Early mortality attributable to PICC-lines in 4 public hospitals of Marseille from 2010 to 2016 (Revised V3).

INTRODUCTION: Peripherally inserted central catheters (PICC-line) are devices inserted through peripheral venous access. In our institution, this technology has been rapidly adopted by physicians in their routine practice. Bacteremia on catheters remains an important public health issue in France. However, the mortality attributable to bacteremia on PICC-line remains poorly evaluated in France and in the literature in general. We report in our study an exhaustive inventory of bacteremia on PICC-line and their 30 days mortality, over a 7 years period. MATERIAL AND METHODS: From January 2010 to December 2016, we retrospectively matched PICC-line registers of the radiology department, blood culture records of the microbiology laboratory and medical records from the Hospital Information Systems. RESULTS: The 11,334 hospital stays during which a PICC-line was inserted were included over a period of 7 years. Among them, 258 episodes of PICC-line-associated bacteremia were recorded, resulting in a prevalence of 2.27%. Hematology units: 20/324 (6.17%), oncology units: 55/1375 (4%) and hepato-gastro-enterology units: 42/1142 (3.66%) had the highest prevalence of PICC-line related bacteremia. The correlation analysis, when adjusted by exposure and year, shows that the unit profile explains 72% of the variability in the rate of bacteremia with a P = .023. Early bacteremia, occurring within 21 days of insertion, represented 75% of cases. The crude death ratio at 30 days, among patients PICC-line associated bacteremia was 57/11 334 (0.50%). The overall 30-day mortality of patients with PICC-line with and without bacteremia was 1369/11334 (12.07%). On day 30, mortality of patients with bacteremia associated PICC-line was 57/258 or 22.09% of cases, compared to a mortality rate of 1311/11076, or 11.83% in the control group (P < .05, RR 2.066 [1.54-2.75]). Kaplan-Meier survival analysis revealed a statistically significant excess mortality between patients with PICC-line associated bacteremia and PICC-line carriers without bacteremia (P < .0007, hazard ratio 1.89 [1307-2709]). CONCLUSION: Patients with PICC-line associated bacteremia have a significant excess mortality. The implementation of a PICC-line should remain the last resort after a careful assessment of the benefit/risk ratio by a senior doctor.

The costs of breast cancer prior to and following diagnosis.

This retrospective incidence-based cost-of-illness analysis aims to quantify the costs associated with female breast cancer in Flanders for the year prior to diagnosis and for each of the 5 years following diagnosis. A bottom-up analysis from the societal perspective included direct health care costs and indirect costs of productivity loss due to morbidity and premature mortality. A case-control study design compared total costs of breast cancer patients with costs of an equivalent standardised population with a view to calculating the additional costs that can be attributed to breast cancer. Total average costs of breast cancer amounted to 107,456 per patient over 6 years. Total costs consisted of productivity loss costs (89% of costs) and health care costs (11% of costs). Health care costs did not vary with age at diagnosis. Health care costs of breast cancer patients converged with those of the general population at 5 years following diagnosis. Patients with advanced breast cancer stadia had higher health care costs. Cost estimates provided by this analysis can be used to determine priorities for, and inform, future research on breast cancer. In particular, attention needs to be focussed on decreasing productivity loss from breast cancer.

Patterns of health care use and expenditure during the last 6 months of life in Belgium: differences between age categories in cancer and non-cancer patients.

OBJECTIVES: We analysed end-of-life care in Belgium to examine potential age variation in place of death, transitions between care settings, health care utilisation and public expenditure in the last 6 months of life. METHODS: Administrative data of one sickness fund were used, and analysed through regression analysis and analysis of variance. The study population comprised 40,794 individuals (age>40 years). RESULTS: Several indicators showed that the end-of-life of older people, especially of those aged >or=90 years, differs from that of younger individuals. Older persons more likely died in a care home, were less transferred between care settings, and stayed less days in hospital. On the other hand, older persons used more home care services, and had more contacts with the general practitioner. Differences between age categories were equally observed for last week's end-of-life care. Opposite to the trend for cancer patients, the odds of having a palliative home care allowance increased with age for non-cancer patients. Public expenditure for the oldest old was lower as compared to the younger decedents, but dependent on place of death. CONCLUSION: Several aspects of end-of-life care in Belgium appear to be influenced by age. In view of ageing of the population, these findings can be of interest to decision makers.