How AI Can Read the Trend Line in Your Chronic-Disease Records
A person with type 2 diabetes for five years has, in most cases, fifteen to twenty HbA1c readings, thirty to forty fasting glucose readings, dozens of BP measurements, several lipid profiles, and a stack of prescriptions. Any single one is a data point; the pattern across all of them is the story. Reading the pattern is what an endocrinologist does in ten minutes at each appointment. AI can produce a plain-language summary of the same pattern for the patient to read in advance.
What AI is good at, in chronic-disease trends
- Summarising direction — is HbA1c trending up, down, or stable across the last two years?
- Identifying inflection points — when did the numbers start changing?
- Aligning trends with events — did the change coincide with a medicine adjustment, an illness, or a life event?
- Producing plain-language explanations — 'your average BP has drifted up 8 mmHg over the last year despite the same medicine'.
- Preparing focused questions for the specialist.
What AI cannot do — and where a doctor still owns the decision
- Diagnose the reason for a trend change (may be treatment failure, may be adherence, may be a new condition).
- Decide the next treatment step.
- Weigh the trend against the specific clinical picture in front of them.
- Distinguish a lab artefact (a bad batch of reagents at one lab) from a real change.
The kinds of trends the tool can highlight
- Condition: Diabetes · Trend to look for: HbA1c drifting up over 6 months · Typical implication: Adherence, treatment intensification, or new insulin resistance
- Condition: Hypertension · Trend to look for: BP rising over 3 months on same medicine · Typical implication: Medicine effect fading, dose adjustment needed
- Condition: Chronic kidney disease · Trend to look for: Creatinine trending up · Typical implication: Rate of decline predicts dialysis timing
- Condition: Thyroid · Trend to look for: TSH cycling above and below target · Typical implication: Under- or over-treatment
- Condition: Lipids · Trend to look for: LDL trending up despite statin · Typical implication: Adherence check or dose escalation
- Condition: Weight · Trend to look for: Slow steady rise over 2 years · Typical implication: Metabolic drift, common precursor to other issues
The specific way to ask
Structured requests get structured answers:
- 'Here are my HbA1c readings for the last 5 years. Summarise the trend and any inflection points.'
- 'These are my BP readings across 12 months on the same medicine. Is the trend stable, improving, or worsening, and by how much?'
- 'Here are three years of lipid profiles. What has happened to LDL and HDL?'
Attach the data (as text if the app allows, or the actual reports if it accepts them). The response is a summary you can take to your specialist.
The privacy consideration
Trend analysis needs data. The same privacy principles apply as with any AI health use:
- A record-integrated AI can analyse trends without exposing identifying information to third parties.
- A general-purpose AI needs the data pasted in; consider anonymising ('a 55-year-old with these readings') rather than including the patient's identifying details.
- For sensitive conditions, discuss with your doctor what tools they use before uploading data.
The pattern of use that works
Before every specialist visit for a chronic condition:
- Pull the last 12 months of relevant readings into a text summary.
- Ask AI to describe the trend and prepare 3-5 questions the trend raises.
- Take the summary and questions to the visit.
- The visit itself becomes focused on the decisions, not the review of numbers.
Fifteen minutes of preparation. A better ten-minute visit. The trend-line, made visible, is often what the visit was supposed to be about anyway.
References
Free for 90 days, no card needed. After that, keeping the record costs ₹349 for the year.
General information, not medical advice. Always talk to a qualified doctor about your own care. Where this and your doctor disagree, your doctor is right.