OBJECTIVES: We hypothesized that one central laboratory would  be more cost-effective than several local laboratories because price of single Chlamylege test depends on number of simultaneous tests performed. Incidence of pneumonia requiring hospitalization in Poland is app.4/1000 persons/year. For hypothetical population of one million people 4 000/ year will require hospitalization for pneumonia (VIP). We established seasonal incidence using clinical database of  Military Medical Institute Warsaw. New statistical method (Queuing Model)was used. METHODS: Two simulation models were constructed for one  million inhabitants: one, where four hospital laboratories exist, performing   10-50 tests. Second, where large central laboratory exists performing  50-200 tests.  Three scenarios of morbidity were established: 1) 3000 outpatients, 1000 inpatients, equal number of patients over year; 2) 3000 outpatients, 1000 inpatients, number of patients/month related to  seasonality  (4 seasons specified); 3) increased influx of patients: 12000 outpatients and 4000 inpatients; number of patients/month related to  seasonality. RESULTS: 1) in central laboratory significantly smaller number of samples were tested (mean 0.25vs0.75); 2) quarter 1  (increased patients influx) % tested in central laboratory was significantly higher  but still smaller than in the local labs (0.92vs.0.95) whereas in quarter 3 (decreased patients influx) less tests performed(0.2777 vs. 0.0005); and 3) % of tests made in central lab is much higher, but still not exceeding number of tests performed in local labs ( 0.82vs.0.83). Central laboratory performed less tests  comparing to local labs and periods  of inactivity were noted which significantly increased  cost of a  single test. CONCLUSIONS: According to  Queuing Model  it was  confirmed that creation of the central laboratory is not     reasonable  in terms of costs. We conclude that Queuing Statistical Model  can be a useful tool for preliminary assessment ofthe  cost-effectiveness  of hypothesized research methodology.