Abstract
Background. Socioeconomic differences in survival after head and neck squamous cell carcinoma (HNSCC) are among the greatest for any malignancy. To improve our understanding of the mechanisms by which socioeconomic position influences HNSCC survival, we investigated the association between socioeconomic position and advanced stage HNSCC at diagnosis.
Material and methods. Men and women with HNSCC diagnosed between 1992 and 2008 were identified in the Danish Head and Neck Cancer Group (DAHANCA) database, which contains detailed information on all cases of HNSCC treated in Denmark. Individual information on the following four socioeconomic indicators: highest attained educational level, cohabitation status, disposable income and degree of urbanisation were obtained from Statistics Denmark. For the 9683 cases on which there was full information, we estimated odds ratios (ORs) for a diagnosis of advanced stage (TNM III-IV) HNSCC in multivariate logistic regression models by site (glottic, non-glottic larynx, oropharynx, hypopharynx and oral cancer), with adjustment for age, gender, period of diagnosis, education, income, cohabitation status, degree of urbanisation and comorbidity in accordance with a causal diagram.
Results. For all HNSCC sites, the ORs for advanced stage at diagnosis were increased for patients with low income and for men living alone. For glottic and oral cancers, the ORs for advanced stage HNSCC increased systematically by decreasing length of education. Increased ORs were found for hypopharynx cancer patients living in rural areas or provincial cities. Having one or more comorbid conditions was associated with an increased OR for advanced stage oral cancer but with a decreased OR for oropharynx cancer.
Conclusion. In this nationwide population-based study, socioeconomic differences in stage at diagnosis were found for all HNSCC subsites. Focus on the high risk for advanced stage HNSCC among vulnerable patients may be beneficial during referral and diagnosis in order to improve HNSCC outcomes.
Cancer care is organised into prevention, diagnostics, treatment and follow-up. In countries with universal access to healthcare services one might expect that cancer patients should achieve the same outcome, no matter of their socioeconomic position. Nevertheless, both short- and long-term cancer survival are influenced by socioeconomic position [Citation1,Citation2], and the gap in survival between affluent and deprived cancer patients has increased in the past decades [Citation3]. In many countries with universal access to healthcare services, socioeconomic inequality in survival is particularly pronounced for head and neck squamous cell carcinoma (HNSCC) [Citation1–4]. In a Danish population-based study, the five-year relative survival was up to 21 percentage points lower among patients in the lowest income quartile, than those in the highest [Citation2]. In order to improve HNSCC outcomes, we must improve our understanding of the mechanisms by which socioeconomic position influences HNSCC survival in order to develop targeted interventions that can address such differences in survival.
Stage at the time of HNSCC diagnosis is the most important clinical prognostic factor [Citation5]. Advanced stage HNSCC results not only in poor survival but is also associated with more severe late effects of the disease and treatment, such as dysphagia [Citation6], with implications for rehabilitation and survivorship. Socioeconomic position refers to the social and economic factors that influence what position an individual holds within the structure of a society [Citation7]. As socioeconomic position is a proxy for various underlying factors, including differences in health behaviour, symptom perception, access to health care (e.g. dentists), communication with healthcare professionals and compliance to health advice [Citation7,Citation8], people in a high socioeconomic position may experience shorter delays in diagnosis and thus pre-sent at an earlier stage, than less advantaged people.
The results of the few previous studies of the association between socioeconomic position and HNSCC stage at diagnosis have addressed mainly racial disparities, involved area level measures of socioeconomic position only [Citation9], had small samples (N = 160–245) [Citation10,Citation11], or measured the effect at combined, heterogeneous HNSCC subsites with different aetiology and initial symptoms [Citation10]. Further, because of shared risk factors (i.e. tobacco and alcohol consumption), HNSCC patients have a high prevalence of comorbid disease [Citation5], which may influence symptom perception, health care-seeking behaviour and health surveillance and detection of HNSCC in general practice. As the prevalence of comorbid disease is generally higher among people in a lower socioeconomic position [Citation8], comorbidity may mediate the association between socioeconomic position and HNSCC stage at diagnosis, if such an association exists.
In Denmark, health services in general practice and hospitals are tax-funded for all residents, and HNSCC treatment is centralised and standardised. By linking clinical data and socioeconomic information from nationwide population-based databases and registries, we investigated the association between several individual level socioeconomic indicators and the risk for advanced stage HNSCC at each subsite, adjusting for relevant confounders and taking into account comorbidity.
Material and methods
Linkage of registry data
Since 1968, all Danish residents have been assigned a 10-digit unique personal identification number and registered in the civil registration system with continuously updated information on date of birth, gender, current and past addresses, spouses, migration and vital status [Citation12]. The personal identification number allows linkage with national administrative registries and clinical databases.
Study population
Since 1992, all treated Danish HNSCC patients have been registered in the national clinical database of the Danish Head and Neck Cancer Group (DAHANCA), which contains detailed information on site, TNM classification, treatment and follow-up [Citation13]. In this database, we identified 10 564 men and women diagnosed with glottic, non-glottic larynx, oropharynx, hypopharynx, or oral squamous cell carcinomas (ICD-10: C01-06, C09-10, C12-14, C32) diagnosed between 1992 and 2008, who were born after 1920, aged ≥ 30 at the date of diagnosis and had resided in Denmark one year before diagnosis. Information on the education of people born before 1920 is limited in the files of Statistics Denmark, and people below the age of 30 may still being educated and thus establishing their socioeconomic position. We excluded 405 persons for whom information on site or stage at diagnosis was missing and 476 persons for whom information on highest attained educational level or disposable income was missing. Thus, 9683 (92%) HNSCC patients were eligible for our analyses ().
Figure 1. Inclusion and exclusion of patients diagnosed with head and neck squamous cell carcinoma (HNSCC), 1992–2008, Denmark.
Socioeconomic position
Individual socioeconomic position was determined by data linkage to various population-based administrative registers operated by Statistics Denmark that contain annually updated data on a broad range of socioeconomic factors [Citation14,Citation15]. We obtained information on highest attained educational level by 1 October of the year before diagnosis and categorized it as short (mandatory school, 7 and 9 years for patients born before and after 1958, respectively), medium (youth education and vocational education, 10–12 years of education) and higher education (> 12 years) [Citation15]. Patients’ disposable income in the year before diagnosis was grouped into quintiles on the basis of that of the entire population with the same gender and in the same birth cohort, calculated per household after taxation and interest per person, adjusted for number of people in the household and deflated according to the 2000 value of the Danish crown, by applying the formula: ‘deflated household income/(no. of person in household × 0.6)’ from the Danish Ministry of Finance. We categorised disposable income as lowest (1st quintile), medium (2nd–3rd quintiles) and highest (4th–5th quintiles) to ensure a sufficient number of cases in each interval [Citation14]. Information on cohabitation status was obtained by 1 January of the year of diagnosis and was defined as living alone or with a partner (married or living with a partner of the opposite sex, over the age of 16, with a maximum age difference of 15 years, at the same address with no other adult in residence) [Citation12]. Finally, we included information on degree of urbanisation, based on the municipality of residence on 1 January of the year of diagnosis [Citation12], grouped into rural area (municipalities with < 10 000 inhabitants), provincial city (municipalities with ≥ 10 000 inhabitants) and greater Copenhagen area.
Stage at diagnosis
From the DAHANCA database, we obtained information on the date of diagnosis, site and clinical stage at diagnosis. Clinical stage was defined according to the UICC TNM-classification, with stages I-II defined as early stage and III-IV as advanced stage.
Comorbidity
Comorbidity was assessed according to a revised Charlson Comorbidity Index specific for HNSCC (HN-CCI) [Citation16]. HN-CCI is based on the six conditions (congestive heart failure, cerebrovascular disease, chronic pulmonary disease, ulcer disease, liver disease and diabetes) that were the only ones in the original Charlson Comorbidity Index that affected survival after HNSCC [Citation16]. Information on all somatic diagnoses of the six HN-CCI conditions, that led to hospitalisation (from 1995, also outpatient visits) between 10 years and 30 days before HNSCC diagnosis, was obtained for each individual from the Danish National Patient Register [Citation17], coded according to the Danish modified versions of ICD8 and ICD10 (from 1994). Information on diabetes was obtained from the National Diabetes Register [Citation18], as diabetes is usually treated and registered in general practice.
Statistical analysis
Odds ratios (ORs) for a diagnosis of advanced stage HNSCC were computed in multivariate logistic regression models for each site. To account for possible clustering within hospital wards, generalised estimating equations with the exchangeable working correlation structure and robust variance estimates were used. Hypothesised confounders and mediators were identified in a causal diagram (). A three-step statistical model was used to ensure the transparency of results and to observe changes. In model 1, the estimates were adjusted for age, gender and period of diagnosis. Model 2 included further adjustment for relevant socioeconomic factors according to the causal diagram (). In model 3, all estimates were further adjusted for comorbidity (HN-CCI), in order to observe a potential mediating role of comorbidity. We tested interactions between age, gender and period of diagnosis and the socioeconomic variables as well as comorbidity. A significant interaction between gender and cohabitation status was found for glottic (p = 0.013) and oral cancers (p = 0.007); however, addition of this interaction term to the model did not change the estimates for the remaining covariates. Thus, we performed separate analyses for cohabitation status for women and men. All other estimates are based on a full model without the interaction term to facilitate interpretation of results. All tests were two-sided. All analyses were performed in SAS 9.3, with the proc genmod procedure.
Figure 2. Hypothesised causal relations between socioeconomic position and head and neck squamous cell carcinoma (HNSCC) stage at diagnosis, 1992–2008, Denmark.
Results
Advanced stage HNSCC was diagnosed in 5575 (58%) patients, the proportion varying from 20% to 84% for the different sites ().
Table I. Descriptive characteristics of 9683 patients aged ≥ 30 years and born ≥ 1920 with head and neck squamous cell carcinoma (HNSCC) diagnosed in Denmark between 1992 and 2008, by site.
The ORs for a diagnosis of advanced stage HNSCC at all sites were increased for men and, except for hypopharynx cancer, for patients with a diagnosis in the most recent calendar periods (). The ORs for glottic and oral cancers increased systematically with decreasing length of education, and that for oral cancer was increased by 80% for patients with short education, compared to those with higher education (OR 1.80, 95% CI 1.58–2.05). The ORs were consistently higher among men living alone, particularly for glottic cancer (OR 1.87; 95% CI 1.76–1.98). Among women, no associations were found between cohabitation status and stage at diagnosis. The ORs for all HNSCC sites were increased for patients with low income, by 328% for glottic cancer, as compared with those in the highest income quintiles (OR 3.28; 95% CI 2.41–4.47). Except for increased ORs for hypopharynx cancer among patients living in rural areas and provincial cities (2.10; 95% CI 1.14–3.89 and 2.26; 95% CI 1.91–2.67, respectively), as compared with those living in the capital area, no associations were found with degree of urbanisation. The OR was increased by 23% for oral cancer patients with one or more HN-CCI comorbid diseases (1.23; 95% CI 1.05–1.44) but decreased by 22% among oropharynx cancer patients (0.78; 95% CI 0.71–0.86). The estimates for cohabitation status and disposable income were more increased for sites where a low proportion of patients were diagnosed at an advanced stage ( and ).
Table II. Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) for diagnosis of advanced stage head and neck squamous cell carcinoma among 9683 patients aged ≥ 30 years and born ≥ 1920, Denmark, 1992–2008, by site.
Only minor changes in risk estimates were observed with stepwise inclusion of covariates (Supplementary Table I to be found online at http://informahealthcare.com/doi/abs/10.3109/0284186X.2014.998279), and the estimates were virtually unchanged when comorbidity was included in the model (Supplementary Table I, model 3 to be found online at http://informahealthcare.com/doi/abs/10.3109/0284186X.2014.998279). As many values for site in 2008 were missing [40% of all (N = 269) missing values for site], a sub-analysis was performed with exclusion of data for 2008. The results did virtually not change the estimates (data not shown).
Discussion
In this large, nationwide, population-based study, we found that patients who attend the clinic for HNSCC at an advanced stage are more likely to have short education, low income and be men living alone. This finding indicates that more effort is required to support these patients through treatment and follow-up care in order to improve prognosis and minimise the likelihood and severity of late effects.
Educational level is associated with health behaviour, cognitive functioning and health literacy [Citation7]. Patients with short education may have less know-ledge about symptoms or more often ignore early symptoms, than patients with higher education, delaying healthcare seeking [Citation7]. A Canadian study found no strong association between education and the stage of HNSCC at diagnosis: patients with a short education had a higher risk for advanced stage than patients with a long education, but not significantly so (OR 1.70, 95% CI 0.49–5.94) [Citation10]. The study was, however, small (N = 160), and, in contrast to our model, the authors adjusted for area-based income level [Citation10].
Income was more consistently and more strongly associated with the stage of HNSCC at diagnosis than educational level. Although income is strongly correlated to education, it can also be considered a measure of available material resources and access to health services, such as dental care [Citation7,Citation19]. Patients referred by their dentist may have tumours at an earlier stage [Citation9]. Further, income better reflects peoples’ current socioeconomic position: while educational attainment is usually stable and fixed for people over 30 years of age, income is more likely to change, especially for people who have retired [Citation7,Citation8]. Accordingly, income has a stronger reversed causal relation with health, as long-term illnesses may lead to a decrease in income [Citation7]. The association between disposable income and advanced stage HNSCC was systematically stronger for sites where more patients were diagnosed with earlier stages (e.g. glottic and oral cancer). This may indicate that socioeconomic differences in diagnostic delays (i.e. time from onset of symptoms to diagnosis) are more influential in these HNSCC sites, which might be characterised by more specific initial symptoms. In contrast to our study, two previous studies found no increased risk for advanced stage head and neck cancer (OR 1.08; 95% CI 0.94–1.25) [Citation10] or cancer of the tongue or floor of the mouth (RR 1.01; 95% CI 0.78–1.31 and 0.96; 95% CI 0.77–1.19, respectively) among patients living in low-income areas [Citation9]. In both studies, however, income was measured only at area level and the samples were relatively small (N = 160 [Citation10] and N = 1221 and 812 [Citation9]), which may have introduced misclassification and limited their power [Citation9,Citation10].
In support of our results, two studies showed decreased risks for advanced stage oral cancer (N = 245) [Citation11] and cancer of the tongue (N = 1221) (but not floor of mouth (N = 812) [Citation9] among cancer patients married or living with a partner. These associations may reflect differences in social networks and social support [Citation7]. Patients living alone may be less prone to discuss their symptoms and encouraged to seek health care than patients living with a partner, and people living alone use dental services less often [Citation19]. We included information on cohabitation (not only marital status), to minimize misclassification of the social impact of non-marital relationships, which accounted for 23% of all couples living together in Denmark in 2001 [Citation20].
In addition to the three individual level socioeconomic indicators we also investigated area-based socioeconomic position in terms of degree of urbanisation. Except for hypopharynx cancer, we found no association between degree of urbanisation and stage of HNSCC at diagnosis. Area-based socioeconomic measures are conceptually different from individual level measures. Although they contribute to and are associated with individual socioeconomic characteristics, they reflect community, neighbourhood and social structural influences to a greater extent [Citation7]. Degree of urbanisation is based on the number of citizens in each municipality; geographical differences may still be present in larger municipalities.
Unexpectedly, adjustment for comorbidity left the estimates virtually unchanged, suggesting that comorbidity does not play a major role in the socioeconomic differences in the stage distribution of HNSCC. As HNSCC patients share risk factors, they may be a relatively homogenous group in respect of the prevalence of the comorbid conditions included in the HN-CCI index. Further, the association between comorbid diseases and HNSCC stage at diagnosis could go both ways: having a comorbid disease might lead to a quicker diagnosis, due to greater surveillance, or it might delay diagnosis, because of shared or hidden symptoms.
The risk for a diagnosis of advanced stage HNSCC increased by period of diagnosis. This might be explained partly by stage migration, i.e. advances in diagnostic technology allow more sensitive detection of tumour spread, which, in and of itself, increase the number of patients with higher stage tumours [Citation21].
An important advantage of this large, nationwide, population-based study over the few previous studies is the inclusion of various individual level socioeconomic indicators. Further, all the clinical and socioeconomic information was collected prospectively and consistently for administrative purposes, independently of this study, thus minimising selection bias and misclassification of disease-related and socioeconomic information. With a large study population and detailed clinical data, we were able to investigate differences in risk for advanced stage HNSCC at each site. HNSCC at some sites are, however, rare, and we might not have had sufficient power to detect significant associations for some covariates. Finally, we had no information on behavioural factors such as smoking and alcohol consumption, symptom perception, the regularity of dental visits, tumour-specific factors (e.g. human papillomavirus status), and diagnostic delays. Further information on these factors would improve our understanding on how socioeconomic position is associated with advanced stage HNSCC.
In conclusion, in this nationwide population-based study, we found significant socioeconomic differences in the stage of HNSCC at diagnosis. Socioeconomic inequality in stage at diagnosis was present for all HNSCC sites, even though they comprise a wide spectrum of distinct neoplasms with different aetiology, initial symptoms and different proportions diagnosed at an advanced stage. These findings are in line with findings of socioeconomic inequality in stage at diagnosis in a number of other cancer types [Citation22–26]. The implications of the observed socioeconomic inequality in stage at diagnosis may reach beyond prognosis to differences in rehabilitation needs [Citation27,Citation28]. Greater attention to persons with short education, low income or who live alone when they enter the healthcare system with symptoms of potential HNSCC in order to facilitate earlier diagnosis. Efforts might include securing timely referrals and navigation of vulnerable patients through treatment, follow-up and rehabilitation in order to improve HNSCC outcomes for all patients.
Supplementary material available online
Supplementary Tables IA–E to be found online at http://informahealthcare.com/doi/abs/10.3109/0284186X.2014.998279.
ionc_a_998279_sm6642.pdf
Download PDF (42.1 KB)Acknowledgements
This work was supported by the Danish Cancer Society's Scientific Committee (KBVU).
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
- Ito Y, Nakaya T, Nakayama T, Miyashiro I, Ioka A, Tsukuma H, et al. Socioeconomic inequalities in cancer survival: A population-based study of adult patients diagnosed in Osaka, Japan, during the period 1993–2004. Acta Oncol 2014;53:1423–33.
- Dalton SO, Schüz J, Engholm G, Johansen C, Kjaer SK, Steding-Jessen M, et al. Social inequality in incidence of and survival from cancer in a population-based study in Denmark, 1994–2003: Summary of findings. Eur J Cancer 2008;44:2074–85.
- Rachet B, Quinn MJ, Cooper N, Coleman MP. Survival from cancer of the larynx in England and Wales up to 2001. Br J Cancer 2008;99(Suppl 1):S35–7.
- McDonald JT, Johnson-Obaseki S, Hwang E, Connell C, Corsten M. The relationship between survival and socio-economic status for head and neck cancer in Canada. J Otolaryngol Head Neck Surg 2014;43:2.
- Bøje CR, Dalton SO, Grønborg TK, Primdahl H, Kristensen C a, Andersen E, et al. The impact of comorbidity on outcome in 12 623 Danish head and neck cancer patients: A population based study from the DAHANCA database. Acta Oncol 2013;52:285–93.
- Mortensen HR, Overgaard J, Jensen K, Specht L, Overgaard M, Johansen J, et al. Factors associated with acute and late dysphagia in the DAHANCA 6 & 7 randomized trial with accelerated radiotherapy for head and neck cancer. Acta Oncol 2013;52:1535–42.
- Galobardes B, Shaw M, Lawlor DA, Lynch JW, Davey Smith G. Indicators of socioeconomic position (part 1). J Epidemiol Community Health 2006;60:7–12.
- Diderichsen F, Andersen I, Manuel C, Andersen A-MN, Bach E, Baadsgaard M, et al. Health inequality – determinants and policies. Scand J Public Health 2012;40: 12–105.
- Groome PA, Rohland SL, Hall SF, Irish J, Mackillop WJ, O’Sullivan B. A population-based study of factors associated with early versus late stage oral cavity cancer diagnoses. Oral Oncol 2011;47:642–7.
- Johnson S, Corsten MJ, McDonald JT, Chun J. Socio- economic factors and stage at presentation of head and neck cancer patients in Ottawa, Canada: A logistic regression analysis. Oral Oncol 2010;46:366–8.
- Scott SE, Grunfeld EA, McGurk M. The idiosyncratic relationship between diagnostic delay and stage of oral squamous cell carcinoma. Oral Oncol 2005;41:396–403.
- Pedersen CB. The Danish Civil Registration System. Scand J Public Health 2011;39:22–5.
- DAHANCA. Danish Head And Neck Cancer Group 2014. [cited 2014 Oct 22]. http://www.dahanca.dk.
- Baadsgaard M, Quitzau J. Danish registers on personal income and transfer payments. Scand J Public Health 2011;39:103–5.
- Jensen VM, Rasmussen AW. Danish education registers. Scand J Public Health 2011;39:91–4.
- Bøje CR, Dalton SO, Primdahl H, Kristensen CA, Andersen E, Johansen J, et al. Evaluation of comorbidity in 9388 head and neck cancer patients: A national cohort study from the DAHANCA database. Radiother Oncol 2014;110:91–7.
- Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Health 2011;39:30–3.
- Carstensen B, Kristensen JK, Marcussen MM, Borch-Johnsen K. The National Diabetes Register. Scand J Public Health 2011;39:58–61.
- Christensen LB, Petersen PE, Steding-Jessen M. Consumption of dental services among adults in Denmark 1994–2003. Eur J Oral Sci 2007;115:174–9.
- Statistics Denmark. Households 2014. [cited 2014 Oct 22]. http://dst.dk/en/Statistik/emner/husstande-familier-boern/husstande.aspx
- Lyhne NM, Christensen A, Alanin MC, Bruun MT, Jung TH, Bruhn MA, et al. Waiting times for diagnosis and treatment of head and neck cancer in Denmark in 2010 compared to 1992 and 2002. Eur J Cancer 2013;49:1627–33.
- Dalton SO, Frederiksen BL, Jacobsen E, Steding-Jessen M, Østerlind K, Schüz J, et al. Socioeconomic position, stage of lung cancer and time between referral and diagnosis in Denmark, 2001–2008. Br J Cancer 2011;105:1042–8.
- Frederiksen BL, Brown P de N, Dalton SO, Steding-Jessen M, Osler M. Socioeconomic inequalities in prognostic markers of non-Hodgkin lymphoma: Analysis of a national clinical database. Eur J Cancer 2011;47:910–7.
- Frederiksen BL, Osler M, Harling H, Jørgensen T. Social inequalities in stage at diagnosis of rectal but not in colonic cancer: A nationwide study. Br J Cancer 2008;98:668–73.
- Dalton SO, Düring M, Ross L, Carlsen K, Mortensen PB, Lynch J, et al. The relation between socioeconomic and demographic factors and tumour stage in women diagnosed with breast cancer in Denmark, 1983–1999. Br J Cancer 2006;95:653–9.
- Ibfelt E, Kjær SK, Johansen C, Høgdall C, Steding-Jessen M, Frederiksen K, et al. Socioeconomic position and stage of cervical cancer in Danish women diagnosed 2005 to 2009. Cancer Epidemiol Biomarkers Prev 2012;21:835–42.
- Dalton SO, Johansen C. New paradigms in planning cancer rehabilitation and survivorship. Acta Oncol 2013;52:191–4.
- Kjær T, Bøje CR, Olsen MH, Overgaard J, Johansen J, Ibfelt E, et al. Affiliation to the work market after curative treatment of head and neck cancer: A population-based study from the DAHANCA database. Acta Oncol 2013;52:430–9.