Summary of “ITdotHealth II” – the 2012 Harvard Health IT Meeting

The following is an overview of the conference, held September 10-11, 2012. In several weeks, we will post a complete executive summary, as well as videos and slide presentations from the event.

The concept of substitutable apps has become a reality, as multiple examples illustrate—such as the blood pressure app, now in live clinical use on the Cerner System at Boston Children’s Hospital:

  • Compelling, useful health care apps can be developed relatively quickly and inexpensively
  • There is interest in the idea of an “app store for health.” Some EMR companies have been exploring this idea.
  • A focus on developers is key: make app development easy
  • Tremendous progress has been made in the development of the SMART platform, with many success stories and organizations adopting/using.
  • Integrating apps with EHRs is difficult and time consuming. (At Boston Children’s Hospital, an app was developed in 8 weeks; integrating with the EHR took 11 months.)
  • There was great enthusiasm for the SMART platform and for SMART apps – the SMART platform is part of creating a new ecosystem in health care.
  • Major organizations committed to explore establishing a consortium to support the SMART API as a robust interface between health system data and an ecosystem of apps

Key ingredients for success

  • Encouragement to accelerate adoption of the SMART platform, particularly among the many EHR vendors
  • Getting many more apps developed (developers go where critical mass and opportunity are)
  • Making apps not just read-only, but with writing capability
  • Creating a common API
  • Creating a common UI; if too many different apps with different UI’s, will cause confusion
  • Ensuring privacy and security
  • Having data available during development for testing

Clay Christensen, in his keynote address, shared the following:

  1. Technological innovations cause industries from being centralized to decentralized. Centralization means a technology (like mainframes) is used only by those with skill and money; decentralization results in simpler, lower cost technologies used by more people. Health care is centralized, with care in hospitals. Technology needs to be used so care can be delivered in lower cost venues by lower skilled labor.
  2. Industries evolved from vertically integrated to modular. Early on a company must do everything (think Ford in the early days, which made its own steel). As a technology matures and exceeds customers’ expectations, the architecture becomes modular and the component makes (like Intel) capture the value. Currently, EHRs are closed systems; they aren’t modular or open.
  3. Don’t just market products to people with certain attributes. Understand the “jobs” that people need to get done and develop a product/service they will hire to do their job.

Eric Horvitz, in his keynote address, emphasized these points:

  1. Data from electronic health records can and is beingaggregated, used to make predictions (like which patients are most likely to be readmitted to a hospital within 30 days), and drive decisions are interventions. (Participants suggested that while CMS plans to penalize hospitals with high readmission rates for delivering inadequate quality, the quality of care may not be the key driver of readmission; it may be social,economic, and personal factors.)
  2. As a result of the capability, Eric Horvitz articularated a vision of a “General Predictive Platform.”
  3. As a result, decision support will become more sophisticated – always running in the background and always predicting
  4. Apps will be developed that look at data for anomalous events