Effectiveness of Psychotherapies and Psychotropic Drugs against Mental Illness

	Effectiveness of Psychotherapies and Psychotropic Drugs against Mental Illness	Psychotherapies for Depression
	Contents

This book analyzes the results of numerous clinical studies conducted over the past forty years to evaluate the effectiveness of psychotherapies and psychotropic drugs against mental illness. This evaluation was conducted for each study by creating samples of people with mental illness undergoing psychotherapy or psychotropic drugs, and samples of people with mental illness not receiving psychotherapy or psychotropic drugs (control groups). The studies were conducted through interviews with self-administered or clinician-completed scales. The effect size was then statistically measured, often expressed as Hedges' g or Cohen's d. This measure indicates, for example, by how many standard deviations the psychotherapy group improved compared to the control group. The interpretation of g is:

`g ≈ 0.2` → small effect
`g ≈ 0.5` → medium
`g ≈ 0.8` → large

Another type of measurement is the Odds Ratio (OR), which is useful for interpreting the probability of clinical success, or better yet, of achieving a positive outcome (treatment response) compared to control groups. The interpretation of OR is:

OR = 1 → no difference
OR > 1 → treatment increases positive response compared to the control group
OR < 1 → treatment worsens the control

Example:

OR = 2 → probability of positive response doubles
OR = 1.5 → +50% probability

So if in a group of 10 people, 3 recover and 7 do not, the OR is 3/7 (0.42).

In addition to g and OR, the Risk Ratio (RR) is also measured, which compares the probability (the risk) of an event (such as recovery) occurring in the treated group compared to the control group.

RR = 1 → no difference
RR > 1 → treatment increases the probability of a positive response compared to the control group
RR < 1 → treatment is less likely than the control group

So if the treatment group has a 30% probability of remission and the control group a 10% probability, the RR is 3 (30/10), meaning the treatment group is three times more likely to be cured.

The calculations of g, OR, and RR are performed using meta-analysis, which is a statistical research procedure that integrates and synthesizes the results of multiple clinical trials to answer a specific research question. Rather than relying on the results of a single study, meta-analysis combines data from many studies to provide a more robust and precise estimate of a treatment's effectiveness. In short, meta-analysis allows us to move from "many small results" to "a solid conclusion," helping to establish, for example, whether psychotherapy is as effective as medication or whether it produces long-lasting benefits.

Finally, network meta-analysis is an advanced statistical methodology that allows for the simultaneous comparison of multiple treatments (interventions) within a single model, even when they have not been directly compared in clinical trials. While a traditional meta-analysis typically compares two options (e.g., a drug versus a placebo), NMA is used to compare and rank, for example, numerous psychotropic drugs and a placebo. One of the main advantages of NMA is that it allows for the creation of a ranking that indicates the probability of a drug being the best for a given outcome.

Implementation in R

The meta-analyses in this book are performed in the R programming language, using the RStudio software and downloading the "dplyr, meta, and metafor" packages and datasets provided by researchers, each containing a clinical study and its associated variables in each row.

Implementation in R

Table of Contents