R/create_param_values.R
In n8thangreen/LTBIdiagTST: Latent TB infection cost-effectiveness analysis
Defines functions
create_param_values
Documented in
create_param_values
#' Create parameter values
#'
#' All parameter values required for the decision tree model.
#'
#' Excel model with structure and labelling:
#' G:/DIDE-PC_2019/MSc-MPH/projects/2017/LTBI-TST_Manabu/decision tree Excel/all_decision_trees.xls
#' @export
#'
#' @examples
#' create_param_values()
#'
#' ## save lists as csv
#' # write.csv(cbind(label_costs), file = "data/cost_inputs.csv")
#' # write.csv(cbind(label_health), file = "data/health_inputs.csv")
#' # write.csv(cbind(label_probs), file = "data/probs_inputs.csv")
#'
create_param_values <- function(save = TRUE) {
# assign values to labels ------------------------------------------
label_costs <-
list(
# "Contact tracing per contact" = 368.9,
# "Mean number of contacts examined per primary case" = 6.5,
# "Total Contact tracing" = 2397.85,
#
# "Cost of inpatient episode for acute TB" = 3325.15,
# "Proportion of patients with acute TB who are admitted" = 0.53,
# "Total Inpatient care" = 1762.3295,
#
# "Cost of culture test" = 8.06,
# "Culture tests per case" = 4,
# "CXR per case" = 2,
# "LFT per case" = 4,
# "Total Cost of tests" = 104.74,
#
# "HRZE tablet (60)" = 39.5,
# "_tablets per month" = 150,
# "HRZE per month" = 98.75,
# "HR tablet 300/150 (56)" = 25.22,
# "tablets per month" = 60,
# "HR per month" = 27.02,
# "Duration of initial therapy" = 2,
# "Duration of continious therapy" = 4,
# "Total Cost of chemotherapy" = 305.58,
#
# "Cost of outpatient consultation (first visit)" = 208,
# "Cost of outpatient consultation (follow-up visit)" = 94,
# "Number of outpatient clinical visits per case" = 4,
# "Visits from TB nurse per case" = 6,
# "Total Out patient care" = 755.86,
"TST" = 25.02,
"QFT" = 29.11,
"TSPOT" = 43.17,
"Ns_cost" = 53, # TB nurse visit
"Out patient consultation (first visit)" = 224, # c_out
"Follow-up via nurses" = 166, # c_fuout
"CXR" = 46,
"LFT" = 3.95,
"Number of CXR" = 1, # Pareek (2013) 2
"Number of LFT" = 1, # Pareek (2013) 4
"Sp_cost" = 274,
"HR tablet (300/150) (56)" = 25.22, # BNF
"drug cost per month" = 27.02,
"drug_cost" = 33.22, # per month, inflate to 2023
"dur_hr" = 3, # duration of chemo HR
"Number of outpatient consultation" = 1,
"n_appts" = 2, # Number of TB nurse appointments
"Hep" = 984, # Pooran (2010) inflated to 2023
"TB_cost" = 6055) # NICE guideline NG33 (2016) inflated to 2023
label_costs$LTBIcompl_cost <- with(label_costs, Ns_cost*n_appts + drug_cost*dur_hr)
label_costs$LTBIincompl_cost <- with(label_costs, Ns_cost*floor(n_appts/2) + drug_cost*dur_hr/2)
label_costs_distns <-
tribble(
~name.cost, ~vals,
"TST", list(distn = "pert", params = c(mode = 25.02, min = 12.51, max = 50.05)), # Pooran 2010
"QFT", list(distn = "pert", params = c(mode = 29.11, min = 15.16, max = 55.67)), # Imperial NHS
"TSPOT", list(distn = "pert", params = c(mode = 43.17, min = 22.21, max = 83.87)), # Imperial NHS
"CXR", list(distn = "pert", params = c(mode = 46, min = 40, max = 52)), # Reference cost 2016
"LFT", list(distn = "pert", params = c(mode = 3.95, min = 2.63, max = 5.27)), # Reference cost 2016
"Hep", list(distn = "pert", params = c(mode = 984, min = 492, max = 1968)), # Pooran 2010 inflated to 2023
"TB_cost", list(distn = "pert", params = c(mode = 6055, min = 3028, max = 12110)), # NICE guideline NG33 (2016) inflated to 2023
"Sp_cost", list(distn = "pert", params = c(mode = 274, min = 274, max = 274)), # positive screening cost
"Ns_cost", list(distn = "pert", params = c(mode = 53, min = 43, max = 63)), # TB nurse visit
)
##TODO: problem with this because the sampling should be dependent on previous sampling so its really a function
## how an we incorporate this in to the sampling function? this isnt actually needed in the sampling step
## so could define as function and run afterwards e.g. label_costs_distns_fns
label_costs_distns <- bind_rows(
label_costs_distns,
tribble(~name.cost, ~vals,
"LTBIincompl_cost", with(label_costs,
list(distn = "unif",
params = c(min = Ns_cost*floor(n_appts/2) + drug_cost*dur_hr/2,
max = Ns_cost*floor(n_appts/2) + drug_cost*dur_hr/2))),
"LTBIcompl_cost", with(label_costs,
list(distn = "unif",
params = c(min = Ns_cost*n_appts + drug_cost*dur_hr,
max = Ns_cost*n_appts + drug_cost*dur_hr)))))
label_probs_distns <-
tribble(
~name.prob, ~prob,
"TST_sens", list(distn = "pert", params = c(mode = 0.79, min = 0.69, max = 0.89)), # Kahwati (2016)
"TST_spec", list(distn = "pert", params = c(mode = 0.59, min = 0.46, max = 0.73)), # Pai (2008)
# "TST_sens", list(distn = "norm", params = c(mean = 0.69, sd = 0.89)), # posterior
# "TST_spec", list(distn = "norm", params = c(mean = 0.46, sd = 0.73)), # posterior
"QFT_sens", list(distn = "pert", params = c(mode = 0.886, min = 0.812, max = 0.944)), # Zhang et al. BMC Infectious Diseases (2023) 23:40
"QFT_spec", list(distn = "pert", params = c(mode = 0.995, min = 0.959, max = 1)), # see above
"TSPOT_sens", list(distn = "pert", params = c(mode = 0.872, min = 0.643, max = 0.991)), # see above
"TSPOT_spec", list(distn = "pert", params = c(mode = 0.998, min = 0.996, max = 1)), # see above
"pAccept_chemo", list(distn = "pert", params = c(mode = 0.95, min = 0.5, max = 1)), # August, Pareek (2013)
"pComp_chemo", list(distn = "pert", params = c(mode = 0.8, min = 0.5, max = 0.9)), # Kowada (2013)
"pHep", list(distn = "unif", params = c(min = 0.001, max = 0.003)), # Kunst, Pareek
"Eff_comp", list(distn = "pert", params = c(mode = 0.65, min = 0.5, max = 0.8)), # Pareek (2013)
"Eff_incomp", list(distn = "pert", params = c(mode = 0.21, min = 0.1, max = 0.3)), # Pareek (2013)
"pTB", list(distn = "pert", params = c(mode = 0.12, min = 0.08, max = 0.19)),
"pAccept_TST", list(distn = "pert", params = c(mode = 0.96, min = 0.94, max = 0.98)), # posterior
"pTSTread", list(distn = "unif", params = c(min = 0.979, max = 0.979)),
"pIGRAread", list(distn = "pert", params = c(mode = 1, min = 1, max = 1)),
"pAccept_IGRA", list(distn = "pert", params = c(mode = 0.99, min = 0.98, max = 1)), # posterior
"pAccept_IGRA_TST+", list(distn = "unif", params = c(min = 0.995, max = 0.995)),
"Dual_sens", list(distn = "pert", params = c(mode = 0.632, min = 0.632, max = 0.632)),
"Dual_spec", list(distn = "pert", params = c(mode = 0.988, min = 0.988, max = 0.988)),
"pLTBI", list(distn = "pert", params = c(mode = 0.26, min = 0.23, max = 0.30)), # posterior, prev
"TSTIGRA_pos", list(distn = "pert", params = c(mode = 0.38, min = 0.32, max = 0.45)), # posterior
"TST_pos", list(distn = "pert", params = c(mode = 0.54, min = 0.50, max = 0.58)),
"TSPOT_pos", list(distn = "pert", params = c(mode = 0.23, min = 0.20, max = 0.26)),
"QFT_pos", list(distn = "pert", params = c(mode = 0.24, min = 0.20, max = 0.27)),
"PPV_TST", list(distn = "pert", params = c(mode = 0.39, min = 0.33, max = 0.45)),
"NPV_TST", list(distn = "pert", params = c(mode = 0.88, min = 0.84, max = 0.92)),
"PPV_QFT", list(distn = "pert", params = c(mode = 0.99, min = 0.97, max = 0.999)),
"NPV_QFT", list(distn = "pert", params = c(mode = 0.96, min = 0.94, max = 0.98)),
"PPV_TSPOT", list(distn = "pert", params = c(mode = 0.99, min = 0.97, max = 0.999)),
"NPV_TSPOT", list(distn = "pert", params = c(mode = 0.95, min = 0.93, max = 0.98)),
"QFT_pos_TST+", list(distn = "pert", params = c(mode = 0.34, min = 0.29, max = 0.41)),
"TSPOT_pos_TST+", list(distn = "pert", params = c(mode = 0.34, min = 0.29, max = 0.40)),
"PPV_QFT_TST+", list(distn = "pert", params = c(mode = 0.99, min = 0.98, max = 0.999)),
"NPV_QFT_TST+", list(distn = "pert", params = c(mode = 0.93, min = 0.89, max = 0.96)),
"PPV_TSPOT_TST+", list(distn = "pert", params = c(mode = 0.99, min = 0.99, max = 0.999)),
"NPV_TSPOT_TST+", list(distn = "pert", params = c(mode = 0.92, min = 0.89, max = 0.96)),
"pReact", list(distn = "pert", params = c(mode = 0.0075, min = 0.005, max = 0.01)), # CDC
##TODO: this should really be function of random Eff_comp/Eff_incomp above
## so outside of sampling step
"pReact_comp", list(distn = "pert", params = c(mode = 0.0026, min = 0.0018, max = 0.0035)),
"pReact_incomp", list(distn = "pert", params = c(mode = 0.0059, min = 0.0040, max = 0.0079))
)
label_probs <-
list(
"TST_pos" = 0.54, # posterior
"QFT_pos" = 0.24, # posterior
"TSPOT_pos" = 0.23, # posterior
# "TSTIGRA_pos" = 0.38, # posterior
"TSPOT_pos_TST+" = 0.34, # posterior
"QFT_pos_TST+" = 0.34, # posterior
"pTSTread" = 0.979,
"pIGRAread" = 1,
"pAccept_TST" = 0.96, # posterior
"pAccept_IGRA" = 0.99, # posterior
"pAccept_IGRA_TST+" = 0.99,
"Dual_sens" = 0.632,
"Dual_spec" = 0.988,
"TST_sens" = 0.79, # prior
"TST_spec" = 0.59, # prior
# "TST_sens" = 0.79, # posterior
# "TST_spec" = 0.59, # posterior
"QFT_sens" = 0.8,
"QFT_spec" = 0.97,
"TSPOT_sens" = 0.90,
"TSPOT_spec" = 0.95,
"pLTBI" = 0.26, # prior
# "pLTBI" = 0.26, # posterior
"PPV_TST" = 0.39, # posterior
"NPV_TST" = 0.88, # posterior
"PPV_QFT" = 0.99, # posterior
"NPV_QFT" = 0.96, # posterior
"PPV_TSPOT" = 0.99, # posterior
"NPV_TSPOT" = 0.95, # posterior
"PPV_QFT_TST+" = 0.99,
"NPV_QFT_TST+" = 0.93,
"PPV_TSPOT_TST+" = 0.99,
"NPV_TSPOT_TST+" = 0.92,
"pAccept_chemo" = 0.95,
"pComp_chemo" = 0.8,
"pHep" = 0.002,
"pReact" = 0.0075, # CDC
"pReact_comp" = 0.0026,
"pReact_incomp" = 0.0059)
hsuv <-
list(
"loss_chemo" = 0.001, # Auguste (2016)
"loss_hep" = 0.14, # Woo (2012)
"loss_tb" = 0.15) # Auguste (2016)
# assign labels to branches -----------------------------------------------
#########
# costs #
#########
TST_cname_from_to <-
rbind.data.frame(
# c("Ns_cost", 1, 2), # remove this since all trees start with nurse visit
c("TST", 2, 3),
c("Ns_cost", 3, 4),
c("Sp_cost", 4, 5),
c("Hep", 7, 8),
c("LTBIincompl_cost", 8, 9),
c("LTBIcompl_cost", 10, 11),
c("LTBIincompl_cost", 10, 12),
c("Hep", 16, 17),
c("LTBIincompl_cost", 17, 18),
c("LTBIcompl_cost", 19, 20),
c("LTBIincompl_cost", 19, 21)
) %>%
setNames(c("name", "from", "to"))
QFT_cname_from_to <-
rbind.data.frame(
c("QFT", 1, 2),
c("Sp_cost", 2, 3),
c("Hep", 5, 6),
c("LTBIincompl_cost", 6, 7),
c("LTBIcompl_cost", 8, 9),
c("LTBIincompl_cost", 8, 10),
c("Hep", 14, 15),
c("LTBIincompl_cost", 15, 16),
c("LTBIcompl_cost", 17, 18),
c("LTBIincompl_cost", 17, 19)
) %>%
setNames(c("name", "from", "to"))
TSPOT_cname_from_to <-
rbind.data.frame(
c("TSPOT", 1, 2),
c("Sp_cost", 2, 3),
c("Hep", 5, 6),
c("LTBIincompl_cost", 6, 7),
c("LTBIcompl_cost", 8, 9),
c("LTBIincompl_cost", 8, 10),
c("Hep", 14, 15),
c("LTBIincompl_cost", 15, 16),
c("LTBIcompl_cost", 17, 18),
c("LTBIincompl_cost", 17, 19)
) %>%
setNames(c("name", "from", "to"))
TST_QFT_cname_from_to <-
rbind.data.frame(
c("TST", 1, 2),
c("Ns_cost", 2, 3),
c("QFT", 4, 5),
c("Sp_cost", 5, 6),
c("Hep", 8, 9),
c("Hep", 17, 18),
c("LTBIincompl_cost", 9, 10),
c("LTBIcompl_cost", 11, 12),
c("LTBIincompl_cost", 11, 13),
c("LTBIincompl_cost", 18, 19),
c("LTBIincompl_cost", 20, 21),
c("LTBIincompl_cost", 20, 22)
) %>%
setNames(c("name", "from", "to"))
TST_TSPOT_cname_from_to <-
rbind.data.frame(
c("TST", 1, 2),
c("Ns_cost", 2, 3),
c("TSPOT", 4, 5),
c("Sp_cost", 5, 6),
c("Hep", 8, 9),
c("Hep", 17, 18),
c("LTBIincompl_cost", 9, 10),
c("LTBIcompl_cost", 11, 12),
c("LTBIincompl_cost", 11, 13),
c("LTBIincompl_cost", 18, 19),
c("LTBIcompl_cost", 20, 21),
c("LTBIincompl_cost", 20, 22)
) %>%
setNames(c("name", "from", "to"))
TST_QFT_fwd_cname_from_to <-
rbind.data.frame(
c("TST", 2, 3),
c("TST", 25, 26),
c("Ns_cost", 3, 4),
c("Ns_cost", 26, 27),
c("QFT", 5, 6),
c("QFT", 28, 29),
c("Sp_cost", 6, 7),
c("Sp_cost", 29, 30),
c("Hep", 8, 9),
c("Hep", 31, 32),
c("LTBIincompl_cost", 9, 10),
c("LTBIincompl_cost", 32, 33),
c("LTBIcompl_cost", 11, 12),
c("LTBIcompl_cost", 34, 35)
) %>%
setNames(c("name", "from", "to"))
#########
# probs #
#########
TST_pname_from_to <-
rbind.data.frame(
c("pAccept_TST", 2, 3),
c("pTSTread", 3, 4),
c("TST_pos", 4, 5),
c("PPV_TST", 5, 6),
c("pAccept_chemo", 6, 7),
c("pHep", 7, 8),
c("pComp_chemo", 10, 11),
c("pAccept_chemo", 15, 16),
c("pHep", 16, 17),
c("pComp_chemo", 19, 20),
c("NPV_TST", 24, 27),
c("pLTBI", 29, 30),
c("pLTBI", 32, 33)
) %>%
setNames(c("name", "from", "to"))
QFT_pname_from_to <-
rbind.data.frame(
c("pAccept_IGRA", 1, 2),
c("QFT_pos", 2, 3),
c("PPV_QFT", 3, 4),
c("pAccept_chemo", 4, 5),
c("pHep", 5, 6),
c("pComp_chemo", 8, 9),
c("pAccept_chemo", 13, 14),
c("pHep", 14, 15),
c("pComp_chemo", 17, 18),
c("NPV_QFT", 22, 25),
c("pLTBI", 27, 28)
) %>%
setNames(c("name", "from", "to"))
TSPOT_pname_from_to <-
rbind.data.frame(
c("pAccept_IGRA", 1, 2),
c("TSPOT_pos", 2, 3),
c("PPV_TSPOT", 3, 4),
c("pAccept_chemo", 4, 5),
c("pHep", 5, 6),
c("pComp_chemo", 8, 9),
c("pAccept_chemo", 13, 14),
c("pHep", 14, 15),
c("pComp_chemo", 17, 18),
c("NPV_TSPOT", 22, 25),
c("pLTBI", 27, 28)
) %>%
setNames(c("name", "from", "to"))
TST_TSPOT_pname_from_to <-
rbind.data.frame(
c("pAccept_TST", 1, 2),
c("pTSTread", 2, 3),
c("TST_pos", 3, 4),
c("pAccept_IGRA_TST+", 4, 5),
c("TSPOT_pos_TST+", 5, 6),
c("PPV_TSPOT_TST+", 6, 7),
c("pAccept_chemo", 7, 8),
c("pHep", 8, 9),
c("pComp_chemo", 11, 12),
c("pAccept_chemo", 16, 17),
c("pHep", 17, 18),
c("pComp_chemo", 20, 21),
c("NPV_TSPOT_TST+", 25, 28),
c("PPV_TST", 30, 31),
c("NPV_TST", 33, 36),
c("pLTBI", 38, 39),
c("pLTBI", 41, 42)
) %>%
setNames(c("name", "from", "to"))
TST_QFT_pname_from_to <-
rbind.data.frame(
c("pAccept_TST", 1, 2),
c("pTSTread", 2, 3),
c("TST_pos", 3, 4),
c("pAccept_IGRA_TST+", 4, 5),
c("QFT_pos_TST+", 5, 6),
c("PPV_QFT_TST+", 6, 7),
c("pAccept_chemo", 7, 8),
c("pHep", 8, 9),
c("pComp_chemo", 11, 12),
c("pAccept_chemo", 16, 17),
c("pHep", 17, 18),
c("pComp_chemo", 20, 21),
c("NPV_QFT_TST+", 25, 28),
c("PPV_TST", 30, 31),
c("NPV_TST", 33, 36),
c("pLTBI", 38, 39),
c("pLTBI", 41, 42)
) %>%
setNames(c("name", "from", "to"))
TST_QFT_fwd_pname_from_to <-
rbind.data.frame(
c("pLTBI", 1, 2),
c("pAccept_TST", 2, 3),
c("pAccept_TST", 25, 26),
c("pTSTread", 3, 4),
c("pTSTread", 26, 27),
c("TST_sens", 4, 5),
c("TST_spec", 27, 42),
c("pAccept_IGRA", 5, 6),
c("pAccept_IGRA", 28, 29),
c("QFT_sens", 6, 7),
c("QFT_spec", 29, 38),
c("pAccept_chemo", 7, 8),
c("pAccept_chemo", 30, 31),
c("pHep", 8, 9),
c("pHep", 31, 32),
c("pComp_chemo", 11, 12),
c("pComp_chemo", 34, 35)
) %>%
setNames(c("name", "from", "to"))
#############
# QALY loss #
#############
t_chemo <- 3/12 # LTBI treatment, years
t_hep <- 1.5/12
# assume that Auguste (2016) detriment is for whole treatment duration
label_health <-
list(
"Hep" = hsuv$loss_hep*t_hep,
"Total (complete)" = hsuv$loss_chemo,
"Total (incomplete)" = hsuv$loss_chemo/2) # assume drop-out half way through
# "Total (complete)" = hsuv$loss_chemo*t_chemo,
# "Total (incomplete)" = hsuv$loss_chemo*t_chemo/2)
label_health_distns <-
tribble(
~name.health, ~vals,
"Hep",
list(distn = "unif", params = c(min = 0.13, max = 0.15)), # Woo (2012)
"Total (complete)",
list(distn = "unif", params = c(min = 0, max = 0.002)), # Auguste (2016)
"Total (incomplete)",
list(distn = "unif", params = c(min = 0, max = 0.002/2))) # Auguste (2016)
# "Total (complete)",
# list(distn = "unif", params = c(min = hsuv$loss_chemo*t_chemo, max = hsuv$loss_chemo*t_chemo)), # Auguste (2016)
# "Total (incomplete)",
# list(distn = "unif", params = c(min = hsuv$loss_chemo*t_chemo/2, max = hsuv$loss_chemo*t_chemo/2))) # Auguste (2016)
TST_hname_from_to <-
rbind.data.frame(
c("Hep", 7, 8),
c("Hep", 16, 17),
c("Total (incomplete)", 8, 9),
c("Total (incomplete)", 17, 18),
c("Total (incomplete)", 10, 12),
c("Total (incomplete)", 19, 21),
c("Total (complete)", 10, 11),
c("Total (complete)", 19, 20)) %>%
setNames(c("name", "from", "to"))
QFT_hname_from_to <-
rbind.data.frame(
c("Hep", 5, 6),
c("Hep", 14, 15),
c("Total (incomplete)", 6, 7),
c("Total (incomplete)", 8, 10),
c("Total (incomplete)", 15, 16),
c("Total (incomplete)", 17, 19),
c("Total (complete)", 8, 9),
c("Total (complete)", 17, 18)) %>%
setNames(c("name", "from", "to"))
TSPOT_hname_from_to <-
rbind.data.frame(
c("Hep", 5, 6),
c("Hep", 14, 15),
c("Total (incomplete)", 6, 7),
c("Total (incomplete)", 8, 10),
c("Total (incomplete)", 15, 16),
c("Total (incomplete)", 17, 19),
c("Total (complete)", 8, 9),
c("Total (complete)", 17, 18)) %>%
setNames(c("name", "from", "to"))
TST_QFT_hname_from_to <-
rbind.data.frame(
c("Hep", 8, 9),
c("Hep", 17, 18),
c("Total (incomplete)", 9, 10),
c("Total (incomplete)", 11, 13),
c("Total (incomplete)", 18, 19),
c("Total (incomplete)", 20, 22),
c("Total (complete)", 11, 12),
c("Total (complete)", 20, 21)) %>%
setNames(c("name", "from", "to"))
TST_TSPOT_hname_from_to <-
rbind.data.frame(
c("Hep", 8, 9),
c("Hep", 17, 18),
c("Total (incomplete)", 9, 10),
c("Total (incomplete)", 11, 13),
c("Total (incomplete)", 18, 19),
c("Total (incomplete)", 20, 22),
c("Total (complete)", 11, 12),
c("Total (complete)", 20, 21)) %>%
setNames(c("name", "from", "to"))
TST_QFT_fwd_hname_from_to <-
rbind.data.frame(
c("Hep", 8, 9),
c("Hep", 31, 32),
c("Total (incomplete)", 9, 10),
c("Total (incomplete)", 11, 48),
c("Total (incomplete)", 32, 33),
c("Total (incomplete)", 34, 49),
c("Total (complete)", 11, 12),
c("Total (complete)", 34, 35)) %>%
setNames(c("name", "from", "to"))
if (save) {
save(label_costs,
label_health,
label_probs,
label_costs_distns,
label_health_distns,
label_probs_distns,
TST_cname_from_to,
TST_hname_from_to,
TST_pname_from_to,
TST_QFT_cname_from_to,
TST_QFT_hname_from_to,
TST_QFT_pname_from_to,
TST_QFT_fwd_cname_from_to,
TST_QFT_fwd_pname_from_to,
TST_QFT_fwd_hname_from_to,
TSPOT_cname_from_to,
TSPOT_hname_from_to,
TSPOT_pname_from_to,
TST_TSPOT_cname_from_to,
TST_TSPOT_hname_from_to,
TST_TSPOT_pname_from_to,
QFT_cname_from_to,
QFT_hname_from_to,
QFT_pname_from_to,
hsuv,
file = here::here("data/params.RData"))}
}
n8thangreen/LTBIdiagTST documentation built on Oct. 29, 2024, 9:23 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions create_param_values
Documented in create_param_values
#' Create parameter values
#'
#' All parameter values required for the decision tree model.
#'
#' Excel model with structure and labelling:
#' G:/DIDE-PC_2019/MSc-MPH/projects/2017/LTBI-TST_Manabu/decision tree Excel/all_decision_trees.xls
#' @export
#'
#' @examples
#' create_param_values()
#'
#' ## save lists as csv
#' # write.csv(cbind(label_costs), file = "data/cost_inputs.csv")
#' # write.csv(cbind(label_health), file = "data/health_inputs.csv")
#' # write.csv(cbind(label_probs), file = "data/probs_inputs.csv")
#'
create_param_values <- function(save = TRUE) {
# assign values to labels ------------------------------------------
label_costs <-
list(
# "Contact tracing per contact" = 368.9,
# "Mean number of contacts examined per primary case" = 6.5,
# "Total Contact tracing" = 2397.85,
#
# "Cost of inpatient episode for acute TB" = 3325.15,
# "Proportion of patients with acute TB who are admitted" = 0.53,
# "Total Inpatient care" = 1762.3295,
#
# "Cost of culture test" = 8.06,
# "Culture tests per case" = 4,
# "CXR per case" = 2,
# "LFT per case" = 4,
# "Total Cost of tests" = 104.74,
#
# "HRZE tablet (60)" = 39.5,
# "_tablets per month" = 150,
# "HRZE per month" = 98.75,
# "HR tablet 300/150 (56)" = 25.22,
# "tablets per month" = 60,
# "HR per month" = 27.02,
# "Duration of initial therapy" = 2,
# "Duration of continious therapy" = 4,
# "Total Cost of chemotherapy" = 305.58,
#
# "Cost of outpatient consultation (first visit)" = 208,
# "Cost of outpatient consultation (follow-up visit)" = 94,
# "Number of outpatient clinical visits per case" = 4,
# "Visits from TB nurse per case" = 6,
# "Total Out patient care" = 755.86,
"TST" = 25.02,
"QFT" = 29.11,
"TSPOT" = 43.17,
"Ns_cost" = 53, # TB nurse visit
"Out patient consultation (first visit)" = 224, # c_out
"Follow-up via nurses" = 166, # c_fuout
"CXR" = 46,
"LFT" = 3.95,
"Number of CXR" = 1, # Pareek (2013) 2
"Number of LFT" = 1, # Pareek (2013) 4
"Sp_cost" = 274,
"HR tablet (300/150) (56)" = 25.22, # BNF
"drug cost per month" = 27.02,
"drug_cost" = 33.22, # per month, inflate to 2023
"dur_hr" = 3, # duration of chemo HR
"Number of outpatient consultation" = 1,
"n_appts" = 2, # Number of TB nurse appointments
"Hep" = 984, # Pooran (2010) inflated to 2023
"TB_cost" = 6055) # NICE guideline NG33 (2016) inflated to 2023
label_costs$LTBIcompl_cost <- with(label_costs, Ns_cost*n_appts + drug_cost*dur_hr)
label_costs$LTBIincompl_cost <- with(label_costs, Ns_cost*floor(n_appts/2) + drug_cost*dur_hr/2)
label_costs_distns <-
tribble(
~name.cost, ~vals,
"TST", list(distn = "pert", params = c(mode = 25.02, min = 12.51, max = 50.05)), # Pooran 2010
"QFT", list(distn = "pert", params = c(mode = 29.11, min = 15.16, max = 55.67)), # Imperial NHS
"TSPOT", list(distn = "pert", params = c(mode = 43.17, min = 22.21, max = 83.87)), # Imperial NHS
"CXR", list(distn = "pert", params = c(mode = 46, min = 40, max = 52)), # Reference cost 2016
"LFT", list(distn = "pert", params = c(mode = 3.95, min = 2.63, max = 5.27)), # Reference cost 2016
"Hep", list(distn = "pert", params = c(mode = 984, min = 492, max = 1968)), # Pooran 2010 inflated to 2023
"TB_cost", list(distn = "pert", params = c(mode = 6055, min = 3028, max = 12110)), # NICE guideline NG33 (2016) inflated to 2023
"Sp_cost", list(distn = "pert", params = c(mode = 274, min = 274, max = 274)), # positive screening cost
"Ns_cost", list(distn = "pert", params = c(mode = 53, min = 43, max = 63)), # TB nurse visit
)
##TODO: problem with this because the sampling should be dependent on previous sampling so its really a function
## how an we incorporate this in to the sampling function? this isnt actually needed in the sampling step
## so could define as function and run afterwards e.g. label_costs_distns_fns
label_costs_distns <- bind_rows(
label_costs_distns,
tribble(~name.cost, ~vals,
"LTBIincompl_cost", with(label_costs,
list(distn = "unif",
params = c(min = Ns_cost*floor(n_appts/2) + drug_cost*dur_hr/2,
max = Ns_cost*floor(n_appts/2) + drug_cost*dur_hr/2))),
"LTBIcompl_cost", with(label_costs,
list(distn = "unif",
params = c(min = Ns_cost*n_appts + drug_cost*dur_hr,
max = Ns_cost*n_appts + drug_cost*dur_hr)))))
label_probs_distns <-
tribble(
~name.prob, ~prob,
"TST_sens", list(distn = "pert", params = c(mode = 0.79, min = 0.69, max = 0.89)), # Kahwati (2016)
"TST_spec", list(distn = "pert", params = c(mode = 0.59, min = 0.46, max = 0.73)), # Pai (2008)
# "TST_sens", list(distn = "norm", params = c(mean = 0.69, sd = 0.89)), # posterior
# "TST_spec", list(distn = "norm", params = c(mean = 0.46, sd = 0.73)), # posterior
"QFT_sens", list(distn = "pert", params = c(mode = 0.886, min = 0.812, max = 0.944)), # Zhang et al. BMC Infectious Diseases (2023) 23:40
"QFT_spec", list(distn = "pert", params = c(mode = 0.995, min = 0.959, max = 1)), # see above
"TSPOT_sens", list(distn = "pert", params = c(mode = 0.872, min = 0.643, max = 0.991)), # see above
"TSPOT_spec", list(distn = "pert", params = c(mode = 0.998, min = 0.996, max = 1)), # see above
"pAccept_chemo", list(distn = "pert", params = c(mode = 0.95, min = 0.5, max = 1)), # August, Pareek (2013)
"pComp_chemo", list(distn = "pert", params = c(mode = 0.8, min = 0.5, max = 0.9)), # Kowada (2013)
"pHep", list(distn = "unif", params = c(min = 0.001, max = 0.003)), # Kunst, Pareek
"Eff_comp", list(distn = "pert", params = c(mode = 0.65, min = 0.5, max = 0.8)), # Pareek (2013)
"Eff_incomp", list(distn = "pert", params = c(mode = 0.21, min = 0.1, max = 0.3)), # Pareek (2013)
"pTB", list(distn = "pert", params = c(mode = 0.12, min = 0.08, max = 0.19)),
"pAccept_TST", list(distn = "pert", params = c(mode = 0.96, min = 0.94, max = 0.98)), # posterior
"pTSTread", list(distn = "unif", params = c(min = 0.979, max = 0.979)),
"pIGRAread", list(distn = "pert", params = c(mode = 1, min = 1, max = 1)),
"pAccept_IGRA", list(distn = "pert", params = c(mode = 0.99, min = 0.98, max = 1)), # posterior
"pAccept_IGRA_TST+", list(distn = "unif", params = c(min = 0.995, max = 0.995)),
"Dual_sens", list(distn = "pert", params = c(mode = 0.632, min = 0.632, max = 0.632)),
"Dual_spec", list(distn = "pert", params = c(mode = 0.988, min = 0.988, max = 0.988)),
"pLTBI", list(distn = "pert", params = c(mode = 0.26, min = 0.23, max = 0.30)), # posterior, prev
"TSTIGRA_pos", list(distn = "pert", params = c(mode = 0.38, min = 0.32, max = 0.45)), # posterior
"TST_pos", list(distn = "pert", params = c(mode = 0.54, min = 0.50, max = 0.58)),
"TSPOT_pos", list(distn = "pert", params = c(mode = 0.23, min = 0.20, max = 0.26)),
"QFT_pos", list(distn = "pert", params = c(mode = 0.24, min = 0.20, max = 0.27)),
"PPV_TST", list(distn = "pert", params = c(mode = 0.39, min = 0.33, max = 0.45)),
"NPV_TST", list(distn = "pert", params = c(mode = 0.88, min = 0.84, max = 0.92)),
"PPV_QFT", list(distn = "pert", params = c(mode = 0.99, min = 0.97, max = 0.999)),
"NPV_QFT", list(distn = "pert", params = c(mode = 0.96, min = 0.94, max = 0.98)),
"PPV_TSPOT", list(distn = "pert", params = c(mode = 0.99, min = 0.97, max = 0.999)),
"NPV_TSPOT", list(distn = "pert", params = c(mode = 0.95, min = 0.93, max = 0.98)),
"QFT_pos_TST+", list(distn = "pert", params = c(mode = 0.34, min = 0.29, max = 0.41)),
"TSPOT_pos_TST+", list(distn = "pert", params = c(mode = 0.34, min = 0.29, max = 0.40)),
"PPV_QFT_TST+", list(distn = "pert", params = c(mode = 0.99, min = 0.98, max = 0.999)),
"NPV_QFT_TST+", list(distn = "pert", params = c(mode = 0.93, min = 0.89, max = 0.96)),
"PPV_TSPOT_TST+", list(distn = "pert", params = c(mode = 0.99, min = 0.99, max = 0.999)),
"NPV_TSPOT_TST+", list(distn = "pert", params = c(mode = 0.92, min = 0.89, max = 0.96)),
"pReact", list(distn = "pert", params = c(mode = 0.0075, min = 0.005, max = 0.01)), # CDC
##TODO: this should really be function of random Eff_comp/Eff_incomp above
## so outside of sampling step
"pReact_comp", list(distn = "pert", params = c(mode = 0.0026, min = 0.0018, max = 0.0035)),
"pReact_incomp", list(distn = "pert", params = c(mode = 0.0059, min = 0.0040, max = 0.0079))
)
label_probs <-
list(
"TST_pos" = 0.54, # posterior
"QFT_pos" = 0.24, # posterior
"TSPOT_pos" = 0.23, # posterior
# "TSTIGRA_pos" = 0.38, # posterior
"TSPOT_pos_TST+" = 0.34, # posterior
"QFT_pos_TST+" = 0.34, # posterior
"pTSTread" = 0.979,
"pIGRAread" = 1,
"pAccept_TST" = 0.96, # posterior
"pAccept_IGRA" = 0.99, # posterior
"pAccept_IGRA_TST+" = 0.99,
"Dual_sens" = 0.632,
"Dual_spec" = 0.988,
"TST_sens" = 0.79, # prior
"TST_spec" = 0.59, # prior
# "TST_sens" = 0.79, # posterior
# "TST_spec" = 0.59, # posterior
"QFT_sens" = 0.8,
"QFT_spec" = 0.97,
"TSPOT_sens" = 0.90,
"TSPOT_spec" = 0.95,
"pLTBI" = 0.26, # prior
# "pLTBI" = 0.26, # posterior
"PPV_TST" = 0.39, # posterior
"NPV_TST" = 0.88, # posterior
"PPV_QFT" = 0.99, # posterior
"NPV_QFT" = 0.96, # posterior
"PPV_TSPOT" = 0.99, # posterior
"NPV_TSPOT" = 0.95, # posterior
"PPV_QFT_TST+" = 0.99,
"NPV_QFT_TST+" = 0.93,
"PPV_TSPOT_TST+" = 0.99,
"NPV_TSPOT_TST+" = 0.92,
"pAccept_chemo" = 0.95,
"pComp_chemo" = 0.8,
"pHep" = 0.002,
"pReact" = 0.0075, # CDC
"pReact_comp" = 0.0026,
"pReact_incomp" = 0.0059)
hsuv <-
list(
"loss_chemo" = 0.001, # Auguste (2016)
"loss_hep" = 0.14, # Woo (2012)
"loss_tb" = 0.15) # Auguste (2016)
# assign labels to branches -----------------------------------------------
#########
# costs #
#########
TST_cname_from_to <-
rbind.data.frame(
# c("Ns_cost", 1, 2), # remove this since all trees start with nurse visit
c("TST", 2, 3),
c("Ns_cost", 3, 4),
c("Sp_cost", 4, 5),
c("Hep", 7, 8),
c("LTBIincompl_cost", 8, 9),
c("LTBIcompl_cost", 10, 11),
c("LTBIincompl_cost", 10, 12),
c("Hep", 16, 17),
c("LTBIincompl_cost", 17, 18),
c("LTBIcompl_cost", 19, 20),
c("LTBIincompl_cost", 19, 21)
) %>%
setNames(c("name", "from", "to"))
QFT_cname_from_to <-
rbind.data.frame(
c("QFT", 1, 2),
c("Sp_cost", 2, 3),
c("Hep", 5, 6),
c("LTBIincompl_cost", 6, 7),
c("LTBIcompl_cost", 8, 9),
c("LTBIincompl_cost", 8, 10),
c("Hep", 14, 15),
c("LTBIincompl_cost", 15, 16),
c("LTBIcompl_cost", 17, 18),
c("LTBIincompl_cost", 17, 19)
) %>%
setNames(c("name", "from", "to"))
TSPOT_cname_from_to <-
rbind.data.frame(
c("TSPOT", 1, 2),
c("Sp_cost", 2, 3),
c("Hep", 5, 6),
c("LTBIincompl_cost", 6, 7),
c("LTBIcompl_cost", 8, 9),
c("LTBIincompl_cost", 8, 10),
c("Hep", 14, 15),
c("LTBIincompl_cost", 15, 16),
c("LTBIcompl_cost", 17, 18),
c("LTBIincompl_cost", 17, 19)
) %>%
setNames(c("name", "from", "to"))
TST_QFT_cname_from_to <-
rbind.data.frame(
c("TST", 1, 2),
c("Ns_cost", 2, 3),
c("QFT", 4, 5),
c("Sp_cost", 5, 6),
c("Hep", 8, 9),
c("Hep", 17, 18),
c("LTBIincompl_cost", 9, 10),
c("LTBIcompl_cost", 11, 12),
c("LTBIincompl_cost", 11, 13),
c("LTBIincompl_cost", 18, 19),
c("LTBIincompl_cost", 20, 21),
c("LTBIincompl_cost", 20, 22)
) %>%
setNames(c("name", "from", "to"))
TST_TSPOT_cname_from_to <-
rbind.data.frame(
c("TST", 1, 2),
c("Ns_cost", 2, 3),
c("TSPOT", 4, 5),
c("Sp_cost", 5, 6),
c("Hep", 8, 9),
c("Hep", 17, 18),
c("LTBIincompl_cost", 9, 10),
c("LTBIcompl_cost", 11, 12),
c("LTBIincompl_cost", 11, 13),
c("LTBIincompl_cost", 18, 19),
c("LTBIcompl_cost", 20, 21),
c("LTBIincompl_cost", 20, 22)
) %>%
setNames(c("name", "from", "to"))
TST_QFT_fwd_cname_from_to <-
rbind.data.frame(
c("TST", 2, 3),
c("TST", 25, 26),
c("Ns_cost", 3, 4),
c("Ns_cost", 26, 27),
c("QFT", 5, 6),
c("QFT", 28, 29),
c("Sp_cost", 6, 7),
c("Sp_cost", 29, 30),
c("Hep", 8, 9),
c("Hep", 31, 32),
c("LTBIincompl_cost", 9, 10),
c("LTBIincompl_cost", 32, 33),
c("LTBIcompl_cost", 11, 12),
c("LTBIcompl_cost", 34, 35)
) %>%
setNames(c("name", "from", "to"))
#########
# probs #
#########
TST_pname_from_to <-
rbind.data.frame(
c("pAccept_TST", 2, 3),
c("pTSTread", 3, 4),
c("TST_pos", 4, 5),
c("PPV_TST", 5, 6),
c("pAccept_chemo", 6, 7),
c("pHep", 7, 8),
c("pComp_chemo", 10, 11),
c("pAccept_chemo", 15, 16),
c("pHep", 16, 17),
c("pComp_chemo", 19, 20),
c("NPV_TST", 24, 27),
c("pLTBI", 29, 30),
c("pLTBI", 32, 33)
) %>%
setNames(c("name", "from", "to"))
QFT_pname_from_to <-
rbind.data.frame(
c("pAccept_IGRA", 1, 2),
c("QFT_pos", 2, 3),
c("PPV_QFT", 3, 4),
c("pAccept_chemo", 4, 5),
c("pHep", 5, 6),
c("pComp_chemo", 8, 9),
c("pAccept_chemo", 13, 14),
c("pHep", 14, 15),
c("pComp_chemo", 17, 18),
c("NPV_QFT", 22, 25),
c("pLTBI", 27, 28)
) %>%
setNames(c("name", "from", "to"))
TSPOT_pname_from_to <-
rbind.data.frame(
c("pAccept_IGRA", 1, 2),
c("TSPOT_pos", 2, 3),
c("PPV_TSPOT", 3, 4),
c("pAccept_chemo", 4, 5),
c("pHep", 5, 6),
c("pComp_chemo", 8, 9),
c("pAccept_chemo", 13, 14),
c("pHep", 14, 15),
c("pComp_chemo", 17, 18),
c("NPV_TSPOT", 22, 25),
c("pLTBI", 27, 28)
) %>%
setNames(c("name", "from", "to"))
TST_TSPOT_pname_from_to <-
rbind.data.frame(
c("pAccept_TST", 1, 2),
c("pTSTread", 2, 3),
c("TST_pos", 3, 4),
c("pAccept_IGRA_TST+", 4, 5),
c("TSPOT_pos_TST+", 5, 6),
c("PPV_TSPOT_TST+", 6, 7),
c("pAccept_chemo", 7, 8),
c("pHep", 8, 9),
c("pComp_chemo", 11, 12),
c("pAccept_chemo", 16, 17),
c("pHep", 17, 18),
c("pComp_chemo", 20, 21),
c("NPV_TSPOT_TST+", 25, 28),
c("PPV_TST", 30, 31),
c("NPV_TST", 33, 36),
c("pLTBI", 38, 39),
c("pLTBI", 41, 42)
) %>%
setNames(c("name", "from", "to"))
TST_QFT_pname_from_to <-
rbind.data.frame(
c("pAccept_TST", 1, 2),
c("pTSTread", 2, 3),
c("TST_pos", 3, 4),
c("pAccept_IGRA_TST+", 4, 5),
c("QFT_pos_TST+", 5, 6),
c("PPV_QFT_TST+", 6, 7),
c("pAccept_chemo", 7, 8),
c("pHep", 8, 9),
c("pComp_chemo", 11, 12),
c("pAccept_chemo", 16, 17),
c("pHep", 17, 18),
c("pComp_chemo", 20, 21),
c("NPV_QFT_TST+", 25, 28),
c("PPV_TST", 30, 31),
c("NPV_TST", 33, 36),
c("pLTBI", 38, 39),
c("pLTBI", 41, 42)
) %>%
setNames(c("name", "from", "to"))
TST_QFT_fwd_pname_from_to <-
rbind.data.frame(
c("pLTBI", 1, 2),
c("pAccept_TST", 2, 3),
c("pAccept_TST", 25, 26),
c("pTSTread", 3, 4),
c("pTSTread", 26, 27),
c("TST_sens", 4, 5),
c("TST_spec", 27, 42),
c("pAccept_IGRA", 5, 6),
c("pAccept_IGRA", 28, 29),
c("QFT_sens", 6, 7),
c("QFT_spec", 29, 38),
c("pAccept_chemo", 7, 8),
c("pAccept_chemo", 30, 31),
c("pHep", 8, 9),
c("pHep", 31, 32),
c("pComp_chemo", 11, 12),
c("pComp_chemo", 34, 35)
) %>%
setNames(c("name", "from", "to"))
#############
# QALY loss #
#############
t_chemo <- 3/12 # LTBI treatment, years
t_hep <- 1.5/12
# assume that Auguste (2016) detriment is for whole treatment duration
label_health <-
list(
"Hep" = hsuv$loss_hep*t_hep,
"Total (complete)" = hsuv$loss_chemo,
"Total (incomplete)" = hsuv$loss_chemo/2) # assume drop-out half way through
# "Total (complete)" = hsuv$loss_chemo*t_chemo,
# "Total (incomplete)" = hsuv$loss_chemo*t_chemo/2)
label_health_distns <-
tribble(
~name.health, ~vals,
"Hep",
list(distn = "unif", params = c(min = 0.13, max = 0.15)), # Woo (2012)
"Total (complete)",
list(distn = "unif", params = c(min = 0, max = 0.002)), # Auguste (2016)
"Total (incomplete)",
list(distn = "unif", params = c(min = 0, max = 0.002/2))) # Auguste (2016)
# "Total (complete)",
# list(distn = "unif", params = c(min = hsuv$loss_chemo*t_chemo, max = hsuv$loss_chemo*t_chemo)), # Auguste (2016)
# "Total (incomplete)",
# list(distn = "unif", params = c(min = hsuv$loss_chemo*t_chemo/2, max = hsuv$loss_chemo*t_chemo/2))) # Auguste (2016)
TST_hname_from_to <-
rbind.data.frame(
c("Hep", 7, 8),
c("Hep", 16, 17),
c("Total (incomplete)", 8, 9),
c("Total (incomplete)", 17, 18),
c("Total (incomplete)", 10, 12),
c("Total (incomplete)", 19, 21),
c("Total (complete)", 10, 11),
c("Total (complete)", 19, 20)) %>%
setNames(c("name", "from", "to"))
QFT_hname_from_to <-
rbind.data.frame(
c("Hep", 5, 6),
c("Hep", 14, 15),
c("Total (incomplete)", 6, 7),
c("Total (incomplete)", 8, 10),
c("Total (incomplete)", 15, 16),
c("Total (incomplete)", 17, 19),
c("Total (complete)", 8, 9),
c("Total (complete)", 17, 18)) %>%
setNames(c("name", "from", "to"))
TSPOT_hname_from_to <-
rbind.data.frame(
c("Hep", 5, 6),
c("Hep", 14, 15),
c("Total (incomplete)", 6, 7),
c("Total (incomplete)", 8, 10),
c("Total (incomplete)", 15, 16),
c("Total (incomplete)", 17, 19),
c("Total (complete)", 8, 9),
c("Total (complete)", 17, 18)) %>%
setNames(c("name", "from", "to"))
TST_QFT_hname_from_to <-
rbind.data.frame(
c("Hep", 8, 9),
c("Hep", 17, 18),
c("Total (incomplete)", 9, 10),
c("Total (incomplete)", 11, 13),
c("Total (incomplete)", 18, 19),
c("Total (incomplete)", 20, 22),
c("Total (complete)", 11, 12),
c("Total (complete)", 20, 21)) %>%
setNames(c("name", "from", "to"))
TST_TSPOT_hname_from_to <-
rbind.data.frame(
c("Hep", 8, 9),
c("Hep", 17, 18),
c("Total (incomplete)", 9, 10),
c("Total (incomplete)", 11, 13),
c("Total (incomplete)", 18, 19),
c("Total (incomplete)", 20, 22),
c("Total (complete)", 11, 12),
c("Total (complete)", 20, 21)) %>%
setNames(c("name", "from", "to"))
TST_QFT_fwd_hname_from_to <-
rbind.data.frame(
c("Hep", 8, 9),
c("Hep", 31, 32),
c("Total (incomplete)", 9, 10),
c("Total (incomplete)", 11, 48),
c("Total (incomplete)", 32, 33),
c("Total (incomplete)", 34, 49),
c("Total (complete)", 11, 12),
c("Total (complete)", 34, 35)) %>%
setNames(c("name", "from", "to"))
if (save) {
save(label_costs,
label_health,
label_probs,
label_costs_distns,
label_health_distns,
label_probs_distns,
TST_cname_from_to,
TST_hname_from_to,
TST_pname_from_to,
TST_QFT_cname_from_to,
TST_QFT_hname_from_to,
TST_QFT_pname_from_to,
TST_QFT_fwd_cname_from_to,
TST_QFT_fwd_pname_from_to,
TST_QFT_fwd_hname_from_to,
TSPOT_cname_from_to,
TSPOT_hname_from_to,
TSPOT_pname_from_to,
TST_TSPOT_cname_from_to,
TST_TSPOT_hname_from_to,
TST_TSPOT_pname_from_to,
QFT_cname_from_to,
QFT_hname_from_to,
QFT_pname_from_to,
hsuv,
file = here::here("data/params.RData"))}
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.