Table 1
Lessons Learned
| Lesson Learned | Details |
| Let the clinical processes drive IT design and not vice versa. | Do not try to automate an incompletely defined clinical process. |
| Prototypes and pilots are our friends. | Involve end-users early and iteratively in the design process using prototypes that will allow them to understand how they might use the solutions. Pilot your solutions early and iteratively to test assumptions and catch mistakes. |
| Respect the power of paper. | The goal of the project is not necessarily to create a paperless system but one that is well integrated into workflow. In certain cases, the medication history may be highly incomplete (e.g., emergent surgical cases) and fragments of the pre-admission medication history are collected by various clinical disciplines with vastly different workflows. In those cases, it may be desirable to ask clinicians to use a paper form to build iteratively a preliminary PAML before the PAML is entered into the electronic application by the admitting clinician. |
| Let user feedback and usage data guide the incremental enhancement plan. Don’t over-anticipate users’ needs. | For example, users may have difficulty defining a priori the full set of communication patterns around the use of a new application. Instead of hard wiring these communication patterns into the application, monitor usage and let the users define the best communication practices. |
| Do not underestimate training needs. | An application that is supposed to support a new type of workflow cannot be deployed without a) setting the expectation around its use, and b) following up on users who do not use the application. |
