文献检索计划(LSP)

在临床评价计划（CEP）中，清晰、系统地阐述文献检索的规划和方法至关重要。这部分内容，无论是在CEP正文中概述并在附录中详述，还是作为CEP的一个独立章节，其核心目的均是预先设定一个科学、严谨、可重复的流程，用以全面、客观地识别和获取与待评价医疗器械相关的已发表科学文献及安全性数据。一个周密的文献检索计划是确保临床评价所依赖的外部证据既充分又无偏倚的关键保障。

LSP节应清晰勾勒出从检索目标设定到最终文献筛选的完整策略。

明确文献检索的目标与范围

首先，需要明确本次文献检索旨在达成的核心目标。这些目标通常与临床评价的整体目的紧密相连，可能包括（但不限于）：

• 确立与待评价器械相关的医学领域、特定临床状况以及相关治疗技术的当前技术发展水平（State-of-the-Art, SOTA）。
• 搜集关于待评价器械本身（如果已有相关公开发表的研究或报告）的临床数据、安全性和性能信息。
• 识别并获取关于类似器械、基准器械或已确定的等同器械的临床数据，以便进行比较分析或支持等同性论证。
• 收集关于潜在的临床危害、不良事件、并发症以及与该类器械或技术相关的已知风险信息。

在明确目标之后，应界定文献检索的范围。这包括确定检索的时间跨度（例如，过去5年、10年，或根据技术发展速度和产品特性确定的特定时间窗口）、文献的语种（通常为英文），以及在检索过程中特别关注的器械特性、临床应用场景或患者亚群。

构建清晰的检索问题 (Research Questions)

为了使文献检索更具针对性和有效性，预先设定明确的检索问题是不可或缺的一步。这些问题应源自临床评价的核心目标、待评价器械的预期用途、关键临床宣称、以及在风险评估或现有数据回顾中发现的任何不确定性或信息缺口。

检索问题可以围绕不同主题进行组织，例如：

• 针对当前技术发展水平（SOTA）的检索问题：这类问题旨在全面了解相关疾病的治疗现状、主流技术、替代疗法及其各自的临床安全性和性能表现。
• 针对特定器械（包括待评价器械、类似/等同/基准器械）的检索问题：这类问题则更侧重于收集直接相关的临床证据，用以评估这些器械的具体安全性和性能，或进行横向比较。

制定周详的检索策略

检索策略是文献检索计划的核心，它详细规定了将如何系统地查找所需信息。

数据来源的选择： 必须明确计划检索哪些类型的数据库和信息来源。详见：文献来源及选择理由 和 临床数据的分析。

检索词与检索式的构建： 这是确保检索全面性和准确性的关键技术环节。应阐述选择和组合检索词的系统方法。

• 初步信息收集与理解： 在正式构建复杂的数据库检索式之前，进行广泛的初步信息调研是非常有帮助的。有助于全面理解相关的疾病背景、当前临床实践、权威指南推荐、关键的安全与性能评价指标以及尚未满足的医疗需求。这种初步探索（可利用通用搜索引擎如Google Scholar）为后续精准检索词的选择和优化奠定了基础。
• 核心检索词的选择与组合： 需要围绕待评价器械的特性（如通用名称、技术原理、关键组件）、预期用途（相关的疾病、适应症、手术过程）、目标人群、关注的临床结局（如安全性指标、性能指标、有效性指标）以及比较对象（如类似/等同器械的商品名、通用名）来系统地构建关键词表。
• 检索逻辑的运用： 说明将如何运用布尔逻辑运算符（AND, OR, NOT）、截词符、通配符以及字段限制（如标题、摘要、关键词字段）等技术来组合这些关键词，构建针对不同数据库语法的具体检索式（Search Strings）。
• 针对不同检索目标的策略调整： 针对SOTA、待评价器械、以及类似/等同/基准器械的检索，其检索词的侧重点和组合方式可能有所不同，应分别予以考虑和规划。

检索限制条件的应用： 明确是否计划以及如何应用特定的检索限制，例如出版日期范围、文献语言、研究类型（如RCTs, 系统综述等）。但需注意，在筛选安全性相关信息时，过度依赖研究类型过滤器可能会遗漏来自非传统研究（如病例报告、不良事件数据库）的重要信号，因此使用时需谨慎权衡。

文献的筛选与选择流程

在执行检索并获得初步结果后，CEP必须规划一个系统、透明的文献筛选流程，以准确识别出真正相关的文献。

• 纳入与排除标准 (Inclusion and Exclusion Criteria)： 强调必须预先制定清晰、具体的文献纳入和排除标准。这些标准应直接源于先前设定的检索问题和评价范围，并应用于文献筛选的各个阶段（如标题和摘要初筛、全文复筛）。
• 筛选过程的执行： 简要描述计划如何执行筛选过程。

详细说明见：文献选择过程、文献选择标准，以及文献评估标准。

数据提取与检索过程的文档化

CEP中还应简要提及，对于最终被纳入的文献，计划如何系统地提取与其相关的数据（例如，可以预先设计数据提取表格，明确需要提取的关键信息字段）。

整个文献检索过程：从数据库的选择、检索日期、实际使用的完整检索式、每个数据库的检索结果数量，到各筛选阶段的纳入/排除文献数量（在CER中以PRISMA流程图等形式呈现）——都将被完整记录下来。这种全面的文档化是确保文献检索过程透明、可重复、可审核的基础。

与数据评估的衔接

所有通过上述系统性文献检索过程所识别并筛选出的相关文献，都将进入下一个环节——即按照CEP中“相关数据的评估与分析”章节所规划的方法学，进行严格的数据评估（Appraisal）。

通过如此规划和阐述文献检索计划，能够确保临床评价所依赖的文献证据是通过一个系统、全面、无偏倚且可重复的过程获得的，从而为后续的数据评估、分析以及最终的临床评价结论提供高质量的输入。

---

Within the Clinical Evaluation Plan (CEP), a clear and systematic articulation of the literature search plan (LSP) and methodology is of critical importance. This section, whether outlined in the main body of the CEP and detailed in an appendix or presented as a standalone chapter, has the core purpose of prospectively defining a scientific, rigorous, and reproducible process for comprehensively and objectively identifying and retrieving published scientific literature and safety data relevant to the device under evaluation. A well-thought-out Literature Search Protocol is a key safeguard for ensuring that the external evidence upon which the clinical evaluation relies is both sufficient and unbiased.

The LSP should clearly delineate the complete strategy, from defining search objectives to the final selection of literature.

Defining the Objectives and Scope of the Literature Search

The core objectives that the literature search aims to achieve must be clearly defined. These objectives typically align closely with the overall goals of the clinical evaluation and may include (but are not limited to):

• Establishing the current State-of-the-Art (SOTA) for the relevant medical field, device type, and specific clinical conditions.
• Gathering clinical data and safety/performance information on the device under evaluation itself (if any publicly available studies or reports exist).
• Identifying and obtaining clinical data on similar devices, benchmark devices, or identified equivalent devices for comparative analysis or to support claims of equivalence.
• Collecting information on potential clinical hazards, adverse events, complications, and known risks associated with this class of device or technology.

After clarifying the objectives, the scope of the literature search should be defined. This includes determining the time period for the search (e.g., the last 5 years, 10 years, or a specific window based on the pace of technological development and product characteristics), language preferences for literature (typically English-dominant), and any particular device characteristics, clinical application scenarios, or patient subgroups that will be prioritized during the search.

Formulating Clear Research Questions

To make the literature search more targeted and effective, prospectively defining clear research questions is an indispensable step. These questions should stem from the core objectives of the clinical evaluation, the intended use of the device under evaluation, its key clinical claims, and any uncertainties or information gaps identified during risk assessment or review of existing data.

Research questions can be organized around different themes, such as:

• Research questions for the State-of-the-Art (SOTA): These questions aim to comprehensively understand the current treatment landscape for the relevant disease, mainstream technologies, alternative therapies, and their respective clinical safety and performance profiles.
• Research questions for device-specific searches (including the device under evaluation, similar/equivalent/benchmark devices): These questions are more focused on gathering directly relevant clinical evidence to assess the specific safety and performance of these devices or to conduct head-to-head comparisons.

Developing a Comprehensive Search Strategy

The search strategy is the heart of the Literature Search Protocol, detailing how the required information will be systematically sought.

Selection of Data Sources: The types of databases and information sources to be searched must be clearly specified.

Development of Search Terms and Search Strings: This is a critical technical aspect for ensuring the comprehensiveness and accuracy of the search. A systematic approach to selecting and combining search terms should be described.

• Foundational Information Gathering: Conducting broad preliminary information gathering before constructing complex database search strings is highly beneficial. This helps to comprehensively understand the relevant disease context, current clinical practices, authoritative guideline recommendations, key safety and performance evaluation metrics, and unmet medical needs. This initial exploration (which can utilize general search engines like Google Scholar) lays the foundation for the subsequent selection and optimization of precise search terms.
• Selection and Combination of Core Search Terms: A systematic list of keywords should be developed around the characteristics of the device under evaluation (e.g., generic name, technological principle, key components), its intended use (relevant diseases, indications, surgical procedures), target population, clinical outcomes of interest (e.g., safety indicators, performance indicators, efficacy indicators), and comparator devices (e.g., trade names, generic names of similar/equivalent devices).
• Application of Search Logic: Explain how Boolean operators (AND, OR, NOT), truncation symbols, wildcards, and field tags (e.g., title, abstract, keyword fields) will be used to combine these keywords into specific search strings adapted for the syntax of different databases.
• Tailoring Strategies for Different Search Objectives: The emphasis and combination of search terms may differ when searching for SOTA, the device under evaluation, or similar/equivalent/benchmark devices. These should be considered and planned separately.

Application of Search Limits and Filters: Specify if and how any particular search limits will be applied, such as publication date ranges, language of literature, or study types (e.g., RCTs, systematic reviews). However, caution is advised when using study type filters for safety-related searches, as important signals from non-traditional studies (like case reports, adverse event databases) might be missed.

Planning the Literature Screening and Selection Process

After executing the searches and obtaining initial results, the CEP must plan a systematic and transparent literature screening process to accurately identify truly relevant publications.

• Inclusion and Exclusion Criteria: Emphasize the necessity of pre-defining clear and specific criteria for including or excluding literature at each stage of screening (e.g., title and abstract review, full-text review). These criteria should be directly derived from the previously established research questions and evaluation scope.
• Execution of the Screening Process: Briefly describe how the screening process is planned to be carried out.

Data Extraction and Documentation of the Search Process

The CEP should also briefly mention the plan for systematically extracting relevant data from the finally included literature (e.g., a pre-designed data extraction form specifying key information fields can be developed).

The entire literature search process—from database selection, search dates, full search strings actually used, number of hits from each database, to the number of articles included/excluded at each screening stage (presented in the CER using a PRISMA flow diagram or similar)—will be fully and meticulously documented. Such comprehensive documentation is fundamental to ensuring the transparency, reproducibility, and auditability of the literature search process.

Linkage to Data Appraisal

All literature identified as pertinent through this systematic search process will proceed to the next step—rigorous data appraisal according to the methodology planned in the “Appraisal and Analysis of Pertinent Data” chapter of the CEP.

By planning and articulating the Literature Search Protocol in this manner, it ensures that the literature evidence relied upon for the clinical evaluation is obtained through a systematic, comprehensive, unbiased, and reproducible process, thereby providing high-quality input for subsequent data appraisal, analysis, and ultimately, the conclusions of the clinical evaluation.