Table 2.
Examples of the top rule for subsets.
| Comorbidities | Keywordsa | Top rule | Lift | Count |
| Cardiovascular disease | ‘x.michd’ | {diffwalk,x.casthm1,x.rfhype5,x.rfsmok3} => {x.michd} | 1.43 | 2783 |
| Chronic pulmonary disease | chccopd1,’ ‘x.casthm1’ | {diffalon,x.casthm1,x.drdxar1,x.rfsmok3} => {diffwalk} | 1.93 | 2643 |
| Rheumatoid arthritis | ‘x.drdxar1’ | {diffalon,x.casthm1,x.drdxar1,x.rfsmok3} => {diffwalk} | 1.93 | 2676 |
| Diabetes | ‘diabete3’ | {diffwalk,x.casthm1,x.rfchol1,x.rfhype5,x.rfsmok3} => {diabete3} | 1.53 | 2726 |
| Cancer | ‘chcocncr’ | {chcocncr,x.casthm1} => {x.age65yr} | 1.2 | 2697 |
aWe used the variable code to represent each variable. The meaning of the code is shown in Multimedia Appendix 1.