Chilling Failure at a Village Milk Collection Centre

Decision: do not load it as normal supply. Segregate the milk, inform the plant before the tanker leaves the depot, and let the plant decide in writing whether it will accept it at a reduced rate for a fully heat-treated use.

Interpreting the results

• The 4-hour morning result describes milk sampled before the breakdown; it is a baseline only and says nothing about growth between 09:10 and 16:00.
• The 40-minute evening result is in the poor class and indicates a heavy bacterial load. It is a proxy measure, not a count, and it cannot identify the organisms or their heat-stable metabolites, so it cannot certify safety.
• The decisive fact is the time-temperature history: about 6.5 hours at 9-14 °C, which is squarely in the growth range for mesophiles and psychrotrophs. Even if the milk is later pasteurised, heat-stable lipases and proteases and possible staphylococcal or Bacillus toxins are not removed.

Why not blend it with tomorrow's milk: blending spreads the load across the whole tank and converts a single rejected consignment into a plant-wide quality and payment problem. Rejection of one tanker is cheaper than a downgraded week.

What to record: the breakdown time, the temperature curve, the MBRT results and the plant's written acceptance or rejection.

Same-day recovery

• Send a worker to the ice plant 9 km away and buy crushed ice to precool the tank contents toward 6 °C; every hour near 14 °C adds bacterial load.
• Restart the cooler and run it continuously until the set point is met; if the plant still refuses the load, sell to the nearest halwai or dairy at a documented discount rather than dumping, subject to the plant's approval.
• Keep the rejected milk in a separate, labelled tank; do not use the same cans for the next collection until they are washed and sanitised.

Preventive changes, ranked

1. Standby power (diesel generator or solar with battery) with automatic changeover and a temperature alarm that sends an SMS to the manager.
2. An ice bank or a larger insulated storage tank to buy 4-6 hours of hold time.
3. Precooling at the collection point: move the collection window earlier, or chill cans on arrival in a plate cooler before they enter the bulk tank.
4. Ingredient and process hygiene: a pre-breakdown MBRT of 4 hours is mediocre for fresh morning milk, so audit can washing, water quality, udder preparation and farmer training.
5. Written cold-chain SOP with hourly temperature logs and a documented escalation route.

Response: the two results measure different things. The morning sample was drawn before the compressor tripped; the evening sample was drawn from milk that had spent 6.5 hours between 9 °C and 14 °C. The evening result is the one that describes the milk being offered for acceptance, and the plant's own specification (2 hours or more, 6 °C or below) is failed on both counts. An acceptable earlier result does not overwrite a later failure; that logic would make every cold-chain specification meaningless.

Also note the limitation in the other direction: MBRT is slow (hours) and insensitive at low counts, so it cannot be used to prove good hygiene.

On 13 June

• Verify the cooler has returned to 4 °C and log it hourly.
• Decontaminate the bulk tank and the cans (the rejected load will have left a heavy biofilm risk).
• Draw a fresh sample and run MBRT and a clot-on-boiling test before loading the tanker.
• Hold a short meeting with farmers and staff, record the incident in the centre's register, and give the plant a written undertaking with the corrective-action timeline.
• If the same fault recurs, expect the plant to suspend collection from the centre; the standby generator is now a commercial necessity, not a luxury.