Research Brief — RB-027

How Much Trial Data Is Non-Essential? Protocol Design and Site Burden

Published June 2026 — An analysis of 105 Phase II and III protocols by TransCelerate and Tufts CSDD found that up to a third of the data collected per patient does not support primary or key secondary endpoints, and that this data accounts for 25–30% of the burden on sites and participants.

Executive Summary

A Third of the Work May Not Be Needed

Protocol complexity has risen for two decades, and sites feel it in longer visits, more procedures, and more data entry. What has been harder to establish is how much of that work is necessary. A study by TransCelerate BioPharma and the Tufts Center for the Study of Drug Development, released in September 2025 and discussed widely through the first half of 2026, puts numbers on the question.

The study classified the procedures and data points in 105 Phase II and III protocols from 14 biopharmaceutical companies according to whether they supported primary and key secondary endpoints. A substantial share did not. This brief summarizes the findings and what they mean for sites reviewing new protocols.

Key Findings

5.96M
Data Points per Phase III Protocol

The average Phase III protocol in the sample generated about 5.96 million data points. The analysis covered 105 Phase II and III protocols from 14 companies.

32.5%
Non-Core or Non-Essential Data

Up to 32.5% of Phase III data collected per patient fell into categories that are non-core or non-essential to the protocol's primary and key secondary endpoints.

29.5%
Share of Phase III Site Burden

Non-core procedures accounted for 16.6% of Phase III site burden and non-essential core procedures for 12.9%. The combined figure for Phase II was 25.1%.

21+ hrs
Source Preparation Time

In a separate Tufts CSDD and CRIO study of 209 site professionals, more than 80% of sites reported spending over 21 hours preparing source documents for a high-complexity study.

What the Analysis Shows

The findings are consistent across phases and point to the same conclusion: a meaningful part of site workload is generated by data that the trial's main questions do not require.

1

The Burden Is Concentrated in Lower-Value Data

The study estimates that non-core and non-essential data contribute 25–30% of the total burden on trial participants and investigative sites. These are procedures performed, recorded, entered, queried, and monitored like any other, although they do not bear on the endpoints that will decide the trial.

2

Non-Oncology Trials Show the Largest Share

The authors report the greatest inefficiency in non-oncology trials, where up to 40% of patient-level data fell into the non-core and non-essential categories.

3

The Work Starts Before the First Patient

The Tufts CSDD and CRIO survey shows that complexity costs sites time well before enrollment. Building source documents from a complex protocol takes most sites more than half a working week, and that effort is rarely reflected in start-up budgets.

4

The Regulatory Direction Supports Change

The authors link their findings to ICH E6(R3), which encourages fit-for-purpose data collection and asks sponsors to avoid unnecessary complexity and burden while maintaining trial quality.

Implications for Sites and Sponsors

Sites cannot rewrite a protocol, but they can use this evidence in feasibility, budgeting, and conversations with sponsors.

Use Feasibility to Flag Burden

When a protocol includes procedures that appear unrelated to its endpoints, say so in the feasibility response. Sponsors are increasingly looking for this feedback before the protocol is final.

Budget for Source Preparation

Time spent building source documents is real start-up cost. Sites should estimate it per study and include it in the start-up fee.

Design Around Core Endpoints

Sponsors can apply the study's classification to their own protocols before finalization, removing procedures that do not support primary or key secondary endpoints.

Conclusions

The TransCelerate and Tufts CSDD analysis gives sites and sponsors a shared, quantified basis for a conversation that has usually been anecdotal. Roughly a third of collected data sits outside the trial's central questions, and it carries a proportionate share of the burden.

The study covers protocols from 14 large companies and may not represent all sponsors or therapeutic areas. Even so, the direction is clear, and it aligns with what regulators are now asking for under E6(R3).

Sources

This brief summarizes published industry research. Figures are reported as stated in the sources below and have not been independently verified by Clinitiative.

Discuss These Findings with Our Team

Talk with us about what this research means for your protocol, site selection, or enrollment plan.