PU520 Unit 7 Screening Test Appraisal Example

Same Test, Two Settings, Four Positives in Five Wrong: A Screening Test Appraisal Placing One Composite Rapid Test With 96 Percent Sensitivity and 98 Percent Specificity in a Community Testing Drive and a Specialty Clinic, and Showing Why Its Positive Predictive Value Collapses

[Student Name]

Purdue University Global

PU520: Principles of Epidemiology

Unit 7 Screening Test Appraisal

[Instructor Name]

[Date]

The test, both settings and every count are composite and bracketed for the writer's own data; the methods and principles cited are real.

The Test

Sensitivity: the share of infected people the test correctly calls positive, [96] percent. Specificity: the share of uninfected people it correctly calls negative, [98] percent. Both belong to the test and do not change between settings.

Table 1: Community Testing Drive, Prevalence [0.5] Percent

Built for a notional [10,000] people, so the counts read as whole persons. Infected: [50]. Test positive [48], true positives; test negative [2], false negatives. Uninfected: [9,950]. Test positive [199], false positives; test negative [9,751], true negatives. Positive predictive value: [48] / ([48] + [199]) = [19.4] percent. Negative predictive value: [9,751] / ([9,751] + [2]) = [99.98] percent.

Table 2: Specialty Clinic, Prevalence [8] Percent

Infected: [800]. True positives [768]; false negatives [32]. Uninfected: [9,200]. False positives [184]; true negatives [9,016]. Positive predictive value: [768] / ([768] + [184]) = [80.7] percent. Negative predictive value: [9,016] / ([9,016] + [32]) = [99.65] percent. In the community drive, four of every five positive results are wrong, though the test is the same one that performs well in the clinic.

Why Predictive Values Change

Predictive values depend on prevalence (Altman & Bland, 1994). When an infection is rare, even a small false-positive rate applied to a large uninfected group produces more false positives than there are true cases. In the community drive, [2] percent of [9,950] uninfected people is [199] false positives against only [48] true positives. In the clinic, the uninfected group is smaller and the infected group larger, so true positives dominate.

Why Natural Frequencies Help

Expressing the tables as counts of people, rather than conditional probabilities, makes the result intuitive: anyone can see that [199] is larger than [48]. Clinicians and the public misjudge predictive values when given only percentages, so reports and patient materials should use counts.

What Happens to the People Behind the Numbers

In the community drive, [199] people out of every [10,000] tested will hear that their screening result is positive when they are not infected. Some will worry for days, tell partners or employers, or avoid future testing. Two infected people will be falsely reassured. In the clinic, [184] false positives occur among a smaller uninfected group, and [32] infected people are missed. These are not abstractions: the program's design decides how quickly the false positives are corrected and how the missed cases are found. A same-day confirmatory sample, collected at the drive for every positive rapid result, would shorten the period of false alarm from a week to hours.

How Prevalence Was Estimated

Prevalence in the community drive comes from the previous year's drives, where [0.5] percent of confirmed results were positive. Clinic prevalence comes from the clinic's own confirmed results. If the community drive moves to a setting with higher prevalence, such as a shelter or a needle exchange, its positive predictive value will rise, and the program should recalculate it for each new site.

Implications for the Community Drive

A positive rapid result in the community must never be reported as a diagnosis. The program should state, in its materials and in what testers say, that a positive result means a confirmatory test is needed and that most positives at this setting turn out negative on confirmation. A two-step approach, rapid screening followed by a confirmatory laboratory test for every positive, preserves the test's high sensitivity while removing most false positives.

Implications for the Clinic

In the clinic, a positive rapid result is correct about four times in five, but confirmation is still needed before treatment decisions. A negative result is highly reliable in both settings, which is the test's main strength and the reason it suits screening, where ruling out infection quickly matters most.

Is the Community Drive Worth Doing?

Screening is justified when the condition is important, early detection helps, a suitable test exists, confirmation and treatment are available and the program is acceptable to the population (Wilson & Jungner, 1968). The drive meets these principles only if confirmation is fast and the anxiety of false positives is handled with care. Testers should be trained to explain results, and confirmatory results should reach people within days.

Summary Table

Community: sensitivity [96], specificity [98], prevalence [0.5], positive predictive value [19.4], negative predictive value [99.98] percent. Clinic: same sensitivity and specificity, prevalence [8], positive predictive value [80.7], negative predictive value [99.65] percent.

Common Errors to Avoid

Assuming a test with high specificity yields reliable positives everywhere. Reporting sensitivity and specificity to patients as if they were the chance their own result is correct. Comparing predictive values across settings without stating prevalence (Celentano & Szklo, 2019).

References

Altman, D. G., & Bland, J. M. (1994). Diagnostic tests 2: Predictive values. BMJ, 309(6947), 102. https://doi.org/10.1136/bmj.309.6947.102

Celentano, D. D., & Szklo, M. (2019). Gordis epidemiology (6th ed.). Elsevier.

Wilson, J. M. G., & Jungner, G. (1968). Principles and practice of screening for disease (Public Health Papers No. 34). World Health Organization.