All possible combinations of disease risk (0 through 1)

[Cruise]

Hi Ya'll,

I'd like to assess how human mobility affects select disease estimates at population level. Number of new leukemia cases each year is known. However, there is no data on how many of new cases moved out of each county and how many of new cases were migrants who moved in from other counties.

*Base population = 980 where 1000+inflow(20)-outflow(30)-new cases(10)

For a demo purpose, Table 1 has four different scenarios of A, B, C and D where rate (probability) of leukemia among movers (either ‘0’ for min or ‘1’ for max). Given these extreme assumptions, If leukemia risk is 1 among outflow then all new 10 leukemia cases must have left the county. If leukemia risk is 1 among inflow then all 10 new cases must have had moved in from other counties. A and B cells in Table 1 shows that leukemia rate wouldn’t be affected in case of equal leukemia risk across inflow and outflow fluxes. In contrast, cell C and D shows differential risk of leukemia across inflow and outflow fluxes would affect the incidence rate.

I’d like come up with continuous leukemia rate over continuous leukemia risk of 0 through 1 associated with inflow and outflow fluxes. I read but failed to apply the approaches in SAS paper “Ten Tips for Simulating Data with SAS®”.

I’ll highly appreciate if you could help me with a demo code if possible? Using SAS 9.4. Sample data attached.

One of plots that would summarize the result would be: (this is to show what my desirable outcome would be)

[Reeza]

Does new case align with total case in your data? Or is that a different metric?

[Cruise]

Sorry for ambiguity. Yes, tot_case aligns with new cases , tot_pop is Census 2000 population estimate