Medication Error and Near-Miss Reporting in Indian Clinics: From Incident to Tested Improvement

Start with two separate questions:

1. Did the event reach the patient?
2. What harm, if any, has been assessed?

The WHO patient-safety framework distinguishes an incident intercepted before reaching a patient from one that reaches the patient without discernible harm, and from an incident that causes harm. These distinctions help avoid classifying every apparently harmless event as a near miss. WHO patient-safety conceptual framework.

For example, a wrong-strength pack caught before handover differs from a wrong-strength pack discovered after it was supplied. Whether the patient used it, and what happened afterwards, require further assessment.

Do not ask a receptionist to conclude that an incident caused “no harm” because the patient has not complained. Use a provisional status when exposure or outcome remains uncertain.

Also distinguish an actual event from an unsafe condition. Two confusingly similar catalogue entries may warrant correction even when no mistaken selection has yet been identified.

Immediate patient needs take priority over completing an incident form. Staff should promptly alert the responsible clinician and follow the clinic’s established escalation process whenever patient exposure or harm may be involved.

The first reporter should not have to establish a root cause.

A useful initial account answers:

• What was observed?
• When and where was it noticed?
• Which medicine, prescription, or transaction was involved?
• Had the medicine or instruction reached the patient?
• Who has been informed?
• What immediate action has already occurred?
• What remains unknown?

Prefer “the selected strength differed from the issued prescription” to “the pharmacist was careless.”

Both statements describe the same concern differently, but only the first begins with an observable fact.

Allow missing information to remain marked as unknown. A reporting process should not encourage staff to invent a batch number, reconstruct a time as if it were exact, or guess what a patient took.

Give each incident its own reference. Link it to the relevant clinical or pharmacy record rather than turning the patient’s entire chart into the incident register.

A suggested minimum structure is:

These are proposed fields, not a requirement to purchase a separate application.

Start with the smallest protected system the clinic can operate reliably. Whether it is a controlled register or dedicated software, someone must review incoming reports and know what to do when the usual reviewer is absent.

The medical record should contain the relevant clinical assessment, care decisions, patient communication, and follow-up.

The learning record should explain the operational investigation and improvement work.

Neither replaces the other.

For example, correcting an issued prescription does not explain why an incorrect selection occurred. Conversely, writing an incident report does not update the patient’s current treatment instructions.

Use references between records. Preserve the original issued document and make later corrections through the authorised process. Do not silently edit an earlier prescription to make it appear that the original error never existed.

The medical-record correction guide explains this distinction in more detail.

When a supplier or software provider needs evidence, prepare a relevant, controlled extract. An entire patient-record export is rarely the right starting point for explaining a catalogue-selection problem.

Begin the review with a timeline:

• What information was available when the decision was made?
• What did the prescription or instruction actually say?
• What appeared on screen or on the shelf?
• Which action followed?
• At what point was the discrepancy recognised?
• What prevented further progression—or failed to do so?

Then examine the conditions around that sequence.

Was the strength hidden in a truncated label? Did two catalogue entries look identical? Was an outdated printout still in use? Was work interrupted? Did one person assume another had completed a check?

These are investigation questions, not conclusions.

Separate evidence from explanation

“The two entries display the same abbreviated name” can be verified.

“The abbreviated name caused this selection” is an explanation that needs supporting evidence.

Record the distinction. If the available information cannot establish a cause, state that limitation and identify which risk can still be addressed.

Do not force a complicated event into one convenient explanation.

“Be more careful” does not specify a changed process.

A stronger action describes what will change, who will change it, and how the clinic will determine whether the change addresses the problem.

For a confusing medicine entry, a proposed action might be:

The authorised catalogue maintainer will review the two entries with the pharmacist, preserve clinically relevant distinctions, and document the approved display change. A second reviewer will test the selection using fictional prescriptions.

This is more useful than marking the incident “staff counselled” without explaining the expected difference.

Training may still be appropriate. If so, specify the skill or decision being taught and how understanding will be checked.

Keep the immediate correction separate from the longer-term action. Supplying the correct pack resolves that transaction; it does not demonstrate that the selection problem has been removed.

The Institute for Healthcare Improvement describes Plan–Do–Study–Act cycles as a way to test changes locally, observe what happens, and adapt the approach. Testing and permanent implementation are different stages. IHI guidance on testing changes.

For a clinic, the practical question is simple: can the revised process handle the situation that exposed the weakness?

Use a safe rehearsal before changing a live clinical workflow. Include the ordinary case and the confusing case.

Document:

• What the team expected.
• What actually happened.
• Whether a different problem appeared.
• What needs adjustment.
• Who will check the next version.

Do not create an intentional error in a real patient’s care to test a safeguard.

A completed software request, training session, or shelf-label change is evidence of an action. It is not, by itself, evidence that the action worked.

An AHRQ-supported primary-care study identified reporting burden, fear of punitive consequences, and doubt that reports would lead to improvement as barriers. Its findings support making reporting practical and connecting it to visible follow-through; they do not establish a guaranteed outcome for Indian clinics. AHRQ primary-care reporting study.

Explain the clinic’s reporting arrangements before an incident occurs:

• Who receives a report?
• Who can read it?
• Can the reporter provide further information privately?
• How will the team hear about resulting changes?
• How are urgent concerns escalated?

Do not promise anonymity if identities can be inferred from the record or the system logs submissions.

A fair review can examine individual actions without assuming that the first person named is the complete explanation. Learning from an event and addressing serious conduct concerns are separate responsibilities; neither should be concealed inside vague “no-blame” language.

An internal medication incident record is not automatically a pharmacovigilance submission.

WHO describes connections between medication-error learning and pharmacovigilance, while recognising their different roles. WHO medication-error reporting guidance.

In India, IPC’s PvPI resources provide routes for healthcare professionals and patients to report suspected medicine-related adverse events. Where a suspected reaction is involved, the responsible professional should assess the appropriate reporting route. IPC PvPI reporting resources.

Use the existing ADR reporting guide for that workflow.

A suspected quality defect or recall may need a separate product and batch process. Follow the clinic pharmacy recall guide where relevant.

Do not fabricate a reaction to fit an ADR form, assume an internal entry satisfies an external obligation, or adopt one universal reporting deadline for every event. Confirm applicable requirements with the relevant authority and qualified advisers.

Consider this fictional exercise. It is not a CliniKite customer incident or an example of medical treatment.

The event

An issued prescription specifies Medicine A in one strength. During preparation, a different strength is selected. The pharmacist identifies the discrepancy before the pack is handed over.

The initial record

The reporter records:

• The fictional prescription reference.
• The intended and selected products.
• The point at which the discrepancy was detected.
• That the pack had not been handed over.
• The immediate correction.
• The person notified for review.

The reporter does not guess why the selection occurred.

The review

The reviewer compares the issued prescription, relevant transaction evidence, and catalogue display.

The team finds that the two strengths are difficult to distinguish in the selection list. It also checks whether the available evidence supports a display-related explanation, rather than assuming that finding explains everything.

The action

The authorised team proposes a clearer display of the existing medicine information. It verifies that the change will not merge distinct products, alter strengths, or obscure stock history.

The action has a named owner and a planned check.

The verification

Using fictional data, another staff member tries the revised selection process, including similar-looking alternatives.

If the revised display remains confusing, the action stays open. If the demonstration works, record that limited result and decide what further observation is needed.

The final learning note should explain the changed process, not celebrate the absence of reported harm as proof that the original system was safe.

WHO cautions that incident-reporting data has important limitations and must be interpreted carefully. A reporting system captures what is noticed and submitted, not necessarily every event that occurs. WHO incident-reporting guidance.

Accordingly, a rise in reports could reflect greater willingness to report. A month with no reports does not establish that no errors occurred.

For internal management, consider questions such as:

• Which reports have not received an initial review?
• Which actions are overdue?
• Which “completed” actions still lack verification?
• Has a similar problem appeared after a change?
• Are unresolved clinical follow-up tasks being tracked separately?

These measures assess the clinic’s handling of reported concerns. They should not be presented as a validated medication-error incidence rate or used to rank individual clinicians without appropriate methodology and context.

Keep the review focused enough that it produces decisions, not merely another monthly total.

CliniKite’s public feature descriptions cover connected consultations, prescriptions, pharmacy dispensing and batch records, alongside access controls and audit-related capabilities. These can provide relevant source information when investigating a medication incident. CliniKite features.

They do not, by themselves, establish a dedicated medication-error register, automated investigation, corrective-action tracker, or automatic external reporting.

During a CliniKite demonstration, use the fictional case above and ask:

• Can we retrieve the issued prescription and relevant dispensing information?
• Can we distinguish the prescribed item from what was actually dispensed?
• What history remains after a correction?
• Which activities identify the responsible user?
• Can access and exports be limited appropriately?
• Where would the incident review and action tracking live?
• Which steps would remain manual?

Review the security and data-control information alongside the demonstration. Confirm essential requirements in the agreed implementation scope rather than treating a general feature label as proof of an entire safety workflow.

A useful medication-error reporting process does more than preserve a description of something that went wrong.

It distinguishes what is known from what remains uncertain, connects the event to the relevant care records, examines contributing conditions, and gives improvement work an owner.

Start with one clear reporting route and one fictional rehearsal. Then check whether the clinic can explain the full sequence: what happened, how it was reviewed, what changed, and what evidence supports closing the action.