1. To present the steps taken and lessons learned from one cancer centre's efforts to capture tumour stage information in a cancer database. 2. To determine the accuracy of the stage data through a chart audit. 3. To describe the potential uses of stage information in a cancer centre.
This is a retrospective review of an initiative to capture tumour stage information at a regional cancer centre in Ontario.
The minutes of the centre's Health Records and Medical Advisory Committees related to staging were reviewed. Data on stage by tumour type was extracted from the centre's Oncology Patient Information System (OPIS). Three hundred and ninety charts were analysed to assess the accuracy of stage information and identify staging errors. Health Information Services workload statistics were reviewed to determine the types and frequency of projects undertaken using stage-related data.
In January 1994, the Ottawa Regional Cancer Centre introduced policies and procedures to capture stage-related information. Standardized staging forms and a physician reminder system encouraged the centre's physicians to record tumour stage within 3 months of new patient registration. Of all qualifying cases in 1994, 92% were staged. A medical audit in 1998 of 390 charts from the 3 previous years of staging data revealed that 71.5% of the charts reviewed had been staged completely. Of the incompletely staged cases, 19% to 57% had TNM recorded, but the stage grouping was not recorded, or the "stage" was the extent of disease at the time of disease progression rather than at initial diagnosis (35% to 71%). Physician-related staging errors occurred in 2% to 5% of cases; data-entry errors occurred in 3% to 6% of cases.
Stage information has enabled the centre to better describe its patient clientele for accreditation purposes and to assist researchers in estimating the number of patients potentially available for prospective and retrospective studies. It is being used to guide targeted educational initiatives to selected populations in the region's catchment area and assists administrators in estimating resource needs. Resistance to the capture of stage information can be overcome with persistence, the development of procedures that facilitate physician compliance, including a reminder system, the development of institutional policies and procedures and by feedback on the uses and availability of stage information.
This study examined the 5-year survival of 2192 breast cancer women diagnosed between 1994 and 1997 in Ottawa, Canada, by age, TNM stage, histology, grade and treatment, including assessment of the independent value of variables in defining prognosis. Our results showed that age, stage, treatment and grade significantly influenced outcome regardless of the confounding factors considered, with histology failing to achieve significant independent prognostic information. The survival rates were highest at ages 50-69 years for stage I and at ages 40-49 years for stages II-IV. The rates were lowest at ages or=70 years for stages III-IV. The differences in survival between grade 1 and grade 3 were 9% in stage I and 20% in stage II. The treatment leading to the best survival was surgery plus radiation for stages I-II and surgery combined with chemotherapy for stages III-IV. Lobular carcinoma had a better prognosis than ductal carcinoma; this can be explained by more grade 1 and less grade 3 cases in lobular carcinoma. The worse prognosis for young patients than other ages can be explained by their higher proportion of poorly differentiated cancers. Stage I patients aged 50-69 years having the best survival is likely due to the earlier diagnosis achieved through screening.
Cites: Cancer. 2000 Jan 15;88(2):369-7410640970
Cites: Stat Med. 1997 Jul 15;16(13):1435-499249917