• Drafting an appropriate question:

To a successful Systematic Review，constructing a good research question is necessary. Hence, how to drafting an appropriate question is the important first step.

First, we should decide what kind of question is possible for researching, which can decide by following question:

1. Is it an answerable question?

2. Is it a question worth researching?

3. Has the same research topic SR been published?

4. Is there enough literature to accomplish the project?

5. 釐清收案的研究設計類型?

• When drafting is done, we can refine our idea by following question:

• Does the question fit into the PICO question format?

  1. Have you defined all of your terms? (Even if your question does not fit into a PICO, it can be helpful to think about whether you have defined your population, intervention, etc.)

     • What age group?

     • What type or types of conditions?

     • What intervention? How else might it be described?

     • What outcomes? How else might they be described?

     • What is the relationship between the different elements of your question?

  2. Do you have several questions lumped into one? If so, should you split them into more than one review? Alternatively, do you have many questions that could be lumped into one review?

• Register for SR research (checking for registeration of same topic at the same time)

• • • PROSPERO

    • Cochrane Library

    • Epistemonikos

    • Preprint servers Embase Classic + Embase

    • BEME: Best Evidence Medical and Health Professional Education

    • Campbell Library of Systematic Reviews

    • healthevidence.org

文獻搜尋是SR研究的重要開端，發展良好與完整的檢索策略有助於後續研究的進行，避免研究文獻蒐集的遺漏與偏差。

NCKU Library provide Search Form for developing seraching statics and recording.

Developing seraching statics cotains following steps:

1. Choosing database:

Commonly used database like Medline, Embase, Cochrane Trials...etc, others like national/regional database, specific topics database, citation database (Web of Science, Scopus), Grey Literature...etc.

Usually a Systematic Review will contain 3 database.

2. Search terms: include Free Text and Controlled Vocabulary

(1) Free Text：the same as synonyms，including: Abbreviation, Broader & Narrower Terms, Singular and Plural/ Part of Speech/ Tense & Situation.

除此之外，同義詞也可以藉由字典、文章、中英對照書目...等去發掘。

(2) Controlled Vocabulary：由特定資料庫指定的制式詞彙，又稱為Thesaurus、Subject Term、Index Term等。MEDLINE資料庫的控制詞彙表明稱為MeSH、Embase資料庫的控制詞彙典為Emtree。在查找適用的控制詞彙的同時，可順便借重控制詞彙典提供的同義詞清單，交叉整理同義詞。

3. 在搜尋時可搭配布林運算子、片語及鄰近字...等方式以擬定檢索策略

4. 在檢索完成時，記得將檢索策略記下以便往後進行資料更新及最後撰寫論文時使用。

When searching is done, we need to screen the data set we searching to filt out the data related to our research.

Endnote, Rayyan and Abstrackr is commonly uesd for screening.

經過篩選的文章要進一步去評估該文章的可信度及實用性，包含該文章控制條件、研究流程...等，找出該篇文章有何優點/缺漏，紀錄文章數量並列出文章不予以採用的原因。

Synthesis 的目的在於彙整文章並對文章進行有效分析和得出結論，其通常會使用於兩組或兩組以上的獨立研究(或稱meta-analysis)，並利用statistical synthesis 的方式對文章進行統計。

Finally，present yuor data and conclution following the guideline, then a Systematic Review is finished.