testing_incidence: Incindence from testing history
In fellstat/TestingHistoryIncidence: Estimate Incidence from Testing History

Description Usage Arguments Details Value Examples

View source: R/incidence.R

Incindence from testing history

testing_incidence(report_pos, hiv, ever_test, last_test,
  last_upper = last_test, biomarker_art = rep(FALSE, length(report_pos)),
  low_viral = rep(FALSE, length(report_pos)), age = NULL,
  testing_debut_age = 0, weights = rep(1,
  length(report_pos))/length(report_pos), distribution = c("weibull",
  "empirical"), test_pop = c("negative", "undiagnosed"), ptruth = NULL,
  ptreated = NULL, aids_dist = function(x) pweibull(x/12, scale = 1/0.086,
  shape = 2.516), age_breaks = NULL, subset = NULL,
  uniform_missreport = FALSE)

`report_pos`	A logical vector indicating whether each subject reported a positive hiv status
`hiv`	A logical vector indicating hiv status
`ever_test`	A logical vector indicating whether the subject had eer had an hiv test.
`last_test`	A numeric vector indicating the time since last hiv test in months. If testing times are binned into buckets, this is the lower bound of the months since last hiv test.
`last_upper`	A numeric vector indicating the upper bound of the months since last hiv test for each individual.
`biomarker_art`	A logical vector indicating whether ART antibodies are present. NA if test not done.
`low_viral`	A logical vector indicating whether viral load is <=1000
`age`	A numeric vector indicating the age of each subject in years.
`testing_debut_age`	The age at which individuals begin engaging in regular testing.
`weights`	Survey weights
`distribution`	Either "empirical", or "weibull." This controls the family of distribution used to model time since last test. "empirical" may not be used with binned testing times.
`test_pop`	If "negative', the time since last negative is calculated amoung the HIV- population, otherwise it is calculated of those who report being undiagnosed.
`ptruth`	The proportion of the diagnosed hiv positive population that would report being hiv positive. If NULL, this is estimated using biomarker_art and low_viral.
`ptreated`	The proportion of hiv positive individuals with postive biomarker_art or low_viral.
`aids_dist`	The distribution function of time from infection to aids in months.
`age_breaks`	age stratification break points. If NULL, this is estimated from the data.
`subset`	An optional vector specifying a subset of observations on which to perform the analysis. If uniform_missreport is TRUE, ptruth and ptreated are calculated over the full data.
`uniform_missreport`	If true, the rate of missreporting of undiagnosed status is conisuered uniform over the age strata.

When time since last test is grouped into bins, last_upper should always be greater than last_test, and may be infinite (e.g. test was > 24 months ago). Those with missing data or who never had an hiv test should be assigned NA for both their lower and upper bound.

A data.frame of class test_inc with elements: 'incidence': the estimated incidence. 'transmission rate': the estimated transmission rate. 'pundiag': The estimated proportion of positive cases that are undiagnosed after removing treated cases and adjusting for miss-reporting. 'psay_undiag': The proportion of positive cases that report being undiagnosed. 'pmiss_class': the proportion whose diagnosis status is incorrectly reported by the individual and are observed as treated due to viral load or art biomarkers. 'phiv': The proportion with a positive diagnosis. 'ptester': The proportion who have ever been tested. 'mean_time_since_last_test': the mean tie since last test in years. 'tid': mean time between infection and diagnosis in years. 'ptruth' the proportion of positive individuals that correctly report their status. 'ptreated': the proportion of positive indivudals who are identified as treated by viral load or biomarker.

  data(tstdat)
  tstdat$age <- rep(15:64, 200)


  inc <- with(tstdat, testing_incidence(report_pos, hiv,
                                        ever_test, last_test,
                                        biomarker_art=biomarker_art, low_viral=low_viral))
  inc

  # Using last test times divided into bins
  tstdat$last_test_lower <- 24
  tstdat$last_test_upper <- Inf
  tstdat$last_test_lower[tstdat$last_test < 6] <- 0
  tstdat$last_test_upper[tstdat$last_test < 6] <- 6
  tstdat$last_test_lower[tstdat$last_test >= 6 & tstdat$last_test < 12] <- 6
  tstdat$last_test_upper[tstdat$last_test >= 6 & tstdat$last_test < 12] <- 12
  tstdat$last_test_lower[tstdat$last_test >= 12 & tstdat$last_test < 24] <- 12
  tstdat$last_test_upper[tstdat$last_test >= 12 & tstdat$last_test < 24] <- 24
  inc <- with(tstdat, testing_incidence(report_pos, hiv,
                                        ever_test, last_test_lower, last_test_upper,
                                        biomarker_art=biomarker_art, low_viral=low_viral))
  inc

  # HIV testing starts at age 13
  inc <- with(tstdat, testing_incidence(report_pos, hiv,
                                        ever_test, last_test,
                                        biomarker_art=biomarker_art, low_viral=low_viral,
                                        age=age,testing_debut_age=13))
  inc

  # Stratify by age
  inc <- with(tstdat, testing_incidence(report_pos, hiv,
                                        ever_test, last_test,
                                        biomarker_art=biomarker_art, low_viral=low_viral,
                                        age=age,testing_debut_age=13,age_breaks=c(25,35,45,55)))
  inc

  # Pooled estimate probability of miss-reporting diagnosis status across groups
  inc <- with(tstdat, testing_incidence(report_pos, hiv,
                                        ever_test, last_test,
                                        biomarker_art=biomarker_art, low_viral=low_viral,
                                        age=age,testing_debut_age=13,
                                        age_breaks=c(25,35,45,55),
                                        uniform_missreport=TRUE))
  inc