What a complainant's advocate asks for first
In a consumer complaint or a council inquiry, the first request is almost never for the clinical note. It is for everything around it: proof of who conducted the consultation and under what registration, proof that the patient was who you say, the consent, the mode, the timestamps, and the audit log showing who accessed the record and when. The clinical reasoning is examined afterwards, once the procedural picture is established, and by then a weak procedural record has already framed the case.
This matters because hospitals invest their documentation effort almost entirely in the clinical content. A teleconsultation note may be clinically excellent and still leave every one of the procedural questions unanswered, because the fields that answer them were never designed into the encounter. The distinguishing feature of telemedicine litigation is that the surrounding facts are contested in a way they rarely are for a physical visit, where attendance itself is not in dispute.
Design backwards from that. Take the list of what a complainant would ask for, check whether your system can produce each item for a consultation conducted last Tuesday, and fix the ones it cannot. This exercise takes an afternoon and it is the single most useful thing a medical administrator can do for a teleconsultation service.
Identity: proving you knew who you were treating
The record must show, without relying on anyone's memory, how the patient was identified. In practice that means storing the evidence rather than an assertion: which UHID the consultation was booked against, whether a one-time password to the registered mobile was verified and at what time, whether the practitioner made a visual confirmation on video, and the identity of anyone else present. A field reading identity verified: yes is a conclusion. What you need is the basis for it.
Caregiver consultations are the weak point. When a son describes his mother's symptoms from another city, the record must say that this is what happened, name him, state the relationship, and note whether the patient was present. Notes written in the first person as though the patient spoke, when they did not, misdescribe the encounter. If that is discovered later it damages the credibility of everything else in the file, including the parts that were accurate.
The practitioner side is equally part of identity. The record and the prescription must carry the practitioner's name, qualification, and registration number, and the audit log must show that this practitioner, not a junior using a shared login, conducted the session. Shared credentials are still common in Indian hospital systems and they destroy attribution completely. There is no documentation fix for a shared login.
Identity evidence the record should hold, not assert
- UHID the consultation was booked against, and how the match was made
- One-time password verification event with its timestamp
- Whether visual confirmation occurred, on which mode, at what point
- Name and relationship of any caregiver who participated or substituted
- Practitioner registration number and an individual, non-shared login attribution
Consent and mode: the two fields most records lack
Consent has to be produced with its context to be worth anything: which version of which notice, in which language, at what time, granted by whom. A stored boolean is evidence of a database state, not of a conversation. The same record should distinguish the clinical consent to be treated remotely from the data-processing consent required under the DPDP Act 2023, because a challenge may go to either and a single undifferentiated flag answers neither cleanly.
Mode is the field that most often decides whether a prescription was permissible. If the consultation is recorded as video and the prescription includes a medicine that may only be prescribed at a first consultation over video, the record supports itself. If mode is absent, or if it records the booked mode rather than the actual session, then the prescription is unsupported and the argument turns on recollection of a call from eighteen months ago. Store the actual session mode, and store any change during the session.
Both fields should be immutable once the encounter closes. A consent record or a mode field that can be edited afterwards without a versioned history invites the suggestion that it was edited afterwards, and the suggestion is enough to be damaging even where nothing improper occurred. Make the amendment path an addendum with its own timestamp and author rather than an overwrite.
Consent and mode fields to store per encounter
- Notice version and language actually displayed to this patient
- Separate clinical and data-processing consent events with timestamps
- Who granted consent, and on what authority if not the patient
- Session mode at start, plus any recorded change and the time it changed
- Immutability, with amendments recorded as timestamped addenda
The clinical note: what teleconsultation adds
A teleconsultation note has to carry everything an in-person note carries, plus an account of the limitation. The examination section should state what could and could not be assessed remotely, because the standard a clinician is held to accounts for the information reasonably available to them, and a note silent on limitation implicitly claims a completeness that did not exist. A line recording that abdominal examination was not possible remotely and the plan accounted for that is worth a great deal later.
Safety-netting advice is the other addition and it is the one most often missing. What the patient was told to watch for, what to do if it occurs, and where to go, recorded as delivered rather than as a template. Where a consultation ends in advice rather than a prescription, the safety-netting is essentially the whole clinical product and should be documented with corresponding care. Generic template text that appears identically on every note carries little weight because it evidences the template, not the conversation.
The clinical reasoning should also record why teleconsultation was appropriate for this presentation. That is a one-line judgement and it addresses the argument that the patient should have been asked to attend physically. If the practitioner considered and rejected physical review, saying so converts a potential omission into a documented clinical decision.
The prescription as evidence
A prescription is the most likely document to be produced against you because the patient has a copy of it. It must carry the practitioner name, qualification and registration number, the patient identification, the date, and complete medicine details, and it should be attributable in a way that resists the claim that it was altered. A digital signature under the Information Technology Act 2000 gives a materially stronger position than a photographed signature on a phone image.
Keep one prescription of record. Where a hospital runs a standalone teleconsultation platform alongside its main system, two versions of the same prescription frequently exist with small differences in formatting, timestamp, or content. Producing two documents that both purport to be the prescription is a genuinely bad position to be in, and it happens without anyone deciding it should.
Where the prescribing decision hit a restriction, record it. A note stating that a controlled medicine could not be prescribed remotely and the patient was advised to attend in person converts a refusal into a documented compliance with the guidelines. Without that line, the record simply shows a patient who asked for something and did not get it, which is the raw material of a complaint.
“The case turned on a single sentence in the note saying the patient had been told to come to emergency if the pain returned. Everything else was contested. That line was not.”
Retention, immutability and the audit trail
Retention for teleconsultation records should follow the same policy as your other outpatient records rather than being decided separately. The ethics regulations set a baseline for indoor case records, state Clinical Establishments Act rules impose their own requirements, and medico-legal cases and paediatric records carry longer practical exposure because limitation periods run differently. Take the longest applicable period, apply it uniformly, and write it in a retention schedule the records department can actually follow.
Electronic records carry an evidential formality that hospitals rarely prepare for. Producing computer output as evidence has long required an accompanying certificate about the system that generated it, under Section 65B of the Evidence Act and carried into Section 63 of the Bharatiya Sakshya Adhiniyam 2023. Someone in your organisation has to be able to speak to how the system works, that it was operating properly, and that the output is a faithful reproduction. Decide who that person is before you need them.
The audit trail is what makes immutability credible. It should show creation, every amendment with author and time, and every access, including views by people who were not part of the care. HealUDoc activity logs can carry that trail against the encounter, which is what allows a hospital to answer who saw this record with a report rather than an assurance.
Audit trail elements that carry evidential weight
- Creation event with author, role and timestamp from a controlled clock
- Every amendment recorded as an addendum, never as an overwrite
- Read access logged, including access by staff outside the care team
- Export and print events, with the user and destination recorded
- A named individual able to certify how the system operates, if required
Where hospitals lose these cases
The recurring failure is not bad medicine. It is a record that cannot describe its own circumstances: no mode, no consent artefact, no identity basis, a template note identical to two hundred others, and a prescription that exists in two versions. Each of these is individually minor and collectively they establish an impression of a service that was not being run carefully, which colours how the clinical decisions are then read.
The second failure is the unmanaged channel. Consultants who advise patients over personal messaging outside the hospital system generate clinical records that the hospital does not hold and cannot produce, while the patient holds a complete transcript. When a complaint arrives, the hospital's file contains a fraction of what actually passed between the doctor and the patient. Either bring the channel into governance or state its exclusion in writing and enforce it.
The third is late correction. A note amended after a complaint is received, even for an entirely innocent reason, is extraordinarily difficult to defend if the system does not clearly show it as a timestamped addendum. Train clinicians that the correct response to noticing an omission is an addendum recorded openly, and make that the easiest thing to do in the software. The path of least resistance is what people will take under pressure, so it needs to be the defensible one.
