Expected 10-year treatment cost of breast cancer detected within and outside a public screening program in Norway

Abstract

Background

The shift towards earlier stages of disease advancement at diagnosis when introducing mammography screening is expected to affect the treatment costs of breast cancer.

Materials and methods

We collected data on hospital resource use in Norway following a breast cancer diagnosis for the period 1 January, 2008 through 31 December, 2009 for women aged 50–69 years, diagnosed with breast cancer during the period 1 January, 1999 through 31 December, 2009. We estimated treatment costs using a function that included the probability of being at risk for receiving treatment, estimated by means of the Cox proportional hazard model.

Results

In total, 16,045 patients were included for the analyses among which 10.5 % died during the study period. The mean 10-year per-person treatment cost was €31,940 (95 % CI €31,030–32,880), and lower for cancers detected within the public screening program (€30,730) than for those detected elsewhere (€36,230). For ductal carcinoma in situ (DCIS) and cancers in stages I thru IV, treatment costs were €15,740, €23,570, €46,550, €55,230 and €60,430, respectively. Interval cancers occurring within the screening program were generally more resource demanding than both cancers detected at screening or elsewhere.

Conclusions

Ten-year treatment costs increased by increasing stage at diagnosis. Patients whose cancer was detected within the public screening program had lower treatment costs than those detected elsewhere. Interval cancers had higher costs than others.

Appendix: Detailed description of the cost model

Based on work by Etzioni and co-workers [28], expected breast cancer treatment costs was calculated as the mean cost incurred during a 2-month period multiplied with the probability of surviving up to, or longer than, that 2-month period:

$${\text{E}}^{D} = \mathop \sum \limits_{t} \mathop \sum \limits_{i\in D} \left[ {1 - {\text{F}}_{i}^{D} \left( t \right)} \right]{\text{E}}_{i}^{D} (t)$$

where D refers to the groups we will compare (i.e., D = 1,...,3 if the groups are screen detected, interval cancers, and non-attenders, or D = 1,...,12 if the groups are all combinations of detection mode and cancer stage). Moreover, i is the index for each combination of other covariates observed in D, hence the sum goes over all combinations of other covariates observed in the data in group D to get the marginal costs for this group. Also, t denotes the 2-month period after diagnosis (t = 1, 2…60), F ^D _i (t) is the cumulative distribution of T for group D and combination of covariates i, 1 − F ^D _i (t) is the probability of survival to month t for group D and combination of covariates i, and E ^D _i (t) is the mean treatment cost incurred in the 2-month period t among those alive at the start of 2-month period t for group D and combination of covariates i. The estimator is usually known as the Kaplan–Meier sample size estimator, but in our case the survival probability was estimated by means of a Cox model as we have several covariates we wanted to include in the analysis.

Lead time bias, meaning that a cancer is detected earlier and thus treated for a longer time even though actual survival time is not increased, should be present in the data. This is irrelevant for our analysis since the outcome is expected treatment costs. In order to estimate the expected treatment costs, one has to adjust for the observed survival time, not survival adjusted for lead time bias, as it is the former that estimates the probability of being at risk for receiving treatment at a given time point.