Talk to us

Query backlogs decide when your database locks

Clinical trial database lock is planned as a single data management milestone, yet it arrives only when tens of thousands of open queries, reconciliations and sign-offs have closed across sites and vendors. Leading sponsors and CROs are managing those obligations as one live, priced backlog from the first patient onward.

Clinical researchSegeta Insights
Clinical research
The number11 days

longer for sponsors than CROs to lock a database

What is at stake

Each month that lock slips on a Phase III program could put about $$26 million at risk.

Where days are lost
  1. Data entry to query
  2. Query to site response
  3. Response to closure, coding and sign-off

Clinical trial database lock is scheduled, reported and reviewed as a data management milestone, with a date in the CTMS and an owner in data management or biostatistics. In practice, the lock date is set by thousands of small obligations held by people who sit outside the data management team. Each open query waits on a site coordinator, a monitor, a medical reviewer, a coder or an external data vendor, and the database can lock only when the last of them has closed.

Query volume has grown with the data behind it. Veeva research published in 2025 estimates that an average Phase III study now generates about 96,980 queries, and that a query raised on non-EDC data, such as imaging, eCOA or central laboratory results, takes 16.4 days to resolve against 9.6 days for a query raised in the EDC.1 The sections that follow trace where those days accumulate on the way to lock, what each day of delay costs a Phase III program and how sponsors can run the final stretch of a study as one managed backlog.

More data sources mean more queries to close before lock

The volume of data behind each lock has grown faster than the teams that clean it. Tufts CSDD and TransCelerate found that the average Phase III protocol collected 929,203 data points in 2012 and 3,560,201 in 2020, and that the average now stands at about 5.9 million, a growth rate of 15.4 percent a year since 2012.2 The same research suggests that up to 32.5 percent of the data collected per patient in a typical Phase III study may be nonessential to efficacy and safety conclusions, yet every one of those data points can still raise a query.3

Data now arrive from electronic clinical outcome assessments, wearables, central laboratories and imaging vendors as well as the electronic case report form. In a foundational Tufts CSDD study of data management practice, sponsors and CROs used approximately six applications per study, and 77 percent reported difficulty loading data into their primary EDC system.4 Veeva's 2025 research found that 97 percent of data managers who name reconciliation as a top inefficiency perform it outside clinical systems or across a mix of systems, and each round of manual review, cleaning and reconciliation takes a data manager more than 12 hours per week per study.15

Clinical trial database lock waits on three handoffs

The foundational Tufts CSDD study remains the most widely cited benchmark for data management cycle times, and its results show where the calendar goes. Sponsors took an average of 38.7 days from last patient last visit to lock, against 27.7 days at CROs, and companies that always released the study database after first patient first visit took 53.8 days to lock, against 31.4 days for those that never did.4 In both cases the difference lies in how quickly open items move between people.

Sponsors took about 11 days longer than CROs to lock a database after the last patient visit.

Exhibit 1Sponsors took 11 days longer than CROs to lock after the last patient visitMean cycle time by data management step, days (2017 survey of sponsors and CROs)
Database build to release68.3
LPLV to lock, sponsors38.7
LPLV to lock, CROs27.7
Visit to EDC entry8.1

Source: Tufts CSDD, 2017 eClinical Landscape Study, Impact Report, January/February 2018

01Data entry to query

The first clock starts at the patient visit, when a site coordinator transcribes source data into the EDC, a step that took an average of 8.1 days in the Tufts study.4 Edit checks fire only after entry, and eCOA, laboratory and imaging transfers are reconciled later still, so a discrepancy may surface well after the visit that produced it.

02Query to site response

Each query then waits in the coordinator's queue alongside scheduled visits, source documents and monitoring follow-ups. A query on non-EDC data takes 71 percent longer to resolve than one raised in the EDC and requires 34 percent more active time, because the data manager, the vendor and the site must agree on which record is correct.1

03Response to closure, coding and sign-off

An answered query still needs a data manager to review and close it, and medical coding, serious adverse event reconciliation and the medical monitor's review must be complete before investigators sign their casebooks. Each of these steps sits with a different person, and the lock date moves with whichever backlog is longest.

A connected operating model manages lock as a live backlog

Leading sponsors and CROs are beginning to treat each open query, reconciliation item and investigator signature as an obligation with an owner, a deadline and a cost for every day it stays open. A live layer reads the systems already in use, including the EDC, CTMS, eCOA and laboratory platforms, the safety database, email and site messaging, and runs a four-step loop:

  1. Detect

    an aging backlog when it first forms, such as a site whose open queries on laboratory data have doubled since its last monitoring visit.

  2. Explain

    which subjects, endpoints and lock tasks the backlog affects, with the source record behind each open item.

  3. Recommend

    a targeted response, such as a query review call with the coordinator or a corrected data transfer requested from the laboratory.

  4. Execute and track

    each agreed action until the query closes, reporting the days recovered against the planned lock date.

The same record can give sites, CRAs, laboratories, eCOA providers and coders one shared view of what is open, who owns it and when it is due, in place of status spreadsheets passed around by email. Artificial intelligence reads the data, links each query to its source and drafts query text and follow-ups, while data managers, medical monitors and investigators remain accountable for every clinical judgment and signature.

A day of delay at lock can cost a Phase III program about $$856,000

Cost of delay at database lock is the daily direct trial cost and deferred product revenue incurred when the last open queries, reconciliations and sign-offs hold back analysis, the clinical study report and the regulatory submission.

Each month that lock slips on a Phase III program could put about $$26 million at risk.

Exhibit 2. Estimated daily cost of a delayed database lock, composite Phase III program

ComponentBasisEstimate per day
Direct cost of running the trialTufts CSDD estimate of $$55,716 in direct cost per day for Phase III trials 6About $$55,700
Deferred prescription drug salesTufts CSDD estimate of about $$800,000 in sales per day of delay; assumes lock sits on the critical path to submission and the product is approved 6About $$800,000
Total cost per day of delaySum of the two componentsAbout $$855,700

Source: Segeta estimate based on Tufts CSDD data; assumptions are illustrative.

On these assumptions, a 30-day slip in lock puts about $$25.7 million at risk and a 60-day slip about $$51.3 million. The sales component applies in full only to a product that reaches approval and whose submission waits on the lock, and part of the late-phase direct cost is fixed and would be incurred in any case. Query handling carries its own bill as well, because at an estimated $$28 to $$225 per query, an average Phase III study's 96,980 queries cost about $$2.7 million to $$21.8 million to resolve.1

What data management and clinical operations leaders can do now

  1. 01Plan lock as a backlog of obligations

    From first patient onward, track every open query, reconciliation and signature against the planned lock date, and name an owner at the site, the CRO or the vendor for each.

  2. 02Price each day of the lock date

    Attach the daily cost of delay to the lock milestone, so an aging query backlog at a high-enrolling site is discussed in dollars at the weekly study meeting.

  3. 03Bring non-EDC data into the query cycle

    Reconcile eCOA, laboratory and imaging transfers on a fixed cadence throughout the study, because queries on these sources take the longest to close.

  4. 04Collect only the data the analysis needs

    Review each protocol for nonessential data points before the database build, since every field that is collected can generate queries before lock.

  5. 05Measure query aging by owner and site

    Report each week the days every open query has spent with the site, the monitor, the data manager and the coder, and set a target for each handoff.

From a follow-up booked after a consultation to an open query at database lock, care pathways and clinical trials run on the same unit of work: an obligation with an owner, a deadline and a cost for every day it stays open. Sponsors and CROs cannot make sites or vendors answer faster by mandate, yet they can see every open obligation, price it and route it to the person who can close it. Those that manage the months before lock this way are likely to reach analysis sooner and to bring evidence to regulators and patients with fewer lost days.

Questions leaders ask

How long does it take to lock a clinical trial database?

In the Tufts CSDD study of data management practice, sponsors and CROs took an average of about 36 days from last patient last visit to database lock, with sponsors at 38.7 days and CROs at 27.7 days. Companies that released their study database after first patient first visit took longer, and growing query volumes from non-EDC data can extend the final weeks further.

How long does it take to resolve a query in a clinical trial?

Veeva research published in 2025 found that queries raised in the EDC took an average of 9.6 days to resolve, while queries on non-EDC data, such as imaging, eCOA and central laboratory results, took 16.4 days. Each query passes from the data manager to the site and back, and the slowest handoff usually sets the pace for the whole study.

What is EDC query management in clinical data management?

EDC query management is the process of raising, answering and closing questions about discrepant or missing data in a trial's electronic data capture system. Edit checks and data managers raise queries, site staff answer them, and data managers review and close them, and every query must be closed or resolved before the clinical trial database can lock.

How much does a day of delay cost in a Phase III trial?

Tufts CSDD estimates the direct cost of running a Phase III trial at $$55,716 per day and the value of a day of delay at about $$800,000 in prescription drug sales. When database lock sits on the critical path to submission, a 30-day slip could put about $$25.7 million at risk on these figures.

Notes and sources
  1. Veeva Systems, "Clinical Data Industry Research," white paper, September 2025, veeva.com.
  2. Kenneth Getz et al., "Insights Informing Strategies for Optimizing the Collection of Clinical Trial Data," Therapeutic Innovation & Regulatory Science, 2025, link.springer.com.
  3. Gareth Macdonald, "A third of data collected in studies is 'nonessential,' Tufts study finds," BioXconomy, September 2025, bioxconomy.com.
  4. Tufts Center for the Study of Drug Development, "2017 eClinical Landscape Study: Assessing Data Management Practices, Performance, and Challenges," Tufts CSDD Impact Report, Vol. 20, No. 1, January/February 2018, veeva.com. Cited as a foundational study.
  5. Veeva Systems, "Industry research finds two-thirds of clinical data managers and CRAs believe current inefficiencies put data quality at future risk," press release, September 2025, finviz.com.
  6. Zachary Smith, Joseph DiMasi and Kenneth Getz, "Quantifying the Value of a Day of Delay in Drug Development," Tufts Center for the Study of Drug Development, white paper, 2024, csdd.tufts.edu.