Chapter 3, Problem 74P

Pulmonary Disease

Research into cigarette-smoking habits, smoking prevention, and cessation programs necessitates accurate measurement of smoking behavior. However, decreasing social acceptability of smoking appears to cause significant underreporting. Chemical markers for cigarette use can provide objective indicators of smoking behavior. One widely used noninvasive marker is the level of saliva thiocyanate (SCN). In a Minneapolis school district, 1332 students in eighth grade (ages 12–14) participated in a study [12] whereby they

1. (1) Viewed a film illustrating how recent cigarette use could be readily detected from small samples of saliva
2. (2) Provided a personal sample of SCN
3. (3) Provided a self-report of the number of cigarettes smoked per week

The results are given in Table 3.10.

Table 3.10 Relationship between SCN levels and self-reported cigarettes smoked per week

Suppose the self-reports are completely accurate and are representative of the number of eighth-grade students who smoke in the general community. We are considering using an SCN level ≥ 100 μg/mL as a test criterion for identifying cigarette smokers. Regard a student as positive if he or she smokes one or more cigarettes per week.

Suppose we regard the self-reports of all students who report some cigarette consumption as valid but estimate that 20% of students who report no cigarette consumption actually smoke 1–4 cigarettes per week and an additional 10% smoke 5–14 cigarettes per week.

Assuming the percentage of students with SCN ≥ 100 μg/mL in these two subgroups is the same as in those who truly report 1–4 and 5–14 cigarettes per week, compute the specificity under these assumptions.