The Highland Haemoglobin Report: Adaptation or Anaemia?
In a highland block of Koraput district (900–1,100 m), an NGO screens 1,240 adults aged 18–60 and finds mean haemoglobin of 11.2 g/dL among men in highland villages against 12.8 g/dL in valley villages. The programme officer writes in the draft report that the highlanders are 'a genetically adapted population for whom low haemoglobin is physiological', and on that basis proposes to cancel iron–folic acid supplementation and redirect the budget to latrines. The field notes record three awkward facts. First, 34% of highland women report persistent fatigue. Second, a 90-person subsample shows ferritin above 100 ng/mL in 21 people and below 12 ng/mL in 29. Third, haemoglobinopathy screening in one highland village found sickle-cell trait in 18% of those tested, against 4% in another village only 6 km away. Malaria was holoendemic in the valley until 2009. Hookworm prevalence was last measured in 1998. You are asked to review the draft before submission.
1. Before accepting or rejecting the 'genetically adapted for low haemoglobin' interpretation, what must the review establish? Show the inferential steps rather than just naming the answer.
Read the worked answer
Step 1 - the dose of the stressor. The claim assumes altitude hypoxia. At 900–1,100 m, barometric pressure is roughly 89–90 kPa and arterial oxygen saturation is typically 95–97% in healthy adults; classic altitude-driven haemoglobin responses are documented at far higher elevations. A mean difference of 1.6 g/dL between two blocks of the same district is therefore unlikely to be a hypoxia signal, and the report must state this explicitly rather than borrow the Tibetan/Andean literature.
Step 2 - the level of explanation. Four different things can produce a low group mean and only one of them is heritable:
- Genetic adaptation: allele frequencies shaped by selection (e.g. HbS, alpha-thalassaemia, G6PD deficiency).
- Acclimatisation: reversible physiological adjustment in an individual.
- Developmental plasticity: permanent calibration during growth (e.g. iron status in childhood).
- Pathology or deprivation: iron deficiency, hookworm, malaria, chronic inflammation.
A population mean cannot distinguish these; a group comparison alone is compatible with all four.
Step 3 - confounding. Highland and valley villages differ in diet, sanitation, water source, malaria history and possibly ethnicity. Until these are measured, altitude is one variable among many.
Conclusion. The draft commits the ecological fallacy (inferring individual or genetic causation from a group average) and misuses 'adaptation' as a synonym for 'normal for them'. The supplementation decision cannot rest on this inference; the ferritin split (both high and very low values) already suggests a mixed picture of iron deficiency and inflammation, not a uniform genetic norm.
2. Specify the smallest additional data collection that would adjudicate between the competing explanations, and state the interpretation rule you would apply to each measurement.
Read the worked answer
Measurements to add (one round, same households):
| Measure | What it discriminates | Interpretation rule |
|---|---|---|
| Ferritin plus CRP | Iron deficiency vs inflammation | Ferritin is an acute-phase reactant; with CRP >5 mg/L, use soluble transferrin receptor or a corrected threshold. Ferritin <15 ng/mL is fairly specific for depleted stores. |
| Hb electrophoresis / HPLC | Structural haemoglobinopathy | HbA2 >3.5% suggests beta-thalassaemia trait; HbAS (sickle trait) does not by itself cause anaemia - only HbSS or compound states do. |
| G6PD assay | Enzyme deficiency | Explains haemolysis risk, not the mean Hb difference; matters for treatment choice. |
| Stool microscopy (hookworm) | Blood loss | Prevalence data are 25 years old; even a 200-person sample is informative. |
| Malaria RDT plus history | Chronic infection | Distinguishes current infection from historical exposure. |
| 24-hour dietary recall, 2 days | Intake | Estimates bioavailable iron, not just iron. |
| GPS altitude and village-level water/sanitation | The stressor itself | Confirms whether the block is really 'highland' in any physiologically meaningful sense. |
Decision logic. If altitude is trivial and ferritin/CRP point to iron deficiency with inflammation, the supplementation programme should continue, ideally with deworming and infection control. If haemoglobinopathy explains most of the variance, anaemia prevalence should be recomputed excluding those genotypes, and the residual burden - not the label 'adapted' - drives policy.
Sampling note. The 6 km pair of villages with 18% and 4% trait frequency is a warning against treating 'highlander' as a single population; stratification by village, not by altitude band, is needed.
3. The programme officer defends the draft by saying 'respecting local adaptation is culturally sensitive'. How would you answer, and what harm could the narrative do?
Read the worked answer
The confusion. 'Adapted' is a technical claim about selection and physiology, not a courtesy. Using it to mean 'leave them alone' converts a hypothesis into a policy and makes disagreement look like disrespect for local culture. Anthropologically, the more respectful move is to take people's fatigue reports seriously as evidence rather than explaining them away.
Three harms of the narrative.
- Naturalising deprivation. If low Hb is 'their normal', a treatable nutritional deficiency becomes invisible and prevalence statistics stop triggering action. The 34% fatigue figure is the population's own testimony against the norm being benign.
- Racialising a population. Labelling an entire block as genetically distinct ignores internal variation (18% vs 4% trait) and shifts attention from environment to ancestry, a pattern with a long and damaging history in Indian health administration.
- Closing the question. The claim is unfalsifiable as used - any low value confirms it. Science and policy both require it to be testable.
Constructive counter-proposal. Reframe as: 'We currently cannot distinguish iron deficiency, haemoglobinopathy and inflammation in this block; here is a 6-week measurement plan; until then, supplementation continues because the risk of untreated deficiency exceeds the risk of mild iron loading in a population with high inflammation.' This honours local reports, keeps the intervention reversible, and does not require anyone to be labelled adapted or maladapted.