Francisci S, Tursini F, Dal Maso L, Gigli A, Guzzinati S.

Front Oncol. 2023 Aug 30;13:1201464. doi: 10.3389/fonc.2023.1201464. eCollection 2023.


Background: In most developed countries, the number of cancer survivors is expected to increase in the coming decades because of rising incidence and survival rates and an aging population. These patients are heterogeneous in terms of health service demands: from recently diagnosed patients requiring first-course therapy to patients with extensive care needs and severe disabilities to long-term survivors who only need minimal care. Therefore, in terms of providing healthcare planners and policymakers with useful indicators for addressing policies according to health service demands, it is worth supplying updated measures of prevalence for groups of patients based on the level of care they require. The aim of this paper is to illustrate a new method for estimating short-term projections of cancer prevalence by phase of care that applies to areas covered by cancer registration.

Methods: The proposed method combines linear regression models to project limited duration prevalence derived from cancer registry data and a session of the freely available software COMPREV to estimate the projected complete prevalence into three distinct clinically relevant phases of care: initial, continuing, and final. The method is illustrated and validated using data from the Veneto region in Italy for breast, colorectal, and lung cancers.

Results: Prevalence is expected to increase in 2015-2026 for all considered cancer sites and sexes, with average annual variations spanning from 2.6% for women with lung cancer to 0.5% for men with colorectal cancer. The only exception is lung cancer prevalence in men, which shows an average annual decrease of 1.9%. The majority of patients are in the continuing phase of care, followed by the initial and final phases, except for lung cancer, where the final phase of care prevails over the initial one.

Discussion: The paper proposes a method for estimating (short-term) future cancer healthcare needs that is based on user-friendly and freely available software and linear regression models. Validation results confirm the applicability of our method to the most frequent cancer types, provided that cancer registry data with at least 15 years of registration are available. Evidence from this method is addressed to policymakers for planning future cancer care, thus improving the cancer survivorship experience for patients and caregivers.