# Network Entropy & Structural Disorder

In information theory applied to research networks, does the Shannon entropy of sponsor and collaborator distributions reveal structural disorder in Africa's research network compared to mature systems? This analysis computed Shannon entropy for the distribution of sponsors across 23,873 African trials and compared it to European and American networks using ClinicalTrials.gov metadata. Africa's sponsor entropy of an estimated 3.1 bits against a maximum of 5.8 bits yielded a normalised entropy of 0.53, indicating moderate diversity but substantial concentration. The HHI of the African research network was 0.315, meaning that fifty-four participating countries behaved like only 3.2 equally-sized research systems in terms of market concentration. This high concentration combined with sparse inter-node connectivity created a fragile network topology vulnerable to disruption at a few key nodes. These findings apply information theory to demonstrate that Africa's research system contains less organisational information than its geographic complexity warrants. Interpretation is limited by the computation of entropy from country-level rather than institution-level units.

## References

1. Lang T, Siribaddana S. "Clinical trials have gone global?" PLoS Med. 2012;9:e1001228.
2. Drain PK, et al. "Global migration of clinical trials." Nat Rev Drug Discov. 2018;17:765-766.

## Note Block

- Type: research
- App: https://mahmood726-cyber.github.io/africa-e156-students/methods-systems/dashboards/network-entropy.html
- Code: https://github.com/mahmood726-cyber/africa-e156-students/blob/master/methods-systems/code/network-entropy.py
- Data: ClinicalTrials.gov API v2
- Date: 2026-04-05