A modified Prevalence Incidence Analysis Model method may improve disease prevalence prediction

Ilya Novikov, Liraz Olmer, Lital Keinan-Boker, Barbara Silverman, Eliezer Robinson, Laurence S. Freedman

Research output: Contribution to journalArticlepeer-review

Abstract

Objectives: The Prevalence Incidence Analysis Model method is used for predicting disease prevalence, using past data on incidence and relative survival. Our objective was to propose and evaluate a modified approach for choosing the Prevalence Incidence Analysis Model. Study Design and Setting: Instead of the standard approach using the likelihood ratio statistic, we find the model that predicts most successfully the prevalence in the last available Y years using data up to but not including those Y years and then use that model to predict future prevalence another Y years ahead using all the data. We also make an “alignment” adjustment using the last known prevalence level. We evaluate the relative performance of the modified and standard methods using data on cancer from Israel in 1983–2013. Results: In this example, the modified approach gave as good or better predictions than the standard. Using the modified approach, we forecast cancer prevalence in Israel for 2014–2024 to increase at a gradually accelerating rate from the current 10,000 per year to 12,000 per year by 2020, reaching a total of 380,000 by 2024. Conclusion: The modified approach may offer improved forecasting, but further methodological work on forecasting cancer prevalence is needed.

Original languageEnglish
Pages (from-to)18-26
Number of pages9
JournalJournal of Clinical Epidemiology
Volume123
DOIs
StatePublished - Jul 2020

Keywords

  • Cancer prevalence
  • Cancer survivors
  • Epidemiology
  • Israel
  • Statistical forecasting
  • Statistical projection

Fingerprint

Dive into the research topics of 'A modified Prevalence Incidence Analysis Model method may improve disease prevalence prediction'. Together they form a unique fingerprint.

Cite this