Conduction of Pragmatic Clinical Trials in Routine Clinical Settings: Guidelines for Development of Protocol

Abstract

Background

The relationships between treatments and real-world healthcare outcomes can be effectively evaluated through the implementation of a pragmatic clinical trial or practical clinical trial (PCT).

Objective

This trial helps to bridge the gap between clinical research and real-world practice by providing valuable insights into treatment effectiveness in routine health care settings. This article aims to present a systematic approach for developing study protocols by combining the author’s real-world experience from study center and knowledge derived from existing literature.

Methods

This brief review article is based on a focused literature review and the author’s academic experience in developing study protocol for academic research using PCT design. The core methodology components, such as research question formation, site selection, feasibility testing, and sample size calculation, are presented in a stepwise manner to guide protocol development.

Results

The study outlines a structured approach that can be applied in routine health care practice. The stepwise approach aids in the systematic planning and execution of PCT, ensuring their feasibility and relevance in real-world clinical settings.

Conclusions

This article serves as a practical guide for future trialists and researchers interested in adopting the PCT design. It also aims to raise awareness among those unfamiliar with this approach, offering an alternative perspective beyond traditional randomized controlled trials.

clinical trial registry of India

CTRI/2025/03/081559.

Introduction

A pragmatic clinical trial, also termed as a practical clinical trial (PCT) or pragmatic randomized clinical trial, was designed in 1967 by Schwartz and Lellouch to assess the relation between intervention effectiveness and its outcomes in real-world health care settings which are identical to routine health care practice setup. Practical clinical trials emphasize external validity by evaluating how interventions work in diverse, heterogeneous population, in contrast to randomized controlled trials (RCTs) which focus on building causation through diligent eligibility criteria.¹ Although RCTs are known for determining efficacy and safety under controlled environments, they often discount the intricacies of real-world clinical settings.^2,3

This gap can be fulfilled by using PCT design with significant insights about treatment effectiveness in routine health care settings. Because of its distinctive design, PCTs offer practical structured configuration for interpreting research outcomes into real-world applications. Despite the abundant existing literature on RCTs based on designs and applications, PCTs gain relatively less consideration.⁴ This current literary work collates existing literature and authors’ experience in developing a study protocol focused purely on PCT design applied in routine clinical settings. The rationale of this study is to offer a practical guide for future inquisitive trialists and researchers in PCT conduction, as well as to promote awareness among those unaccustomed to the characteristics of this study design. By presenting our experiences and insights, we aim to make the future study process more accessible and feasible, advocating the implementation and adoption of PCT design in real-world clinical research.

In any study, identifying future research gaps often stems from analyzing the limitations of previously published literature. Most existing trials are based on RCTs with strict inclusion and exclusion criteria, controlled settings, and extensive resource utilization. In contrast, the application of PCTs differs significantly from RCTs or explanatory clinical trials. Practical clinical trials can be conducted with more flexible eligibility criteria, involving all patients who visit routine clinical settings, such as large multispecialty and tertiary care hospitals, thereby enhancing relevance to real-world scenarios. This gap informed our approach to designing a trial that can be implemented in typical real-world settings, enabling the results to be generalized to the broader patient population. The following are the steps that we took while preparing our study protocol (Figure 1).

Fig. 1.

Ladder step approach for pragmatic clinical trial protocol development. PRECIS-2 = Pragmatic Exploratory Continuum Indicator Summary.

Methods

While preparing the protocol, we must adopt and follow Consolidated Standards of Reporting of Trials pragmatic extension guidelines alongside the checklist.⁵

The research question must comply Feasibility, Interesting, Novel, Ethical, Relevant and Population of interest, Intervention, Comparison, Outcome, Time criteria

The first ladder step in PCT design is to recognize the relevant research question (RQ), emphasizing whether pertinent interventions administered in routine or usual clinical practice is impactful in enhancing patient health outcomes compared with standard available treatments. This method depends largely on flexibility, generalizability, and adherence. Research questions in PCT design are usually framed for swift implementation in usual practice, and they emerge from a range of stakeholders such as patients, health care professionals, sponsors, and health care delivery systems.

The RQ should be designed to determine whether the intervention is practical and feasible in routine clinical practice. For example, we formulated the following RQ.

“Does addition of statin therapy to antidiabetic and antihypertensive therapy offer more effective management of T2DM patients with hypertension as compared to antidiabetic and antihypertensive therapy alone in real-world practice?”

We then applied the Feasibility, Interesting, Novel, Ethical, Relevant criteria to evaluate its feasibility and practicality in real-world settings (Table 1).

Table 1.

FINER criteria.

F	Feasible	Feasible to conduct study comparing antidiabetic and antihypertensive therapy versus antidiabetic, antihypertensive, and statins therapy in patients with T2DM and hypertension, as these patients are frequently find in routine health care settings.
I	Inquisitive	It seeks to determine the effect of antidiabetic and antihypertensive drugs with or without statins on BP.
N	Novel	There are no comparative studies that assess the effectiveness of antidiabetic and antihypertensive with or without statins therapy in patients with T2DM and hypertension in real-life clinical settings.
E	Ethical	The study will conduct in compliance with ethical guidelines, patient consent, privacy, and data confidentiality will maintain.
R	Relevance	T2DM and hypertension are most prevalently found patients in routine clinical practice; hence there is potential implication for improving clinical management and outcomes in these patients.

BP = blood pressure; T2DM = type 2 diabetes mellitus.

Feasibility, Interesting, Novel, Ethical, Relevant criteria are used to assess the significance and practicality of the study question, which further confirms that it is not only scientifically important but also feasible to apply in real-world clinical practice. This criterion improves overall quality and efficiency of PCTs in real setup. In case of Population of interest, Intervention, Comparison, Outcome, Time, while using these criteria, it is feasible to assess the intervention impact while comparing it with standard therapy; the comparison group is always a standard treatment which is usually given in the routine settings without overt effort.

Ensure feasibility of the study before launching of PCTs

There is a need for sustainable design, a plan to understand the process and significant attributes to trial demonstration, dissemination, feasibility, and clinical relevance assessment prior to trial application. These attributes also involve the outcome effects, stakeholders’ requirements, ideal trial outlines, and the resources needed for intervention delivery. For this purpose, the Pragmatic Exploratory Continuum Indicator Summary (PRECIS-2 tool) is utilized, which has 9 domains. These domains ensure that the trial design corresponds with the trial objectives and RQ, making it appropriately pragmatic and significant to real-world scenario. This tool application helps researchers to understand how much the current study reflects routine settings, thereby enhancing the chances of generalizability and practical applicability.

The feasibility of clinical trials (PCTs) is crucial for linking the design to their real-world performance. This involves evaluating the logistics of clinical trials within the existing health care system and conducting pilot tests on key aspects of PCTs. The logistics assessment involves documenting a comprehensive record of PCTs and investigating stakeholder training. Pilot testing considers methods like access to and use of electronic health records data, collaborating with statistical experts to predict cluster randomization problems, and validly defining outcomes of interest. Factors related to PCTs, such as informed consent forms, should also be assessed. This process confirms feasibility by exploring potential problems and solutions when PCTs are performed.⁶

The process of implementing the PRECIS-2 tool varies from RCT application, therefore understanding the PRECIS-2 tool and its application needs conclusive steps.

Application of PRECIS 2 tool

The PRECIS-2 tool^7,8 upholds multidisciplinary teams in the trial plan, assimilating expertise in the intervention, health care background, and statistical assessment. The “four-step approach” we used involved to improve the design, resolution of discrepancies of the initial design alternatives, and trial purpose. With the PRECIS-2 tool’s assistance, researchers can assure the intervention’s effectiveness is improved for certain real-world settings.

Step 1: Choose right intention of trial

Prior to designing our study, we clarified our attitude and intention towards it by understanding the difference between explanatory and pragmatic approaches. We ensured our study approach is to check whether an intervention works under usual conditions, relevant to a pragmatic clinical trial.⁷

Step 2: Considering trial choices for each domain

While considering the pragmatic design, we the authors decided how to implement 9 domain choices and their real-world relevancy. We analyzed every domain as per the study centers, patient feasibility, and health care providers.⁷

Step 3: Scoring domains

We started scoring each domain from 1 to 5 by answering every domain based on feasibility and applicability. Position each choice on the PRECIS-2 wheel (Figure 2) to record its pragmatic or explanatory nature, using a 5-point Likert scale.⁷

• 1.Very explanatory
• 2.Rather explanatory
• 3.Equally pragmatic and explanatory
• 4.Rather pragmatic
• 5.Very pragmatic.

Fig. 2.

Pragmatic Exploratory Continuum Indicator Summary (PRECIS-2) wheel.

Step 4: Review the PRECIS 2 wheel

The assessment of 9 domains (Table 2) is needed to examine whether a trial is pragmatic or exploratory. The table below is the assessment of these 9 domains conducted by subject matter experts (SMEs) for our study. The background of our study was to compare antihypertensive and antidiabetic drugs, with or without statins, in the management of patients with type 2 diabetes mellitus (T2DM) and hypertension. The main eligibility criteria included patients aged ≥18 years with either newly diagnosed or previously diagnosed T2DM and hypertension. To assess feasibility, we considered 3 study centers, where SMEs conducted a thorough analysis.⁹ After the final assessment using PRECIS-2, the trial received a score of 38, signifying its classification as a pragmatic clinical trial (Table 2).

Table 2.

Assessment of Pragmatic Exploratory Continuum Indicator Summary (PRECIS-2) 9 domains by the subject expert matters.

Sr No	Domain	Score	Rationale
1	Eligibility criteria	3	Patients with age group of 18 to ≥60 years with the condition of T2DM, and hypertension, irrespective of dyslipidemia were included. That means study has predefined criteria for exclusion
2	Recruitment	5	The patients with T2DM and hypertension, ie, condition of interest, were included without any overt recruitment approach, and also patient presenting to the general practitioner or physician in tertiary care and other private hospital attended or diagnosed by the physician
3	Setting	5	Usual care setting, where patients usually go for advice and treatment.
4	Organization	5	Patient information sheet and informed consent were taken from the patients which is not daily usual practice. The study is conducted in routine health care setup and use existing health care staff and resources
5	Flexibility delivery	5	Full flexibility in delivery of the intervention
6	Flexibility adherence	5	Participants free to “not-comply” with advice. No special measure were required to ensure the compliance of the patient to the intervention
7	Follow-up	4	Routine follow-up as per the physician recommendation
8	Primary outcome	4	Relevant to the patients. Routinely measured clinical parameters were advised
9	Primary analysis	4	Intention-to-treat, including all the data collected in the analysis
	Total score	40/45	Pragmatic clinical trial

T2DM = type 2 diabetes mellitus.

Calculation of sample size

The effect size relies on what is essential in routine clinical practice rather than statistical significance in PCTs. The effect size can be obtained from the existing literature or studies conducted on a similar population of interest, pilot study, meta-analysis, real-world evidence studies, and observational data from registries or health records. Then select an appropriate effect size based on the primary outcome of the study such as “Cohen’s d” for continuous outcomes, risk and odd ratios for binary outcomes, and hazard ratios for time-to-event outcomes. Then use software tools like “G*Power,” “RStudio,” or other statistical software to estimate sample size. Smaller effect sizes need a larger sample size to ensure adequate power. PCTs involve more diverse, heterogeneous population and settings, effect size can be diluted. This can be resolved by including a subgroup analysis plan. Then perform sensitivity analysis such as small versus moderate, which demonstrates the robustness of the design.

Selecting study sites identical to the actual clinical settings

It is important to consider the study centers or sites that are similar to the actual or routine clinical settings for the PCTs. We precisely selected sites based on the prevalence of diseases or conditions and the availability of medical care which may differ largely based on the type and regions of the institution. The sites naturally depend on the heterogeneity, patient characteristics, and disease prevalence. We confirmed the feasibility and capacity of the study by aligning the study objectives, study-level experience, and the number of patients accessible. The characteristic of PCTs is inclusion of heterogeneous and multiple sites rather than homogeneous or single sites to apply results to patients in various setups and declare generalizability.^6,10 In our study, we selected the 3 study centers, one tertiary care hospital, and other 2 were multispeciality hospital; these centers are where the patients come for routine checkup.

Independent evaluation by SME

Before initiating the study, it is essential to determine whether the trial is routing towards pragmatic (real-world) or explanatory (ideal-world) in nature on PRECIS-2 continuum scale, and to assess trial feasibility in routine clinical setup.⁴ To ensure its objectivity, the appointment or involvement of SME plays a crucial role to independently evaluate 9 domains of the PRECIS-2 tool. Each SME will independently review predefined domain criteria and assign score ranging from 1 to 5 based on practical application. Based on this score, the nature of trial will be decided whether it is pragmatic or explanatory.

For instance in our study, after the evaluation of all 9 domains, the SME allotted a total score of 40 points, confirming that the study design is pragmatic in nature (Table 2).

Estimating sample size

Two methods for integrating real-world data into power and sample size estimates were proposed by Shortreed and Simon.¹¹ The first method estimates the amounts needed for typical sample size or power calculations using the data that are currently available. The second method estimates power under various conditions using a simulated study based on actual data.

Otherwise, apply the sample size formula and all practical trial parameters in classic sample size calculation techniques, along with the application of power analysis to detect a significant impact with a certain degree of probability. To determine how alternative sample sizes affect the study's capacity to identify a real effect, use simulation tools. In our study, we used Lehr’s equation to estimate sample size.

We assumed a mean systolic blood pressure difference of 6 mm Hg with a standard deviation (SD) of 12 mm Hg, corresponding to an effect size of Cohen’s d = 0.5. This assumption is based on prior clinical study evaluating statins in hypertensive and diabetic populations although meta-analyses report smaller effects.¹² The study is aimed to assess effect of statins on BP in T2DM and hypertensive patients, the participants will compare statins with antidiabetic and antihypertensive drug (statin user/intervention group) versus antidiabetic and antihypertensive drug (nonstatin user/control group).

From previous literature, we assumed Cohen’s d/effect size

• Expected difference in systolic blood pressure = 6 mm Hg

• SD = 12 mm Hg

• Cohen’s d (d) = 6/12 = 0.5 (medium effect)

Formula used for sample size calculation Lehr Equation

• n = 16/ d² (for α = 0.05, power = 0.80)

• n = 16/(0.5)² = 64 per group

• Total (without attrition) = 128

• Add 10%-15% attrition = approximately 128–150 participants

Using this effect size (d = 0.5), a 2-tailed significance level of 0.05, and 80% power, the required sample size was calculated to be 64 participants per group (128 total). To account for potential attrition and to ensure adequate power for secondary outcomes, the final sample size included will be 300 participants.⁹

Statistical analyses

Understanding statistical analysis methods in PCTs is crucial for creating an analysis plan. Randomized controlled trial methods can be applied to PCTs, but considerations are needed due to their uniqueness. Intention-to-treat analysis and per-protocol analysis are necessary for clinical trial results. Intention-to-treat is the principal analysis in PCTs, but potential validity issues may arise if patients switch treatments. Intention-to-treat is generally recommended to reduce selection bias and unbalanced covariates in per-protocol analysis. The intention-to-treat analysis can be confounded by large amounts of missing data. To address this, a follow-up history and possible reasons for missing data should be recorded for all patients. Causal inference tools like marginal structural models or additional data like patient-reported outcomes from other sources can address bias. Practical clinical trials have more heterogeneous populations than RCTs, so subgroup responses to interventions may differ. A priori subgroup analysis with a large sample size is recommended for valid treatment effect estimates and avoiding false positive results.¹³ As per our study experience, we also utilized IIT analysis as patient interventions were decided by the attending physician which are usually done in routine clinical settings.

Based on the intervention and treatment group, subgroup analysis was performed, forming an important aspect of PCT statistical analyses for drawing meaningful results. Our subgroup analyses were evaluated according to the type of drug therapies administered, age categories, gender, and other factors. This subgroup data set provided us with how the treatment works on the real patients depending on their demographic and clinical characteristics.

Discussion

This study proposes a methodical framework for designing and carrying out pragmatic clinical trials (PCTs) in practical settings. Because it is step by step, researchers and academics may efficiently plan and design studies that are applicable to real-world settings. Prior review studies have emphasized the importance of significant criteria in PCT design, including feasibility, research site selection, benefits, and stakeholder engagement, use of the PRECIS-2 tool, randomization strategies, and statistical analysis.¹⁴

Experts from the United States and Canada have argued about whether, in the period of big data and advanced analytics, RCTs continue to serve as the gold standard for medical evidence.¹⁵ Although RCTs have historically been essential to medical research because they can eliminate bias from confounding variables, thus guaranteeing internal validity. Later, the panel discussion highlighted an array of limitations: RCTs are time consuming and expensive, and participant dropouts and noncompliance may skew the results. In addition, findings tend to be only applicable in tightly controlled settings or precise replication conditions due to their restricted generalizability.¹³

The RCTs hold advantages as the random assignment guarantees comparable research groups and reduces baseline confounding. They continue to be the most effective way to assess the effectiveness of pharmaceutical interventions, are acknowledged as the gold standard, and, in the right circumstances, produce an efficient intervention effect when carried out in accordance with accepted protocols.^16,17 Nonetheless, questions still surround the applicability for use in actual clinical settings. It is challenging to extrapolate RCT results to a multitude of patient populations as different patients have different diseases, genetic predispositions, tolerance levels, and treatment responses.¹⁸

In order to overcome these constraints, PCTs play a crucial role as they strengthen real-world applicability and generalizability.¹ Practical clinical trials are intended to be integrated into routine clinical treatments, bridging the gap between routine clinical practice and controlled trial research. They are especially useful for multifaceted disorders and for researching intricate drug therapies.¹⁹

However, there are numerous challenges in conducting trials with actual or routine practice involving the selection of study participants, assessing treatment adaptability, incorporating data on synchronous treatments and patient compliance, and choosing appropriate comparator conditions and outcome measures. Practical clinical trials are randomized studies that evaluate the effectiveness, risks, and costs of interventions in clinical practice. They use active comparators, broad inclusion criteria, diverse patient populations, and clinically relevant endpoints. Practical clinical trials minimize eligibility restrictions, making results more applicable to routine clinical settings.⁴ This guidance helps researchers unfamiliar with PCT design and encourages its widespread adoption.

Challenges and ethical considerations

Since PCTs are conducted in real-world settings that pose complexities and vulnerabilities that must be handled to safeguard participant autonomy and well-being, hence ethical considerations and informed consent processes demand particular attention.

In PCTs, participant vulnerability is a concern, as they often involve diverse populations with limited health care access or health status.²⁰ The real-world setting may present unclear benefits and unpredictable risks. Participants should have autonomy to make informed decisions, and the consent process should be tailored to the trial context. In the case of informed consent, it should be clear, concise, respectful of autonomy, and transparent about risks and benefits, considering the complexity of real-world settings.²¹

A central ethical consideration for PCTs is the necessity of informed consent, which is distinct from consent for treatment. The PCT framework allows for streamlined consent models approved by institutional review boards, such as integrated discussions and brief verbal permissions, or waivers of written documentation under specific criteria. This flexibility helps reduce logistical barriers, minimizes selection bias by including diverse patient populations, and enables the generation of real-world evidence to improve care, highlighting a key methodological strength of PCTs.²²

Stakeholder engagement

Stakeholder involvement is essential in the PCT research process for ensuring clinical significance, feasibility, real-world usage, improved adaptation, sustainability, and enhanced outcome measures.²³ Stakeholder interaction and joint efforts are needed to resolve the complexities of real-world situations to generate real-world evidence. This aids in incorporating research findings more efficiently into the clinical practice by identifying potential barriers and developing strategies to overcome such situations. In our study, stakeholder engagement such as health care professionals, laboratory teams, patients, and statistical analysts played a major role in keeping the PCT mirroring routine practice.^1,24

Future implication

Our study serves as a guide for interested future researchers for conducting and organizing PCTs in their settings. Practical clinical trials also pave a way for the pharmaceutical companies to conduct postmarketing surveillance studies and generate real-world data. By adopting this design, subsequent research could use real-time patient-reported outcomes to enhance findings, examine new trials incorporating digital health technologies, and detect long-term outcomes of already marketed therapies. By adhering to this narrative approach of this article and our study, researchers may effectively design and carry out PCTs that provide significant, practical information for “clinical decision making.”

Conclusion

Aiming to provide practical guidance on PCTs, this brief review is based on pioneering Indian study that had employed a PCT design. By combining authors’ practical based methodology with existing literatures, authors’ developed a stepwise approach for designing protocols in routine clinical settings. This method, grounded in direct author involvement and self-analysis, enhances practical application and leads to more meaningful research outcome.

Funding

None.

CRediT Authorship Contribution Statement

Smruti Besekar: Literature review, Manuscript drafting and preparation. Sangita Jogdand: Conceptualization, Study design, Overall supervision. Satywam Singh Patel: Conceptualization, Critical review of the manuscript, Final approval.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.