The hearing is in a borrowed conference room, four months after the incident. The employee is present with a union representative. Nobody in the room disputes that a bac test was conducted, and nobody is arguing about whether the man looked unsteady that day. The questions are narrower and colder than that: which device was used, what its serial number was, when it was last calibrated, who watched him before he blew, whether he was tested twice, and where the signed slip is.
The safety officer has a photocopy of a handwritten register entry and a memory of the conversation. That is the whole file. The reinstatement follows.
That is the real subject of a BAC test at a workplace. Not the number on the screen, but whether the number can be defended long after the shift it came from. Screening is easy. Defending it is the discipline.
This article is for the person who will have to explain the result in that room: the safety head, the HR lead, the transport manager, the site in-charge whose signature is on the incident register.
What a BAC test actually measures
BAC stands for blood alcohol concentration: the quantity of ethanol present in the blood, usually stated as milligrams per 100 millilitres of blood, or as a percentage. In India the reference point in the Motor Vehicles Act, 1988 is 30 mg of alcohol per 100 ml of blood, which is the same thing as 0.03%.
A breath analyser does not sample blood. It samples deep lung air and converts the alcohol found there into an equivalent blood figure using a fixed ratio. That conversion is standard, accepted and used by enforcement worldwide, but it is a conversion, and it is the first thing a defence will poke at. Which is why the procedure around the breath sample matters more than the arithmetic inside the device.
Blood and urine testing exist and are more direct, but they are laboratory events: consent, a qualified collector, transport, turnaround time. No factory holds a production line for that. Breath screening on site, with a documented escalation path, is the workable arrangement.
Breath, blood and urine compared
Method What it tells you Practical use in an Indian workplace
Breath analyser Current blood alcohol equivalent, instantly Primary screening tool; repeatable, non-invasive, on the spot
Blood test Direct concentration at the time of draw Escalation and confirmation, at a recognised laboratory
Urine test Recent consumption, not current impairment Poor fit for fitness-for-duty decisions; lags actual BAC
What counts as a normal or acceptable BAC level
A normal BAC test result is 0.00%. Alcohol is not naturally present in blood in any meaningful quantity, so anything the device reports above zero is something you have to account for.
There is no such thing as a good BAC score for someone about to work. The 0.03% road limit is a threshold for prosecution, not a licence for a forklift operator or a crane rigger to arrive at 0.02%. Most Indian industrial and logistics operations that run screening set the standard at zero for anyone on duty, on a vehicle, on scaffolding or in a control room, and let the road limit be what it is: a separate, lower bar set by law for the general public.
Be explicit in the policy about which figure you enforce. A dispute frequently turns on the fact that the company acted against someone who was under the legal driving limit, and the only answer to that is a written, communicated, uniformly applied internal standard.
Reading What it means at a workplace screening point
0.00% Clear. Log it and move on.
Any reading above 0.00% Wait, retest, follow the escalation path. Not automatically a positive.
At or above your internal threshold on confirmation Positive. Remove from duty, document, follow disciplinary procedure.
At or above 0.03% (30 mg/100 ml) Also a road-law issue if the person then drives.
The fifteen-minute wait decides more disputes than the device does
Alcohol lingers in the mouth and throat after a drink, after mouthwash, after certain medicines, after a burp or reflux. That residue is far more concentrated than what is in the blood, and if it reaches the sensor it inflates the reading badly.
The control is a deprivation period. Nothing to eat, nothing to drink, no smoking, no gutkha, no mouth freshener, no inhaler for fifteen minutes before the sample, under observation. Not fifteen minutes on the person's word. Fifteen minutes where somebody watched.
Write the observation into the form. Time in, time of sample, name of the observer. A reading taken three minutes after a chai break is not a result you can take to a hearing.
Mouth alcohol and the other things that inflate a reading
Know the challenges before they are made to you:
• Mouth alcohol from recent drinking, mouthwash, alcohol-based hand sanitiser used near the face, ayurvedic tonics or cough syrups.
• Residual vapour in the device from the previous subject, if units are used back to back without a purge interval or a fresh mouthpiece.
• Ambient alcohol in the air near a paint shop, a solvent store or a chemical dosing area. Screen away from the source.
• Improper blow — a short, weak or interrupted breath that never delivers deep lung air. Good devices reject it; cheap ones report something anyway.
• Sensor drift in an uncalibrated unit, which is the single easiest attack on your evidence.
Each of these has an answer, and the answer is procedural, not technological. Wait, observe, purge, change the mouthpiece, retest.
Never act on a single reading
One reading is a screen. A decision needs confirmation. Build a two-step sequence and make it non-negotiable:
• Initial screen before task allotment or vehicle handover.
• If not zero, a fifteen-minute observed wait, then a second test on a fresh mouthpiece.
• If the second test also reads positive, that is the actionable result. Consider a third test on a different device where the consequence is termination-grade.
• Offer the person the option of a blood test at a recognised laboratory, at a stated timeframe, and record whether they accepted or declined.
The second reading does two things. It clears people who were caught out by mouth alcohol, which protects your credibility. And it converts a disputed number into a pattern, which is much harder to argue with.
Calibration history is the first document an auditor asks for
A breath analyser is a measuring instrument. Fuel cell sensors are stable but they drift; semiconductor sensors drift faster. Calibration against a known alcohol standard resets the device to truth, and the certificate is what proves the reading was trustworthy on the day it was taken.
Follow the manufacturer's stated interval, whether that is expressed in months or in number of tests, whichever comes first. Do not let a unit run past due and keep it in service. A single expired calibration in a spreadsheet can undo every result that device produced after that date.
Keep, for each device: serial number, date of purchase, calibration certificates in sequence, service records, and the date it was withdrawn from use. When a result is challenged, you should be able to produce the certificate covering that exact date within minutes.
What a chain of records looks like
A chain of records is the paper trail that connects a person, a moment, a device and a decision without a gap. If any link is missing, the whole thing is arguable.
Link What it must show
Policy The written rule, the threshold, the consequences, and proof it was communicated
Consent and notice That the employee knew screening applied, usually via appointment letter or standing order
Test log Date, time, location, employee name and ID, operator name
Device identity Serial number of the unit used, plus mouthpiece changed
Procedure Observed deprivation period, time of first and second readings
Result Both readings, printed or photographed, not just handwritten
Employee acknowledgement Signature or refusal to sign, witnessed
Action taken Removal from duty, transport home, medical referral, disciplinary reference
Retention Where the record is stored and for how long
Devices that print a slip or store readings against a serial number make this chain far shorter to assemble. Where you screen at volume, a unit with onboard memory and printer support turns record-keeping from a clerical burden into a by-product of testing.
Handling the moment the result is challenged
Stay procedural. Do not argue about whether the person seems drunk. Impairment is an opinion; a documented two-stage reading on a calibrated device is a fact.
Give the person the retest. Record their explanation in their own words. Arrange transport home rather than letting them drive, and note that you did. Refer to the policy clause by number in the incident note. Nothing in that sequence requires you to be certain about anything except what the procedure produced.
If the person refuses to blow
Refusal is not an absence of evidence, provided your policy says what refusal means. State in the policy that declining a lawful screening test is itself a disciplinary matter, name the consequence, and record the refusal with a witness signature and the time. A refusal that was never defined in writing is the one gap that turns a clear case into a negotiation.
Screening at home and outside the workplace
Families and individual drivers ask how to run a BAC test at home. A personal breath analyser is the practical option: wait at least fifteen to twenty minutes after the last drink, blow steadily until the device signals completion, and read the figure.
Treat the result as a warning, not a permission slip. A personal-grade device is built to tell you that you should not drive, not to prove that you can. If it shows anything above zero and you have keys in your hand, the answer is already no. The same logic applies at a wedding venue or a banquet hall late at night: the device is there to stop a car leaving, not to certify that one may.
Choosing equipment you can stand behind
For site screening and any result that may end in disciplinary action, use a fuel cell sensor device with a documented calibration schedule, replaceable mouthpieces and stored or printable results. Semiconductor units are acceptable for informal self-checks and low-stakes awareness, not for evidence.
Buy from a supplier who will handle calibration on schedule and issue certificates you can file. The device is the smaller part of the purchase. What you are actually buying is the ability to walk into that borrowed conference room months later with one folder that answers every question in it.