Relation between socioeconomic status and tumour stage in patients with breast, colorectal, ovarian, and lung cancer: results from four national, population based studiesBMJ 2001; 322 doi: http://dx.doi.org/10.1136/bmj.322.7290.830 (Published 07 April 2001) Cite this as: BMJ 2001;322:830
- David H Brewster, director of cancer registration in Scotland ()a,
- Catherine S Thomson, senior statisticiana,
- David J Hole, deputy directorb,
- Roger J Black, heada,
- Paul L Stroner, coordinator of the Scottish Cancer Therapy Networka,
- Charles R Gillis, directorb the Scottish Cancer Therapy Network.
- a Scottish Cancer Intelligence Unit, Information and Statistics Division, Trinity Park House, Edinburgh EH5 3SQ
- b West of Scotland Cancer Surveillance Unit, Department of Public Health, University of Glasgow, Glasgow G12 8RZ
- Correspondence to: D H Brewster
- Accepted 13 December 2000
Although there is evidence that patients with cancer from deprived communities in Scotland might present with more advanced disease,1–3 this finding has not been replicated in every study.4 Using data from four population based audits, we investigated whether there is any relation between socioeconomic status and tumour stage at presentation in patients with breast, colorectal, ovarian, and lung cancer.
Participants, methods, and results
For each audit potential study populations were identified from the Scottish Cancer Registry, and data were abstracted from medical records. The years of diagnosis for patients with breast, colorectal, ovarian, and lung cancer were 1993, 1993, 1992-4, and 1995, respectively.
The staging details examined were: pathological size, pathological nodal status, and metastatic status for breast cancers; Dukes' stage for colorectal cancers; Fédération Internationale de Gynécologie et d'Obstétrique stage for ovarian cancers; and a simple extent of disease classification based on clinical findings or investigations for lung cancers (as only 11% of cases had surgical resection), or both.
Medical records were available for 2518 patients with breast cancer, 2778 with colorectal cancer, 1387 with ovarian cancer, and 3855 with lung cancer, representing more than 90% of potentially eligible cases for each of the four cancer sites. A higher proportion of records was unobtainable for deprived than for affluent patients with lung cancer (11.7% v 7.0%, P<0.001) (see table on website). No significant differences were found in availability of medical records by deprivation grouping for the other cancers.
The table shows the distributions of variables for tumour staging by cancer site and deprivation grouping. P values for associations between the staging variables and deprivation were similar when unknown stages were excluded. We found no evidence that patients from deprived communities were likely to present with more advanced disease for breast or colorectal cancer. For ovarian cancer there was a possibility that deprived patients may have more advanced disease (see table); however, deprived patients with lung cancer were more likely to present with localised disease (see table). Multivariate analyses, performed using log linear modelling, showed no evidence of age dependent relations between stage of disease and deprivation. The four age groups used for these analyses differed by cancer site because they were predefined by different specialist groups—for example, the breast cancer specialists included a category for screening age group (50-64 years).
We found no consistent evidence that patients from deprived communities present with more advanced disease for breast, colorectal, ovarian, or lung cancer. Despite the introduction of breast screening and differential uptake by socioeconomic status,5 our results for breast cancer remain similar to those of Carnon et al.4 However, another recent but smaller study from the west of Scotland found that women from deprived areas were more likely to present with locally advanced or metastatic disease.3 Our result for colorectal cancer conflicts with those of a study from Tayside, although the latter was based on cases for which there was a record of disease, and so not population based, and excluded patients with distant metastases.2 The result for ovarian cancer was of borderline significance, although patients with stage unknown had a generally worse prognosis than those with stage IV disease, presumably reflecting inoperability. The greater likelihood of deprived patients with lung cancer presenting with localised disease could be an artefact resulting from differential availability of medical records across the socioeconomic groupings or, owing to comorbidity, less intensive investigation of these patients leading to less accurate data for staging. Alternatively, it might reflect a lower threshold for investigation and referral in patients from deprived communities presenting with suspicious symptoms, because they are more likely to be current smokers. Further research is needed to investigate the contribution of tumour, host, and treatment related factors to outcome.
We thank Professor Colin McArdle and his surgical colleagues in the west of Scotland who allowed their data from a previous audit to be used in the Scottish Cancer Therapy Network's national audit of colorectal cancer and the many other clinicians in Scotland who contribute to the network's audits.
Contributors: DHB initiated the analysis, but all the authors were involved in specifying the analysis and interpretation of the results. CST and DJH carried out the analyses. DHB wrote the first draft of the manuscript, which was modified to incorporate comments from all the other authors; he will act as guarantor.
Funding The Scottish Cancer Therapy Network is funded by grants from the Clinical Resource and Audit Group and the Chief Scientist Office, both of the Scottish Executive Health Department. The views expressed, however, are those of the authors.
Competing interests None declared.
- A table showing the availability of medical records by socioeconomic status appears on the BMJ's website