This repository is a fork of the development repository stulacy/rprev-dev.
The upstream project implements the Monte Carlo prevalence framework of Crouch et al. (2014), where prevalence is estimated from incidence and survival using simulated incident populations and bootstrapped survival models.

The original implementation estimates prevalence for a single index date.
This fork extends the framework to support multiple index dates in one coherent simulation pipeline.

Key features:

• The index argument accepts either a single date or a vector c(t1, ..., tK).
• Within each Monte Carlo / bootstrap replicate, the incident population is generated once and then evaluated at all index dates.
• Methodological deviation from the original single-index sampling (analogue, not contradiction): for each simulated case i, one shared latent threshold is sampled per replicate, U_i ~ Unif(0,1), and alive indicators are set as A_{ik} = 1{U_i <= S_i(t_k)} for each index date t_k. This preserves the original marginal target P(A_{ik}=1)=S_i(t_k) while enforcing cross-time coherence (A_{ik} evolves monotonically with S_i(t_k)), avoiding trajectories that can be non-coherent under independent per-time Bernoulli draws.

The output is a timepoint-wise set of prevalence estimates with uncertainty summaries.

This codebase is part of an Applied Mathematics MSc thesis completed at Óbuda University (Óbudai Egyetem).

Thesis title:
Extending the Simulation-Based Prevalence Estimation of Crouch to Continuous Temporal Modeling

The thesis provides:

• a methodological background on prevalence estimation from incidence and survival,
• a literature-grounded treatment of the Crouch et al. simulation framework,
• a formal and implementation-level extension to the multi-index setting.

The central practical motivation is computational efficiency: one simulated population and one fitted survival structure can be reused across several index dates, rather than rerunning full simulation workflows independently for each date.

Additional thesis materials (exploratory analyses, accuracy checks, runtime evaluations) are available in:

• notebooks/

Install from GitHub with devtools:

# install.packages("devtools")
devtools::install_github("lrado1/rprev-ext", ref = "master")

library(rprev)
library(survival)
data(prevsim)

results <- prevalence(
  index = c("2010-01-01", "2011-01-01", "2012-01-01"),
  num_years_to_estimate = c(5, 10, 20),
  data = prevsim,
  inc_formula = entrydate ~ sex,
  surv_formula = Surv(time, status) ~ age + sex,
  dist = "weibull",
  population_size = 1e6,
  death_column = "eventdate"
)

print(results)
summary(results)