Promise and Pitfalls: A Look at California’s Regional Health Information Organizations JANUARY 2019 Contents About the Author 3Introduction Walter Sujansky, MD, PhD is the principal consul- 4What Is an HIO? tant at Sujansky & Associates, a California-based consulting firm that specializes in the analysis 6Types of HIOs and design of EHRs, disease registries, and solu- 1. Regional tions for health information exchange. 2. Enterprise or Private 3. EHR Systems That Enable Data Exchange Acknowledgment 4. National Vendor-Sponsored The author thanks the many HIO leaders and other experts who provided valuable insights 5. HIOs That Connect HIOs and data used in this report. Thanks also to 6. Niche Commercial Data-Exchange Services Ross Martin of 360 Degree Insights and Melissa Schoen of Schoen Consulting who helped with 7High-Value Use Cases the research and interview processes. 10Current Challenges and Potential Paths Forward About the Foundation 14Types of Entities That May Participate in The California Health Care Foundation is Regional HIOs dedicated to advancing meaningful, measur- 17Snapshot of Nine California Regional HIOs able improvements in the way the health care delivery system provides care to the people of 27How Regional Market Dynamics Shape the Role California, particularly those with low incomes of Regional HIOs and those whose needs are not well served by Key Factors the status quo. We work to ensure that people Case Studies: Real-World Implications of Regional Market have access to the care they need, when they Dynamics on HIO Strength need it, at a price they can afford. 29Looking Ahead: Policy and Technology Trends CHCF informs policymakers and industry lead- to Watch ers, invests in ideas and innovations, and New HITECH Funds on the Horizon for California’s connects with changemakers to create a more Regional HIOs responsive, patient-centered health care system. The Potential of TEFCA For more information, visit www.chcf.org. Fast Healthcare Interoperability Resources (FHIR) Blockchain Consumer-Mediated Health Information Exchange 31 Conclusion 32 Endnotes California Health Care Foundation 2 Introduction Some providers participated in efforts to create nonprofit alternatives to private data-exchange networks, known E ach time a person comes in contact with a health as community-based or regional HIOs, and at least nine care or social service entity, some amount of new such entities operate across various parts of the state data about that person is created. It could be as today. Participation in these networks, however, has simple as their current weight or employment status, or been variable and, in many regions, has not yet reached as complex as a summary of a two-week hospital stay. the critical mass needed to provide maximal value and All too often, entities must share in the care of a patient achieve financial self-sustainability. without actually being able to share much of the valu- able data they hold about that patient. The inability to At the same time, the EHR vendor market has been exchange information can result in care rife with some of consolidating, with fewer vendors serving an increasing the industry’s worst flaws, including wasteful spending, proportion of provider organizations in the state. Certain poor coordination, and reactive rather than preventive of these vendors have created capabilities to enable care. This report examines the various types of health robust data sharing among the customers of their own information exchange (HIE) resources available to products, and also collaborated with each other to create provider organizations in California, the value that stake- basic data-sharing networks across their products. These holders are seeking to realize from such resources, and developments have created new avenues for interopera- the specific role of nonprofit regional health information bility among the provider organizations using EHRs from organizations (HIOs) within this landscape. these largest of vendors. However, they have also further marginalized provider organizations that continue to use Government and private enterprise have both recog- EHRs not yet participating in these vendor-based data- nized the value of improving the ability of entities inside sharing networks or using older versions of EHRs that and outside of the health care system to easily exchange do not have these data-sharing features. In many cases, data that could inform patient care. On the government these providers, especially in the outpatient setting, side, initiatives and incentives to promote the exchange comprise smaller, independent physicians and commu- of health information have ranged from sweeping federal nity clinics that serve the safety-net population. efforts to more limited local ones. On the federal level, the Health Information Technology for Economic and Hence, for independent providers and safety-net clinics, Clinical Health (HITECH) Act, passed in 2009, has been nonprofit regional HIOs remain an important means to one of the most influential legislative efforts. It has offered connect and exchange data with collaborators in their billions of dollars in financial incentives focused on two communities. Because they are community-run and aim primary goals: first, increasing adoption of electronic to achieve total regional connectivity through inclusivity, health record (EHR) technologies; and second, enabling regional HIOs offer a healthy counterbalance to trends entities to share this newly created wealth of electronic in the private market. They guard against any one EHR health information through the creation of HIOs. vendor or other corporate entity gaining too much con- trol over vital data-exchange capabilities. They are also In California, EHR adoption flourished uniformly, while especially well suited to meet the needs of safety-net the growth of HIOs was more fragmented. State gov- patients and the providers who care for them. For exam- ernment efforts to standardize and coordinate HIO ple, a regional HIO can include nontraditional service development across the large California marketplace providers, such as housing agencies or substance-use were largely superseded by local market dynamics and treatment facilities, that are vital to the well-being of development trajectories. Parts of the California provider vulnerable populations but are otherwise excluded market are dominated by large private health systems, from data exchange occurring via EHRs or within private many of which could afford to purchase robust EHR health systems. Also, the nonprofit nature of regional systems and develop private, exclusive HIOs to enable HIOs fosters collaboration and communication among the exchange of data within their health systems. Many members of competing health systems and EHRs that smaller provider entities, often members of the safety- private-market forces might otherwise inhibit. This col- net or independent physician communities, were left out laboration is especially critical for the care of safety-net of the more robust EHR and private HIOs and, as a result, patients, whose frequent use of emergency services1 often lacked early access to data-exchange capabilities. and specialty care referrals2 makes them more likely to Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 3 see providers belonging to multiple health systems and Technical Components using multiple EHRs. $$ Implemented data interfaces. The HIO’s means of sending and receiving patient data; sometimes While regional HIOs have come a long way in California includes user-interface features integrated within the since the first one was founded in Santa Cruz in 1996, existing EHR systems of participating enterprises. they have a long road still to travel before they fully real- ize their potential to help create more connected and $$ Master patient index (usually). Consolidates patient coordinated care systems within the state. Regional HIOs demographic information and unique identifiers currently touch an estimated 22 million lives in the state, across participating enterprises for the purpose of but only about half of California’s hospitals participate, matching a person’s clinical information held by and 23 of California’s 58 counties still lack any significant different providers. regional HIO presence. Many regional HIOs are strug- $$ Record-locator service (sometimes). Tracks the pres- gling to find sustainable financial footing and to prove ence and location of specific patients’ data among their value in the face of well-funded private alternatives. the participating enterprises. Investing in their success offers a tangible path to tack- ling the fragmentation in California’s health care system, $$ Patient-data repository (sometimes). Centrally which remains a persistent source of frustrations, ineffi- aggregates, normalizes, and stores patient data ciencies, and disparities. submitted by participating enterprises. Many HIOs, however, just transmit patient data from point A to point B and do not maintain a persisted copy of the data. What Is an HIO? $$ Data-sharing applications (usually). Provide vari- In general, health information exchange organiza- ous functions, including search, document retrieval, tions, also known as HIOs, are entities that facilitate the alerts, and data analysis, for the patient data that are exchange of patient health information among the enter- accessible via the HIO. The most commonly included prises comprising a health care delivery system. They application is a web-based portal for the search and/ can be either community-based and nonprofit, known in or retrieval of patient documents. Other applications California as regional HIOs, or owned and operated by a may include a subscription and routing mechanism private enterprise. for event notifications (e.g., inpatient admissions) or a bulk data-export capability to populate analytical databases and population-health tools. Components of HIOs No two HIOs are exactly alike, but they typically have sim- ilar organizational and technical components to enable the sharing of patient data among their participants. Organizational Components $$ Documented data-exchange standards. Agreed- upon formats for the exchange of health information that all participating enterprises will support. $$ Participation agreement. Formalized relationship between the HIO and the enterprises that participate in it, including payment terms and legal obligations. $$ Data-use agreement. Agreed-upon allowed uses of data received via the HIO — for example, limiting use to treatment purposes or prohibiting the bulk aggregation of data for insurance-contracting pur- poses. California Health Care Foundation 4 Distinct Technology Models Although HIOs share many components, a key distinction among many of them lies in the technology models that underlie their data infrastructures. The technology model that an HIO chooses fundamentally shapes how it collects, organizes, and exchanges its data, and therefore what use cases it can offer its members. There are three commonly used technology models: federated, hybrid, and centralized models. Centralized Model BENEFITS The centralized model operates like a hub and spoke $$ Quickly scalable whereby data are physically aggregated and managed $$ Lower cost to implement centrally. An HIO is responsible for operating the cen- tralized technology and making that clinical information CONSTRAINTS available to HIO participants through it for permitted $$ Limited potential for data consolidation and analysis purposes agreed to by those participants. $$ Lower likelihood of matching patients’ data between BENEFIT organizations $$ Rich set of aggregated and consolidated patient data, enabling more analytical use cases Hybrid Model The hybrid model is similar to the federated model in CONSTRAINTS that it mostly relies on legal and governance agree- $$ Difficult to normalize and standardize data ments, but it has a thin layer of technology that $$ More difficult to scale centralizes some patient data, like identities and record- locator services. This thin layer of technology and $$ Requires greater trust among participating members centralized data storage serve to improve the coordina- tion of data exchange. Federated Model In the federated model, data are stored and man- BENEFIT aged by a distributed network of HIO members. These $$ More scalable than fully centralized model peer organizations adopt standards and processes for sharing information under a common legal agreement CONSTRAINT among participants. If each participant adopts the abil- $$ Limited potential for use cases that require ity to communicate by those standards, participants can data analysis query one another to search for information on com- mon patients without relying on any central technology operator. Note: For a more detailed comparison of these technical models, refer to Douglas B. McCarthy et al., “Learning from Health Information Exchange Technical Architecture and Implementation in Seven Beacon Communities,” eGEMs (Generating Evidence & Methods to improve patient outcomes) 2, no. 1 (May 5, 2014), accessed December 14, 2018, www.ncbi.nlm.nih.gov (PDF). Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 5 Types of HIOs accountable care organization (ACO). While these HIOs can include many different participants, such as hospitals, Generally, any entity facilitating some form of HIE activity, clinics, laboratories, and even payers, they are typically regardless of the underlying technology or governance open only to organizations contractually partnered with model it uses, can be considered an HIO. Such enti- the business entity that built the HIO. That purchasing ties can generally be grouped into one of the following entity has sole control over the exchange’s data and categories: available features. Examples of enterprise or private HIOs in California 1. Regional include those operated by Kaiser Permanente, Sharp Regional HIOs are distinct from other HIE resources in HealthCare, Dignity Health, and Monarch HealthCare. that they (1) serve defined geographical regions, ranging from a single county to an entire state; (2) are open to any While private HIOs are often perceived as presenting health care enterprise that serves patients in a region, fewer legal and business liabilities than regional HIOs, regardless of its business affiliations or choice of infor- they do have limitations. For example, provider organi- mation technology vendors; and (3) are nonprofit entities zations can exchange data only with other organizations primarily concerned with improving the quality and cost- that are part of the same business entity served by the effectiveness of health care in a region through greater private HIO. Data generated at “outside” provider orga- availability and sharing of patients’ health information. nizations — at which a patient currently “in network” may have been seen in the past or in an emergency — are Examples of regional HIOs in California include Manifest therefore not available. Furthermore, business entities MedEx, Santa Cruz HIO, and North Coast Health must finance the entirety of the private HIO licensing and Improvement and Information Network. operations and cannot share those costs with outside organizations that might otherwise be part of and con- While regional HIOs can offer a healthy counterbalance tribute financially to a more inclusive exchange, such as to more exclusive private-market options for exchanging a regional HIO. data, they do face unique constraints. Chief among them is financial sustainability. In California, most regional HIOs rely on a mix of subscription fees and philanthropic or 3. EHR Systems That Enable Data government grants. If one or more large hospitals in a Exchange given region opts to build a private HIE network instead When a single EHR system has been widely adopted in of participating in the regional HIO, then that HIO may a particular region and it contains robust data-exchange lose critical subscription revenues and may be forced to features, that EHR can act in some ways like an HIO. Data raise rates on smaller entities. In turn, those entities may exchanged through the EHR has the advantage of always opt out of the regional HIO themselves, due to budget- being integrated directly into the EHR user interface. ary constraints. Ultimately, the HIO may become overly Provider organizations can also import patient records dependent on grant funding, the long-term availabil- from other facilities that use the same EHR and have ity of which can be unpredictable. Another challenge enabled its data-exchange features. for regional HIOs is that many provide access to data primarily through a web portal rather than via full EHR By far the most prominent example in the state is Epic integration. EHR integration is more costly for regional and its Care Everywhere network. The Epic EHR is widely HIOs to implement, but far more attractive for busy pro- used in California by many hospital systems (e.g., Sutter, vider users. Providence, Memorial Care), academic medical cen- ters (e.g., Stanford, UCSF, UC San Diego, UCLA), IDNs (e.g., Kaiser Permanente, Scripps Health, Cedars-Sinai), 2. Enterprise or Private and community clinic networks (e.g., OCHIN [Oregon An enterprise HIO is built specifically to meet both the Community Health Information Network], Community financial and clinical objectives of a distinct business Medical Centers). entity such as a hospital system, independent physician association (IPA), integrated delivery network (IDN), or California Health Care Foundation 6 On the downside, provider organizations cannot access patient records from facilities that use other EHRs or 6. Niche Commercial Data-Exchange have not enabled their EHR’s data-exchange features. Services Furthermore, there is no centralization or curation of There are numerous commercial, for-profit companies patient identities, so matching rates can be poor and that provide specific data-exchange services to medi- depend on the quality of patient demographic informa- cal communities. Services tend to focus on a particular tion provided by the two facilities attempting to exchange aspect of care, such as the sharing of controlled substance data. EHRs also often lack the more robust features that prescribing data across hospital emergency departments certain regional and private HIE networks provide, such (EDs), the sharing of patients’ Physician Orders for Life- as encounter notifications, referral management, results Sustaining Treatment (POLST) directives, or the sharing delivery (i.e., “pushing” patient data), or the ability to of care plans for care coordination. aggregate and analyze patient data in bulk across mul- tiple EHR instances. Examples of vendors providing such services include Collective Medical Technologies, ACT.md, and Vynca. 4. National Vendor-Sponsored This type of HIO is funded and operated by a consor- tium of commercial vendors who have the shared goal of High-Value Use Cases enabling interoperability among their respective health From high hospital readmission rates to the costly order- information technology (IT) products, such as EHRs. ing of duplicative tests, there are many major pain points Access to the network is typically tightly integrated into in health care that persist at least in part because pro- each vendor’s respective IT product and available to its viders are unable to easily exchange information with customers with minimal custom development or configu- one another about the patients they share. The ability ration. Since these networks’ members tend to be EHR to exchange information will only grow more important vendors, they present benefits (data integrated directly as the United States health care system increasingly into the EHR) and challenges (less robust features, inabil- embraces value-based payment models. These new ity to access data from facilities that have not joined the models will require payers and providers to better under- network or use a nonmember EHR) similar to the afore- stand which of their patients are at risk for poor outcomes mentioned individual EHRs offering HIO-like exchange. and then effectively target resources, from both within and outside of the traditional health care system, to Examples of national vendor-sponsored HIOs include prevent those outcomes. That shift will require many Carequality, whose network is available to users of Epic, new capabilities, including robust data collection and athenahealth, eClinicalWorks, and NextGen Healthcare analysis, and proactive coordination with nontraditional EHRs; CommonWell Health Alliance, whose network is entities that have an outsized effect on health, such as available to users of Cerner, Meditech, Evident, athena- food banks and housing authorities. health, eClinicalWorks, and Greenway Health EHRs; and DirectTrust. The good news is that research has shown that HIOs can help providers meet these new expectations. Substantial HIO adoption has been associated in some studies with a 5. HIOs That Connect HIOs lower rate of hospital readmissions, fewer ED admissions, These HIOs serve as “gateways” between other existing fewer duplicated procedures, improved medication rec- networks, including enterprise HIOs and regional HIOs. onciliation, greater immunization and health record They provide services to normalize searches for and completeness, better identification of drug-seeking delivery of patient data across distinct HIOs, which can behaviors, and reduced total cost of care.3 Whether have differing data formats and standards. California will eventually reap similar benefits depends on both increasing HIO participation rates and strength- Examples nationwide include eHealth Exchange and ening the capabilities of the HIOs to ensure they deliver Strategic Health Information Exchange Collaborative the utmost value. (SHIEC). Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 7 The use cases outlined in Table 1 comprise the six areas members. Although the different regional HIOs are sup- in which regional HIOs in California are focusing the porting these use cases to varying degrees today, most majority of their efforts related to enhancing their capa- have acknowledged and prioritized them as high-value bilities and ultimately delivering greater value to their goals to pursue. Table 1. Six Use Cases of Regional HIOs in California, continued USE CASE / DESCRIPTION CLINICAL VALUE BUSINESS VALUE DATA REQUIRED Longitudinal patient record. HIOs $$ Reduces potential for $$ Enables provider entities $$ Medication allergies; can provide access to patients’ health errors caused by poor that bear financial risk results of past labora- information originating from numer- information about aller- to avoid poor clinical tory, imaging, and other ous sites of care either by aggregating gies, prior treatments, outcomes and waste- diagnostic procedures; data from across sites into a single and other informa- ful utilization, such as previously diagnosed physical patient record or by enabling tion critical to clinical redundant testing. and treated disorders; retrieval of data on demand from decisions. currently or previously such sites, effectively creating a single taken medications; and $$ Increases clinicians’ virtual patient record. sites and frequencies chances of making of previous medical well-informed and encounters evidence-based care decisions. Real-time event notification. HIOs $$ Enables proactive $$ Enables stakeholders to $$ Relevant health care can establish “publish/subscribe” intervention, timely divert patients to more events warranting infrastructures, in which certain clini- outpatient follow-up cost-effective sites of notification may include cal events are always reported to the after ED visits and care and to prevent ED visits, hospital HIO, which then forwards information hospital discharges, and costly avoidable hospital admissions, hospital about the events to parties that have tracking of patients’ admissions by interven- discharges, and appoint- expressed interest in being notified attendance at important ing proactively and ments for specialist of them. This mechanism can be specialist visits. arranging alternative visits. configured on an event-specific and care arrangements. $$ Hospitals provide the patient-specific basis. For example, data for ED visits, hospi- the care-management team at a health tal admissions, and insurer could be notified each time hospital discharges, a high-risk patient is seen in an ED, typically via HL7 ADT or a primary care physician could be (“admit/discharge/ notified upon the discharge of one of transfer”) messages. The her patients from the hospital. referring and/or consult- ing physician provide(s) the data for scheduled specialist visits. Results reporting and document $$ Recovers time other- $$ Saves time and money $$ Test results and clini- delivery to ambulatory providers. wise spent by clinical spent on the staff and cal documents from An HIO can provide a central “hub” and administrative staff technologies required hospitals sent via HL7 for receiving, translating, and forward- translating, faxing, and to maintain numerous interfaces to outpatient ing diagnostic results and clinical receiving diagnostic electronic data inter- providers, who receive documents between hospitals and results and clinical faces to different trading and integrate the data outpatient providers. In this model, documents. partners. into their EHRs each hospital and outpatient provider $$ Moves test results and need only maintain a single interface clinical documents to the hub, which translates the data that otherwise exist in formats appropriately to accommo- fragmented faxes into date all senders and recipients. This the EHR, allowing them approach replaces the highly ineffi- to more easily inform cient and costly process of having each clinical decisionmaking. hospital and outpatient organization within a given health care ecosystem maintain numerous electronic data interfaces and perform many separate data translations. California Health Care Foundation 8 Table 1. Six Use Cases of Regional HIOs in California, continued USE CASE / DESCRIPTION CLINICAL VALUE BUSINESS VALUE DATA REQUIRED Data submission to public health $$ Ensures more complete $$ Saves time and money $$ Diagnosis, immuniza- agencies. For provider organizations records of patients’ spent monitoring test tion, and other clinical that are already submitting lab results immunizations and results for those that data required by health and immunization data via some reportable diseases, require reporting to agencies and submitted means of data exchange, the HIO facilitating and improv- public agencies. by hospitals and outpa- could analyze, appropriately format, ing their future care. tient providers $$ Saves time and money and transmit these data to the public spent building and $$ In California, the health department on behalf of the maintaining the separate CalREDIE (reportable provider organizations. This model interfaces required to diseases) and CAIR2 obviates the need for provider organi- complete electronic (immunizations) public zations to build separate interfaces submissions, or spent health registries receive, to public health agencies and can manually submitting via store, and process automatically monitor all test results to web portals. these data. identify and transmit those that require reporting. Data aggregation for population- $$ Enables proactive identi- $$ Enables proactive identi- $$ Clinical data in longitu- health and utilization-management fication of patients at fication of patients at dinal patient records; analytics. HIOs with data connections risk for certain poor risk for costly and avoid- claims records from to numerous health care providers outcomes and the able outcomes and the health insurers can receive, integrate, and normalize proper allocation of care proper allocation of care clinical data pertaining to individual management resources management resources patients in a physical data repository. needed to avoid those needed to avoid those These data can then be made available outcomes. outcomes. for analysis to interested stakeholders, $$ Enables the proac- $$ Enables retrospec- either by exporting the consolidated tive identification and tive analysis of care records for all applicable patients correction of patient outcomes and costs to the stakeholders, or by providing care not aligned across a population to analytical software to process the data with evidence-based identify patterns associ- directly on the data repository. practices. ated with higher-value care. Coordinating with nonmedical $$ Enables care managers $$ Better targets resources, $$ Data on behavioral providers to address patients’ social to better facilitate care both medical and social, health care, substance needs. An HIO that is coordinating coordination and follow to avoid more costly and use disorder treat- care between clinical provider organi- up on necessary refer- avoidable outcomes, ment, and use of social zations and delivering longitudinal rals for both social and such as hospitalization. services — contrib- patient records, event notifications, medical needs. $$ Better tracks and uted by hospitals, and data aggregation can facilitate outpatient clinics $$ Enables providers to measures the cost-effec- “whole person care” by including data (including community more effectively screen tiveness of social service on behavioral health care, substance- health clinics), county for and address the referrals with respect to use treatment, and use of social mental health facilities, social factors that can medical utilization. services. An HIO can add additional substance use disor- harm a person’s health. value by providing built-in tools for der treatment centers, care coordination, referral manage- homeless shelters, food- ment, and patient tracking. assistance agencies, employment agencies, and corrections facilities Notes: CAIR2 is California Immunization Registry; CalREDIE is California Reportable Disease Information Exchange; HL7 is Health Level Seven International. Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 9 regional HIOs include the nearly 200 California hospitals Current Challenges and not currently participating in a regional HIO, other poten- Potential Paths Forward tial participant entities, vendors whose technologies could improve HIO capabilities, state and federal officials While community-based regional HIOs have shown who can clarify and revise regulations to be more sup- promise in other parts of the country, their potential portive of data exchange, and philanthropies and other has yet to be fully realized in California. Regional HIOs funders that can help remove up-front financial barriers can certainly do more work to prove their financial and to HIOs’ achievement of greater scale. Table 2 describes clinical value, especially in the face of steep competition some of the key obstacles that regional HIOs, together from private-market alternatives, but they alone cannot with other key stakeholders, must navigate in order to close California’s data-connectivity gap. Other stakehold- continue progressing toward a more connected, coordi- ers critical to the long-term success and sustainability of nated system of care throughout the state. Table 2. Key Challenges and Potential Paths Forward for Regional HIOs Limited Participation by Commercial Health Systems CHALLENGE POTENTIAL PATHS FORWARD For executives at commercial hospital systems and health $$ HIOs can add additional services that provider organizations systems, the business case for participating in regional HIOs is need and cannot obtain elsewhere, such as encounter notifi- often unclear. These organizations are increasingly using data- cation, data aggregation, referral management, etc. exchange services provided through their EHR vendors (such $$ Reduce the subscription fees charged to participating as Carequality, CommonWell Health Alliance, and Epic Care provider organizations whether through state or philan- Everywhere) and investing in private HIE solutions to achieve thropic subsidies, the expansion of the number of HIO the levels of clinical integration that they require, rather than participants that pay subscription fees, and/or a reduction in participating in regional HIOs. In this environment, the value the overall operational costs of the HIO (perhaps by sharing proposition of regional HIOs is generally decreasing and the technology platform or staffing with other HIOs). remains greatest only in fragmented medical communities, where significant provider consolidation has not yet occurred $$ Mandate or incentivize participation in a regional HIO via (in California, typically rural communities). As commercial directives issued either by the state or payers. hospitals and health systems are often major contributors of funding to regional HIOs, this trend may reduce available funding over time if regional HIOs are not able to provide additional services of value to these organizations or if partici- pation in regional HIOs is not somehow mandated. Cumbersome Provider Workflow Due to Limited EHR Integration CHALLENGE POTENTIAL PATHS FORWARD Access to the comprehensive patient record is often available $$ Whenever possible, regional HIOs should aim to integrate to users of regional HIOs only via a web-based portal applica- their services directly into the EHR products used by HIO tion, which requires clinicians to leave their EHR tool, log in to participants. This is especially true for ambulatory EHR a different application, and reenter the patient’s demographic products, for which the integration process is often more data before accessing the patient’s record. When a separate difficult and less well supported by the provider organiza- application is the only means for accessing HIO data, actual tions that use them. use of the HIO is significantly diminished. While integration of $$ At the same time, HIOs would benefit from an attempt HIO services directly into the EHR is preferred, it also tends to to standardize the APIs and other integration features of be far more costly. EHR products, so that they more consistently support the existing and envisioned functionalities of HIOs. A working group of EHR vendors and HIOs could convene to address this issue, possibly in collaboration with the HIMSS EHR Association and the Strategic HIE Collaborative (SHIEC), respectively, or similar bodies. California Health Care Foundation 10 Transition to Centralized Data Storage Models CHALLENGE POTENTIAL PATHS FORWARD Most regional HIOs have historically not physically aggregated $$ HIOs could develop centralized data repositories for aggre- the clinical data to which they provide access. Typically, much gating clinical data submitted by participants. If the HIO’s of the data remains stored locally at participating provider core technology does not support this function, a separate organizations and is retrieved only on demand in the context data-repository technology can be procured and integrated of a specific patient search. HIOs are now more aggressively with the core technology. This data repository should pursuing the strategy of physically aggregating and storing include a standardized/normalized data model that supports patient data within their own data repositories, as this allows relevant data analytics. them to provide additional data-delivery and analytic services $$ HIO participants could be incentivized to contribute their and to differentiate themselves from the data-exchange data to the centralized repository. Incentives could include capabilities that are increasingly built into EHR products. free data-normalization services or discounts on subscrip- This transition, however, requires HIOs to upgrade their tion fees. technologies, their data-normalization capabilities, and their governance documents. Also, not all participants in HIOs wish $$ HIOs should seek to make participants feel as comfort- to submit all of their clinical data to an external, centralized able contributing their data to the centralized repository repository due to privacy or data-ownership concerns. Hence, as possible. Steps may include (a) requiring only a subset it may be time-consuming and/or not always possible for HIOs of structured clinical data initially, such as encounter dates, to fully transition to this model of centralized data storage. primary diagnoses, lab results, prescribed medications, performed procedures, blood pressure, and weight; (b) implementing robust security and access controls on the aggregated data to minimize risk of unauthorized disclo- sure, as well as performing formal penetration testing on the data repository; and (c) developing specific policies regarding access to data in the repository and formalizing these policies in all participation agreements such that they cannot be changed without each participant’s consent. Normalizing Data in Centralized Models CHALLENGE POTENTIAL PATHS FORWARD As noted above, HIOs are increasingly seeking to provide $$ Regional HIOs can engage third-party data-cleansing and additional value by physically aggregating data from their data-normalization firms, such as Diameter Health, InteropX, multiple participants in central data repositories. Centralized and Redox. aggregation allows HIOs to consolidate and deliver relevant $$ Stakeholders can lobby federal regulators to increase the data in batch mode to payers, ACOs, and other participants level of data standardization required of EHR vendors, clini- for analysis. It also allows HIOs to, themselves, perform data cal laboratories, and other contributors of data to HIOs. analytics for risk stratification, chronic disease management, and quality improvement. However, data aggregation, consoli- dation, and analysis require that heterogeneous clinical data from multiple sources be standardized and normalized, which remains a complex and time-consuming task. For example, lab results from hospitals are often represented using the hospi- tals’ own coding systems, rather than the standardized LOINC coding system, necessitating code mapping and translation when the data are aggregated. Also, the representation and completeness of clinical data transmitted using the C-CDA document standard can still vary considerably depending on the specific hospital or ambulatory provider that generates these data. Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 11 Matching Patient Data in Centralized Models CHALLENGE POTENTIAL PATHS FORWARD Correctly matching data received from different organiza- $$ Outside experts or HIOs themselves could more rigor- tions to the appropriate patient remains a major challenge ously study the performance of HIOs’ MPI technologies and for not just HIOs, but many payers and providers as well. identify any weaknesses and their root causes. Although the master patient index (MPI) technologies used by $$ Technology vendors, such as Verato, can continue honing HIOs have improved, certain HIOs have reported that wholly their solutions for improving patient-matching performance reliable patient matching still requires a considerable degree and consider partnering with regional HIOs as customers. of manual curation, i.e., manual review of potential duplicate records or uncertain match results. In the absence of sufficient $$ State officials who oversee the CAIR2 and CURES databases manual effort devoted to this task, multiple identities may exist could more rigorously analyze the patient-matching for individual patients within an MPI, which results in fragmen- techniques used and the accuracy of results delivered by tation of these patients’ data and incorrect or incomplete their databases. They could identify any weaknesses and results in response to data queries. Such errors undermine work to resolve them in order to improve the match rate and clinicians’ confidence in the HIO’s data and can reduce their reduce the likelihood of incorrect matches. The state may use of the HIO. A similar patient-matching problem exists consider procuring a more advanced matching technology. when clinicians access the state’s immunization registry and prescription drug-monitoring program (i.e., CURES) database. This problem undermines the value being provided by HIOs that directly interface to these state databases as a conve- nience for their users. Need for Centralized Consent Management CHALLENGE POTENTIAL PATHS FORWARD Currently, a patient’s consent to have her data shared via $$ Regional HIOs could implement centralized consent- an HIO is collected and stored separately by each provider management systems that can be populated by all provider organization participating in the HIO. Each participant organi- organizations participating in a given HIO and accessed zation’s local interface is configured accordingly, so that by the HIO’s technology at the time that data requests only the data of patients who have consented to participate are issued. Such a technology could, for example, enable in the HIO are made available in response to requests from patients to specify global consent preferences regarding other organizations. This model makes it difficult for HIOs to HIO data access across all their health care provider organi- ascertain whether the absence of a patient’s data in response zations, as well as enable an HIO to cross-reference its MPI to a request is because there are no data for that patient at to its consent records to identify and contact patients who the responding organization or because the patient has not have not consented to sharing some or all of their data. consented to have her data at that organization shared with $$ Regulators could also further clarify and align statutes and the given HIO. This leads to situations, for example, in which regulations affecting patient privacy to make centralized a provider knows that a patient has received services at an consent management simpler. As outlined in the challenge organization, but retrieves no data on that patient from that below regarding privacy regulations, the nuances and organization, leaving uncertainty as to whether the cause is a discrepancies that exist among the state and federal regula- consent issue or an error in the HIO (for example, a patient- tions are incredibly complex. matching error). Such uncertainty can undermine confidence in the HIO’s data among provider organizations and reduce their use of the HIO. Need for Referral-Management Capabilities for Population-Health Management CHALLENGE POTENTIAL PATH FORWARD Most HIOs do not yet provide referral-management and HIOs could implement closed-loop referral-management care-management tools. Such applications enable referral capabilities that are tightly integrated with their core requests to be made and consult notes to be delivered via technologies and, if possible, with the EHRs of participating the HIO technology, as well as oversight to be provided for organizations. If the HIO’s core technology does not support the referral process (e.g., referral authorization, appointment this function, a separate referral-management technology reminders, transportation assistance). Certain commercial could be procured and integrated. Available third-party vendors offer third-party referral-management solutions, commercial products include ACT.md, Netsmart, and CrossTx. but they require technical integration with the HIO’s core technology and may entail the use of separate interfaces or applications by HIO participants. California Health Care Foundation 12 Need for Robust Real-Time Encounter-Notification Capability CHALLENGE POTENTIAL PATH FORWARD Regional HIOs are just beginning to provide real-time event- HIOs could develop or procure publish/subscribe mechanisms notification services. However, the necessary real-time HL7 and patient-attribution data for routing event notifications to ADT data submissions from all hospitals and referring physi- appropriate recipients. Third-party commercial vendors such cians are generally not yet in place, nor are the mechanisms as Audacious Inquiry and Collective Medical Technologies for stakeholders to subscribe to and receive alerts regard- provide such services. ing events of interest. One key challenge in implementing this feature is maintaining an up-to-date mapping between patients and the providers and insurers who are interested in and authorized to receive relevant alerts about them. Privacy Regulations Inhibiting Behavioral Health and Nonclinical Data Sharing CHALLENGE POTENTIAL PATHS FORWARD HIPAA and state regulations impose additional consent $$ To address restrictions on the sharing of mental health and requirements for the sharing of mental health and substance- substance-use data, HIOs and the participating provider use data by provider organizations. These regulations typically organizations that serve patients with those needs could require an affirmative (“opt-in”) consent model for these collaborate to better streamline the process of consent- specific types of data, even when the default consent model ing patients specifically to share these types of data. for an HIO is “opt-out.” The likelihood of an HIO obtaining Streamlining would primarily require workflow and policy that much more burdensome level of consent from all partici- changes at those provider organizations but could also pants is far lower. This dynamic can result in significant gaps in involve HIOs implementing centralized consent-manage- the mental health and substance-use data available via HIOs. ment systems (see above). HIPAA also prohibits the sharing of protected health informa- $$ To address the HIPAA proscriptions on the sharing of PHI tion with organizations that are not “covered entities” in the with noncovered entities, HIOs could explore the creation absence of explicit patient consent. Such organizations include of business-associate agreements (BAAs) with social service various social service agencies, such as housing agencies, agencies, which may allow data sharing without explicit employment agencies, food-assistance agencies, and correc- patient consent. Given that provider organizations already tional facilities, whose data are relevant to coordinating “whole have BAAs in place with their HIOs, this approach may person care” to underserved populations. require also modifying those BAAs. Stakeholders may also consider lobbying federal regulators for additional clarity on these legal complexities. Notes: API is application programming interface; CAIR2 is California Immunization Registry; C-CDA is Consolidated-Clinical Document Architecture; CURES is Controlled Substance Utilization Review and Evaluation System; HIMSS is Healthcare Information and Management Systems Society; HIPAA is Health Information Portability and Accountability Act; HL7 is Health Level Seven International; LOINC is Logical Observation Identifiers Names and Codes; PHI is protected health information. Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 13 Types of Entities That regional HIO. Each entity will weigh the benefits of partic- ipating against the costs of doing so. It will also examine May Participate in the types of data its regional HIO offers and expects Regional HIOs members to contribute — and whether those data align with the entity’s needs and capabilities. Ultimately, all of Nearly every type of health care entity, no matter how these factors amount to a question of value: Compared small or how specialized, could improve the care it offers to other data-exchange solutions available to a given patients by participating in data exchange. However, due entity, does its regional HIO meet enough unique needs to a host of business, financial, and technological rea- to justify the cost and effort required to join? sons, not every entity is equally likely to participate in its Table 3. HIO Participation Considerations for Various Health Care Entities, continued PARTICIPATION DATA THE ENTITY MAY… Reasons to participate Barriers to participation Contribute to the HIO Seek from the HIO HIGH LIKELIHOOD OF PARTICIPATION Private Hospitals $$ Exchange clinical information $$ Most use major EHR platforms $$ Inpatient and ED $$ Clinical data contributed with ambulatory centers as with built-in HIE capabilities admissions (ADT) by other hospitals part of referral network via the national HIE networks. $$ Lab, radiology, and $$ Outpatient encounter $$ Participation in risk-bearing or $$ The value of receiving data other results (HL7) notifications (ADT) pay-for-performance contracts from ambulatory centers has $$ Structured encounter $$ Outpatient lab, that require data exchange for been limited, especially when summary documents radiology, and care coordination the number of ambulatory (C-CDA) other results participants in the HIO is low. $$ Especially high use among $$ Unstructured summary EDs and outpatient clinics documents (e.g., within hospitals discharge summary) Public Hospitals $$ Participation in risk-bearing or $$ Costs associated with both $$ Same as private hospitals $$ Same as private hospitals pay-for-performance contracts up-front IT integration work that require data exchange for and ongoing subscription and care coordination maintenance fees $$ Less likely to benefit from the $$ The value of receiving data private HIO resources avail- from ambulatory centers has able to hospitals in larger been limited, especially when systems the number of ambulatory participants in the HIO is low. Large Outpatient Providers (multispecialty, community health centers, IPAs) $$ Participation in risk-bearing $$ Costs associated with both $$ Outpatient encounter $$ Notifications of hospital or pay-for-performance up-front IT integration work notifications (ADT) and ED encounters contracts that require data and ongoing subscription and $$ Outpatient lab, $$ Result data from hospi- exchange for care coordina- maintenance fees radiology, and other tals, other outpatient tion and population health $$ Alternatives for data results (HL7) providers, and lab/ $$ Especially need ED or exchange through participa- imaging centers $$ Medication lists and inpatient information from tion in private HIOs (e.g., in medication allergies $$ Outpatient medication hospitals, including encounter IDNs, IPAs, or ACOs) lists notifications and discharge $$ Immunizations and $$ Lack of integration with the summaries reportable diseases $$ Structured and providers’ EHRs may make unstructured summary access to HIO data cumber- $$ Structured encounter documents from hospi- some and time-consuming. summary documents tals, EDs, and outpatient (C-CDA) providers $$ Unstructured summary $$ Immunization registry documents (specialty records consult notes) California Health Care Foundation 14 Table 3. HIO Participation Considerations for Various Health Care Entities, continued PARTICIPATION DATA THE ENTITY MAY… Reasons to participate Barriers to participation Contribute to the HIO Seek from the HIO MEDIUM LIKELIHOOD OF PARTICIPATION Laboratory and Imaging Centers $$ Provides a single interface $$ Costs associated with both $$ Lab results and $$ Lab and radiology orders hub for delivering results to up-front IT integration work radiology reports many providers and ongoing subscription $$ Radiology images and maintenance fees $$ Aggregates and delivers results to disease registries, $$ Existing alternative channels population-health programs, to deliver results to ordering etc. providers (via dedicated HL7 interfaces) Payers (including county, state, and commercial plans) $$ Contribute claims data to $$ Subscription fees charged $$ Claims data $$ Real-time notification supplement missing clini- by HIOs of hospital, ED, and $$ Membership and cal data due to incomplete outpatient encounters $$ Data sharing agreements PCP-assignment data participation by provider can sometimes restrict payer $$ Structured clinical data to organizations access to clinical data, reduc- drive population-health $$ Access clinical data to ing the value of participation. and quality-measurement facilitate population-health, activities $$ Reluctance to share claims care-coordination, quality, data or membership data and pay-for-performance with competitors initiatives Small FQHCs, Community Health Clinics, and Small Physician Practices $$ Participation in risk-bearing or $$ Costs associated with both $$ Same as large outpatient $$ Same as large outpatient pay-for-performance contracts up-front IT integration work providers, although often providers that require data exchange for and ongoing subscription limited because of costs care coordination and maintenance fees and difficulties of EHR integration $$ More likely to use less $$ Lack of integration with the expensive EHRs that do not providers’ EHRs may make yet include access to other access to HIO data cumber- vendor-centric HIE networks some and time-consuming. $$ Less likely to be participating in a private HIO through an IDN, IPA, or ACO EMS Providers $$ Facilitate clinical care $$ Limited resources for $$ Clinical status during $$ Medication lists during patient treatment technical integration of transport to ED, includ- $$ Medication allergies and transport EMS information systems ing chief complaint, vital with HIO signs, and acuity $$ Problem lists $$ Prepare receiving ED for patient arrival $$ Limited business drivers $$ POLST/DNR forms for EMS integration $$ Track patient outcomes with HIOs subsequent to transport Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 15 Table 3. HIO Participation Considerations for Various Health Care Entities, continued PARTICIPATION DATA THE ENTITY MAY… Reasons to participate Barriers to participation Contribute to the HIO Seek from the HIO LOW LIKELIHOOD OF PARTICIPATION Urgent Care Centers $$ Facilitate diagnosis and treat- $$ Costs associated with both $$ Outpatient encounter $$ Problem lists ment up-front IT integration work notifications (ADT) $$ Medication lists and ongoing subscription $$ Assist in arranging appropri- $$ Lab results and radiol- and maintenance fees $$ Medication allergies ate follow-up care ogy reports for locally $$ Urgent care centers unaffili- performed studies $$ Past lab results and $$ Facilitate referrals to affiliated ated with IDNs or ACOs may radiology images provider organizations $$ Structured encounter have little financial incentive Record of past inpatient summary documents $$ $$ Participation in risk-bearing or to join an HIO. (C-CDA) and outpatient encoun- pay-for-performance contracts ters, including specialists that require data exchange for $$ Unstructured discharge care coordination summaries Long-Term Care (LTC) and Skilled Nursing Facilities (SNFs) $$ Participation in risk-bearing or $$ Some do not yet have $$ Medication lists $$ Structured and unstruc- pay-for-performance contracts sophisticated EHRs that can tured transition of care $$ Problem lists that require data exchange for interface with an HIO. documents from hospi- care coordination and avoid- $$ Progress notes tals and EDs $$ The EHRs that are used are ance of readmissions Lab results and different from those used $$ $$ Especially need ED or hospi- by hospitals and outpatient radiology reports tal discharge information providers, and are not likely for returning or incoming to have the interoperability residents features required under the meaningful use EHR certifica- $$ Need to transmit patient data tion program. to hospitals for patient trans- fers to the ED or inpatient $$ Initial funding provided wards, to prevent readmis- under HITECH to assist HIO sions, facilitate clinical care, onboarding did not include and reduce length of stay LTC facilities and SNFs. Social Service Agencies $$ Enable county-driven initia- $$ Social service agencies are $$ Various social $$ Mental and physical tives started through federal not covered entities under determinants of health, health problems waiver programs, such as HIPAA regulations, so infor- including employment, $$ Treatment and appoint- Whole Person Care Pilots or mation sharing from medical housing, and food- ment schedules (for Health Homes, or other local providers requires explicit security status transport assistance) service providers to coordi- (“opt-in”) patient consent. nate social services with $$ Medi-Cal status and medical services PCP assignment Inpatient Mental Health and Substance-Use Treatment Facilities $$ Exchange clinical information $$ Stringent and complex $$ Typically limited because $$ Medication lists about shared patients restrictions on sharing of mental health data are $$ Problem lists data related to mental health subject to additional and substance use constrain state and federal restric- ability to contribute data. tions on sharing, and substance-use treatment facilities must obtain explicit patient consent for any data sharing Notes: C-CDA is consolidated-clinical document architecture; DNR is do not resuscitate; EMS is emergency medical services; FQHC is Federally Qualified Health Center; HIPAA is Health Information Portability and Accountability Act; HL7 is Health Level Seven International; PCP is primary care provider. California Health Care Foundation 16 Snapshot of Nine California’s regional HIOs touch 35 of 58 counties state- California Regional HIOs wide and about 22 million lives. An estimated 187 hospitals statewide participate in these initiatives, but Federal funding provided by the HITECH Act, passed in close to 200 hospitals do not. Many other types of provid- 2009, spurred substantial growth in the number of both ers, ranging from skilled nursing facilities to laboratories, public and private HIOs in California. With minimal state- have far lower participation rates (see Table 3 above). level coordination or regulation of this burgenoning new patchwork of HIE organizations, its growth unfolded both What follows are high-level snapshots of nine major organically and unevenly. On the one hand, that organic regional HIOs in California, including the areas they growth has allowed regional dynamics to shape solutions serve, the types of data exchange they facilitate, the ser- truly tailored to local needs. On the other hand, some vices they provide, the members they serve, and some of regions and entities have been left behind or left vulner- the historical and geographic context that has led them able to changing market forces. to evolve in the unique ways that they have. (This infor- mation is valid as of Q2 2018.) Figure 1. Nine Regional HIOs Operating in California Del Norte Siskiyou Los Angeles Network for Enhanced Services (LANES) Modoc Manifest MedEx North Coast Health Improvement and Information Network (NCHIIN) Shasta Lassen Orange County Partnership Regional Health Information Organization (OCPRHIO) Trinity Humboldt Redwood MedNet (RWMN) SacValley MedShare Tehama Plumas San Diego Health Connect (SDHC) Mendocino Glenn Butte Sierra San Joaquin Community HIE (SJCHIE)* Nevada Santa Cruz Health Information Organization (SCHIO) Lake Colusa Yuba Placer No HIO Sutter El Dorado Yolo Sonoma Napa Alpine Sacra- mento Amador Solano Cala- Marin Tuolumne Contra Joaquin veras San Mono Costa San Francisco Alameda Stanislaus Mariposa San Mateo Santa Clara Merced Santa Cruz Madera Fresno San Inyo Benito Tulare Monterey Kings San Luis Obispo Kern Santa Barbara San Bernardino Ventura Los Angeles Orange Riverside San Diego Imperial *Uses Manifest MedEx’s technology platform and services. Note: Dots indicate counties in which multiple HIOs have participants. Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 17 Los Angeles Network for Enhanced Services (LANES) Participants Del Norte Siskiyou Total participants: ~50 Modoc $$ Hospitals: 4 county, 8 private Shasta Lassen $$ Clinics: 24 county, 5 to 10 community Trinity Humboldt $$ County behavioral health Tehama Plumas Mendocino Glenn Butte Sierra Services Nevada Lake Colusa Yuba Placer $$ Longitudinal patient record, with Sutter Yolo El Dorado web-portal viewer Sonoma Napa Alpine Sacra- mento Amador $$ Real-time event-notification services Solano Marin San Cala- Contra Joaquin veras Tuolumne Mono $$ Direct secure messaging San Francisco Costa Alameda Stanislaus Mariposa $$ Analytics reports (population health, San Mateo Santa Clara Merced pay-for-performance metrics) Santa Cruz Madera Fresno San Inyo Benito Types of Data Tulare Monterey Kings $$ Lab results $$ Medication lists San Luis Obispo Kern $$ Clinical notes Santa Barbara San Bernardino Ventura Los Angeles Pricing Subscription fees from participating hospitals Orange Riverside and health plans $$ Hospitals: Fixed fees, based on net revenue San Diego Imperial ($25K to $125K/year) $$ Health plans: $1 PMPY fee $$ IPAs: Connection and subscription fees (fee structure not available) About LANES $$ FQHCs: Connection and subscription fees subsidized LANES is seeking to reestablish itself after several years of for 2 years (fee structure thereafter not available) dormancy as a regional HIO for all provider organizations and payers in LA County. Due to the predominance of past and present funding by LA County, LANES is fighting the misperception that it is an HIO primarily for safety-net provid- ers (i.e., county medical facilities, FQHCs, and LA Care). In fact, LANES is striving to connect all care delivery providers in Los Angeles, including primary care providers, specialists, hospitals, health plans, and long-term care facilities. LANES physically aggregates data in a central data repository, which positions it well to serve customers seeking population-health analytics, bulk data feeds to payers and ACOs, and other data-dependent services. Notes: ACO is accountable care organization; FQHC is Federally Qualified Health Center; HIO is health information [exchange] organization; IPA is indepen- dent physician association; PMPY is per member per year. California Health Care Foundation 18 Manifest MedEx Participants Del Norte Siskiyou Total participants: 243 Modoc $$ 65 hospitals Shasta Lassen $$ 58 IPAs or medical groups Trinity Humboldt $$ 54 community clinics Tehama Plumas $$ 49 small practices Mendocino Butte Glenn Sierra $$ 17 others (ACOs, health plans, Nevada Lake Colusa Yuba Placer county agencies) Sutter Yolo El Dorado Sonoma Napa Alpine Sacra- mento Amador Services Marin Solano San Cala- Contra Joaquin veras Tuolumne Mono CURRENT Costa San Francisco Alameda Stanislaus Mariposa $$ Longitudinal patient record, with web-portal San Mateo Santa Clara Merced viewer and EHR-integrated document delivery Madera Santa Cruz Fresno $$ Real-time event-notification services San Inyo Benito Data submission to public health registries Tulare $$ Monterey Kings (immunizations and reportable events) San Luis FUTURE Obispo Kern $$ Data aggregation and normalization for bulk-data Santa Barbara San Bernardino distribution, real-time access via APIs, and analytical Ventura Los Angeles services in support of population health, care coordination, etc. Orange Riverside Types of Data Imperial San Diego $$ ADT feeds $$ Medication lists $$ Lab results $$ Immunization records $$ Claims data $$ C-CDA documents via the eHealth Exchange network About Manifest MedEx Manifest MedEx has a uniquely large footprint in the state of California. Manifest MedEx operates the technical infrastruc- Pricing ture for three other HIOs (Central Valley HIE, San Joaquin $$ Hospitals: A sliding-scale annual subscription fee Community HIE, and Inland Empire HIO), each of which ranging from $10K to $100K (based on hospital’s net handles its own local recruitment, training, and other on-the- patient revenue) plus a $35K flat implementation fee ground support. Manifest MedEx also boasts substantial participation by health plans, including commercial payers $$ Ambulatory providers: No annual subscription or — Blue Shield is a particularly strong participant — and implementation fees Medi-Cal managed care plans (Inland Empire Health Plan, Health Plan of San Joaquin). Notes: ACO is accountable care organization; ADT is admission, discharge, and transfer; API is application program interface; C-CDA is consolidated- clinical document architecture; EHR is electronic health record; HIE is health information exchange; HIO is health information [exchange] organization; IPA is independent physician association. Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 19 North Coast Health Improvement and Information Network (NCHIIN) Participants Del Norte Siskiyou Total participants: > 9 Modoc $$ 4 hospitals Shasta Lassen $$ 1 health plan Trinity Humboldt $$ 2 labs Tehama Plumas $$ 90% of PCPs in Humboldt County Mendocino Butte Glenn Sierra $$ County agencies (Public Health Branch, Nevada Lake Colusa Yuba Placer Mental Health Services) Sutter Yolo El Dorado Sonoma Napa Alpine Sacra- mento Amador Services Marin Solano San Cala- Contra Joaquin veras Tuolumne Mono $$ Patient data summaries San Francisco Costa Alameda Stanislaus Mariposa (faxed to hospital EDs when San Mateo Santa patients seek care there) Clara Merced Santa Cruz Madera $$ Patient mental health summaries San Fresno Inyo Benito $$ Real-time event-notification services Monterey Tulare Kings $$ Results and document delivery San Luis $$ Care-management platform Obispo Kern $$ Data submission to public health registries San Bernardino Santa Barbara (CalREDIE and CAIR2) Ventura Los Angeles Types of Data Orange Riverside $$ Lab results San Diego Imperial $$ Clinical notes $$ ADT feeds $$ Immunization records $$ Demographic data $$ Health plan eligibility About NCHIIN $$ Contact information for case managers North Coast Health Improvement and Information Network has one of the smallest budgets of any regional HIO in the $$ Contact information for probation officers state and is highly focused on supporting a “social model” $$ Jail entry and release data (name and dates) of health care, which depends on the integration of medical services and nonmedical social services. As a result, NCHIIN has significant experience in sharing patient data between Pricing medical providers and social service agencies, including navigating complex HIPAA and state regulatory challenges. $$ $35/user/month NCHIIN is also very focused on supporting and integrating $$ Fee charged to hospitals and labs for each result care-coordination capabilities. NCHIIN has significant mission delivered (28,000 results delivered per month) and operational staffing overlap with the Humboldt IPA. Notes: ADT is admission, discharge, and transfer; CAIR2 is California Immunization Registry; CalREDIE is California Reportable Disease Information Exchange; ED is emergency department; HIO is health information [exchange] organization; HIPAA is Health Insurance Portability and Accountability Act; IPA is indepen- dent physician association; PCP is primary care provider. California Health Care Foundation 20 Orange County Partnership Regional Health Information Organization (OCPRHIO) Participants Del Norte Siskiyou Total participants: > 20 Modoc $$ 17 hospitals (out of 20) Shasta Lassen $$ 3 large IPAs or medical groups Trinity Humboldt $$ 30 clinics (some using secure Tehama messaging only) Plumas Mendocino Glenn Butte Sierra Nevada Services Lake Colusa Yuba Placer Sutter El Dorado Longitudinal patient record, Yolo $$ Sonoma Napa Alpine Sacra- with web-portal viewer and mento Amador Solano San Cala- Marin EHR-integrated viewers Contra Joaquin veras Tuolumne Mono Costa San Francisco $$ Results and document delivery Alameda Stanislaus Mariposa San Mateo Santa (lab, radiology, pathology) Clara Merced Santa Cruz Madera $$ Direct secure messaging Fresno San Inyo Benito Tulare Types of Data Monterey Kings $$ ADT feeds San Luis Obispo Kern $$ Lab results $$ Radiology reports Santa Barbara San Bernardino Ventura Los Angeles $$ Pathology results $$ C-CDA documents Orange Riverside $$ Radiology images San Diego Imperial Pricing $$ Hospitals: ~$50,000/year $$ Medical groups: Variable, depending on number of interfaces (starts at ~$7,500/year) $$ Sites using web-portal viewer only: $40/physician/year About OCPRHIO A significant proportion of the patient population served by OCPRHIO lives in or receives medical care in adjoining counties (LA, Riverside, San Bernardino, and San Diego). OCPRHIO members therefore highly prioritize its ability to exchange data with HIOs in neighboring counties. Many of OCPRHIO’s participating community clinics and small physician practices do not yet have data interfaces, and are accessing HIO data exclusively via a web-portal viewer. Notes: ADT is admission, discharge, and transfer; C-CDA is consolidated-clinical document architecture; EHR is electronic health record; HIO is health information [exchange] organization; IPA is independent physician association. Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 21 Redwood MedNet (RWMN) Participants Del Norte Siskiyou Total participants: > 8 Modoc $$ 6 hospitals Shasta Lassen $$ 600 outpatient providers Trinity Humboldt (IPA and medical groups, community clinics) Tehama Plumas Mendocino Glenn Butte Sierra $$ County behavioral health Nevada Lake Colusa Yuba Placer Sutter Services Sonoma Napa Yolo El Dorado Alpine Sacra- CURRENT mento Amador Solano San Cala- Marin $$ Results and document delivery Contra Joaquin veras Tuolumne Mono Costa San Francisco (lab, radiology) Alameda Stanislaus Mariposa San Mateo Santa Clara Merced FUTURE Madera Santa Cruz $$ Longitudinal patient record, with web- San Fresno Inyo portal viewer and EHR-integrated viewers Benito Tulare Monterey Kings $$ Analytics reports (population health, pay-for-performance metrics) San Luis Obispo Kern $$ Access to social determinants of health Santa Barbara San Bernardino Types of Data Ventura Los Angeles $$ ADT feeds Orange Riverside $$ Lab results $$ Radiology reports San Diego Imperial $$ Health plan eligibility Pricing Details unavailable About RWMN Redwood MedNet, which is built on open-source software with cloud-hosted infrastructure, has a narrower focus than other regional HIOs. Rather than aiming to support general patient-data exchange, RWMN focuses on facilitating HL7 result delivery and supporting targeted projects to improve the coordination of care for the local safety-net population. RWMN has relied heavily on one-off development projects as a revenue source and is now focused on growing its recur- ring subscription revenue in order to strengthen its financial sustainability. Notes: ADT is admission, discharge, and transfer; EHR is electronic health record; HIO is health information [exchange] organization; HL7 is Health Level Seven International; IPA is independent physician association. California Health Care Foundation 22 SacValley MedShare Participants Del Norte Siskiyou Total participants: 30 Modoc $$ 15 hospitals Shasta Lassen $$ 24 clinics Trinity Humboldt $$ 1 health plan Tehama Plumas $$ 2 imaging centers Mendocino Butte Glenn Sierra Nevada Lake Colusa Yuba Placer Services Sutter Yolo El Dorado $$ Longitudinal patient record, Sonoma Napa Alpine Sacra- mento Amador with web-portal viewer and Solano Marin San Cala- EHR-integrated viewers Contra Joaquin veras Tuolumne Mono Costa San Francisco Alameda Stanislaus Mariposa $$ Real-time event-notification services San Mateo Santa Clara Merced $$ Access to national health information Santa Cruz Madera exchange networks (eHealth Exchange) San Fresno Inyo Benito Tulare Monterey Kings Types of Data $$ ADT feeds San Luis Obispo Kern $$ Lab results Santa Barbara San Bernardino $$ Radiology reports Ventura Los Angeles $$ Medication lists $$ Clinical notes Orange Riverside $$ C-CDA documents San Diego Imperial Pricing $$ Hospitals: Based on number of beds $$ Clinics and practices: Based on number of FTE MDs and DOs $$ Imaging centers: Based on number of radiologists reading image About SacValley MedShare SacValley MedShare is the only regional HIO operating in $$ Health plans: Based on number of covered lives many of California’s rural counties. It faces a unique consent- management challenge because its patients are sometimes referred to providers in Nevada, which has a statewide “opt-in” consent policy — in contrast to the HIO, which has a default “opt-out” policy for California providers. SacValley is in the process of integrating with a Medi-Cal managed care plan that has members in roughly half of the counties of this region, and has actively supported HIE as a means to improve the quality and control the costs of the care it provides. Notes: ADT is admission, discharge, and transfer; C-CDA is consolidated-clinical document architecture; DO is doctor of osteopathic medicine; EHR is electronic health record; FTE is full-time equivalent; HIO is health information [exchange] organization; MD is medical doctor. Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 23 San Diego Health Connect (SDHC) Participants Del Norte Siskiyou Total participants: 40 Modoc $$ 21 hospitals Shasta Lassen $$ 1 IPA Trinity Humboldt $$ 17 community clinics Tehama Plumas $$ 1 ambulance company Mendocino Butte Glenn Sierra Nevada Partial participants Placer Lake Colusa Yuba (secure messaging only): Sutter El Dorado Yolo Sonoma Napa Alpine 2 counties Sacra- $$ mento Amador Solano San Cala- Marin $$ 80 SNFs Contra Joaquin veras Tuolumne Mono Costa San Francisco $$ 5 hospices Alameda Stanislaus Mariposa San Mateo Santa Clara Merced $$ 28 ambulatory sites Madera Santa Cruz Fresno San Inyo Services Benito Tulare Monterey Kings $$ Longitudinal patient record, with web-portal viewer and EHR-integrated viewers San Luis Obispo Kern $$ Results and document delivery (lab) $$ Real-time event-notification services Santa Barbara San Bernardino Ventura Los Angeles $$ Data submission to public health registries $$ Prehospital EMS reporting Orange Riverside $$ Secure messaging San Diego Imperial Types of Data $$ C-CDA documents $$ HL7 v2 messages About SDHC San Diego Health Connect was founded and initially $$ Radiologic and cardiologic images operated largely through generous funding from a federal $$ POLST documents government grant. In 2013, when those grant funds ran out, the HIO sought and began receiving most of its funding $$ NEMSIS electronic patient care reports from the county government and the large health systems operating in the county. The HIO is now under pressure to demonstrate tangible value, particularly as alternative HIE Pricing resources provided directly by providers’ EHR products gain Details unavailable more capabilities and wider reach throughout the county. One of SDHC’s most high-value opportunities is providing hospital and ED encounter-notification services, given the broad participation of regional hospitals in its network. One challenge of note for SDHC is the heterogeneity of consent policies among its participants. Some still use an opt-in consent policy, which sometimes results in the appearance that patient data are missing from these participants. Notes: C-CDA is consolidated-clinical document architecture; ED is emergency department; EHR is electronic health record; EMS is emergency medical services; HIE is health information exchange; HIO is health information [exchange] organization; HL7 is Health Level Seven International; IPA is independent physician association; NEMSIS is National Emergency Medical Services Information System; POLST is Physician Orders for Life-Sustaining Treatment; SNF is skilled nursing facility. California Health Care Foundation 24 San Joaquin Community HIE (SJCHIE) Participants Del Norte Siskiyou Total participants: 6 Modoc $$ 1 hospital Shasta Lassen $$ 3 clinics Trinity Humboldt $$ 2 health plans Tehama Plumas Mendocino Glenn Butte Sierra Services Nevada Lake Colusa Yuba Placer CURRENT Sutter Yolo El Dorado $$ Longitudinal patient record, with Sonoma Napa Alpine Sacra- mento Amador web-portal viewer Solano Marin San Cala- Contra Joaquin veras Tuolumne Mono $$ Real-time event-notification services San Francisco Costa Alameda Stanislaus Mariposa $$ Care-management platform* San Mateo Santa Clara Merced Santa Cruz Madera FUTURE Fresno San Inyo Data aggregation and normalization for: Benito Tulare $$ Bulk-data distribution Monterey Kings $$ Real-time access via APIs San Luis Obispo Kern $$ Analytical services in support of population health, care coordination Santa Barbara San Bernardino Ventura Los Angeles Types of Data Orange Riverside $$ ADT feeds $$ Lab results Imperial San Diego $$ Immunization records $$ C-CDA documents $$ Medication lists Pricing About SJCHIE Details unavailable Through a unique partnership, Manifest Medex provides the technical infrastructure underlying the San Joaquin Community HIE, while the HIE provides the local recruiting, consulting, and operational support services. Members have access to data from other regional HIOs that have similar “affiliate” relationships with Manifest Medex. SJCHIE is lever- aging its data exchange and analytics and reporting features to actively support San Joaquin County’s Whole Person Care Pilot initiative. SJCHIE is also in the process of contract- ing with a care-management vendor to integrate with the Manifest Medex technology. *Data exchange platform served by Manifest Medex. Community engagement and care-management platform offered specifically by SJCHIE. Notes: ADT is admission, discharge, and transfer; API is application program interface; C-CDA is consolidated-clinical document architecture; HIE is health information exchange; HIO is health information [exchange] organization. SJCHIE uses Manifest Medex’s technology platform and services. Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 25 Santa Cruz Health Information Organization (SCHIO) Participants Del Norte Siskiyou Total participants: 46 Modoc $$ 7 hospitals Shasta Lassen $$ 12 clinics Trinity Humboldt $$ 2 IPAs or large medical groups Tehama Plumas $$ 12 small practices Mendocino Butte Glenn Sierra $$ 11 labs or imaging centers Nevada Lake Colusa Yuba Placer $$ 2 health plans Yolo Sutter El Dorado Sonoma Napa Alpine Sacra- mento Amador Solano Services Marin San Cala- Joaquin veras Tuolumne Mono Contra Costa $$ Longitudinal patient record, with web- San Francisco Alameda Stanislaus Mariposa portal viewer and EHR-integrated viewers San Mateo Santa Clara Merced $$ Results and document delivery Santa Cruz Madera Fresno (lab, radiology, transcribed notes, C-CDAs) San Inyo Benito Tulare $$ Real-time event-notification services Monterey Kings $$ Access to national health information exchange San Luis networks (Carequality, eHealth Exchange) Obispo Kern $$ Data submission to public health registries San Bernardino Santa Barbara (CalREDIE and CAIR2) Ventura Los Angeles $$ Analytics reports (population health, pay-for-performance) Orange Riverside $$ Referral management San Diego Imperial $$ Communication with nonmedical service organizations Types of Data $$ ADT feeds $$ Lab results $$ Radiology reports About SCHIO $$ Clinical notes Formed in 1995, Santa Cruz HIO is California’s oldest regional HIO, and it has, over those 20+ years, built significant $$ C-CDA documents momentum and traction within the community. It serves a county that is small and lightly populated compared with most other population centers that have a regional HIO. Pricing Nonetheless, it has become financially stable and self- Details unavailable sustaining. SCHIO has substantial participation from both the local county’s safety-net providers and its Medi-Cal managed care plan. Notes: ADT is admission, discharge, and transfer; API is application program interface; CAIR2 is California Immunization Registry; CalREDIE is California Reportable Disease Information Exchange; C-CDA is consolidated-clinical document architecture; EHR is electronic health record; HIO is health informa- tion [exchange] organization; IPA is independent physician association. SCHIO is located primarily in Santa Cruz County with locations at Stanford, Palo Alto Medical Foundation (PAMF), and small practices in Santa Clara County. California Health Care Foundation 26 How Regional Market 3.Presence of a supportive dominant health insurer that believes in the value of the HIO for purposes of Dynamics Shape the Role improving care coordination and population health for its members: A dominant health insurer has the of Regional HIOs business incentive to invest in the regional HIO, as the financial benefits of improved population health and reduced health care utilization within the region Key Factors accrue largely to itself. In regions with highly frag- A regional HIO is generally only one of many health care mented payer markets, it is more difficult for individual players operating in a given local market. Each player insurers to calculate the benefits and justify the costs has a unique set of clinical and financial obligations and of their investments in a regional HIO. incentives, and of course, not all of the players are equal. They can differ in any number of ways, including market 4.Capabilities of the available regional HIOto pro- share, profitability, and customer segment. The one thing vide data-exchange services needed by a significant they all share is that they are operating within the same number of provider organizations and payers in the region, which includes a unique history, geography, and region and not otherwise provided through their EHR demographics. With so many stakeholders and forces at vendors: The more interoperability services that a play, it can be hard to trace the path of how a particu- regional HIO can effectively provide and the better lar market’s health care ecosystem has evolved. When organized and operated the HIO is, the more likely it comes to the evolution and success of regional HIOs, provider organizations will be to benefit from and however, four particular market dynamics seem to be support its services. As the base of participating pro- especially influential: vider organizations increases due to this perception of competence, the services of the regional HIO will 1.Degree of consolidation and competitionamong become that much more valuable, further perpetuat- provider organizations in the region: Larger, more ing participation and support. The converse dynamic consolidated provider organizations seek to and can is equally true. provide a broader range of health care services, which incentivizes them to create private HIO infrastructures to streamline cross-disciplinary care, to attract and Case Studies: Real-World retain patients within their enterprises with the con- Implications of Regional Market venience of a single longitudinal health record, and to support business analytics. If a region is dominated by Dynamics on HIO Strength one or more such entities, they may be fiercely com- petitive and less inclined to share patient data via a Northern Central Valley regional HIO, as such sharing may blunt the competi- (north of Sacramento to Oregon border) tive advantages of their achieved scale. Rural, unconsolidated market with a small but well- supported regional HIO 2.Heterogeneity of EHR systemsin use among pro- This geographically sizable region* of the state is largely vider organizations: If an EHR vendor achieves rural and home to only 1.7 million residents. Although it substantial market share among the provider organi- includes some hospitals and outpatient clinics from large, zations in a region and that vendor provides its own multihospital systems, such as Dignity Health, most of the proprietary mechanism for data sharing across its provider organizations in the region remain independent, customers, then there is less need for a regional HIO. including a host of critical access hospitals, community Conversely, regions with fragmented EHR markets clinic networks, and small, independent practices and depend more on a third party, such as a regional HIO, medical groups. There is no academic medical center to organize and enable data sharing across provider in the region, although the UC Davis Medical Center organizations. in the adjoining Sacramento County is a referral hub. *This region corresponds roughly to the region served by the SacValley MedShare HIO (profiled on page 23). Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 27 The region’s spread-out geography and lack of intense San Diego County competition among provider organizations for the same Consolidated market with a regional HIO facing stiff patient populations limit the degree of resistance on competition from EHRs and other alternatives their parts to exchanging data. San Diego County is a relatively densely populated and affluent region, with over 3 million residents. Although Partnership HealthPlan of California (PHC) is a Medi-Cal its payer market is fragmented, its health care systems, managed care plan that has members in roughly half of including Sharp HealthCare, Scripps Health, Kaiser the counties of this region and has actively supported HIE Permanente, Palomar Health, and the UC San Diego to improve the quality and control the costs of care for Medical Center, are highly consolidated. These systems this population. PHC is currently integrating with the local use a small number of EHR systems, predominantly Epic regional HIO, SacValley MedShare. This HIO provides a and Cerner, which provide certain native data-exchange longitudinal medical record and event-notification ser- capabilities (such as Epic Care Everywhere and con- vices based on data from 15 hospitals, 24 clinics, and 2 nectivity with the Carequality and CommonWell Health imaging centers, to date. Although this remains a minor- Alliance networks). Certain of these health care systems ity of the provider organizations and payers in the region, also use additional private HIE technologies. the HIO is in active negotiations with other organizations. The overall budget of the HIO and, as a result, its sub- As the EHR systems and other private HIE technolo- scription fees for any individual participant are relatively gies used by these health care systems gain increasing low, helping the HIO to maintain stable operations. capabilities and reach throughout the county, the local regional HIO, San Diego Health Connect, is facing San Francisco Bay Area heightened pressure to deliver unique value. The intense Highly consolidated and competitive market with no competition among those larger health systems may regional HIO make it challenging for the HIO to encourage the types of data sharing needed to demonstrate additional value, Despite being home to 7 million residents, the nine- such as contribution of providers’ data into a central data county Bay Area has no regional HIO in operation. repository. The regional HIO does, however, enjoy sup- The area is dominated by a few large hospital systems, port and funding from the local government, which is such as Kaiser Permanente, Sutter Health, and Dignity active in providing and coordinating care for the safety- Health, as well as two large integrated delivery networks net populations treated by several sizable community anchored at academic medical centers (Stanford and health centers and other providers. UCSF). Outpatient care is also consolidated, with Kaiser Permanente operating about a dozen facilities, the Palo Alto Medical Foundation (also affiliated with Sutter Health) having 15 sites and 900 physicians, the Santa Clara County IPA having over 800 physicians, and Brown & Toland Physicians offering a network of more than 2,500 physicians across 38 Bay Area cities. All of these consolidated organizations are also highly competitive with each other, and each is trying to achieve greater consolidation and clinical integration unto themselves. This likely impedes a willingness to openly cooperate and freely share patient data among themselves. Health information exchange does occur across the major provider organizations in the Bay Area, but it is enabled by capabilities made available through EHR ven- dors, such as Carequality, CommonWell Health Alliance, and Epic Care Everywhere, and connectivity among certain of the private HIOs operated by these provider organizations. California Health Care Foundation 28 Looking Ahead: and governance framework that enables nationwide patient-data sharing across the large number of HIOs Policy and Technology (regional and otherwise) that currently operate around Trends to Watch the country. If successful, TEFCA would establish a single national “coordinating entity” that would develop and operationalize policies and procedures for the designa- The larger health care market within which regional HIOs tion of “Qualified Health Information Networks” (the operate is incredibly dynamic. While many policy and participating HIOs), for the terms of participation for indi- technology trends have the potential to shape the fate of vidual provider organizations in such networks, and for HIOs, the following ones loom especially large. the exchange of data among such networks. TEFCA is in its early stages and still undergoing the New HITECH Funds on the Horizon rule-making process, including revisions in response to for California’s Regional HIOs initial public comments. It is also currently formulated The California Department of Health Care Services as an entirely voluntary framework, in which HIOs and (DHCS) is applying to the Centers for Medicare & their participating provider organizations can choose to Medicaid Services (CMS) for $45 million in federal participate or not. However, were TEFCA to be widely HITECH Act funds, matched by $5 million in general state adopted in its currently proposed form, it would sig- funds to onboard more hospital systems and ambulatory nificantly extend the breadth and scope of patient-data providers into regional HIOs. Specifically, the funding will sharing that is available to participants in HIOs, as well as subsidize one-time implementation costs for providers to standardize the rules under which existing HIOs operate. connect to qualified regional HIOs. Up-front onboarding costs for new HIO participants can run anywhere from $5,000 to $10,000 for an individual provider to $150,000 Fast Healthcare Interoperability for a complex hospital system, and have proved to be Resources (FHIR) a meaningful barrier to HIOs’ ability to scale, especially FHIR is a set of standards for creating web-based applica- in the safety net. This incentive funding will help lower tion programming interfaces (APIs) to read data from and the barrier for providers to participate in data-exchange write data to clinical information systems. The standards efforts across the state. The funds are also earmarked to define what clinical data may be read from and written to connect all regional HIOs to the Controlled Substance systems, the degree to which those data are structured Utilization Review and Evaluation System (CURES) data- and coded, and how commands to read and write data base, another activity that can enhance the value HIOs should be formulated. offer potential participants. Although FHIR is a potent new technology to facilitate Additionally, CMS has also approved another $40 mil- health-data interoperability, it is important to under- lion of funding to onboard emergency medical services stand that FHIR only provides a tool kit for building data (EMS) providers into regional HIOs. The availability of interfaces. It does not provide anything close to plug- this approved federal funding is contingent upon The and-play interfaces for exchanging patient data. Health California EMS Authority finding matching state funds. Level Seven International (HL7)’s FHIR specifications are both quite generic and greatly extensible, so two pro- vider organizations that both implement FHIR-compliant The Potential of TEFCA interfaces may do so in very different ways that are not In 2018, the Office of the National Coordinator for Health consistent and do not result in semantic interoperability. Information Technology within the Department of Health and Human Services proposed the Trusted Exchange To address this lack of specificity, certain “profiles” Framework and Common Agreement (TEFCA), a regu- have been developed based on the FHIR standard, and latory instrument intended to implement certain of the these profiles put greater constraints on how clinical interoperability requirements of the 21st Century Cures data that are exchanged using FHIR must be structured, Act of 2016. TEFCA is a detailed and complex pro- populated, and coded. Prominent examples of such posal, but, in essence, it seeks to establish a technical profiles include the “Argonaut” profile (developed by Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 29 a consortium of EHR vendors) and the “HL7 US Core” DL technologies have interesting properties with poten- profile (developed to support interoperability needed tial for applicability in health care, but the IT and health under the federal meaningful use program). These pro- care industries are still very much in the early stages of files help to ensure that interfaces using FHIR will be exploring how these technologies may be used. To date, able to exchange patient data meaningfully, and are an the hype surrounding DL technologies in health care far important step in the right direction, but they still have exceeds the number of demonstrably useful applications. significant shortcomings. It very much remains to be seen if DL technologies can solve problems related to HIE in a better way than alterna- First, these profiles still leave room for substantial varia- tive, preexisting technologies, such as traditional database tion in the way that data interfaces are implemented by models, security models, data standards, APIs, etc. different organizations, which continues to limit semantic interoperability. Furthermore, very few commercial EHR vendors have implemented the profiles faithfully (i.e., in Consumer-Mediated Health full compliance with their technical specifications). Such Information Exchange compliance is not now required under the meaningful Efforts to give patients a greater role and greater con- use EHR certification program or any other regulatory trol in the sharing of their health information have been mandate. As such, despite the existence of the pro- underway for over 10 years. These forays into “consumer- files, the FHIR implementations of leading EHR vendors mediated health information exchange” have primarily remain technically different. taken the form of personal health records (PHRs), such as Microsoft HealthVault, Google Health (now defunct), With regard to HIE, the FHIR standard and FHIR profiles and various PHR initiatives sponsored by health insurers. provide only part of the functionality needed to support To date, these approaches have enjoyed limited uptake patient-data exchange among independent health care among consumers, largely due to interoperability barriers organizations. FHIR is particularly well suited to enable that impede consumers’ ability to conveniently aggre- one computer application to access structured data gate and integrate their health data from the numerous from another over the internet and to manage the user- provider organizations where these data originate. A authentication and access-control aspects of doing that. recently launched initiative in consumer-mediated HIE However, it lacks facilities for supporting other important from Apple, however, may be able to overcome these requirements for HIE across independent organizations, barriers. such as patient-identity management, patient record- locator services, provider directories, data-aggregation Apple’s Health Records enables patients to consolidate and normalization resources, and facilities for the type their medical data on their iPhones by directly interfac- of “push” messaging needed for event-notification and ing these devices to the EHR systems of participating referral-management applications. health care organizations using the FHIR API standards. To date, several hundred health care facilities have imple- mented interfaces to support Health Records. Patients Blockchain authenticate to the health care facilities using their exist- “Distributed ledger” (DL) technologies, such as block- ing passwords for the organizations’ PHR systems. Once chain, have received a great deal of attention recently, authenticated, patients can download specified data including in the health care industry. At their core, these available in their PHRs to their iPhones and can instruct technologies define a distributed database system that Health Records to update these data automatically when (1) is not controlled or maintained by any single entity, (2) additions or changes are made in their PHRs. maintains a single, chronologically sequenced record of events, and (3) includes facilities to verify the authenticity Notably, the medical data accessed by Health Records is of recorded events and to prevent the creation of alter- transmitted directly from a provider organization to the native, inauthentic records, despite the decentralized patient’s iPhone. It does not pass through nor is it stored nature of the system. in any servers operated by Apple or other third parties (other than backed up in encrypted form in iCloud, if desired by the patient, just as other iPhone data may California Health Care Foundation 30 be). Health Records is also able to interface with other to their medical data through Health Records. The capa- compatible health care applications that patients down- bilities enabled by Health Records may even necessitate load to their iPhones. Such applications may be obtained additional consumer protections to prevent abuses of from any third-party app developer. After downloading this new channel for patient-data access and sharing. such an app, a patient must explicitly authorize the app to access her medical data from the Health Records data- Given the potential benefits of enabling patients to base and agree to any terms of service and terms of data aggregate and control their own health data, other tech- usage presented by that app. After the patient does so, nology firms, such as Seqster and Ciitizen, are currently the app may then access her data and process, store, or pursing similar strategies. transmit the data in whatever ways are authorized by the app’s terms of service. Health Records is a potentially significant development Conclusion for several reasons. First, many provider organizations In California, at-risk organizations, including commer- and their EHR vendors seem willing to technically sup- cial payers, Medi-Cal managed care plans, and ACOs, port interoperability with the Health Records platform, are increasingly recognizing that health information which may eventually allow most patients to aggregate exchange is a critical functionality for care coordination clinical data from most of their health care providers and population-health management. The degree to on their iPhone. This would be a first, despite long- which provider organizations move away from fee-for-ser- standing efforts by PHR vendors, insurance companies, vice models and toward risk-sharing arrangements, such medical record-banking organizations, and others to as ACOs and bundled-payment programs, will influence build patient-centric health record systems. Apple’s the demand for clinical integration and data analytics market share in smart phones combined with EHR ven- across independent organizations. Such demand, in turn, dors’ widespread support for FHIR APIs has apparently will increase demand for the types of HIE services that enabled Apple to build data interfaces to a great many can be provided by regional HIOs. This is particularly true distinct provider organizations. This capability could sig- in the safety net and in regions where there is less pro- nificantly empower patients to better manage their own vider consolidation and entities are therefore less likely and their family’s health, to more easily change health to have EHR-mediated or other private data-exchange care providers, and to more easily seek second opinions, options. among other benefits. To be well positioned to meet that demand, regional Second, the ability of third-party apps to access data on HIOs need to be ready with robust features, and large and transmit data from patients’ iPhones also creates the and diverse participant networks, as well as stable finan- potential for iPhones to serve, effectively, as universal cial footing. While regional HIOs have plenty of work to data-interface devices. Using Health Records, any third do — for example, developing high-value features and party could collect and store clinical data on a large prioritizing integrating their services with EHR products cohort of (consenting) patients without having to build — they also require support from the many other stake- any interfaces to individual provider organizations. Such holders that will benefit from a successful and sustainable a universal interface could enable a host of new services HIO network. Without that network, California’s delivery and data-analytic capabilities with benefits for patients, system will likely remain unevenly and inequitably con- providers, insurers, researchers, drug developers, public nected — a costly prospect for both the state and its health agencies, marketers, and others. patients. Third, once downloaded onto a patient’s iPhone, medi- cal data are no longer subject to any protections under the Health Insurance Portability and Accountability Act (HIPAA) and other privacy regulations, leaving consum- ers responsible for understanding and assessing the risks as well as the benefits of granting third parties access Promise and Pitfalls: A Look at California’s Regional Health Information Organizations 31 Endnotes 1.Paul T. Cheung et al., “National Study of Barriers to Timely Primary Care and Emergency Department Utilization Among Medicaid Beneficiaries,” Annals of Emergency Medicine 60, no. 1 (July 2012); 4 –10.e2, accessed December 14, 2018, https://doi.org/10.1016/j.annemergmed.2012.01.035. 2.Gail Patrick and John Hickner, Four Models Bring Specialty Services to the Safety Net: Enhancing Scope of Practice and Referral Efficiency, California HealthCare Foundation, July 2009, www.chcf.org (PDF). 3.Nir Menachemi et al., “The Benefits of Health Information Exchange: An Updated Systematic Review,” Journal of the American Medical Informatics Association 25, no. 9 (September 1, 2018): 1259– 65. California Health Care Foundation 32