed to be representative of the general population, the participation rate was 18% , and comparisons with a representative Doxorubicin health survey  and all cancer patients in NSW  indicate that cohort participants are on average healthier than the general population. Even so, risk estimates calculated from within-cohort comparisons are expected to be valid. Our statistical power was constrained by a relatively small number of incident CUP cases. We had no information on the location or extent of metastatic disease, and nor were we able to conclusively differentiate CUP subgroups (for example: confirmed and inadequately evaluated) using the cancer registry data alone. We also had no direct measure of health literacy for the cohort participants.
Conclusions It is exceptionally challenging to balance the early diagnosis of cancer against over-investigation. Increasing the difficulty for clinicians and patients, the median age at CUP diagnosis is 70–80 years, and CUP is described as exhibiting rapid and aggressive dissemination . If replicated in other cohorts, our findings may help in the identification of high-risk patient subgroups that may benefit from assistance in understanding their health priorities and health care needs, in particular encouraging and empowering effective communication and self-management [18,25]. The strong relationship with educational attainment is further evidence of the many benefits of investment in education, and completion of compulsory education, over the human lifecycle.
Authorship contribution statement
Conflict of interest statement
Acknowledgments This work was supported by a Cancer Institute of New South Wales (NSW) Epidemiology Linkage Program Grant (10/EPI/2-06). The funder had no role in the study design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review or approval of the manuscript; or the decision to submit the manuscript for publication. We thank the 45 and Up Study, the Department of Human Services (DHS), the NSW Ministry of Health, and the NSW Cancer Registry for providing the data used for this study.