chapter_10.qmd

---
title: "Systematic review and meta-analysis of Real-World Evidence"
authors:   
  - name: Dimitris Mavridis
    affiliations:
      - ref: ioannina
  - name: Thomas Debray
    orcid: 0000-0002-1790-2719
    affiliations:
      - ref: smartdas
affiliations:
  - id: smartdas
    name: Smart Data Analysis and Statistics B.V.
    city: Utrecht
  - id: ioannina
    name: University of Ioannina
    city: Ioannina
format:
  html:
    toc: true
    number-sections: true
execute:
  cache: true
bibliography: 'https://api.citedrive.com/bib/0d25b38b-db8f-43c4-b934-f4e2f3bd655a/references.bib?x=eyJpZCI6ICIwZDI1YjM4Yi1kYjhmLTQzYzQtYjkzNC1mNGUyZjNiZDY1NWEiLCAidXNlciI6ICIyNTA2IiwgInNpZ25hdHVyZSI6ICI0MGFkYjZhMzYyYWE5Y2U0MjQ2NWE2ZTQzNjlhMWY3NTk5MzhhNzUxZDNjYWIxNDlmYjM4NDgwOTYzMzY5YzFlIn0=/bibliography.bib'
---

```{r}
#| include: false
#| echo: false
#| message: false
#| warning: false

# List of required packages
required_packages <- c("kableExtra", "dplyr", "gemtc", "netmeta", "remotes")

# Install required packages
for (pkg in required_packages) {
    if (!requireNamespace(pkg, quietly = TRUE)) {
        install.packages(pkg)
    }
}
library(kableExtra)

formatCI <- function(mean, lower, upper, ndigits = 2) {
  
  str <- data.frame(mean, lower, upper) %>%
    mutate(ci = ifelse(!is.na(mean),
                       paste0(sprintf("%.2f", mean), " (", sprintf("%.2f", lower), "; ", sprintf("%.2f", upper), ")"),"")) %>% pull(ci)
  
  str
}
```


## Introduction

We first load the required packages

```{r}
#| warning: false
library(dplyr)
library(gemtc)
library(netmeta)
```

## Pairwise meta-analysis of clinical trials

### Toculizumab for coronavirus disease 2019
In this example, we consider the results from a systematic literature review of clinical trials investigating any pharmacological in hosptialized patients with coronavirus disease 2019 [@selvarajan_efficacy_2022]. A total of 23 randomized controlled trials were included and studied seven different interventions: dexamethasone, remdesivir, tocilizumab, hydroxychloroquine, combination of lopinavir/ritonavir, favipiravir and interferon-β. We here focus on the synthesis of 7 trials that comparted toculizumab (`TOCI`) to standard care (`STD`)
and collected mortality data. 

```{r}
#| echo: false
#| message: false
#| warning: false
studlab <- c(rep("Tomazini et al",2), 
             rep("Horby et al",2), 
             rep("Biegel et al",2), 
             rep("Wang et al",2),
             rep("Spinner et al",2),rep("Solidarity trial - Remdesivir",2),
             rep("Hermine et al",2),rep("Salvarini et al",2),
             rep("Stone et al",2), rep("Rosas et al",2), rep("Salama et al",2),
             rep("Veiga et al",2), rep("Ulrich et al",2),
             rep("Lyngbakken et al",2), rep("Solidarity trial - HCQ",2), 
             rep("Abd-Elsalam et al",2),rep("Cavalcanti et al",2),
             rep("Recovery group",2), rep("Solidarity trial-LR",2), 
             rep("Cao et al",2), rep("Davoudi-Monfared",2), 
             rep("Solidarity trial - Interferon",2),rep("Rohmani",2))

treat <- c("DEXA", "STD", "DEXA", "STD", "REM", "STD", "REM", "STD", "REM", 
           "STD", "REM", "STD", "TOCI", "STD", "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "TOCI", "STD", "TOCI", "STD", "HCQ", "STD", "HCQ", 
           "STD", "HCQ", "STD", "HCQ", "STD", "HCQ", "STD", "LOP+RIT",  
           "STD", "LOP+RIT", "STD", "LOP+RIT", "STD", "INT-B", "STD", "INT-B",
           "STD", "INT-B", "STD")    

event <- c(85,91,482,1110,59,77,22,10,5,4,602,643,7,8,2,1,9,3,58,28,26,11,14,6,13,6,1,1,104,84,6,5,1,1,374,767,148,146,19,25,8,15,243,216,2,6)
n <- c(151,148,2104,4321,541,521,158,78,396,200,4146,4129,63,67,60,66,161,82,294,144,249,128,65,64,67,61,27,26,947,906,97,97,221,227,1616,3424,1399,1372,99,100,42,39,2050,2050,33,33)

covid <- pairwise(treat = treat, event = event, n = n, studlab = studlab, sm = "OR")

tocilizumab <- covid %>% filter(treat1 == "TOCI")
remdesvir <- covid %>% filter(treat1 == "REM")
```


```{r}
#| echo: false
#| message: false
#| warning: false


tocilizumab %>% 
  dplyr::select(studlab, treat1, treat2, event1, n1, event2, n2) %>% 
  kable(
      align = "lrrrrrr",
      booktabs = TRUE,
      longtable = TRUE,
      linesep = ""
    ) %>%
    kableExtra::kable_styling(
      latex_options = c("repeat_header"),
    )
```

We now conduct a pairwise meta-analysis to assess the pooled effect of tocilizumab versus standard care. For each study, the log odds ratio and corresponding standard error is derived after which the corresponding estimates are pooled using the  Mantel-Haenszel method.

```{r}
#| fig-width: 6
#| fig-height: 5
results.TOCI <- metabin(event1,n1,event2,n2,studlab,data=tocilizumab,
                        sm = "OR", main = "tocilizumab vs standard care", 
                        prediction = TRUE)
forest(results.TOCI, leftcols = "studlab", rightcols = "effect.ci")
```

Altough a random effects meta-analysis was conducted, no heterogeneity was found ($\tau$=`r results.TOCI$tau`, with a 95\% confidence interval ranging from `r  results.TOCI$lower.tau` to `r round(results.TOCI$upper.tau,2)`).


### Remdesivir for coronavirus disease 2019
In aforementioned example, a total of 4 trials compared remdesivir to standard care:

```{r}
#| echo: false
#| fig-width: 6
results.REM <- metabin(event1,n1,event2,n2,studlab,data=remdesvir,
                        sm = "OR", main = "remdesvir vs standard care", 
                        prediction = TRUE)
forest(results.REM, leftcols = "studlab", rightcols = "effect.ci")
```


## Network meta-analysis of clinical trials

We here use the R packages `netmeta` for conducting a frequentist network meta-analysis. A detailed tutorial on the use of  `netmeta` is available from the book [Doing Meta-Analysis with R: A Hands-On Guide](https://bookdown.org/MathiasHarrer/Doing_Meta_Analysis_in_R/).



### Interventions for coronavirus disease 2019
We here consider data from a study which aimed to assess the comparative effectiveness of remdesivir and tocilizumab for reducing mortality in hospitalised COVID-19 patients. 80 trials were identified from two published network meta-analyses [@selvarajan_efficacy_2022], [@Siemieniuk_2020], a living COVID-19 trial database (COVID-NMA Initiative) [Covid-NMA.com], and a clinical trial database [clinicaltrials.gov]. Trials were included in this study if the patient population included hospitalized COVID-19 patients, active treatment was remdesivir or tocilizumab, comparator treatment was placebo or standard care, short-term mortality data was available, and the trial was published. 21 trials were included. For included trials, a risk of bias score was extracted from the COVID-NMA Initiative.

```{r}
#| echo: false
#| message: false
#| warning: false

studlab <- c("Ader", "Ader", "Beigel (ACTT-1)", "Beigel (ACTT-1)", "Criner", "Criner",
             "Mahajan", "Mahajan", "Pan (WHO Solidarity)", "Pan (WHO Solidarity)",
             "Spinner", "Spinner", "Wang", "Wang", "Hermine (CORIMUNO)", 
             "Hermine (CORIMUNO)", "Horby (RECOVERY)", "Horby (RECOVERY)", 
             "Rosas (COVACTA)", "Rosas (COVACTA)", "Salama (EMPACTA)", 
             "Salama (EMPACTA)", "Salvarani", "Salvarani", "Soin (COVINTOC)",
             "Soin (COVINTOC)", "Stone (BACC-BAY)", "Stone (BACC-BAY)", "Talaschian",
             "Talaschian", "Veiga (TOCIBRAS)", "Veiga (TOCIBRAS)", "Declerq (COV-AID)", 
             "Declerq (COV-AID)", "Rutgers", "Rutgers", "Broman", "Broman", "Islam",
             "Islam", "Gordon (REMAP-CAP)", "Gordon (REMAP-CAP)")

treat <- c("REM", "STD", "REM", "STD",
           "REM", "STD", "REM", "STD",
           "REM", "STD", "REM", "STD", 
           "REM", "STD", "TOCI", "STD", 
           "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "REM", "STD",
           "TOCI", "STD")    

event <- c(34, 37, 59, 77, 4, 4, 5, 3, 602, 643, 5, 4, 22, 10, 7, 8, 621, 729, 58,
           28, 26, 11, 2, 1, 11, 15, 9, 4, 5, 4, 14, 6, 10, 9, 21, 34, 1, 0, 0, 0, 
           83, 116)

n <- c(414, 418, 541, 521, 384, 200, 34, 36, 4146, 4129, 384, 200, 158, 78, 63, 
       67, 2022, 2094, 294, 144, 249, 128, 60, 63, 92, 88, 161, 82, 17, 19, 65,
       64, 81, 74, 174, 180, 57, 29, 30, 30, 353, 358)

covid <- pairwise(treat = treat, event = event, n = n, studlab = studlab, sm = "OR")

covid %>% dplyr::select(studlab, treat1, treat2, event1, n1, event2, n2) %>% 
  kable(
      align = "lrrrrrr",
      booktabs = TRUE,
      longtable = TRUE,
      linesep = ""
    ) %>%
    kableExtra::kable_styling(
      latex_options = c("repeat_header"),
    )
```

The corresponding network is displayed below:

```{r}
#| echo: false
#| message: false
#| warning: false
#| fig-width: 10
#| fig-cap: Evidence network of the 21 coronavirus-19 trials

NMA.covid <- netmeta(TE = TE, seTE = seTE, treat1 = treat1, treat2 = treat2,
                     studlab = studlab, data = covid, sm = "OR", ref = "STD",
                     comb.random = TRUE, warn = FALSE)

netgraph(NMA.covid)
```

We use the following command to calculate the log odds ratios and corresponding standard errors for each study:

```{r}
#| eval: false
covid <- pairwise(treat = treat, 
                  event = event, 
                  n = n, 
                  studlab = studlab, 
                  sm = "OR")
head(covid)
```

```{r}
#| echo: false
kable(head(covid))
```

Below, we conduct a random effects network meta-analysis where we consider standard care (`STD`) as the control treatment. Note that we have one study where zero cell counts occur, this study will not contribute to the NMA as the log odds ratio and its standard error cannot be determined.

```{r}
NMA.covid <- netmeta(TE = TE, seTE = seTE, treat1 = treat1, treat2 = treat2,
                     studlab = studlab, data = covid, sm = "OR", ref = "STD",
                     comb.random = TRUE, common = FALSE, warn = FALSE)
NMA.covid 
```

A league table of the treatment effect estimates is given below:

```{r}
netleague(NMA.covid)
```

We can also present the results in a forest plot:
  
```{r}
#| echo: false
netmeta:::forest.netmeta(NMA.covid, leftcols = "studlab", rightcols = "effect.ci")
```

We now consider a Bayesian random effects network meta-analysis that analyzes the observed event counts
using a binomial link function. 

```{r}
bdata <- data.frame(study = studlab,
                    treatment = treat,
                    responders = event,
                    sampleSize = n)

network <- mtc.network(data.ab  = bdata)

model <- mtc.model(network,
                   likelihood = "binom",
                   link = "log",
                   linearModel = "random",
                   n.chain = 3)
```

```{r}
#| message: false
# Adaptation
mcmc1 <- mtc.run(model, n.adapt = 1000, n.iter = 1000, thin = 10)

# Sampling
mcmc2 <- mtc.run(model, n.adapt = 10000, n.iter = 100000, thin = 10)
```

We can extract the pooled treatment effect estimates from the posterior distribution. When using `STD` as control group, we have:

```{r}
summary(relative.effect(mcmc2, t1 = "STD"))
```

The corresponding odds ratios are as follows:

```{r}
#| echo: false
bresults.STD <- summary(relative.effect(mcmc2, t1 = "STD"))
bOR <- round(exp(bresults.STD$summaries[[2]][c("d.STD.REM", "d.STD.TOCI"), c("2.5%", "50%", "97.5%")]),2)

out <- data.frame("Comparison" = character(), "Risk Ratio (95% CrI)" = character())
out <- out %>% add_row(data.frame("Comparison" = ifelse(rownames(bOR)[1] == "d.STD.REM", "REM vs. STD", "TOCI vs. STD"), "Risk Ratio (95% CrI)" = paste0(bOR[1,"50%"], " (", bOR[1, "2.5%"], "; ", bOR[1, "97.5%"], ")")))
out <- out %>% add_row(data.frame("Comparison" = ifelse(rownames(bOR)[2] == "d.STD.REM", "REM vs. STD", "TOCI vs. STD"), "Risk Ratio (95% CrI)" = paste0(bOR[2,"50%"], " (", bOR[2, "2.5%"], "; ", bOR[2, "97.5%"], ")")))

bresults.TOCI <- summary(relative.effect(mcmc2, t1 = "TOCI"))
bOR <- round(exp(bresults.TOCI$summaries[[2]]["d.TOCI.REM", c("2.5%", "50%", "97.5%")]),2)
out <- out %>% add_row(data.frame("Comparison" = "REM vs. TOCI", "Risk Ratio (95% CrI)" = paste0(bOR["50%"], " (", bOR["2.5%"], "; ", bOR["97.5%"], ")")))

out %>% knitr::kable(
      align = "lc",
      booktabs = TRUE,
      longtable = TRUE,
      linesep = "",
      col.names = c("Comparison", "95% CrI")
    )
```

Finally, we expand the COVID-19 network with trials investigating the effectiveness of hydroxychloroquine (HCQ), lopinavir/ritonavir (LOPI), dexamethasone (DEXA) or interferon-$\beta$ (INTB) [@selvarajan_efficacy_2022]. The corresponding network is displayed below:

```{r}
#| echo: false
#| message: false
#| warning: false
#| fig-width: 10
#| fig-cap: Evidence network of the 33 coronavirus-19 trials

studlab <- c("Abd-Elsalam et al", "Abd-Elsalam et al",
             "Ader", "Ader", 
             "Beigel (ACTT-1)", "Beigel (ACTT-1)", 
             "Cao et al", "Cao et al", 
             "Cavalcanti et al", "Cavalcanti et al",
             "Davoudi-Monfared", "Davoudi-Monfared",
             "Horby (RECOVERY; DEXA)", "Horby (RECOVERY; DEXA)",
             "Lyngbakken et al", "Lyngbakken et al", 
             "Horby (RECOVERY; LOPI)", "Horby (RECOVERY; LOPI)", 
             "Rohmani", "Rohmani", 
             "Pan (WHO Solidarity; LOPI)", "Pan (WHO Solidarity; LOPI)", 
             "Pan (WHO Solidarity; HCQ)", "Pan (WHO Solidarity; HCQ)", 
             "Pan (WHO Solidarity; INTB)", "Pan (WHO Solidarity; INTB)", 
             "Tomazini et al", "Tomazini et al",
             "Ulrich et al", "Ulrich et al",
             "Criner", "Criner",
             "Mahajan", "Mahajan", 
             "Pan (WHO Solidarity; REM)", "Pan (WHO Solidarity; REM)",
             "Spinner", "Spinner", 
             "Wang", "Wang", 
             "Hermine (CORIMUNO)", "Hermine (CORIMUNO)", 
             "Horby (RECOVERY; TOCI)", "Horby (RECOVERY; TOCI)", 
             "Rosas (COVACTA)", "Rosas (COVACTA)", 
             "Salama (EMPACTA)", "Salama (EMPACTA)", 
             "Salvarani", "Salvarani", 
             "Soin (COVINTOC)","Soin (COVINTOC)", 
             "Stone (BACC-BAY)", "Stone (BACC-BAY)", 
             "Talaschian","Talaschian", 
             "Veiga (TOCIBRAS)", "Veiga (TOCIBRAS)", 
             "Declerq (COV-AID)", "Declerq (COV-AID)", "Rutgers", "Rutgers", "Broman", "Broman", "Islam",
             "Islam", "Gordon (REMAP-CAP)", "Gordon (REMAP-CAP)")

treat <- c("HCQ", "STD",  # Abd-Elsalam
           "REM", "STD", # Ader
           "REM", "STD", # Beigel
           "LOPI", "STD", # Cao et al.
           "HCQ", "STD", # Cavalcanti et al
           "INTB", "STD", # Davoudi-Monfared
           "DEXA", "STD", # Horby et al
           "HCQ", "STD", # Lyngbakken et al
           "LOPI", "STD", # Recovery group
           "INTB", "STD", # Rohmani
           "LOPI", "STD", #Pan (WHO Solidarity; LOPI)
           "HCQ", "STD", # Pan (WHO Solidarity; HCQ)
           "INTB", "STD", # Pan (WHO Solidarity; Interferon)
           "DEXA", "STD", # Tomazini et al
           "HCQ", "STD", # Ulrich et al
           "REM", "STD", # Criner
           "REM", "STD", # Mahajan
           "REM", "STD", # Pan (WHO Solidarity; REM)
           "REM", "STD",  # Spinner
           "REM", "STD", # Wang
           "TOCI", "STD", # Hermine (CORIMUNO)
           "TOCI", "STD", # Horby (RECOVERY; TOCI)
           "TOCI", "STD",
           "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "TOCI", "STD",
           "TOCI", "STD", "REM", "STD",
           "TOCI", "STD")    

event <- c(6, 5, # Abd-Elsalam
           34, 37, # Ader
           59, 77, # Beigel
           19, 25, # Cao
           1, 1, # Cavalcanti et al
           8, 15, # Davoudi-Monfared
           482, 1110, # Horby et al
           1, 1, # Lyngbakken et al
           374, 767, # Recovery group
           2, 6, # Rohmani
           151, 153, # Pan (WHO Solidarity; LOPI)
           104, 89, # Pan (WHO Solidarity; HCQ)
           316, 266, # Pan (WHO Solidarity; Interferon)
           85, 91, # Tomazini et al
           13, 6, # Ulrich et al
           4, 4, # Criner
           5, 3, # Mahajan
           602, 643, # Pan (WHO Solidarity; REM)
           5, 4, # Spinner
           22, 10, # Wang
           7, 8, # Hermine (CORIMUNO)
           621, 729, # Horby (RECOVERY; TOCI)
           58,
           28, 26, 11, 2, 1, 11, 15, 9, 4, 5, 4, 14, 6, 10, 9, 21, 34, 1, 0, 0, 0, 
           83, 116)

n <- c(97, 97, # Abd-Elsalam
       414, 418, # Ader
       541, 521, # Beigel
       99, 100,  # Cao
       221, 227, # Cavalcanti et al
       42, 39, # Davoudi-Monfared
       2104, 4321, # Horby et al
       27, 26, # Lyngbakken et al
       1616, 3424, # Recovery group
       33, 33, # Rohmani
       1404, 1368, # Pan (WHO Solidarity; LOPI)
       948, 900, # Pan (WHO Solidarity; HCQ)
       2144, 2147, # Pan (WHO Solidarity; Interferon)
       151, 148, # Tomazini et al
       67, 61, # Ulrich et al
       384, 200, # Criner
       34, 36, # Mahajan
       4146, 4129, # Pan (WHO Solidarity; REM)
       384, 200, # Spinner
       158, 78, # Wang
       63, 67, 2022, 2094, 294, 144, 249, 128, 60, 63, 92, 88, 161, 82, 17, 19, 65,
       64, 81, 74, 174, 180, 57, 29, 30, 30, 353, 358)

covidfull <- pairwise(treat = treat, event = event, n = n, studlab = studlab, sm = "OR")

NMA.covidfull <- netmeta(TE = TE, seTE = seTE, treat1 = treat1, treat2 = treat2,
                     studlab = studlab, data = covidfull, sm = "OR", ref = "STD",
                     comb.random = TRUE, warn = FALSE)

netgraph(NMA.covidfull)
```


We conducted a random effects network meta-analysis, results are depicted below:

```{r}
#| echo: false
NMA.covidf <- netmeta(TE = TE, seTE = seTE, treat1 = treat1, treat2 = treat2,
                      studlab = studlab, data = covidfull, sm = "OR", ref = "STD",
                      prediction = TRUE,
                      comb.random = TRUE, common = FALSE, warn = FALSE)
NMA.covidf
```

We can calculate the P score for each treatment as follows:

```{r}
netrank(NMA.covidf)
```


### Pharmacologic treatments for chronic obstructive pulmonary disease 
In this example, we consider the resuls from a systematic review of randomized controlled trials on pharmacologic treatments for chronic obstructive pulmonary disease [@Baker_2009]. The primary outcome, occurrence of one or more episodes of COPD exacerbation, is binary (yes / no). For this outcome, five drug treatments (fluticasone, budesonide, salmeterol, formoterol, tiotropium) and two combinations (fluticasone + salmeterol, budesonide + formoterol) were compared to placebo. The authors considered the two combinations as separate treatments instead of evaluating the individual components.

```{r}
data(Baker2009)
```

```{r}
#| echo: false
kable(head(Baker2009))
```

```{r}
Baker <- pairwise(treat = treatment,
                  event = exac,
                  n = total,
                  studlab = id,
                  sm = "OR",
                  data = Baker2009)

NMA.COPD <- netmeta(TE = TE, seTE = seTE, treat1 = treat1, treat2 = treat2,
                    studlab = studlab, data = Baker, sm = "OR", ref = "Placebo",
                    comb.random = TRUE)

netgraph(NMA.COPD)
```
We can test for local network inconsistency as follows:

```{r}
#| eval: false
netsplit(NMA.COPD)
```

By default, indirect estimates are derived using the back-calculation method. A comparison of direct and indirect treatment effect estimates (depicted as odds ratio) from the random effects network meta-analysis model is given in the table below:

```{r}
#| echo: false
SIDES.COPD <- netsplit(NMA.COPD)
out <- data.frame("comparison" = SIDES.COPD$comparison,
                  "k" = SIDES.COPD$k,
             "nma" = sprintf("%.2f",exp(SIDES.COPD$random$TE)), 
             "direct" = sprintf("%.2f",exp(SIDES.COPD$direct.random$TE)), 
             "indirect" = sprintf("%.2f",exp(SIDES.COPD$indirect.random$TE)),  
             "RoR" = sprintf("%.2f",exp(SIDES.COPD$compare.random$TE)),
             "p.RoR" = sprintf("%.2f",SIDES.COPD$compare.random$p))

# Replacing all NA values with "."
out <- out %>% 
  mutate(across(everything(), ~ifelse(.=="NA", ".", .)))

out %>% 
  kable(
      align = "lcccccc",
      booktabs = TRUE,
      longtable = TRUE,
      linesep = ""
    ) %>%
    kableExtra::kable_styling(
      latex_options = c("repeat_header"),
    )
```



### Advanced Therapies for Ulcerative Colitis
In this example, we consider a systematic literature review of Phase 3 
randomized controlled trials investigating the following advanced therapies: 
infliximab, adalimumab, vedolizumab, golimumab, tofacitinib, ustekinumab, 
filgotinib, ozanimod, and upadacitinib [@panaccione_efficacy_2023]. This review 
included 48 RCTs, from which 23 were found eligible for inclusion in a 
network meta-analysis. The included RCT populations were largely comparable in their baseline
characteristics, though some heterogeneity was noted in weight, 
disease duration, extent of disease, and concomitant medications. 
A risk of bias assessment showed a low risk of bias for all included RCTs, 
which were all industry sponsored.

We here focus on the synthesis of 18 trials that contributed efficacy data for 
induction in bio-naive populations. The following FDA- and/or EMA-approved 
biologic or SMD doses were investigated: 

* Adalimumab subcutaneous 160 mg at week 0, 80 mg at week 2, and 40 mg at week 4 (`ADA160/80`)
* Infliximab intravenous 5 mg/kg (`INF5`) at weeks 0, 2, and 6 then every 8 weeks
* Infliximab intravenous 10 mg/kg (`INF10`) at weeks 0, 2, and 6 then every 8 weeks
* Filgotinib oral 100 mg once daily (`FIL100`)
* Filgotinib oral 200 mg once daily (`FIL200`)
* Golimumab subcutaneous 200 mg at week 0 and 100 mg at week 2 (`GOL200/100`)
* Ozanimod oral 0.23 mg once daily for 4 days, 0.46 mg once daily for 3 days, then 0.92 mg once daily (`OZA0.92`)
* Tofacitinib oral 10 mg twice daily for 8 weeks (`TOF10`)
* Upadacitinib oral 45 mg once daily for 8 weeks (`UPA45`)
* Ustekinumab intravenous 6 mg/kg at week 0 (`UST6`)
* Vedolizumab intravenous 300 mg at weeks 0, 2, and 6 (`VED300`)

The reference treatment is placebo (`PBO`).


```{r}
#| echo: false
#| message: false
#| warning: false
#| tbl-cap: Efficacy outcomes (i.e., clinical remission) data of induction bio-naïve populations
# Efficacy outcomes data of induction bio-naïve populations
studlab <- c("ACT-1", "ACT-1", "ACT-1", "ACT-2", "ACT-2", "ACT-2", "GEMINI 1",
             "GEMINI 1", "Japic CTI-060298", "Japic CTI-060298", "Jiang 2015", 
             "Jiang 2015", "M10-447", "M10-447", "NCT01551290", "NCT01551290",
             "NCT02039505", "NCT02039505", "OCTAVE 1", "OCTAVE 1", "OCTAVE 2",
             "OCTAVE 2", "PURSUIT-SC", "PURSUIT-SC", "SELECTION", "SELECTION",
             "SELECTION", "TRUE NORTH", "TRUE NORTH", "U-ACCOMPLISH", 
             "U-ACCOMPLISH", "U-ACHIEVE Study 2", "U-ACHIEVE Study 2", 
             "ULTRA-1", "ULTRA-1", "ULTRA-2", "ULTRA-2", "UNIFI", "UNIFI")

treat <- c("INF10", "INF5", "PBO", "INF10", "INF5", "PBO", "VED300", "PBO", 
           "INF5", "PBO", "INF5", "PBO", "ADA160/80", "PBO", "INF5", "PBO", 
           "VED300", "PBO", "TOF10", "PBO", "TOF10", "PBO", "GOL200/100", 
           "PBO", "FIL100", "FIL200", "PBO", "OZA0.92", "PBO", "UPA45", "PBO",
           "UPA45", "PBO", "ADA160/80", "PBO", "ADA160/80", "PBO", "UST6",
           "PBO")


event <- c(39, 47, 18, 33, 41, 7, 30, 5, 21, 11, 22, 9, 9, 11, 11, 5, 22,
           6, 56, 9, 43, 4, 45, 16, 47, 60, 17, 66, 10, 54, 3, 41, 4, 24,
           12, 32, 16, 27, 15)

n <- c(122, 121, 121, 120, 121, 123, 130, 76, 104, 104, 41, 41, 90, 96, 50,
       49, 79, 41, 222, 57, 195, 47, 253, 251, 277, 245, 137, 299, 151, 166,
       81, 145, 72, 130, 130, 150, 145, 147, 151)

UlcerativeColitis <- pairwise(treat=treat,event=event,n=n,studlab=studlab, sm="OR")

UlcerativeColitis %>% dplyr::select(studlab, treat1, treat2, event1, n1, event2, n2) %>% 
  kable(
      align = "lrrrrrr",
      booktabs = TRUE,
      longtable = TRUE,
      linesep = ""
    ) %>%
    kableExtra::kable_styling(
      latex_options = c("repeat_header"),
    )
```

The corresponding network is displayed below:

```{r}
#| echo: false
#| message: false
#| warning: false
#| fig-width: 10
#| fig-cap: Evidence network of 18 trials that contributed efficacy data for induction in bio-naive populations

NMA.uc <- netmeta(TE = TE, seTE = seTE, treat1 = treat1, treat2 = treat2,
                  studlab = studlab, data = UlcerativeColitis, sm = "OR", 
                  ref = "PBO", comb.random = TRUE)

netgraph(NMA.uc)
```

Below, we conduct a random effects network meta-analysis of the reported study 
effects (expressed as odds ratio) and consider placebo (`treat = "PBO"`) as 
the control treatment.

```{r}
#| eval: false
NMA.uc <- netmeta(TE = TE, seTE = seTE, treat1 = treat1, treat2 = treat2,
                  studlab = studlab, data = UlcerativeColitis, sm = "OR", 
                  ref = "PBO", common = FALSE, comb.random = TRUE)
NMA.uc
```

All treatments except `FIL100` and `UST6` are significantly more efficacious than `PBO` at inducing clinical remission. We can now estimate the probabilities of each treatment being at each possible rank and the SUCRAs (Surface Under the Cumulative RAnking curve):

```{r}
sucra.uc <- rankogram(NMA.uc, nsim = 100, random = TRUE, common = FALSE, 
                      small.values = "undesirable")

# Exctract the SUCRA values
sucra.uc$ranking.random
```

These results indicate that `r round( sucra.uc$ranking.random["UPA45"]*100,1)`% of the evaluated treatments are worse than `UPA45`.





## Version info {.unnumbered}
This chapter was rendered using the following version of R and its packages:

```{r}
#| echo: false
#| message: false
#| warning: false
sessionInfo()
```

## References {.unnumbered}