Thread by @MushtaqBilalPhD

How to conduct a systematic review: A thread written with the help of ChatGPT

Conducting a Systematic Review is an important skill in biomedical research.

I am a literary scholar and don't know much about systematic reviews.

So, I read up on on it. Below is what I learned.

This thread is a summary of an article titled "Systematic Reviews and Meta-Analyses" by Lindsay S. Uman published in the Journal of Canadian Academy of Child Adolescent Psychiatry.

PMID: 21286370

Why are systematic reviews needed?

Since a lot of studies are published in health and biomedical sciences, it's difficult for researchers to keep up with recent developments.

A systematic review summarizes the outcomes of various studies conducted on a given topic.

Systematic reviews help researchers find out easily what works and what doesn't.

Systematic reviews require you to identify, appraise, and analyze all relevant studies to minimize bias.

A systematic review is different from a narrative review.

A narrative review is mainly descriptive with little analysis whereas a systematic review is meant to be analytic.

Systematic reviews also include a meta-analysis component.

This involves using statistical techniques to synthesize data from several studies into a single quantitative estimate or summary effect size.

Uman divides the process of a systematic review into 8 stages.

1. Formulate the review question

Start by defining the review question and forming hypotheses.

Use the formula: "Intervention" for "population" with "condition"

E.g. Dialectical behavior therapy for adolescent females with borderline personality disorder

2. Define inclusion and exclusion criteria

Use the Cochrane acronym PICO (population, intervention, comparison, outcomes) to decide on key components before starting the review.

3. Develop search strategy and locate studies

Work with a reference librarian to develop and run electronic searches.

Generate a comprehensive list of key terms related to each component of PICO to identify all relevant trials.

4. Select studies

Review abstracts and obtain full-text articles for studies meeting inclusion criteria.

Keep a log of all reviewed studies with reasons for inclusion or exclusion. Contact study authors if necessary for missing information.

5. Extract data

Create a data extraction form or table to organize info extracted from each reviewed study.

At least, two reviewers should extract data to ensure inter-rater reliability and to avoid data entry errors.

6. Assess study quality

Assess the quality of each randomized controlled trial (RCT) included in the review using.

7. Analyze and interpret results

You can use various statistical programs like "Review Manager" available to calculate effects sizes for meta-analyses.

Summarize your findings and provide recommendations for future research.

8. Disseminate findins

Publish your findings in simple and easy-to-read language without jargon for the benefit of patients.

Want to learn how you can supercharge your academic writing with AI apps?

I have a complete tutorial of 170+ slides for you.

More than 1,100 academics across the world are using it.

You can get it here: efficientacademicwriter.carrd.co

How I used ChatGPT to help me write this thread.

I downloaded the article and read it in Zotero.

I annotated all the sentences I thought were important.

I converted the annotations into a note in Zotero.

I took the notes and asked ChatGPT-4 to convert them into a coherent article.

Then I asked ChatGPT to convert the article into a thread.

As you can see, this thread is not the same as what ChatGPT gave me.

I read ChatGPT's version, and reread my notes.

I wrote the thread myself, but ChatGPT definitely helped me save a few hours of work.

Found this thread helpful?

1. Scroll to the top and retweet the first tweet to share it with your friends.

2. Bookmark it so you can come back to it later.

3. Follow me @MushtaqBilalPhD for regular threads on how to use AI-powered apps to supercharge your academic writing.