Medicines and results

Lab Results Interpreter

Patterns across a panel, and what to test next

The Lab Results Interpreter reads a set of results the way a clinical pathologist would. It first asks whether the sample itself could explain an odd value, lists anything that needs action today, and then reads the panel as a pattern and as a trend. It ranks the likely causes, orders the next tests by how well they separate them, and says what these results leave unanswered.

Who it is for

  • Physicians and residents on the wards with an abnormal panel that does not point to one diagnosis
  • Emergency doctors deciding which results need action before the patient is admitted or sent home
  • Consultants and registrars in clinic reviewing results ordered elsewhere, or a trend over several months

What it does

  • The sample is questioned first

    Considers haemolysis, a delayed sample, EDTA contamination, blood taken above an infusion, lipaemia and assay interference before an unexpected value is treated as real, and suggests a repeat when the result does not fit the clinical picture.

  • Liver tests read as a pattern

    Classifies derangement as hepatocellular, cholestatic or mixed using the R factor, comments on the AST to ALT ratio, and separates liver injury from loss of synthetic function using albumin and the prothrombin time.

  • Anaemia and iron studies in context

    Classifies anaemia by red cell size and the reticulocyte count, and interprets ferritin and transferrin saturation alongside CRP, because ferritin rises with inflammation and a ferritin in the reference range does not exclude iron deficiency.

  • Endocrine results and calcium

    Reads thyroid function tests as a pattern rather than a single TSH, pairs calcium with PTH to separate parathyroid from non-parathyroid causes, notes when HbA1c is unreliable, and interprets cortisol by the time it was drawn.

  • Body fluids and cardiac markers

    Applies Light's criteria to pleural fluid and the serum-ascites albumin gradient to ascitic fluid, reads troponin as a change between samples, and takes account of age when interpreting D-dimer and natriuretic peptides.

  • Ranked causes and the next tests

    Lists the likely causes with the evidence for and against each, then the next investigations in the order that best separates them, each with what a positive or a negative result would mean.

How it works

  1. The clinician enters the results with units and reference ranges, or attaches the laboratory report, together with earlier results and the reason the tests were requested.
  2. The engine looks for pre-analytical explanations and uses the reporting laboratory's ranges, adjusted for age, sex, pregnancy and kidney function; where no range is given it says it is using a typical one.
  3. Anything that needs action today is listed first.
  4. The results are read together and against earlier values, with every calculation shown, such as the R factor, anion gap or corrected calcium.
  5. Likely causes are ranked, the next tests are ordered, and the report closes with the information still needed and a list of verified references.

What to include

  • The results with units and reference ranges, or an uploaded report
  • Previous results for comparison
  • Symptoms, medications and alcohol intake
  • Why the tests were requested

Example cases

  • ALT 62, AST 148, GGT 410, ALP 120, bilirubin 34, MCV 104, platelets 118. Pattern and next tests?
  • Hb 86, MCV 71, ferritin 68, CRP 54, transferrin saturation 9% in rheumatoid arthritis. Iron deficiency or chronic disease?
  • Corrected calcium 2.94 mmol/L, PTH 1.1 pmol/L, phosphate 1.3, creatinine 142. Differential and work-up?

What the report contains

  1. Conclusion and Recommendation, with the urgency
  2. Results at a Glance
  3. The Pattern
  4. Likely Causes
  5. Next Tests
  6. Action and Urgency
  7. Information Needed
  8. Key Evidence & References

Limits

  • It works only from the numbers it is given. It cannot detect a mislabelled sample or an assay problem the laboratory has not reported, so an unexpected result should be repeated before it drives treatment.
  • When no reference range is supplied it falls back on typical adult ranges, which can differ from the reporting laboratory's, particularly in children, in pregnancy and for method-specific assays such as troponin.
  • A pattern suggests causes; it does not make a diagnosis. Results still have to be read with the history and examination, and critical values still need the hospital's usual escalation.

Frequently asked questions

How is the R factor calculated for liver function tests?

Divide the ALT by its upper limit of normal, then divide that by the ALP divided by its upper limit of normal. A value of 5 or more suggests a hepatocellular pattern, 2 or less a cholestatic pattern, and a value in between a mixed pattern. The engine shows this calculation whenever liver tests are supplied, using the laboratory's own upper limits when they are given.

Can ferritin be normal in iron deficiency anaemia?

Yes. Ferritin is an acute-phase protein, so infection, inflammatory disease, liver disease and malignancy can lift it into the reference range while iron stores are depleted. Transferrin saturation, CRP and the clinical setting help, and several guidelines accept a higher ferritin threshold when inflammation is present. The engine reads iron studies against CRP for this reason and says when the picture remains uncertain.

What does the engine do with a result that looks wrong?

It asks first whether the specimen could explain it: haemolysis raising potassium, EDTA contamination raising potassium and lowering calcium, a sample taken above an infusion, a long delay before analysis or a lipaemic sample. If the result does not fit the clinical picture, the report recommends a repeat sample before anything is acted on.

Does it use the reference ranges printed on my laboratory report?

Yes, when they are included in the question or the attached report, and it adjusts its reading for age, sex, pregnancy and kidney function. If no range is given, the report says that a typical adult range has been assumed, so that the reader can check it against the local laboratory's values.