Why are common diseases missed in the people most exposed?
Kidney disease, asthma and lung scarring are among the most familiar diagnoses in medicine. In workers exposed to heat, dust and chemicals, they can still be missed, misnamed or never recorded.
A 2026 review in the journal Diagnosis maps six points where those cases can be lost, from the first clinic visit to the death record, and charts the evidence for each.
Diagnosis starts with a question
A clinic finds what it looks for. It decides whom to test using known risk factors. For chronic kidney disease, that usually means diabetes or high blood pressure.
Farm workers exposed to repeated heat, dehydration and agrochemicals can develop kidney disease with neither, so they fall outside the pathway built to find it (Johnson et al., 2019). The disease is the same. What differs is how it arrives.
Who enters kidney screening
Screening by usual risk factors
7of 12
people with kidney disease enter screening. The five whose risk comes from their work are never asked to test.
Adding exposure history
12of 12
enter screening, because their work is part of the first conversation. Whether that improves diagnosis is what studies must now test.
Six points where common disease can be lost
The review charted 61 sources onto six stages, from the clinic to the death record, borrowing the structure of the HIV care cascade (Frimpong et al., 2026). Each stage counts a different group of people. For each one, it asked what evidence shows that cases are lost there.
Bar lengths are schematic. The review did not measure how many cases are lost at each stage.
A case passes through six stages
Three happen in the clinic, one in reporting systems and two in death records. Each bar’s colour shows the strongest evidence the review found that cases are lost at that stage.
Detected: is the disease found at all?
Counts everyone who has the disease. If a person’s risk isn’t on the clinic’s list, nobody looks. In sub-Saharan Africa, about 73% of people with high blood pressure are undiagnosed, though that study did not single out exposed workers (Ataklte et al., 2015).
Context onlynot specific to exposed workers
Named correctly: is it the right diagnosis?
Counts patients whose disease was detected. In one specialist centre, 20 of 46 patients told their lung scarring had no known cause were reclassified after a detailed exposure investigation (Morell et al., 2013).
Direct, patient levelthe only stage with it
Cause recorded: does the chart say it came from work?
Counts patients with a diagnosis. Work exposures may account for about 21% of idiopathic pulmonary fibrosis (Gandhi et al., 2024). Applied to US death records, that suggests roughly 14,000 of 67,843 deaths from 2020 to 2022 may have been work related (Mazurek et al., 2025).
Population estimateno individual record checked
Reported or claimed: does it reach a register?
Counts diagnosed or eligible cases. Between 50% and 95% of occupational-disease cases went unreported in the studies that measured it (Khoe et al., 2026).
Relatedwork-specific disease
Death registered: is the death recorded at all?
Counts all deaths. Large populations, mainly in low- and middle-income countries, have no system able to register deaths (Mikkelsen et al., 2015). A death that is never recorded can never be counted.
Context only
Cause coded: is the right cause written down?
Counts registered deaths. Verbal autopsy misclassified about 31% of non-communicable disease deaths in one urban African setting (Misganaw et al., 2012).
Context only
Only one of six points has direct evidence
One stage has direct, patient-level evidence, one a population estimate, one related evidence and three context only. That thin evidence is the review’s main finding, and the gap it asks future studies to fill.
An exposure investigation changed 20 of 46 lung disease diagnoses
Patients diagnosed with idiopathic pulmonary fibrosis, scarring with no known cause, were re-examined with a detailed exposure history and further tests. Twenty turned out to have chronic hypersensitivity pneumonitis, most traced to bird antigens in feather bedding (Morell et al., 2013).
Patients first told their lung scarring had no known cause
As first diagnosed
46of 46
carry the label “no known cause”.
After an exposure investigation
20of 46
reclassified as chronic hypersensitivity pneumonitis, a disease linked to a specific exposure.
Work histories go unasked, and work-related cases go unreported
Loss is better documented around the clinic visit than inside the diagnosis. Exposure information is often never asked for (Siegel et al., 2025), and recognised work-related disease frequently never reaches a register or a compensation claim (Reilly et al., 2018; Khoe et al., 2026).
Three measured gaps around diagnosis
Statistics can correct a vague cause of death, but not a missing one
Deaths recorded with a vague cause can be statistically reassigned later. A person who was never diagnosed, or a death never registered, leaves no record to correct (Monasta et al., 2022; Mikkelsen et al., 2015). That is why the review argues modelling cannot replace an exposure history at the first visit.
What a statistical correction can reach
Records as kept
After a statistical correction
What the review does not show
It is a framework synthesis, not a systematic review: one reviewer charted 61 sources from non-exhaustive searches, with no pooled estimate. It does not show how often these failures happen, or that asking about exposure improves care. That is the hypothesis it hands on: compare routine with structured exposure assessment, and measure new diagnoses, treatment changes, time to diagnosis, trust and cost. In settings with few doctors, that assessment will need testing with trained health workers rather than doctors alone.
Questions people ask
Why is chronic kidney disease missed in farm workers who don’t have diabetes or high blood pressure?
Screening usually targets those two conditions. Farm workers exposed to repeated heat, dehydration and agrochemicals can develop kidney disease without either, so they can fall outside the groups clinics decide to test.
Can an unrecognised exposure cause lung disease to be misdiagnosed as idiopathic pulmonary fibrosis?
In one specialist centre, 20 of 46 patients with that diagnosis were reclassified after a detailed exposure investigation and further tests. It is a single study and may not apply to other diseases.
How much occupational disease goes unreported?
In studies that measured it, between 50% and 95% of cases. Most of that evidence comes from high-income countries; none came from lower-middle- or low-income countries.
Does asking about work exposure improve diagnosis?
Not yet known. The review proposes it as a hypothesis for prospective studies to test.
Making the hardest workers visible
The review sets out a question worth testing. bealehorton ai’s screening is built to answer it, with exposure history in the very first conversation.