Health Metrics Dashboard: Blood Test Trends to Track

What a health metrics dashboard adds to lab reports

Most lab portals show a result, a reference range, and maybe a red arrow; they rarely show whether your ferritin has fallen 45% in 9 months or whether your ALT rises every time your triglycerides rise. A good dashboard should display the current value, prior value, date interval, absolute delta, percentage delta, and whether the change exceeds expected biological variation.

I see this pattern weekly: a patient brings 4 reports from 3 laboratories, each with different units and layout. Once those results become a single blood test history, the conversation changes from “is this high?” to “why did this marker move now?”

Kantesti’s neural network reads blood test PDFs and photos, standardizes units where appropriate, and places values into a blood test timeline in about 60 seconds. That does not replace diagnosis; it gives you a cleaner starting point before a rushed 10-minute appointment.

Why trends beat single red and green flags

Creatinine may shift from 0.82 to 1.03 mg/dL after dehydration or creatine use, while hemoglobin can change 0.5–1.0 g/dL from plasma volume alone. A dashboard should mark likely noise differently from repeated directional change, which is the part clinicians tend to trust.

A 52-year-old marathon runner can show AST of 89 IU/L after a hard race, and that number looks alarming until CK is also high and ALT is only 34 IU/L. For readers who want the mechanics, our lab trend graph guide explains how slopes, swings, and drift behave across visits.

The practical cut I use is simple: one unexpected result asks for context; two similar results ask for a plan. Three results moving in the same direction over 6–18 months deserve a direct question, even if the laboratory has not put anything in red.

CBC trends patients should not ignore

Adult hemoglobin is commonly about 13.5–17.5 g/dL in men and 12.0–15.5 g/dL in women, but your personal baseline matters. A drop from 15.1 to 13.4 g/dL in a man is often more meaningful than a one-time “normal” label.

MCV below 80 fL usually suggests microcytosis, often iron deficiency or thalassemia trait, while MCV above 100 fL suggests macrocytosis from B12, folate, liver disease, alcohol exposure, thyroid disease, or medication. If the differential looks confusing, our CBC differential guide separates absolute counts from percentages, which matters far more than many patients realize.

RDW above about 14.5% can rise early in mixed anemia, sometimes before hemoglobin falls below range. In my experience, a rising RDW plus falling ferritin is one of those “do not wait a year” patterns, especially in heavy menstrual bleeding, endurance athletes, and people on long-term acid suppression.

Kidney and electrolyte signals that need context

eGFR above 90 mL/min/1.73 m² is often reported as normal, 60–89 can be normal for some older adults, and below 60 for 3 months is the threshold clinicians take seriously. The KDIGO 2024 guideline also treats urine ACR of 30 mg/g or higher as kidney damage, even when creatinine looks reassuring (KDIGO, 2024).

BUN rises with dehydration, high protein intake, gastrointestinal bleeding, steroids, and kidney impairment, so a BUN/creatinine ratio above 20:1 is a clue rather than a diagnosis. For a plain-English explanation of filtration estimates, see our eGFR guide.

Potassium deserves less casual interpretation than most dashboard numbers. Potassium below 3.0 mmol/L or above 6.0 mmol/L can be dangerous, particularly with weakness, palpitations, kidney disease, ACE inhibitors, ARBs, spironolactone, or potassium supplements.

Lipid trends show risk better than one cholesterol result

LDL-C below 100 mg/dL is often considered reasonable for lower-risk adults, but targets become stricter after diabetes, known cardiovascular disease, chronic kidney disease, or high coronary calcium. LDL-C at 70 mg/dL may be acceptable for one person and too high for another with prior myocardial infarction.

Triglycerides below 150 mg/dL are generally normal, 150–499 mg/dL are elevated, and 500 mg/dL or higher raises pancreatitis concern as well as metabolic risk. When triglycerides exceed 200 mg/dL, the AHA/ACC guideline recognizes ApoB as a useful secondary marker because particle number may reveal risk that LDL-C alone misses (Grundy et al., 2019).

I often watch non-HDL-C because it does not collapse when the patient forgot to fast. Patients comparing particle-related risk can read our ApoB blood test guide before asking whether ApoB or Lp(a) belongs in their next panel.

Glucose, A1c, and insulin trends catch risk early

Fasting plasma glucose below 100 mg/dL is usually normal, 100–125 mg/dL suggests impaired fasting glucose, and 126 mg/dL or higher can diagnose diabetes if confirmed. The ADA Standards of Care also allow diagnosis with a 2-hour glucose of 200 mg/dL or higher on oral glucose tolerance testing (ADA Professional Practice Committee, 2024).

The evidence here is honestly mixed on fasting insulin cutoffs because assays vary, but I pay attention when fasting insulin is above 10–15 µIU/mL alongside triglycerides above 150 mg/dL and HDL-C below 40 mg/dL in men or below 50 mg/dL in women. Our A1c versus fasting sugar article explains why anemia, kidney disease, hemoglobin variants, and recent blood loss can make A1c misleading.

Kantesti AI pairs glucose markers with CBC clues because A1c is a red-cell-based test, not just a sugar test. A 6.1% A1c with ferritin of 8 ng/mL and MCV of 76 fL deserves a different conversation from the same A1c in a person with normal iron markers.

Liver enzyme patterns are more useful than ALT alone

ALT is more liver-specific than AST, while AST also rises from muscle injury, hemolysis, and strenuous exercise. A 2:1 AST-to-ALT ratio can raise concern for alcohol-associated liver injury, but I would still check CK, platelets, GGT, and the history before making that leap.

ALP above about 120 IU/L with GGT above 60 IU/L points more toward hepatobiliary or bile duct sources than bone. A deeper pattern-based walk-through is in our liver function test guide.

Bilirubin deserves its own lane. Total bilirubin of 1.3–2.5 mg/dL with normal ALT, AST, ALP, and CBC often suggests Gilbert syndrome, particularly if it rises during fasting or illness; bilirubin with dark urine, pale stools, or high ALP is a different matter.

Nutrient trends reveal depletion before deficiency is obvious

Ferritin below 15 ng/mL is highly suggestive of iron deficiency, but many symptomatic patients with hair shedding, restless legs, or heavy periods feel unwell at 15–30 ng/mL. A fall from 80 to 28 ng/mL over 12 months is not “fine” just because the lab range starts at 12.

Vitamin B12 below 200 pg/mL is often treated as deficient, while 200–350 pg/mL can still be borderline if neuropathy, glossitis, high MCV, or high methylmalonic acid appears. Our low ferritin with normal hemoglobin guide shows why iron stores often decline before anemia arrives.

25-OH vitamin D below 20 ng/mL is commonly deficient, 20–29 ng/mL is often insufficient, and 30 ng/mL or higher is adequate for many bone-health goals. Clinicians disagree on whether routine targets should exceed 40 ng/mL; in my practice, I avoid chasing high numbers unless there is a clear bone, malabsorption, or parathyroid reason.

Thyroid and hormone timelines need timing notes

TSH is often reported normal around 0.4–4.0 mIU/L, though some European laboratories and pregnancy protocols use lower upper limits. A TSH of 4.8 mIU/L with low free T4 means something different from TSH 4.8 with normal free T4 and positive TPO antibodies.

Biotin supplements at 5–10 mg daily can distort some thyroid immunoassays, often making TSH look falsely low and free T4 or T3 falsely high. Patients comparing thyroid values should read our TSH range guide before assuming a medication dose has suddenly stopped working.

For sex hormones, timing is everything. Testosterone is usually highest between 7 and 10 a.m., progesterone is interpreted about 7 days after ovulation, and estradiol can swing several-fold across a menstrual cycle.

Inflammation markers need a reason and a retest date

ESR is slower and less specific than CRP; it rises with age, anemia, pregnancy, kidney disease, and immunoglobulin changes. A high ESR with normal CRP is not rare, and it often sends clinicians toward chronic inflammatory, autoimmune, protein, or anemia patterns rather than acute infection.

CRP can fall by roughly 50% per day after a bacterial infection starts improving, though this varies widely by source and treatment. Our guide to inflammation blood tests compares CRP, ESR, ferritin, fibrinogen, and CBC changes without pretending any one marker is perfect.

One pattern I do not ignore is CRP plus platelets rising together over several tests. Platelets above 450,000/µL can be reactive after infection or iron deficiency, but persistent thrombocytosis needs a clinician’s eyes on the whole timeline.

Family dashboards can reveal repeated risk patterns

Lp(a) is mostly inherited and changes little after childhood; a level above 50 mg/dL or 125 nmol/L is commonly treated as a cardiovascular risk enhancer. If a parent had a heart attack at 49 and a child has LDL-C 155 mg/dL, I interpret that LDL differently than I would in a low-risk adult.

Ferritin patterns also cluster in families, but not only through genetics. Diet, heavy periods, endurance sport, acid-suppressing medication, and donation habits can repeat inside a household; our family records guide shows how to track relatives safely without turning one person’s result into another person’s diagnosis.

Kantesti AI includes Family Health Risk features because family context often changes the priority of follow-up. A borderline TSH in a person with two relatives with autoimmune thyroid disease deserves a different follow-up cadence than the same TSH in isolation.

Data quality decides whether a trend is real

Glucose can be reported in mg/dL or mmol/L, cholesterol in mg/dL or mmol/L, and vitamin D in ng/mL or nmol/L. A vitamin D result of 50 nmol/L equals 20 ng/mL, not 50 ng/mL, and that conversion mistake changes the clinical interpretation completely.

Fasting changes triglycerides, glucose, insulin, and sometimes liver enzymes after a heavy meal or alcohol exposure. Our lab units guide is worth checking when a dashboard shows a sudden “jump” that might be a unit mismatch.

Kantesti’s blood test analytics flag likely inconsistencies, but no software can know whether you had a viral illness 3 days earlier unless you record it. In my own reviews, I ask patients to add four notes to each test date: fasting hours, exercise in the prior 48 hours, new medications, and acute symptoms.

Questions your dashboard should help you ask

Instead of saying “my labs are weird,” try: “My ferritin fell from 62 to 18 ng/mL in 10 months while hemoglobin stayed 12.7 g/dL; should we look for blood loss, absorption issues, or retest iron studies?” That question gives the clinician enough structure to act.

For kidney trends, a better question is: “My eGFR moved from 92 to 68 mL/min/1.73 m² across 2 years; should we check urine ACR, blood pressure, medications, and repeat creatinine with cystatin C?” If a result is abnormal once, our repeat testing guide gives reasonable retest windows by marker type.

For cholesterol, ask about risk rather than shame. “My LDL-C rose from 112 to 154 mg/dL after menopause and my ApoB is 118 mg/dL; does my family history change the treatment threshold?” is a far more useful line than “is my cholesterol bad?”

When small shifts deserve faster review

Potassium above 6.0 mmol/L, sodium below 125 mmol/L, calcium above 12 mg/dL, glucose above 300 mg/dL with symptoms, or platelets below 50,000/µL generally deserves same-day medical advice. These thresholds are not for panic; they are for triage.

Troponin is a trend-based emergency marker. A rising troponin above the assay’s 99th percentile, especially with chest pressure, breathlessness, sweating, or radiation to the arm or jaw, should not sit inside a wellness dashboard waiting for interpretation.

If you are unsure whether a result is urgent, involve a clinician rather than guessing. Our critical values guide lists common result patterns that should move from dashboard review to direct medical care.

Privacy and medical oversight matter in trend tracking

A dashboard may contain reproductive hormones, HIV testing, cancer markers, genetic risk clues, medication monitoring, and family history in one place. That is not casual wellness data; it can affect insurance, employment anxiety, and family conversations.

Our clinicians and scientific reviewers work with defined medical standards rather than improvising from isolated values. You can read more about our physician oversight through the Medical Advisory Board and our clinical approach through medical validation.

Thomas Klein, MD, reviews dashboard logic with a simple rule: the output should make the next clinician conversation safer and clearer. If an interpretation encourages a patient to ignore chest pain, stop medication, or self-treat severe abnormalities, it has failed clinically.

How Kantesti turns uploads into trend-ready questions

Our AI-powered blood test interpretation does not diagnose you; it organizes the evidence, highlights trend signals, and suggests focused clinician questions. If you want to test the workflow, you can upload a report through our free blood test analysis page and see how a scattered PDF becomes a visit-to-visit view.

Kantesti LTD is a UK company, and our engineering, clinical, and governance details are described on About Kantesti. Readers who want benchmark detail can review the AI blood test benchmark before deciding how much weight to place on AI-supported longitudinal blood test analysis.

Research publications: Kantesti LTD. (2026). Multilingual AI Assisted Clinical Decision Support for Early Hantavirus Triage: Design, Engineering Validation, and Real-World Deployment Across 50,000 Interpreted Blood Test Reports. Figshare. https://doi.org/10.6084/m9.figshare.32230290. ResearchGate. Academia.edu.

Research publications: Kantesti LTD. (2026). Urobilinogen in Urine Test: Complete Urinalysis Guide 2026. Zenodo. https://doi.org/10.5281/zenodo.18226379. ResearchGate. Academia.edu.

Frequently Asked Questions