Abstract
Objectives: To identify possible clusters of cancer in the Negev kibbutzim.
Methods: We used cancer incidence data in the 45 kibbutzim over 1972-1996 years from Israel Cancer Registry and kibbutz local clinics. As a population of comparison we exploited the whole country population (picking up age, sex and site specific data for all Jewish and also for groups by country of birth). Statistical analysis was performed with Stata by methods based on the inter-event time interval [1] using so called "relative interval" (R.I.:the expected number of diagnosis within the specific time interval), "q interval" (the proba?
bility of no event within the actually observed inter-event time interval), "interval" (q interval under the gamma distribution assumption ? the waiting time until the third case, rather than the first one) and "cuscore test" of significance (the interval is considered short based on statistics R.l. < k.; where k depends on the number of events and level of signif? icance).
Results: In 14 kibbutzim we detected clusters for all population and all sites of disease with variety years of onset. Among these settlements we also found clustering of: (1) all sites cancer for different age groups (40-59 and 60+ years old), for men and women or both gender; (2) digestive system. And for two certain kibbutzim we identified clusters for: (1) oldest age group only and urinary tract cancer among all residents; (2) oldest age group and among women only.
Conclusions: q and X intervals techniques is capable to detect cancer clusters in small closed communities.
Methods: We used cancer incidence data in the 45 kibbutzim over 1972-1996 years from Israel Cancer Registry and kibbutz local clinics. As a population of comparison we exploited the whole country population (picking up age, sex and site specific data for all Jewish and also for groups by country of birth). Statistical analysis was performed with Stata by methods based on the inter-event time interval [1] using so called "relative interval" (R.I.:the expected number of diagnosis within the specific time interval), "q interval" (the proba?
bility of no event within the actually observed inter-event time interval), "interval" (q interval under the gamma distribution assumption ? the waiting time until the third case, rather than the first one) and "cuscore test" of significance (the interval is considered short based on statistics R.l. < k.; where k depends on the number of events and level of signif? icance).
Results: In 14 kibbutzim we detected clusters for all population and all sites of disease with variety years of onset. Among these settlements we also found clustering of: (1) all sites cancer for different age groups (40-59 and 60+ years old), for men and women or both gender; (2) digestive system. And for two certain kibbutzim we identified clusters for: (1) oldest age group only and urinary tract cancer among all residents; (2) oldest age group and among women only.
Conclusions: q and X intervals techniques is capable to detect cancer clusters in small closed communities.
| Original language | English |
|---|---|
| Pages (from-to) | S50-S50 |
| Journal | Epidemiology |
| Volume | 12 |
| Issue number | 4 |
| State | Published - 1 Jul 2001 |
