Table 1
Grandjean’s (2020) taxonomy of network types.
| Network Type | Description |
|---|---|
| Affiliation | Maps connections between companies, institutions, and groups of peoples. |
| Character | Maps relationships between fictional characters, character groups, and temporalities. |
| Circulation | Maps transportation means, places, goods, correspondence routes, etc. |
Table 2
Integrated nodes and edges table derived from a COI statement.
| Source | Target | Relationship Type |
|---|---|---|
| Indegene, Inc | Frank Ernst | Employment |
| Indegene, Inc | Peri Barr | Employment |
| Indegene, Inc | Riad Elmor | Employment |
| Akrimax | Indegene, Inc | Fee for Services |
| etc | etc | etc |
Figure 1
PML Logic Model.
Figure 2
HMA Parser Logic Model.
Table 3
A nodes and edges table generated from the parsing of a simple COI statement.
| Target | Relationship Type | Source | Conflict Weight |
|---|---|---|---|
| Simon Knight | fees | OrganOx UK | 1 |
Table 4
Edges and nodes table generated from HMA-parsing of a complex COI statement.
| Source | Target | COI Weight |
|---|---|---|
| Personalis Inc. | Russ Altman | 1 |
| Personalis Inc. | Russ Altman | 3 |
| Pfizer | Russ Altman | 3 |
| Karius | Russ Altman | 3 |
| Rxight Pharmacogenetics | Teri Klein | 3 |
| Rxight Pharmacogenetics | Michelle Whirl Carrillo | 3 |
Table 5
PML and HMA parser reliability measured by average ICC, with lower and upper bounds for the 95% confidence interval.
| COI Level | PML Parser | HMA Parser | ||||
|---|---|---|---|---|---|---|
| ICC | Lower | Upper | ICC | Lower | Upper | |
| Low | .772 | .745 | .797 | .722 | .69 | .751 |
| Medium | .834 | .814 | .852 | .773 | .747 | .797 |
| High | .506 | .458 | .656 | .618 | .578 | .656 |
Table 6
Number of COI identified by human rater or parser.
| Hi | Medium | Low | |
|---|---|---|---|
| Human | 345 | 505 | 1046 |
| HMD | 192 | 351 | 552 |
| PML | 203 | 446 | 530 |
Figure 3
Network map of COI in opioid (left) and HIV (right) research.