Background: The availability of information on profile and trends of cancer in South
African populations is important for the development of appropriate cancer control
strategies, as well as monitoring the efficacy of the existing cancer control programmes.
Yet, generally there is a scarcity of systematically analysed reports on hospital cancer
cases in South Africa, even for urban hospitals. The aim of this study was to describe the
cancer profile of patients diagnosed at Frere Hospital?s Oncology and Radiation
Department and estimate the incidence of cancer among Buffalo City (BFC) urban area
residents, for the 19-year period 01 January 1991 to 31 December 2009 based on the
clinical administrative data system maintained by the department.
Methodology: The study was a descriptive case series study based on a retrospective
review of Frere Hospital?s Oncology and Radiation Department patient records from 1991
to 2009. Permission was obtained to retrieve records of cancer cases for the 19-year
period from the database. Data were extracted from the customized administrative system
to an excel spread sheet. Variables for each case retrieved included: socio-demographic
details; age at diagnosis, sex, race, place of residence and medical aid information, tumor
information; site and date of diagnosis. Data cleaning incorporated techniques such as
checking of completeness and accuracy of patient information details. Dates were
formatted into month-day-year sequence and checked so that the date of birth precedes the
date of diagnosis of the patient and the date last seen. Age less than zero and greater than
ninety nine was replaced as missing. Geographical areas were coded according to the
South African Population Census. Duplicates and cases with missing diagnosis were
excluded.
STATA 12.0 analysis software was used to determine the proportion of cancer cases by
age, sex and site. Frequency distribution tables and graphs were drawn and minimal
cancer incidence rates were estimated based on population estimates from the 2007
Community Survey and the 2011 Population Census. The rates were regarded as
“minimal” and “working” estimates as the database reviewed was from only one of the
two public hospitals in the study area and private hospitals? databases were not used. The
other public hospital, Cecilia Makiwane, does not have an oncology department and
private hospitals in the study area with radiology departments have issues of patient ix
information confidentiality. Age standardized rates were calculated using the World
Standard Population generally applied to cancer incidence data. Poisson regression
analysis model was used to assess trends in selected cancer rates over time adjusted for
age and sex. Results: 19 737 (89.0%) of the 22 173 records retrieved were malignant cases. A total of
7 656 (38.8%) were males and 12 081 (61.2%) were females. The average age at
diagnosis for males was 56 years and for females 54 years. In males the top 5 cancers
were lung (18.6%), larynx (8.3%), mouth (7.7%), prostate (7.1%), tongue (5.2%) whereas
in females were cervix (36.6%), breast (22.0%), lung (3.5%), ovary (2.8%) and corpus
uteri (2.7%). A total of 360 childhood (=14 years) cancers was observed which accounted
for 1.8% of all cancers. The leading childhood cancers were brain (20.8%), kidney
(19.8%), eye (16.1%), Hodgkin?s disease (7.8%) and lymphoid leukaemia (7.3%) in boys.
Brain (21.4%), kidney (18.5%), eye (14.9%), bones, joints & articular cartilage (12.5%)
and lymphoid leukaemia (10.1%) were leading cancers in girls. Estimates of cancer
incidence rates for the BFC urban area sub-population showed that the overall age
standardised rates (ASRs) for males were 83.2 per 100 000 population and for females
83.3 per 100 000 population. Leading cancers in males were lung [22.5%, ASR 21.0 per
100 000], prostate [14.7%, ASR 9.2 per 100 000], larynx [5.8%, ASR 5.0 per 100 000],
mouth [4.4%, ASR 3.7 per 100 000] and colon [3.1%, ASR 2.9 per 100 000]. In females
cervical [20.9%, ASR 23.0 per 100 000], breast [23.6%, ASR 20.2 per 100 000], lung
[3.4%, ASR 4.7 per 100 000], ovary [2.1%, ASR 3.0 per 100 000] and corpus uteri [3.4%,
ASR 2.8 per 100 000] were the leading cancers.
Poisson regression analysis results showed an overall decreasing trend over the period in
selected common cancers: lung, prostate and cervical with an exception of breast cancer
trends which were observed as stable over the period. Compared to males, females were
0.37 [95% CI: 0.05-0.40] less likely to have lung cancer. Analysis of race distribution of
the cancers showed that lung, prostate and breast cancer incidence rate ratios were higher
in the White population compared to the Black African population. White males were
twice more likely to have lung cancer and almost five times more likely to have prostate
cancer compared to Black African males. White females were almost three times more
likely to have breast cancer but 0.57 [95% CI: 0.12-0.66] less likely to have cervical
cancer when compared to Black African females. x

Conclusion: This study demonstrates that data retrieved at Frere Hospital?s Oncology and
Radiation Department were useful for describing the cancer characteristics of patients
presented at this department, as well as in estimating the cancer incidence in an urban
population of the Eastern Cape (EC) Province. However, the observed cancer pattern in
BFC urban area was strikingly different from the profile which was previously reported
for this area when oesophageal cancer was the leading cancer and lung cancer did not
feature. Therefore, further studies to investigate oesophageal cancer cases that might not
have been referred for radiotherapy in this hospital should be done.