rob: Risk of bias assessment

In guido-s/meta: General Package for Meta-Analysis

Risk of bias assessment

Description

Create table with risk of bias assessment or add table to existing meta-analysis

Usage

rob(
  item1,
  item2 = NULL,
  item3 = NULL,
  item4 = NULL,
  item5 = NULL,
  item6 = NULL,
  item7 = NULL,
  item8 = NULL,
  item9 = NULL,
  item10 = NULL,
  studlab = NULL,
  overall = NULL,
  weight = NULL,
  data = NULL,
  tool = gs("tool.rob"),
  domains = NULL,
  categories = NULL,
  cat1 = categories,
  cat2 = categories,
  cat3 = categories,
  cat4 = categories,
  cat5 = categories,
  cat6 = categories,
  cat7 = categories,
  cat8 = categories,
  cat9 = categories,
  cat10 = categories,
  cat.overall = categories,
  col = NULL,
  col1 = col,
  col2 = col,
  col3 = col,
  col4 = col,
  col5 = col,
  col6 = col,
  col7 = col,
  col8 = col,
  col9 = col,
  col10 = col,
  col.overall = col,
  symbols = NULL,
  symb1 = symbols,
  symb2 = symbols,
  symb3 = symbols,
  symb4 = symbols,
  symb5 = symbols,
  symb6 = symbols,
  symb7 = symbols,
  symb8 = symbols,
  symb9 = symbols,
  symb10 = symbols,
  symb.overall = symbols,
  legend = TRUE,
  overwrite = FALSE,
  warn = TRUE
)

## S3 method for class 'rob'
print(x, legend = attr(x, "legend"), details = TRUE, ...)

Arguments

item1	Risk of bias item 1 or a meta-analysis object of class meta with information on risk of bias assessment.
item2	Risk of bias item 2.
item3	Risk of bias item 3.
item4	Risk of bias item 4.
item5	Risk of bias item 5.
item6	Risk of bias item 6.
item7	Risk of bias item 7.
item8	Risk of bias item 8.
item9	Risk of bias item 9.
item10	Risk of bias item 10.
studlab	Study labels.
overall	Overall risk of bias assess.
weight	Weight for each study.
data	A data frame or a meta-analysis object of class meta.
tool	Risk of bias (RoB) tool.
domains	A character vector with names of RoB domains.
categories	Possible RoB categories.
cat1	Possible categories for RoB item 1.
cat2	Possible categories for RoB item 2.
cat3	Possible categories for RoB item 3.
cat4	Possible categories for RoB item 4.
cat5	Possible categories for RoB item 5.
cat6	Possible categories for RoB item 6.
cat7	Possible categories for RoB item 7.
cat8	Possible categories for RoB item 8.
cat9	Possible categories for RoB item 9.
cat10	Possible categories for RoB item 10.
cat.overall	Possible categories for overall RoB.
col	Colours for RoB categories.
col1	Colours for categories for RoB item 1.
col2	Colours for categories for RoB item 2.
col3	Colours for categories for RoB item 3.
col4	Colours for categories for RoB item 4.
col5	Colours for categories for RoB item 5.
col6	Colours for categories for RoB item 6.
col7	Colours for categories for RoB item 7.
col8	Colours for categories for RoB item 8.
col9	Colours for categories for RoB item 9.
col10	Colours for categories for RoB item 10.
col.overall	Colours for categories for overall RoB.
symbols	Corresponding symbols for RoB categories.
symb1	Corresponding symbols for RoB item 1.
symb2	Corresponding symbols for RoB item 2.
symb3	Corresponding symbols for RoB item 3.
symb4	Corresponding symbols for RoB item 4.
symb5	Corresponding symbols for RoB item 5.
symb6	Corresponding symbols for RoB item 6.
symb7	Corresponding symbols for RoB item 7.
symb8	Corresponding symbols for RoB item 8.
symb9	Corresponding symbols for RoB item 9.
symb10	Corresponding symbols for RoB item 10.
symb.overall	Corresponding symbols for overall RoB.
legend	A logical specifying whether legend with RoB domains should be printed.
overwrite	A logical indicating whether an existing risk of bias table in a meta-analysis object should be overwritten.
warn	A logical indicating whether warnings should be printed.
x	An object of class rob.
details	A logical indicating whether to print details on categories and colours.
...	Additional printing arguments.

Details

This function can be used to define a risk of bias (RoB) assessment for a meta-analysis which can be shown in a forest plot (forest.meta), summary weighted barplot (barplot.rob) or traffic light plot (traffic_light). It is also possible to extract the risk of bias assessment from a meta-analysis with RoB information.

The risk of bias table contains

• study labels;

• variables for individual RoB domains (with variable names A, B, ...);

• an overall RoB assessment if argument overall is provided;

• weights for individual studies used in summary weighted barplots.

Note, an overall RoB assessment is mandatory to create a summary weighted barplot or a traffic light plot.

The RoB table is directly returned if argument data is a data frame or argument item1 is a meta-analysis with risk of bias assessment. The RoB table is added as a new list element 'rob' to a meta-analysis object if argument data is a meta-analysis.

The user must either specify the categories and (optionally) domains of the RoB tool (using the eponymous arguments) or one of the following RoB tools.

Argument	Risk of bias tool
tool = "RoB1"	RoB 1 tool for randomized studies (Higgins et al., 2011)
tool = "RoB2"	RoB 2 tool for randomized studies (Sterne et al., 2019)
tool = "RoB2-cluster"	RoB 2 tool for cluster-randomized trials
tool = "RoB2-crossover"	RoB 2 tool for crossover trials
tool = "ROBINS-I"	Risk Of Bias In Non-randomized Studies - of Interventions
	(Sterne et al., 2016)
tool = "ROBINS-E"	Risk Of Bias In Non-randomized Studies - of Exposures
	(ROBINS-E Development Group, 2023)

These RoB tools are described on the website https://www.riskofbias.info/.

Risk of bias domains

By default, i.e., if argument domains is not provided by the user, the following names are used for RoB domains.

• RoB 1 tool for randomized studies (RoB1):

  1. Random sequence generation (selection bias)

  2. Allocation concealment (selection bias)

  3. Blinding of participants and personnel (performance bias)

  4. Blinding of outcome assessment (detection bias)

  5. Incomplete outcome data (attrition bias)

  6. Selective reporting (reporting bias)

  7. Other bias

• RoB 2 tool for randomized studies (RoB2):

  1. Bias arising from the randomization process

  2. Bias due to deviations from intended intervention"

  3. Bias due to missing outcome data

  4. Bias in measurement of the outcome

  5. Bias in selection of the reported result

• RoB 2 tool for cluster-randomized trials (RoB2-cluster):

  1. Bias arising from the randomization process

  2. Bias arising from the identification or recruitment of participants into clusters

  3. Bias due to deviations from intended intervention

  4. Bias due to missing outcome data

  5. Bias in measurement of the outcome

  6. Bias in selection of the reported result

• RoB 2 tool for crossover trials (RoB2-crossover)

  1. Bias arising from the randomization process

  2. Bias arising from period and carryover effects

  3. Bias due to deviations from intended intervention

  4. Bias due to missing outcome data

  5. Bias in measurement of the outcome

  6. Bias in selection of the reported result

• Risk Of Bias In Non-randomized Studies - of Intervention (ROBINS-I):

  1. Risk of bias due to confounding

  2. Risk of bias in selection of participants into the study

  3. Risk of bias in classification of interventions

  4. Risk of bias due to deviations from intented interventions

  5. Risk of bias due to missing outcome data

  6. Risk of bias in measurement of the outcome

  7. Risk of bias in the selection of the reported results

• Risk Of Bias In Non-randomized Studies - of Exposures (ROBINS-E):

  1. Risk of bias due to confounding

  2. Risk of bias arising from measurement of the exposure into the study (or into the analysis)

  3. Risk of bias due to post-exposure interventions

  4. Risk of bias due to deviations from intented interventions

  5. Risk of bias due to missing outcome data

  6. Risk of bias in measurement of the outcome

  7. Risk of bias in the selection of the reported results

• User-defined RoB assessment:

  1. First item

  2. Second item

  3. ...

It is possible to define additional bias domains for the available RoB tools. In this case, only the names for new RoB domains have to be provided in argument domains. If argument domains is not used to specify new domains, the names "Additional item 1" etc. will be used. It is also possible to modify the pre-defined domain names using argument domains.

The maximum number of bias domains / items is ten (see arguments item1, ..., item10).

Risk of bias categories, colours and symbols

By default, the following settings are used.

RoB 1 tool:

Argument	Values
categories	"Low risk of bias", "Unclear risk of bias", "High risk of bias"
col	"green", "yellow", "red"
symbols	"+", "?", "-"

RoB 2 tools:

Argument	Values
categories	"Low risk of bias", "Some concerns", "High risk of bias"
col	"green", "yellow", "red"
symbols	"+", "?", "-"

ROBINS tools:

Argument	Values
categories	"Low risk", "Some concerns", "High risk", "Very high risk", "NI"
col	"green", "yellow", "red", "darkred", "darkgrey"
symbols	none

User-defined RoB tools:

Argument	Values
categories	Must be specified by the user
col	1, 2, ...
symbols	none

If colours (col) and symbols (symbols) are provided, they must be of the same length as the number of categories.

Value

A data frame with study labels and risk of bias items and additional class "rob".

Author(s)

Guido Schwarzer [e-mail]

References

Higgins JPT, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD et al. (2011): The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. British Medical Journal, 343: d5928

ROBINS-E Development Group (Higgins J, Morgan R, Rooney A et al.) (2023): Risk Of Bias In Non-randomized Studies - of Exposure (ROBINS-E) Available from: https://www.riskofbias.info/welcome/robins-e-tool.

Sterne JA, Hernán MA, Reeves BC, Savović J, Berkman ND, Viswanathan M, et al. (2016): ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. British Medical Journal, 355: i4919

Sterne JAC, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al. (2019): RoB 2: a revised tool for assessing risk of bias in randomised trials. British Medical Journal, 366: l4898.

See Also

forest.meta, barplot.rob, traffic_light

Examples

# Use RevMan 5 settings
oldset <- settings.meta("RevMan5", quietly = FALSE)

data(caffeine)

m1 <- metabin(h.caf, n.caf, h.decaf, n.decaf, sm = "OR",
  data = caffeine, studlab = paste(study, year))

# Add risk of bias assessment to meta-analysis
m2 <- rob(D1, D2, D3, D4, D5, overall = rob, data = m1, tool = "rob2")

# Print risk of bias assessment
rob(m2)

# Forest plot with risk of bias assessment
forest(m2)

# Use previous settings
settings.meta(oldset)

guido-s/meta documentation built on April 18, 2024, 7:11 p.m.