Table 4.
Distribution of number of encounters by physician reported and patient estimated levels of respect for three different analyses
| Main Analysis* | |||
|---|---|---|---|
| Physician reported level of respect | |||
| Patient estimated level of respect | Low | High | |
|
|
|||
| Low | 2 | 2 | |
| High | 73 | 122 | |
| Sensitivity Analysis A** | ||||
|---|---|---|---|---|
| Physician reported level of respect | ||||
| Low | Medium | High | ||
|
|
||||
| Patient estimated level of respect | Low | 2 | 1 | 1 |
| Medium | 50 | 47 | 38 | |
| High | 23 | 25 | 12 | |
| Sensitivity Analysis B*** | ||||
|---|---|---|---|---|
| Physician reported level of respect | ||||
| Low | Medium | High | ||
|
|
||||
| Patient estimated level of respect | Low | 2 | 1 | 1 |
| Medium | 50 | 47 | 38 | |
| High | 23 | 25 | 12 | |
Black shaded areas represent encounters where patients accurately perceived their physicians level of respect. Gray shaded areas represent encounters where patients overestimated physician respect. White shaded areas represent encounters where patients underestimated physician respect.
The variables are dichotomized and then used to define accurate estimation, overestimation, and underestimation.
The variables are categorized and then used to define accurate estimation, overestimation, and underestimation, which were modeled after Beach et al [2].
The variables are categorized and then used to create more restrictive definitions of overestimation and underestimation, as detailed in section 2.4.4.