In The News
-Sathish Raman
Behind the public inbox where seriously ill patients ask to join a trial
A pharma expert exposes the hidden challenges of clinical trial access through public patient inboxes. Discover how compliance-driven processes, not eligibility, often block desperate patients from life-saving research. This candid account reveals the surprising solutions and ethical dilemmas in connecting patients with trials.
Every clinical trial listing carries a way to make contact. The specialist who ran it and documented what happens next argues that access to research fails long before anyone checks whether a patient is eligible.
Every clinical trial posted on ClinicalTrials.gov carries a way to make contact. The registry is a compliance record, written for regulators and structured for their purposes. The people who read it most urgently are neither regulators nor researchers. They are patients who have run out of standard options, and the families searching on their behalf.
At the other end of that contact route, inside a large pharmaceutical sponsor, sits an inbox. Yaswanth Sai Pogadapula ran it.
“A registry listing is a front door with no receptionist,” he says.
The public argument about access to trials is almost always an argument about eligibility. Criteria are too narrow, too rigid, too slow to change. Pogadapula does not dispute that, but the failure he watched most often arrived earlier and was considerably more mundane.
“The bottleneck in trial access is rarely eligibility. It is contactability.”
What arrived ranged from a single line asking whether a study was still open to long medical histories written by people who had clearly been assembling them for months. Much of it was not a request to join anything. Side effect reports landed there, along with product complaints, physicians asking to become investigators, and academics seeking study data. Each belongs to a different function inside a drug company, and several start a legal clock on arrival.
When Pogadapula took the inbox on, the working practice had never been consolidated into a single documented procedure.
“Anyone could answer it.”
He supported and documented a decision tree that opened with a single question, whether the message was a request to join a study or something else, and branched from there. Safety reports went to pharmacovigilance, the function that monitors drug side effects. Physician enquiries went to an industry investigator database. Data-sharing requests went to an external body that reviews requests for trial data. Requests to participate were checked against the study’s actual recruitment status in the sponsor’s trial management data before any reply went out, because registry updates and site-level recruitment decisions do not always move at the same speed.
The harder problem was the reply itself, and the constraint on it surprises almost everyone outside the industry. A response may not name the person’s disease. It may not express sympathy for their situation. It may not offer any view on whether they might qualify.
Each prohibition is individually defensible. An unsolicited clinical detail can constitute a reportable safety event, and as Pogadapula puts it, “a well-meant sentence about a condition or a prognosis is an unapproved communication, and in some cases a reportable event.” Encouragement can be read as a promise. Any assessment of eligibility, however gentle, is “practising medicine from an inbox.”
The difficulty is that the person receiving the reply is frightened, and a correct message can land as a cold one. His answer was to stop treating tone as a compliance risk to be minimized and start treating clarity as the form compassion takes here. Tell the person exactly what is true, exactly what they can do next, and where else to look, with no false hope and no filler.
“Dignity, in that inbox, is precision.”
The same reframing underpins his defence of something usually read as institutional coldness. Standardized replies, he argues, are the only fair way to answer at scale. An approved response set is what guarantees that the articulate patient and the frightened one, the first email of the day and the fortieth, receive the same accurate information. “Improvisation feels more human and delivers less equitable results,” he says.
The contactability failures showed up further down the process, after a candidate had already agreed to be referred. Self-referrals went through a short pre screening with a third-party provider, which confirmed the chosen site was recruiting before transmitting contact details. By Pogadapula’s account, the most common reason that chain broke was not clinical. The contact route recorded for the site had decayed. An investigator moved. A coordinator left. A number was retyped wrong once and stayed wrong.
Referrals also reached the correct site and stopped there, because nobody opened them. Non-receipt was made a tracked event, routed back through the sponsor to the study team, rather than a silence nobody owned. For decayed details, the rule ran against instinct: staff never researched a replacement themselves, because sending patient information to an unverified destination is worse than sending it late. The correction was requested through the sponsor and tracked to closure. “Slower, and correct.”
For years, the compliant way to move that information was a fax machine, because it satisfied the privacy requirement and sites could receive it.
“A referral could fail because of a busy signal.”
He is unusually candid about what his own design costs. A meaningful share of messages, he says, arrived with no study number and no title, usually because the sender had typed the address in manually rather than clicking through from the listing, which would have carried the identifiers automatically. Those people wanted help with one specific trial and there was no way to know which, and no way to ask again later. The inbox kept anonymized per-day identifiers rather than names, and held them for a fixed window. “That is the part of the job that stays with you,” he says.
He treats that as a genuine tension rather than a solved problem. Keeping as little as possible about an inquirer, for as short a time as possible, is the correct privacy posture. It also means that when you cannot help someone, you cannot check back.
“I do not think our industry has been honest enough that privacy-protective design carries this cost.”
The piece of the work he is most confident was right concerns messages that should never have received a referral at all. Some studies involve conditions where an inbound message describing personal symptoms is itself a signal of acute risk. For those, the correct response directs the person to seek immediate care. That exception was written into the process as a path of its own, with explicit recognition triggers, rather than leaving it to an individual’s judgement at eight in the morning. Most designs he has seen treat that class of message as an edge case to be escalated, alongside inquiries from minors and adverse event reports arriving in disguise. “The cost of missing one is not a metric,” he says.
Documenting the operation to that level made it transferable. Daily work moved to a contracted operations team while accountability stayed with the sponsor, on the principle that a contracted team should never have to improvise on a patient-facing message.
None of it produces feedback. You confirm a site is recruiting, send the referral, close the record. Whether the person was screened, enrolled or helped never comes back, and the process is deliberately built so that it should not.
“Nobody in that inbox ever learns the ending.”
It is a strange asymmetry. The machinery for publishing what a trial is and what it found has become sophisticated. The machinery for answering the person who reads it has not.
“Trial recruitment infrastructure has been extensively optimized for the sponsor’s data obligations and comparatively little for the last hundred meters to a human being.”
