Table 2. We summarize key challenges attached to the best practices identified in this paper, which must be addressed to realize full value from a RWD life cycle.

Best Practice Recommendation	Key Challenges
1. Compatibility with internationally recognized data standards enables data aggregation at scale	Overcoming limitations imposed by proprietary vendor software and lack of API support. Commercial stakeholder collaboration may be difficult or impossible to obtain.
2. QA must be considered in advance and tailored for use case	Lack of gold standard QA frameworks for different use cases can be overcome with careful multidisciplinary and expert consideration of processes.
3. Incentivize detailed data entry at source to maximize value	Return of value to direct patient care must be demonstrated, to incentivize RWD collection
4. Deploy natural language processing to mobilize unstructured data sources	NLP platforms must be deployed to interface with EHR dataflows. In general, more algorithmic training on medical specific text corpuses required to improve real-world performance and utility.
5. Implement platform solutions that enable rapid-cycle and flexible analytics	Solutions may require greater up-front investment in cost, time, and expertise to accrue long-term benefits.
6. Protect and return value to patients through transparency, engagement, and a focus on data privacy	Providing clear, transparent, and balanced information to the public on the benefits and risks in use of RWD is difficult. Systematically collecting and analyzing public opinion, and setting up citizen juries, can be costly and introduces lag times into decision-making.
7. Prioritize diversity in RWD to reduce bias and maintain equity	Investment required into digital health infrastructure in deprived communities to rebalance the unequal health data map. Opportunity cost of this investment, versus immediate clinical care, must be considered.

API, application programming interface; NLP, natural language processing; RWD, real-world data.