Clinical Study Report-E3 structure
Clinical Study Report First, is important to understand the definition, requirements, and potential uses of a CSR. The report is a comprehensive look at all the data produced in a clinical study, presented in text, tables, and figure formats. It will often include discussions and conclusions that provide context to the findings regarding the drug, device, biological product, surgical method, counseling practice, or any other type of therapeutic product or practice under study and where it may contribute to an improvement on the state of the art for treating or preventing a particular health condition.
Most likely you already appreciate the ethical responsibility a clinical study team has to clinical study data transparency, which for that reason alone would make the production of some sort of CSR necessary. Indeed, the preparation and representation of study progress is prescribed in the aforementioned ICH GCP E6(R2) guideline,{1} which states that study sponsors should ensure that clinical trial reports are prepared and provided to regulatory agencies as they are required.
Further, the guideline recommends study sponsors to rely on a subsequent guideline on Structure and Content of Clinical Study Reports (ICH E3).{2} Lastly, adhering to this ethical responsibility and following GCP have become mandated both in the U.S. and in Europe, where study data are expected to be recorded on ClinicalTrials.gov and the EudraCT database, respectively, for the sake of transparency and in support of further scientific inquiry, thus making the organization and preparation of study data in a prespecified format necessary.
Prepare to write by educating yourself about the study, as you would for any scientific manuscript. If you understand the trial well, research the condition being treated, and study the results closely, the CSR will almost write itself. Don’t be frightened by the list of documents below. You need not read them all closely; many can be skimmed. With some experience, you will be able to read everything in a day or so. Keep in mind that some of these documents can change slightly during the course of the study, so remember to track versions and dates of implementation.
Before writing a CSR, I read these study related materials, in roughly this order:
The study protocol: Clinical Study Report
The protocol is a detailed plan for conducting the study. It describes all study procedures and defines how all study end points are presented and analyzed. The protocol is approved by regulators prior to initiating the study. The writer needs to understand this document in detail. When reading a protocol, I focus throughout on the objectives and end points. I try to understand the rationale for choosing the experimental hypotheses and how each study procedure contributes to answering them.
- The disease process: By knowing the pathophysiology of the condition being treated, I can anticipate the adverse events and outcomes that trial subjects are likely to experience. I consult recent textbooks, review papers, references in the study protocol and investigator brochure, and Web sites such as http://emedicine.medscape.com. When using online materials to research a disorder, make sure sources are scholarly.
- The investigator brochure (IB): The IB is written for clinical site investigators who will administer the study product to patients and follow them clinically during the trial. The IB summarizes previous studies of the product conducted in animals and humans. It may include recent unpublished results. IBs are updated frequently to incorporate significant new study findings. Skim the IB to learn the history of the product and the safety concerns identified in previous studies.
Competing products or other treatments for the same disease:
Peruse Drugs@FDA or www.rxlist.com to learn about currently marketed products. If your product lacks the side effects or deficiencies of a competing treatment, demonstrate that in the CSR.
The case report forms (CRFs): CRFs are electronic or paper forms used by clinical study sites to record information about each subject. Each screen or 8 by 11 inch paper form is called a CRF page, and the collected pages for a particular patient are called a CRF book. In a lengthy or complex trial, each CRF book is hundreds of pages long. Because CRFs show the exact questions presented to investigators, they sometimes give a bit more detail than the protocol by showing the exact questions and the possible answers. Skim these and map each page to the corresponding study procedure in the protocol.
Manuals of operations or other instructional materials: I quickly skim all written instructions or training information given to clinical sites, radiology reviewers, clinical event committee members, or study subjects. These documents and webinars occasionally offer a little more detail than the protocol, allowing me to describe study methods accurately.
The statistical analysis plan (SAP): The SAP is a plan for analyzing study results. It is written by statisticians and approved by regulators prior to study conclusion. Read the SAP, but don’t be concerned abut comprehending statistical minutia. Instead, map the calculations and analyses in the SAP to the end points in the protocol. I create a section heading in the CSR for each study end point in the SAP. This helps me double check whether I am discussing all important analyses in the CSR.
The randomization and blinding schemes: Randomization methods in study protocols are often sketchy. However, all details about how subjects were randomized to treatment groups should be included in the CSR, so I often check with the statistical group to fill in blanks. The actual randomization codes are placed in an appendix to the CSR.
The tables, listings, and figures (TLFs): These are created by the study statistical group to present the data. The writer must understand how these correspond to the study’s objectives and end points. Some TLFs require extensive discussion. Others can be summarized in a sentence, and some can be omitted entirely if they are duplicative (eg, results in a table that are also in a figure). Often the numerical order of the TLFs is not the order in which I will discuss them in the CSR. I rearrange them to correspond to the CSR sections: disposition, demographics, efficacy, and safety.
Data monitoring committee (DMC) minutes and recommendations: During some trials, a committee of independent clinical and statistical experts reviews study data on one or more planned occasions while the study is ongoing. After each data review meeting, DMCs issue recommendations to the sponsor that are also reported to institutional review boards. Their recommendations sometimes impact the study. For example, DMCs can recommend protocol amendments, additional study procedures to assess safety or efficacy, or early termination of a study for reasons of safety or efficacy. Minutes and recommendations from DMC proceedings help me in two ways:
DMC minutes sometimes explain reasons for unusual or unplanned changes in a study. For example, if a trial is stopped early based on a DMC recommendation, the DMC minutes document why. I also check the SAP, the protocol and the DMC charter to find the monitoring plan and communication procedures required for stopping the study. In the CSR, I document these details and state whether they were properly followed.
DMC minutes offer insight into how independent subject matter experts interpret the study data and what findings and issues they consider important, occasionally providing a fresh perspective.
Study Subjects
Describe the study population and disposition of subjects in detail. I use a table or figure to illustrate entry and exit of subjects from the study. Bias can be introduced by improper screening procedures or inclusion of ineligible subjects. Subjects may withdraw from the study, cross over to another treatment arm, or discontinue study treatment because of adverse events. Report the frequency of and reasons for such events by treatment arm.
In addition to comparing treatment groups to one another or to a reference group, look for other relevant comparisons dictated by the study design. For example, in Phase II trials, subjects are sometimes dosed in consecutive increasing dose cohorts. This allows a safety assessment after each dose level so investigators may determine the advisability of using a higher dose for the next cohort. Since cohorts are not treated at different times and clinical site procedures can change subtly from one cohort to the next, cohort differences can introduce bias. Therefore, even if formal statistical comparisons among cohorts are not planned, summarize subject characteristics by cohort. Whenever treatment groups are pooled or historical controls are used, report the baseline characteristics of the individual groups.
Assess subject compliance with study treatment procedures. If all randomized subjects completed all procedures and took all doses, mention that fact. Subject noncompliance is most important when it differs among treatment groups or results in discontinuation of study treatments or study withdrawal.
Study Populations
At least two populations are usually defined in the protocol and SAP. The intention to treat population (ITT) is comprised of all randomized subjects, whether or not they actually received study treatment. The safety population includes subjects who received one or more study treatments. In a perfect world, these two populations contain the same subjects. Other populations might be used for certain subgroup analyses. For example, some analyses may be conducted only in subjects who possess a certain baseline characteristic or who demonstrate a minimum level of compliance with study treatments.
The ITT is included in analyses of demographic and baseline findings and efficacy results. In ITT population analyses by treatment arm, subjects are included in the treatment arm to which they were randomized. The safety population is included in analyses of adverse events, vital signs, clinical laboratory findings, and other safety data. In safety population analyses, subjects are included in the treatment arm corresponding to the treatment they actually received.
These are the usual definitions of the ITT and safety populations, but many variations are possible. Clearly state which study population is described in text sections and in-text tables and figures within the CSR.
Efficacy
Statistical members of the study team often assist with writing the efficacy section. First discuss the primary study results and then describe other results in decreasing order of importance.
Safety
Safety results usually include clinical laboratory values, adverse events, vital signs, medical history, and examination findings, and sometimes other types of data. Clinical members of the study team often assist with writing the safety sections. The level of detail reported in the CSR varies with the study indication, study population, condition being treated, expected risks of the study product, and therapeutic alternatives, but I always discuss adverse events that are serious or precipitate study discontinuation or drug withdrawal.
Sponsors want to present their products in the best possible light, and the safety section is the best place to do so. There is little latitude in the efficacy section because we specify measurement and statistical analytical methods prior to the trial. But the safety section is squishier: safety objectives and end points tend to be less precise and more open to interpretation. The safety section therefore presents an opportunity for “spin.”
It is never correct to mislead or fail to report important information, but placing undesirable study results into context is acceptable. Carefully study the adverse events, then dig through other data to see whether explanatory factors are present for subjects experiencing concerning adverse events. I use the medical history and physical examination, including those performed at baseline, eligibility violations, compliance data, concomitant medications, vital signs, clinical laboratory values, other influential adverse events (eg, vomiting or diarrhea possibly causing incomplete absorption of a dose), and any other relevant data I find. Present all data that support a relationship between an adverse event and a factor other than the study product.