workbook.Rmd

---
title: "Vignette Title"
output:
  github_document: 
    toc: true
---

```{r, setup, echo = FALSE, message = FALSE, warning = FALSE}
library(mrgsolve)
library(tidyverse)
```

# Problems

## Warm-up 

Choose a `PKPD` model from the internal model library (`?modlib_pkpd`)
to explore

- Check the parameter values (`param`)
- Check the compartments and initial values (`init`)
- Review the model code (`see)

```{r, warm_up_example, eval = FALSE}
mod <- mread_cache("", modlib())
```

Extra credit: can you match up the output what what is going 
on in the code?

```{r, eval = FALSE}
mod %>% mrgsim()
```

## Meropenem PK

- Load the `meropenem` model from the model directory
- Simulate the following scenarios:
    - 100 mg IV bolus q8h x 3
    - 100 mg IV over 3 hours q8h x3
    
Look at the `CP` output

```{r}

```


## Z-Pak

You've been sick for the last two weeks and can't take it any more. Finally, 
you decide to go to the doctor, who gives you a diagnosis of 
walking pneumonia. When you get home with your azithromycin 
prescription, you start wondering about the directions: take 
500 mg as a single dose on Day 1, followed by 250 mg once daily 
on Days 2 through 5.

Explore this regimen using the following model:

- Model name: `azithro`
- Model location: `model`

Simulate out to at least day 14 to see what is happening.

```{r, z_pak_example, eval = FALSE}
mod <- mread("", "") %>% zero_re

```

## Output

Run the following code and check the output

```{r, eval = FALSE}
mod <- mread_cache("azithro", "model")

out <- 
  mod %>% 
  ev(amt = 500) %>%
  mrgsim(end = 24, delta = 4)


out

class(out)

head(out)

out$CP

as.data.frame(out)

as_data_frame(out)

filter(out, time==12)

mutate(out, success = TRUE) %>% class

```

## Secukinumab for psoriasis

You are the pharmacometrician supporting secukinumab program for psoriasis, 
and the team wants an update.  You start telling them about 
this awesome indirect response model you developed for week-5 PASI scores. 
It has log-normal IIV and combined additive / prortional error model.  Standard
errors look good and the condition number was low.  The pred-corrected VPC 
and NPDE plots showed showed nothing to worry about. 

As you wrap up your report, you notice the blank stares coming back from the 
faces of the team members. 

The clin pharm lead breaks the uncomfortable slience: We're looking at 
a 100 mg dose in the 5 week induction period. We'd like to push the dose
higher, but also worried about candida infection.  If we increase to 125
mg or 150 mg, are we likely to see lower PASI scores?  How close are
we to reaching the maximum effect for end of induction?

- Use the `secukinumab` PK/PD model
- Dose once weekly for 5 weeks
- Generate a plot of PASI as fraction of baseline after 5 weekly
doses between 0 and 150 mg


```{r}
mod <- mread_cache("secukinumab", "model") %>% zero_re()

see(mod)

out <- mod %>% ev(amt = 25) %>% mrgsim(end = 35) 

plot(out)

tail(out)
```


## Statin/CsA DDI

- Load the model `yoshikado`

- Examine the the following drug-drug interaction
  - pitavastatin 30 mg x1
  - cyclosporine 2000 mg x1; wait 1 hr; then pitavastatin 30 mg x1
- What parameters is the pitavastatin profile sensitive to?


```{r}

```


## GCSF dosing

- Simulate 2.5, 5, and 10 mcg/kg assuming 50, 70, 90 kg individual
- Simulate 2.5, 5, and 10 mcg/kg assuming log(WT) ~ N(log(80),0.1)

- Do daily SC dosing x 7d

```{r}

```


## Sim EBE

```{r}
data <- read_table("data/patab", skip = 1)

```


# Answers

## Warm-up

```{r, warm_up_answer}
mod <- mread_cache("irm4", modlib())

param(mod)

init(mod)

see(mod)
```

```{r, eval = FALSE}
mod %>% mrgsim()
```

## Meropenem PK

```{r}
mod <- mread_cache("meropenem", "model")

mod %>% 
  ev(amt = 100, ii = 8, addl = 2) %>% 
  mrgsim() %>%
  plot
```

```{r}
mod <- mread_cache("meropenem", "model")

mod %>% 
  ev(amt = 100, ii = 8, rate = 100/3, addl = 2) %>% 
  mrgsim() %>%
  plot
```



## Z-Pak
```{r, z_pak_answer}
mod <- mread("azithro", "model") %>% zero_re

param(mod)

param(mod)$TVV1 + param(mod)$TVV2

```

Set up an dosing event
```{r}
load <- ev(amt = 500, ii = 24,  addl = 0)
maint <- ev(amt = 250, ii = 24, addl = 3)
dose <- seq(load, maint)

mod %>% 
  ev(dose) %>%
  mrgsim (end = 24*21) %>% 
  plot(CP + PER2 + PER3 ~time/24)
```


## Output


```{r}
mod <- mread_cache("azithro", "model")

out <- 
  mod %>% 
  ev(amt = 500) %>%
  mrgsim(end = 24, delta = 4)

out

class(out)

head(out)

out$CP

as.data.frame(out)

as_data_frame(out)

filter(out, time==12)

mutate(out, success = TRUE) %>% class

```

## Statin/CsA DDI

```{r}
mod <- mread_cache("yoshikado", "model", end = 14, delta = 0.1)

e1 <- ev(amt = 2000, cmt  = 2)
mod %>% ev(e) %>% mrgsim() %>% plot(CSA~.)

e2 <- ev(amt = 30, time  = 1)
mod %>% ev(e2) %>% mrgsim() %>% plot(CP~.)

out1 <- mod %>% ev(e2) %>% mrgsim_df(delta = 0.1) %>% mutate(ddi = FALSE)
out2 <- mod %>% ev(e1+e2) %>% mrgsim_df(delta = 0.1) %>% mutate(ddi = TRUE)

out <- bind_rows(out1, out2)

ggplot(out, aes(time, CP, col = ddi)) + geom_line(lwd = 1) + 
  scale_y_continuous(trans = "log10", limits = c(0.1, 100), 
                     breaks = 10^seq(-4,4)) + 
  scale_color_brewer(palette = "Set1")


out %>% group_by(ddi) %>% summarise(Cmax = max(CP))

#out %>% group_by(ddi) %>% summarise(auc = PKPDmisc::auc_partial(time, CP))
```


## Sim EBE

```{r, message = FALSE}

patab <- read_table("data/meropenem/patab", skip = 1)

patab <- distinct(patab, .keep_all = TRUE)

head(patab)

mod <- mread_cache("pk2", modlib())

inventory(mod, patab)

patab <- rename(patab, V3 = V2, V2 = V1)

inventory(mod,patab)

data <- read_csv("data/Simulated_DatasetMeropenem.csv", na = '.')

head(data)

data <- filter(data, EVID==1) %>% mutate(DUR = round(AMT/RATE,1), DOSE = AMT)

data <- mutate(data, II = 8, ADDL = 11, CMT = "CENT")

count(data,CMT,AMT,RATE,DUR)

dosing <- left_join(data, patab, by = "ID")

tg <- tgrid(0,24,0.25) + 3*24

out <- 
  mod %>% 
  data_set(dosing) %>%
  carry_out(DOSE,DUR) %>%
  mrgsim(tgrid = tg, obsonly = TRUE) %>% 
  as.data.frame

ggplot(out, aes(TIME,CP, group = ID)) + 
  facet_wrap(DUR~DOSE) + geom_line()
```