I have secondary data on doctor mistakes (outcome variable) for all 50 states in USA by doctor specialty. I mean real numbers per state. I want to assess the relationship between doctor mistakes (outcome variable) and state of doctor practice (predictor #1) and doctor specialty (Predictor #2) (internist, Surgeon and OBGYN).I want to adjust for the following confounders: provider factor, hospital factor, patient factor and miscellaneous factor. I don't have real data for these 4 confounders.

Hypotheses:

1. no association or interaction between doctor mistake and state of doctor practice

2. no association or interaction between doctor mistake and doctor specialty

I want to do 2-way ANOVA using SPSS and my Excel table currently looks like this (other states not shown here):

Doctor Doctor Doctor Hospital Patient Miscellaneous State Specialty Mistake Factor Factor Factor Factor Alabama Internist 100 Alabama Surgeon 345 Alabama OBGYN 502 Alaska Internist 270 Alaska Surgeon 150 Alaska OBGYN 39 Arizona Internist 562 Arizona Surgeon 393 Arizona OBGYN 191 Arkansas Internist 534 Arkansas Surgeon 76 Arkansas OBGYN 75