pith. sign in

arxiv: 2604.23154 · v1 · submitted 2026-04-25 · 📊 stat.ME

A bivariate cure copula model with zero-inflated gamma frailty: dependence in both cure fractions and survival times

Masaki Hino , Shogo Kato , Takeshi Emura This is my paper

Pith reviewed 2026-05-08 07:38 UTC · model grok-4.3

classification 📊 stat.ME
keywords bivariate survivalcure fractionfrailty modelcopulazero-inflated gammapaired datadependence modelingsurvival analysis
0
0 comments X

The pith

A new bivariate model uses zero-inflated gamma frailty and a copula to capture dependence in both cure fractions and survival times for paired data.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a statistical model for paired survival data in which some subjects may never experience the event of interest. It introduces a zero-inflated gamma frailty that simultaneously produces a cure fraction and continuous unobserved heterogeneity among the uncured, an odds-ratio parameter that links the two cure indicators, and a copula that links the two survival times conditional on frailty. The resulting framework contains several existing bivariate cure models as special cases obtained by restricting parameters. For common choices of marginal distributions and copula, the joint survival function takes a closed form that permits direct maximum likelihood estimation and likelihood-ratio tests. The approach is intended for biomedical settings where paired event times, such as those observed in two organs of the same individual, exhibit both possible cures and complex dependence.

Core claim

The authors establish a bivariate cure frailty-copula model that incorporates a zero-inflated gamma frailty. This frailty simultaneously generates a point mass at cure for each margin and a continuous random effect for the uncured subjects. Dependence between the binary cure indicators is expressed through an odds-ratio parameter, while dependence between the conditional survival times is expressed through a copula. The model reduces to several prior bivariate cure specifications when the frailty variance or the odds ratio is set to boundary values. Population-level rank correlations, including tie-corrected Kendall's tau and Spearman's rho, are derived in closed form. When suitable marginal

What carries the argument

The zero-inflated gamma frailty that jointly produces a cure probability and continuous heterogeneity, combined with an odds-ratio link for cure dependence and a copula link for survival-time dependence.

If this is right

  • Existing bivariate cure models arise as special cases when the frailty variance or odds-ratio parameter is fixed at boundary values.
  • Explicit formulas for Kendall's tau and Spearman's rho give direct measures of overall dependence without simulation.
  • Maximum likelihood estimation is feasible whenever the joint survival function is closed form.
  • Likelihood ratio tests can assess whether the cure-fraction dependence or the survival-time dependence is statistically significant.
  • The model supplies a single likelihood for paired biomedical data that may contain both cured and uncured subjects.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The closed-form expressions could support routine inclusion of the model in software for large paired datasets where numerical integration would be slow.
  • One could examine whether the odds-ratio dependence and the copula dependence differ systematically across disease types or study designs.
  • Adding subject-level covariates that affect both cure probability and survival would allow joint prediction of cure status and time to event for new pairs.
  • The framework might be adapted to settings with more than two paired observations by extending the frailty and copula structure.

Load-bearing premise

The zero-inflated gamma distribution must correctly represent both the cure fraction and the unobserved heterogeneity among uncured subjects for the joint model to be correctly specified.

What would settle it

If data simulated from a non-zero-inflated frailty or from a different dependence structure produce biased estimates of the odds ratio or copula parameter, or if the closed-form likelihood fails to recover the true parameters, the model's ability to separate and identify both types of dependence would be falsified.

read the original abstract

In biomedical studies, paired survival data arise naturally when two event times are observed within the same subject. Existing statistical models seldom accommodate both cure fractions and complex dependence structures. In this paper, we propose a novel bivariate cure frailty-copula model for paired survival data with a cure fraction. By incorporating a zero-inflated gamma frailty, the proposed framework simultaneously accommodates a cure fraction and continuous unobserved heterogeneity among uncured subjects. Dependence between cure statuses is modeled naturally via an odds-ratio parameter, while dependence between survival times conditional on frailty is captured through a copula. We show that the proposed model includes existing bivariate cure models as special cases. Population-level rank correlation coefficients are derived for the proposed model, namely tie-corrected versions of Kendall's tau and Spearman's rho. For suitable choices of marginal distributions and copula, the joint survival function admits a closed-form expression, enabling maximum likelihood estimation and likelihood ratio testing. Simulation studies and a real data application demonstrate the practical utility of the proposed approach. An R package, curecopula, implementing the proposed methods is publicly available on GitHub.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The manuscript proposes a bivariate cure frailty-copula model for paired survival data with cure fractions. It uses a zero-inflated gamma frailty to simultaneously capture a point mass at cure and continuous heterogeneity among uncured subjects, an odds-ratio parameter for dependence between cure statuses, and a copula for dependence between conditional survival times. The model is shown to nest existing bivariate cure models as special cases via boundary values on the dependence parameters. Population-level tie-corrected Kendall's tau and Spearman's rho are derived, closed-form joint survival functions are obtained for suitable marginal distributions and copulas (enabling MLE and LRT), and the approach is illustrated via simulations, a real-data application, and a publicly available R package curecopula.

Significance. If the central derivations hold, the work provides a flexible and practically useful extension for bivariate survival analysis in biomedical settings where both cure fractions and complex dependence structures must be accommodated. The explicit nesting of prior models, derivation of rank correlations, and closed-form likelihoods for selected specifications are strengths that facilitate implementation and comparison. The open-source R package further supports reproducibility and adoption.

minor comments (3)
  1. [Abstract and Section 3] The abstract states that the joint survival function admits a closed-form expression 'for suitable choices of marginal distributions and copula,' but the main text should explicitly list the combinations (e.g., specific copula families and marginals) that yield closed forms, including the resulting likelihood expressions, to allow readers to verify the claim without additional derivation.
  2. [Section 4] In the derivation of the population-level rank correlations (Kendall's tau and Spearman's rho), the handling of ties induced by the cure fraction (i.e., the point mass at infinity) should be stated more explicitly, including any adjustments to the standard formulas and whether the expressions remain closed-form under the zero-inflated gamma frailty.
  3. [Section 5] The simulation study section should report the exact data-generation mechanism (including the specific parameter values used for the zero-inflated gamma, odds-ratio, and copula) and the criteria for excluding or censoring observations, to enable exact replication of the reported bias and coverage results.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of our manuscript and the recommendation for minor revision. The referee's summary accurately captures the key elements of the proposed bivariate cure frailty-copula model, including the zero-inflated gamma frailty, odds-ratio dependence for cure statuses, copula for conditional survival times, nesting of existing models, derivation of population-level rank correlations, closed-form likelihoods, and the accompanying R package.

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The proposed bivariate cure frailty-copula model is constructed from standard components (zero-inflated gamma frailty, odds-ratio for cure dependence, copula for conditional survival times) with the joint survival function and rank correlations (Kendall's tau, Spearman's rho) derived mathematically from the model definition. Special cases of existing models arise by setting dependence parameters to boundary values, which is a direct consequence of the parameterization rather than a fitted prediction. Maximum likelihood estimation is performed on observed data, and closed-form expressions hold only for chosen marginals/copulas without reducing any claimed result to its inputs by construction. No load-bearing self-citations or ansatzes are invoked that would make the central claims tautological; the framework remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

3 free parameters · 2 axioms · 1 invented entities

The central claim rests on the introduction of the zero-inflated gamma frailty to jointly handle cure probability and heterogeneity, plus standard frailty-model assumptions and the existence of closed-form expressions for chosen marginals and copulas. Several parameters must be estimated from data.

free parameters (3)
  • odds-ratio parameter
    Controls dependence between the two cure indicators; fitted from data.
  • copula dependence parameter
    Controls dependence between survival times conditional on frailty; fitted from data.
  • frailty distribution parameters
    Shape and rate parameters of the gamma component plus the zero-inflation probability; fitted from data.
axioms (2)
  • domain assumption Conditional independence of the two survival times given the shared frailty
    Standard assumption in shared-frailty models for paired survival data.
  • ad hoc to paper Zero-inflated gamma distribution for the frailty term
    Introduced in the paper to simultaneously produce a point mass at zero (cure) and continuous heterogeneity among the uncured.
invented entities (1)
  • zero-inflated gamma frailty no independent evidence
    purpose: To induce both a cure fraction (via the zero mass) and continuous unobserved heterogeneity among uncured subjects while allowing dependence modeling
    This is an extension of the classical gamma frailty; the paper provides no external empirical validation beyond the model fit itself.

pith-pipeline@v0.9.0 · 5500 in / 1586 out tokens · 71214 ms · 2026-05-08T07:38:41.091354+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

2 extracted references · 2 canonical work pages

  1. [1]

    sn-basic.bst

    FUNCTION identify.basic.version "sn-basic.bst" " [2024/07/19 v1.1 bibliography style]" * top ENTRY address archive author booktitle chapter doi edition editor eid eprint howpublished institution journal key keywords month note number organization pages publisher school series title type url volume year archivePrefix primaryClass adsurl adsnote version lab...

    work page 2024

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION add.period duplicate empty 'skip "." * add.blank if FUNCTION if.digit duplicate "0" = swap duplicate "1" = swap duplicate "2" = swap duplicate "3" = swap duplicate "4" = swap duplicate "5" = swap duplicate "6" = swap duplicate "7" = swap duplicate "8" = swap "9" = or or or or or or or or or FUNCTION ...

    work page