Electronic Medical Records for Public Health The challenge of scaling up under local power A presentation given at the Public Health Information Network conference, September 2009, in Atlanta, Georgia, USA

Many HIS documents include a diagram similar to this one. It is designed to show that health data are collected and reported from one level to the next until reaching the national ministry or department. The feedback loops shown on the left of the diagram rarely exist in reality. Most data are collected at the bottom of the pyramid. Since there is rarely any feedback to bottom of the pyramid, there is little if any motivation to report accurate data on time. The structure narrows towards the top to indicate that data are increasingly aggregated at higher levels. Important information can be lost when data are aggregated and cannot be easily disaggregated for further analysis. Manual aggregation takes time, introduces errors, obscures potentially important information, and delays reporting. Reporting time lags accumulate in such a system such that reports are increasingly late at each level. For example, in the Philippines, the FHSIS report for 2008 has not yet been published as of June 2009. Electronic medical records focus on the bottom of the pyramid, the point of patient care.

Paper-based patient records often start out well But as records begin to accumulate, storage, access, and analysis become increasingly difficult Eventually we are forced to “archive” records, where they decay and information is lost forever

Nurses can spend hours each day updating vertical program registers (on the left) from individual patient records. At the end of the month, nurses spend more time tallying data from program registers and transferring totals to official HIS reporting forms. The process takes time from patient care and results in a many tallying and transcription errors. Because clinicians rarely receive feedback, they see no benefit from their reporting efforts.

And data consolidation in many paper-based HIS systems squeezes out valuable information. Let’s look at a simple illustration. In the Philippines patient records in health units are aggregated to rural health unit (RHU)-level totals before being submitted to the provincial health office. Information about prenatal visits in barangay health stations is lost, and cannot be easily recovered. Reports submitted by the provincial health office to the regional center for health development include only totals for each local government unit (LGU). Information about individual RHUs and barangay health stations is lost, and cannot be easily recovered. In part, this consolidation is being done because good electronic data management and reporting systems do not exist in most health units, and higher level systems are not designed to receive data electronically from lower levels.

Let’s return to the HIS pyramid. An electronic patient record system tightens up the feedback loop at the point of patient care. Clinicians receive immediate feedback and use the data they enter in the course of patient care. Clinicians who use the data are motivated to improve the quality and completeness of the data. More timely and accurate reporting to the national HIS becomes a byproduct.

But an EMR can to much more than just improve the timeliness and quality of reporting to the national HIS. An EMR can guide clinicians through best practice clinical protocols, integrate laboratory test results automatically, alert clinicians to conditions that require follow up, recommend patient referral and refer patients electronically remind clinicians and patients to improve adherence to treatment, integrate point of care testing and behavioral change communications, and even integrate short, targeted, in-service training for clinicians. We have only scratched the surface of this potential. Here we see a graph from an electronic first EMR. It shows that the rate of testing of new antenatal patients for HIV increased 15 percentage points. The EMR was used to identify the problem: low testing rate for new antenatal patients. The data were used for “supportive supervision” that encouraged clinicians to improve their practice in this area. The EMR was also used to measure the results of the intervention. In the next graph we see what is possible when a group of facilities share the same EMR, and in the case share the same patient record database. The graph shows HIV prevalence for each clinic. These data are in real-time, and can be used to target interventions by clinic or clinic catchment area.

The evidence indicates that an EMR can have direct and positive impact on the quality of care and health outcomes However, to realize this potential on a large scale the use of EMRs must be scaled up and sustained The challenges to scale up and sustainability are immense I wrote the words on this slide in 1997 as part of a national assessment of health information systems in Egypt We see these challenges to varying degrees in every country

Despite these challenges there are reasons for optimism. Ten years ago many experts in health informatics would have considered it impossible to sustain an EMR in less developed countries. Today there are several successful examples. The Baobab Antiretroviral Treatment (BART) system in Malawi has been used to register more than 800,000+ patients and is being used to manage treatment for more than 18,000 ART patients. The Community Health Information Tracking System (CHITS) in the Philippines has been used successfully for 3.5 years in one health center, and is now used by 25 facilities. An implementation of OpenMRS in Haiti has been used for four years and is now used to manage more than 670,000 patients. SmartCare, an ART patient record system in Zambia, has been deployed to 200 facilities and is now being used to manage more than 200,000 patients on ART. The Zambian Electronic Perinatal Record System (ZEPRS) in Lusaka has been in use for four years by 25 interconnected facilities. ZEPRS is being used to manage records for more than 270,000 perinatal patients and has been used in more than three million patient encounters.

Successful projects are encouraging, but we should not underestimate the scale of the national health information system. We need to think well beyond pilot projects. We need to think carefully about how to scale up the use of EMRs from the beginning. For example, let’s look at the health information system in the Philippines. The reporting system is hierachical. It has many levels. Patient level data are collected from Barangay Health Stations (BHS) and Rural Health Units (RHU) Data are consolidated at each level, and are reported to the next level It is a large system. Nearly all data consolidation and reporting are done manually. Between 12 percent and 17 percent of nursing time is spent manually compiling reports. This takes time from patient care, and introduces many errors in the data. There are reporting delays or “lags” at each level. Reporting delays and time required for data consolidation and encoding results in late reports to decision makers. Introducing an EMR has great potential to improve this situation, but it will take time. And we must recognize that, while advancements in technology and improvements in our approach make EMRs sustainable in more locations, we may not be able to push them to the very edge of the health care system in every location. The feasible edge is determine by the carrying capacity of the environment, including local government. We must also recognize the respective roles of local and national government in this decentralized system. While encouraging central government to adopt and promote interoperability standards, We must also lay a firm foundation that makes it possible for local governments to adopt, deploy, and sustain an EMR.

In a decentralized system EMR deployment cannot be forced from the top down. It must be adopted, deployed, and sustained by local governments. The role of national government is to adopt and promote interoperability standards, and to encourage local governments to improve their information systems. We must help to lay a firm foundation that enables local governments to scale up the use of EMRs under local power. That means within local financial, human, organizational, and technical support capacity. The following guidelines for technical assistance experts should be useful. 1. Make sure it’s not your idea. If the demand is not indigenous, the system will not be sustained. 2. Make it work for clinicians. If the system does not meet the needs of clinicians it will fail. 3. Know where your system fits into the bigger national health information system picture. 4. Build a broad national partnership designed to long out-live the project. Include local, provincial, regional, and national government, academic institutions, the private sector, international donors, and international standards organizations. 5. Encourage national agencies to adopt and promote necessary national and useful international standards 6. Make sure the system is affordable for local governments. If it is not, it will not be sustained and will not scale. 7. Make sure local governments understand all the requirements and implications of introducing and sustaining an EMR. Make sure local governments have some skin in the game from the beginning. They should at least make a collateral investment in initial cost and should be responsible for most if not all operating costs. If this is not the case, there should be no confidence that the system can be sustained. Genuine sustainability cannot be based entirely on continued operating funding from donors. 8. Finally, as in a successful military operation, shorten the lines of supply and technical support. Local governments need sources of supply for materials and technical support to be as close as possible. Local businesses, universities, and NGOs can all be good sources of supply.

We began with one pyramid and will finish with another. This structure is more than 4,500 years old. For more than 3,800 years it was the tallest man-made structure in the world. It is the result of a common vision, a plan, a good foundation, the work of 20,000 team members over 20 years, and was built one stone at a time. We should approach the scale up of EMRs in health information systems the same way.

Electronic Medical Records for Public Health The challenge of scaling up under local power A presentation given at the Public Health Information Network conference, September 2009, in Atlanta, Georgia, USA

Electronic Medical Records for Public Health The challenge of scaling up under local power PHIN 2009 3 September 2009, Panel 3 Gordon M. Cressman, Senior Director Center for Information, Communication, and Technology International Development Group

The challenge It does not work this way

The challenge Because this is the bottom of the pyramid It starts out well. But records accumulate. And we are forced to archive them.

The challenge And clinicians see no benefit from their reporting efforts

The challenge And consolidation squeezes out valuable information Patient Records One patient with 5 prenatal visits at 1 barangay health station. Rural Health Unit (RHU) reports Local Government Unit (LGU) reports Data for patients are talied. 5 prenatal visits for 1 patient = 1 for this barangay health station. Data from barangay health stations are consolidated. Useful information about barangay health stations and patients are lost. Data from RHUs are consolidated.Useful information about RHUs, barangay health stations, and patients are lost.

The challenge But an EMR tightens the feedback loop and preserves information

The challenge And with intelligence and supportive supervision an EMR can do so much more

The challenge But the challenges to scaling and sustaining an EMR are immense … Equipment waits for repair and personnel ration paper sheet by sheet. … There is little prospect of keeping persons with good, marketable technical skills. In many positions, employees are so poorly paid they merely go through the motions required to collect a check… Record keeping systems in most facilities are rudimentary, consisting of ledger books and hand-drawn tables. Paper filing systems are crude and disorganized. Prescriptions are hand-written on odd scraps of paper and impaled on a pin. The environment for computer equipment is hostile. Electrical systems are crude and ungrounded. Plumbing systems leak. Buildings lack environmental control and are open to weather and dust. Getting people, materials, and data from one place to another can be difficult. Travel time to some facilities makes it hard to distribute and collect forms, monitor data collection, and provide the necessary technical support. … -- Egypt 1997

The challenge Fortunately there are reasons for optimism Community Health Information Tracking System (CHITS) 3.5 years, 25 facilities (Philippines) OpenMRS 4 years, 670,000+ patients (Haiti) SmartCare 200 facilities, 200,000+ patients (Zambia) Zambian Electronic Perinatal Record System (ZEPRS) 4 years, 25 facilities, 270,000+ patients, 3+ million encounters (Zambia) Baobab Antiretroviral Treatment (BART) 800,000+ patients registered, 18,000+ ART patients (Malawi)

The challenge But we should not underestimate the scale Barangay Health Station (BHS) Rural Health Unit (RHU) Each BHS submits reports monthly to a Rural Health Unit (RHU). On average each RHU consolidates data manually for 20 BHS each month. (This diagram shows only 5 for simplicity.) Consolidation and reporting takes nurses and midwives from 2 to 5 days each month. This time could be spent caring for patients. A Barangay Health Station (BHS) is the smallest point-of-care facility in the public health system. Each RHU submits consolidated reports monthly to a Local Government Unit (LGU). In this diagram, LGUs are municipalities and cities. Local Government Unit (LGU) Each LGU submits monthly reports to the Provincial Health Office (PHO). These reports may be consolidated at the LGU level, or may be disaggregated at the RHU level. Provincial Health Office (PHO) Regional Center for Health Development (CHD) Each province submits consolidated reports quarterly to the Regional Center for Health Development (CHD). This diagram shows only a portion. There are: 17 Regions 86 Provinces 1,628 LGU 2,300 RHU 16,000+ BHS Nearly all data consolidation and reporting are done manually. 12% - 17% of nursing time is spent compiling reports Each CHD submits consolidated reports quarterly to the Department of Health (DOH).

The challenge We must help lay a foundation that scales under local power Make sure it’s not your idea Make it work for clinicians Know where it fits into the big picture Build a broad national partnership Encourage national agencies to adopt and promote standards Keep it affordable for sub-national governments Make sure they are informed and have committed their own resources Shorten the lines of supply and technical support

Common vision + a plan + a good foundation + teamwork + one stone at a time gmc@rti.org

Electronic Medical Records for Public Health The challenge of scaling up under local power For more information contact: Gordon M. Cressman, Senior Director Center for Information, Communication, and Technology International Development Group gmc@rti.org