Public Health Genetics and Screening Principles

Screening a whole population is a different activity from testing a patient who came in with a problem, and it is governed by different reasoning. In clinical testing the patient has a question; in population screening the program approaches people who feel well, which raises the bar for justifying it. The criteria that decide whether a condition belongs in a screening program are the foundation for newborn screening, carrier screening, and every proposal to add a condition to either.

• The Wilson and Jungner criteria (1968) remain the reference framework. A condition should be an important health problem, have a recognizable latent or early symptomatic stage, and have a natural history that is understood. There should be an accepted treatment, facilities for diagnosis and treatment available, a suitable and acceptable test, an agreed policy on whom to treat, and costs balanced against benefit. Screening should be a continuing process, not a one-time project.
• The treatment criterion is the contested one in genomics. Sequencing can identify conditions long before symptoms and for many of them nothing can be done, which puts pressure on the classical requirement that early treatment must improve outcome. Proposals to screen for untreatable conditions are where the framework is actively argued over.
• Screening is not diagnosis. A positive screen is a probability statement requiring confirmatory diagnostic testing. Programs that blur this produce families who believe their child has a disease during the interval before confirmation.
• Sensitivity and specificity are set by policy, not physics. Where a program sets its cutoff determines the tradeoff between missed cases and false positives, and screening programs generally accept many false positives to avoid missing a treatable case. That is a deliberate choice with a cost borne by the families who screen positive and are then fine.
• Positive predictive value falls as prevalence falls. A test with excellent sensitivity and specificity applied to a rare condition still produces mostly false positives. This is the single most misunderstood point in screening and the reason a positive screen must never be communicated as a diagnosis.
• The screening program is the system, not the assay. Sample collection, transport, laboratory analysis, result delivery, short-term follow-up, confirmatory testing, and treatment access all have to work. A program with an excellent test and no follow-up mechanism fails.
• Harms are real and must be counted: false-positive anxiety that persists after resolution, overdiagnosis of conditions that would never have caused symptoms, the burden of confirmatory testing, and the identification of carriers or milder variants nobody planned to find.
• Consent in population screening is usually attenuated compared with clinical testing, often opt-out or presumed, which is defended on public health grounds and remains ethically contested, particularly as panels expand.
• Equity is a program design question. A screening program that identifies conditions people cannot afford to treat converts a health problem into a documented, untreated health problem.

• A state considers adding a condition to its newborn screening panel. The debate turns on the treatment criterion: an effective therapy exists but only if started before symptoms, which is precisely the argument for screening. The counterargument is that confirmatory testing capacity and treatment access are not in place statewide, which is a program readiness question rather than a test performance question.
• A family is told their newborn "screened positive" and understands it as a diagnosis. Two weeks of acute distress follow before confirmatory testing is normal. The failure is in communication design: the program's script did not distinguish screen from diagnosis.
• An expanded panel begins detecting a mild variant form of a condition that would likely never have caused symptoms. The program now faces overdiagnosis, and families receive a label and surveillance for something that may never have mattered.
• A proposal to add an untreatable but severe condition is argued on the grounds that families value the information and can avoid a diagnostic odyssey. This does not meet the classical treatment criterion, and it is a live example of where the framework is under revision rather than a settled question.

• Do not describe a positive screen as a diagnosis, in speech or in written material. This is the most common and most damaging error in screening communication.
• Do not assume high sensitivity and specificity produce a high positive predictive value. At low prevalence they do not, and most positives will be false.
• Do not evaluate a screening proposal on test performance alone. Follow-up, confirmatory capacity, and treatment access decide whether the program helps anyone.
• Do not treat false positives as costless. Parental anxiety after a false-positive newborn screen is measurable and can persist well past resolution.
• Do not ignore that expanded panels find more of everything, including carriers and mild variants nobody set out to detect.