From Day 1 of the protocol, Patrick removed Galvus Met entirely β under clinical supervision, with daily monitoring to ensure his glucose did not drop dangerously. He adopted a 16:8 intermittent fasting schedule: eating window from 12pm to 8pm, fasting for the remaining sixteen hours. No breakfast. No morning snacking. Water, black coffee, and nothing else until noon.
The dietary shift was not a generic low-carbohydrate template imported from a European clinical trial. It was adapted specifically for the Ghanaian food environment β the ingredients available in Kasoa markets, the cost constraints, the family cooking context, the social expectations around food. Less rice, less processed carbohydrate, more protein, more vegetable matter, less sugar. Nothing that required a foreign supermarket or a specialist nutrition budget.
Every evening before bed, Patrick logged his blood glucose reading. Every morning before eating, he logged his fasting reading. The data was reviewed by his clinical team and formed the basis of weekly adjustments to his protocol. This is not a generic plan followed in isolation. It is a living protocol, adjusted in real time by people who can read the numbers and respond to them.
Scientific integrity β the discovery nobody else would report
When the numbers lied β and the protocol caught it
During the protocol period, Patrick observed an anomaly. On evenings when he drank a small quantity of Guinness stout β considered by many in Ghana to be a βhealthyβ drink β his morning fasting blood glucose was paradoxically lower than on evenings when he did not drink. This is a documented pharmacological phenomenon: alcohol inhibits gluconeogenesis in the liver, artificially suppressing fasting glucose readings while the liver is metabolising alcohol.
The reading appeared better. The underlying metabolic situation was not better. If Patrick had not been logging evening readings alongside morning readings β and if his clinical team had not been watching the pattern β this could have been misread as a positive dietary signal.
Patrick identified the correlation, reported it to his clinical team, and verified the mechanism against the published literature on alcohol and gluconeogenesis inhibition. He then removed alcohol from his protocol entirely. His readings stabilised. The data became accurate again.
Patrick stopped immediately. He reported the discovery publicly β because if one person in Ghana is unknowingly misreading their glucose data because of Guinness, there are almost certainly others.
The protocol is supervised. Patrick's nurse contacts him regularly β not when something goes wrong, but proactively, as standard practice. His PA reviews his trends. If a reading is unusual, the response comes before Patrick has had to worry about it. This is what clinical oversight looks like when it is genuinely implemented β not a six-monthly appointment with a ten-minute window, but a continuous relationship with people who are watching your numbers alongside you.
The frequency of monitoring is part of why the protocol works. Most people with Type 2 Diabetes in Ghana have their HbA1c measured every three to six months. Patrick measures his fasting blood glucose every morning. That daily data loop β between patient, readings, and clinical team β creates the feedback mechanism that makes real-time adjustment possible.