Elsevier

The Lancet

Volume 360, Issue 9340, 12 October 2002, Pages 1131-1135
The Lancet

Articles
Long-term survival rates of cancer patients achieved by the end of the 20th century: a period analysis

https://doi.org/10.1016/S0140-6736(02)11199-8Get rights and content

Summary

Background

Long-term survival rates for many types of cancer have substantially improved in past decades because of advances in early detection and treatment. However, much of this improvement is only seen many years later with traditional cohort-based methods of survival analysis. I aimed to assess achievements in cancer patients' survival by an alternative method of survival analysis, known as period analysis, which provides more up-to-date estimates of long-term survival rates than do conventional methods.

Methods

The 1973–98 database of the Surveillance, Epidemiology, and End Results (SEER) programme of the US National Cancer Institute was analysed by period analysis.

Findings

Estimates of 5-year, 10-year, 15-year, and 20-year relative survival rates for all types of cancer were 63%, 57%, 53%, and 51%, respectively, by period analysis. These estimates were 1%, 7%, 11%, and 11% higher, respectively, than corresponding estimates by cohort-based survival analysis. By period analysis, 20-year relative survival rates were close to 90% for thyroid and testis cancer, exceeded 80% for melanomas and prostate cancer, were about 80% for endometrial cancer, and almost 70% for bladder cancer and Hodgkin's disease. A 20-year relative survival rate of 65% was estimated for breast cancer, of 60% for cervical cancer, and of about 50% for colorectal, ovarian, and renal cancer.

Interpretation

Timely detection of improvements in long-term survival rates might help to prevent clinicians and their patients from undue discouragement or depression by outdated and often overly pessimistic survival expectations. It also adds to the value of cancer surveillance as a basis for appropriate public-health decisions.

Introduction

One of the most widely used global measures of achievement in health is life expectancy, which is derived from population life tables. These tables can be regarded as a special type of survival analysis, because they quantify survival of newborn babies. Two main approaches are used to derive such tables. The first quantifies survival of a cohort of newborn babies. Although the conception and interpretation of this approach is very straightforward, its use is limited by the fact that it takes a human lifespan before the life table can be completed and life expectancy can be calculated. The second approach uses period life tables, which are based on mortality of a population within a recent period, such as a recent year. With this approach, survival data for various ages are obtained from cohorts of people born in different years. Although the interpretation of period life tables is slightly less straightforward than that for cohort life tables, they are widely used in health statistics because they provide much more up-to-date estimates of life expectancy than do cohort life tables.

Up to now, long-term cancer survival statistics have almost always been calculated from cohorts of patients diagnosed many years ago,1, 2 and they could therefore be quite out of date when they were derived. The idea to use the period principle to obtain estimates of long-term cancer survival rates was proposed a few years ago.3, 4 Meanwhile, extensive empirical evaluation studies have shown that period analysis provides much more up-to-date estimates of long-term survival of cancer patients than does cohort-based survival analysis.5, 6, 7 Furthermore, period estimates for a particular time quite accurately predict long-term survival rates of patients diagnosed in that period.6, 7

Up to now, period estimates of long-term survival rates of cancer patients have been reported by only a few European cancer registries, which cover populations that are limited in size8, 9, 10 or age range.11 In this study, I applied period analysis to obtain up-to-date estimates of long-term survival rates in cancer patients in the USA.

Section snippets

Methods

All data were obtained from the 1973–98 public-use database of the Surveillance, Epidemiology, and End Results (SEER) programme of the US National Cancer Institute.12 Although the SEER programme is not a true nationwide population-based cancer registry, it is the most comprehensive source of information on cancer incidence and survival in the USA, and it is judged the standard for quality in cancer registries around the world. Quality control has been an integral part of SEER since its

Results

The most recent 5-year, 10-year, 15-year, and 20-year period estimates of relative survival rates for all invasive cancers combined (all races, both sexes) were about 1%, 7%, 11%, and 11%, respectively, higher than those calculated by cohort-based analysis (table 2). Cohort estimates of long-term survival rates were much lower for men than for women, whereas period estimates for both sexes were closely similar.

One of the reasons why survival rates have in the past been so much lower for men

Discussion

These results show that long-term survival expectations of patients with many types of cancer are substantially better than those suggested by conventional cohort-based estimates, which refer to cohorts of patients diagnosed many years ago. Although survival rates and their changes over time vary strongly by cancer site, period estimates of 10-year, 15-year, and 20-year relative survival are about 7%, 11%, and 11% higher, respectively, than traditional cohort estimates for all cancers.

References (30)

  • SJ Cutler et al.

    Maximum utilization of the life table method in analysing survival

    J Chron Dis

    (1958)
  • H Brenner et al.

    Deriving more up-to-date estimates of long-term patient survival

    J Clin Epidemiol

    (1997)
  • H Brenner et al.

    A computer program for period analysis of survival

    Eur J Cancer

    (2002)
  • EL Kaplan et al.

    Nonparametric estimation from incomplete observations

    J Am Stat Assoc

    (1958)
  • H Brenner et al.

    An alternative approach to monitoring cancer patient survival

    Cancer

    (1996)
  • H Brenner et al.

    More up-to-date monitoring of long-term survival rates by cancer registries: an empirical example

    Meth Inf Med

    (2001)
  • H Brenner et al.

    Up-to-date long-term survival estimates of patients with cancer by period analysis

    J Clin Oncol

    (2002)
  • H Brenner et al.

    Use of period analysis for providing more up-to-date estimates of long-term survival rates: empirical evaluation among 370 000 cancer patients in Finland

    Int J Epidemiol

    (2002)
  • H Brenner et al.

    Recent improvement in survival of breast cancer patients in Saarland, Germany

    Br J Cancer

    (1998)
  • H Brenner et al.

    Trends in survival of patients with ovarian cancer in Saarland, Germany, 1976-1995

    J Cancer Res Clin Oncol

    (1999)
  • H Brenner et al.

    Long-term cancer survival achieved by the end of the 20th century: most up-to-date estimates from the nationwide Finnish Cancer Registry

    Br J Cancer

    (2001)
  • H Brenner et al.

    Long-term survival of children with leukaemia achieved by the end of the 2nd millennium

    Cancer

    (2001)
  • Surveillance, Epidemiology, and End Results (SEER). Program Public-Use Data (1973-1998)

    (April 2001)
  • F Ederer et al.

    The relative survival rate: a statistical methodology

    Natl Cancer Inst Monogr

    (1961)
  • T Hakulinen

    Cancer survival corrected for heterogeneity in patient withdrawal

    Biometrics

    (1982)
  • Cited by (0)

    View full text