Characteristics and pattern of mortality in cancer patients at a tertiary care oncology center: report of 259 cases.

BACKGROUND

Little is known about mortality statistics of hospitalised cancer patients from developing countries. This paper describes the distribution of causes of death in various malignancies, status of malignancy at the time of death, type and intent of therapy received by the cancer patient prior to death and nature of infections in terminal cancer patients who died in hospital. We also aimed to study discrepancies in mortality reporting in terms of death certificate at our center and tried to analyse possible causes.

RESULTS

Data for 259 consecutive deaths in hospitalized cancer patients in a calendar year were analysed. Of all these, 147 (57%) were cases of solid tumors, 107 (41%) were cases of hematological malignancies and 5 (2%) were other or undiagnosed cases. Median duration of hospital stay prior to death was 7 (1-106) days. Sepsis/multi organ dysfunction syndrome (MODS) was commonest immediate cause of death 118/259 (45.2%) followed by progressive malignancy in 64/259 (24.7%) cases. Only 13/267 (5%) patients died with controlled cancer. Some 184 (71.3%) deaths occurred within 90 days of any form of anticancer treatment of which more than three fourths (77.2%) occurred after chemotherapy. Among these chemotherapy related deaths, 63 were febrile neutropenic deaths, with the commonest site of infection in the lungs, and positive blood culture was found in 18 (28%) cases. There were discrepancies in information derived from death certificates and from case records in 84 (32%) cases. Most of these were due to the use of ambiguous terms like cardio-respiratory arrest as a cause of death in the death certificate.

CONCLUSION

It is important to audit mortality data on a regular basis as this can provide valuable insight into hospital practice and may help to identify preventable causes of mortality. Mortality record keeping is another important aspect as variable practices in this area may have implications for cancer mortality reporting and this may ultimately lead to erroneous cancer epidemiology.