Note: As promised during the session, Isaac Holeman set up a Point of Care page where we can focus our efforts to adapt OpenMRS for use at the point of care.

Definitions

• Point of care: Data entry during clinical visit/encounter.

• Retrospective: Data entry after encounter, usually from paper form.

Expectations for Point-of-care systems:

• Clinics already overburdened - reduce how many times data is written down. No data entry team available. Enter data only once.

• Neal Lesh: 95% of OpenMRS data is retrospective.

• Why is the retrospective data entry happening? What are the barriers?

  • Speed is so important - so many patients - typical systems cannot keep up

  • OpenMRS not designed for Point-of-care environment - need system designed to flow through the patient encounters

  • Problem w/ retroactive entry - data not available throughout the clinic - paper in the wrong place

  • Point-of-care system opportunity to improve quality of care - patients do not have to wait so long for their appointments, reduce time to wait from a day to an hour or two

At the moment aggregate data is pushing downwards - external factors are collecting - data is not used or available by clinicians.

What are circumstances that cause us to use Retrospective rather than Point-of-care?

• Connectivity and uptime/consequence of downtime

• Legal requirements - must have paper copy for legal reasons

• Perceived extra hardware complexity - using mobile phones not as big a barrier as pc's - more accepted by indigenous community

  • Dr. Aamir Khan, Indus Hospital Research Center, Karachi, Pakistan - They made a move towards thin clients to reduce maintenance issues. It is important to note that as you move to Point-of-care systems, you need more sophisticated tech support.

• Security: Look out for security issues:

  • Western cape system w/ 2.5 million patients - some clinics completely cleaned out by thieves. They are considering the use of VMware type systems and thin clients - When thieves steal equipment, it is worthless to them.

• Power - Challenges of thin clients - do not have batteries

  • Pakistan - at clinics - use cellphones rather than pc's, since the clinic is less secure than the hospital - but important to close the feedback loop - return data to the clinic

  • Daniel's Andriod OpenMRS for mobiles - cellphone app example

Important design points for a Point-of-case system

• Runs on a browser and can run in a thin client model - do not need a ton of servers - especially at the point of care.

• Process are important part of Point-of-care. How to change business processes at the clinics? Does OpenMRS support modeling business processes?

• More than interface: need to be able to track delivery of service

  • Baobab very successful at doing this - can see where patients are waiting

  • Baobab - some clinics have different roles - flexible roles where anyone can do anything. This is very useful in resource-poor environments where nurses double as pharmacists.

  • XForms include skip logic capability - can this (i.e., use of XForms' skip logic) be a useful tool for Point-of-care?

  • MOH Uganda: How to develop system that can easily configure workflow to support the local processes?

• Time issue: System must be inherent to the processes so user does not notice it. Do not ask too much of the user, unnecessary fields.

• Performance is very important - cannot wait for a long screen refresh

Types of Point-of-care

• Centralized system

• Distributed - sync issues -

Which type does OpenMRS support?

Sync

OpenMRS has a sync feature built-in

• Use of guid to identify records that go up all the way - human resources system example in uganda

• Boabab: If server can connect to other clinics, it can grab data from other peers

• Importance of unique patient identifiers

  • Baobab - have unique identifiers for patients. Not using uuid's for baobab - just nat'l unique id.

Access to patient information

Is having instant access to patient record more of a requirement of a Point-of-care system or is this typical of all systems including Point-of-care

• Very important for Point-of-care in order to track patient

• Northern Nigeria - silos of data of maternal health data - now to get this data in front of clinicians? paper system doesn't work - no one knows the utility of some of these fields.

• Opportunity is instant access to data vs. not having historical data avaialble

Patient Health Card - smart card w/ last 10 visits on it.

Separate Point of care data from administrative data - sometimes the data sytems collect too much admin data that is not particularly useful to clinicians.

What does OpenMRS need in order to implement a Point-of-care system?

Broad issues:

• Where

• When and

• What

Is it a ui framework or more modularized?

• Example of mdr-tb module - Model the Point-of-care process and enable programmers to create appropriate modules

• Needs to accommodate different types of clients - pc's, phones, etc.

Hardware:

• Promise for mobile phones and thin clients

• Need analysis for the different contexts - 'device strategy' - rural health centres have different needs from urban hospitals

  • In a rural setting they may need a bed rather than a pc, but they need a communication device to send the data to the people who can provision the bed.

  • Need wiki pages on equipment recommendations, stories on what works

Working groups

• Device strategy

• Software design/data model

Case Studies

• Pakistan: one Point-of-care system was being scaled up to over 40 sites, but OpenMRS did not yet have the features to do this.

• Who has deployed Point-of-care systems? - page on wiki to publish success cases

• Look at these stories to filter out some best practices, not just OpenMRS deployments. Don't expect our community to supply all of the solutions.

• Patient registration system in Rwanda is a Point-of-care system currently in development.

Point of Care System Examples

Paper review of HMIS systems

System development

• While developing system, the clinicians worked 3-4 hours a day w/ the programmers in developing each module - therefore they had a strong sense of involvement - they owned it.

• Build flexibility in system - enable clinicians to add extra fields themselves - they may work on a special form of TB and collect extra data.

How to adapt OpenMRS for primary care?

• Don't always know what type if data will be needed when someone comes into the clinic.

• Should the team concentrate on primary care system, or on a subset like HIV?

• Think of the issue on a more general level, on health outcome.

Need for reference implementations. AMPATH has stimulated alot of OpenMRS developments.

Next Steps

• Sign up to the workgroups - Issac will help organise this activity.

• Please add any useful reference material to Point of Care System Examples.

• Idea of funding study of 5-6 Point-of-care sites and gather info for the team.