How ACOs can Utilize Precision Medicine to Improve Quality of Care and Reduce Cost

1.    Problem Statement

How can Accountable Care Organizations improve the quality of care and reduce cost by utilizing a precision medicine program?  

Accountable Care Organizations (ACOs) are comprised of a variety of care providing entities including primary care, specialty clinics, hospitals, and home health services. Their aim is to provide patient centered care with improved health outcomes at a reduced overall cost (Golden, 2015). This is reiterated by payment models where these organizations are reimbursed for their outcomes, rather than for the services they provide (Innovation Models, n.d.). Because the focus is on health outcomes it becomes imperative that prescribed treatments are accurate, have the best possible impact, and with the least amount of adverse effects.

In February 2020, a team led by Dr. Lin at Nationwide Children's Hospital in Ohio, published a study performed on a cohort of over 250,000 patients, where average annual costs were associated with various disease groups as categorized by the level of involvement of a genetic component (see Table 1) (McCandless, Brunger, & Cassidy, 2004). As the genetic component becomes more of a driver for a disease, the cost rise quickly. In addition, the team also noted that in- and out-patient services are utilized more frequently, the duration of stay increases, and cost of prescription drugs increase (Miller, Hoyt, Rust, Doerschuk, Huang, & Lin, 2020). The non-genetic category of disease shows relatively low cost, due to the fact that over the past century diagnostics for infectious diseases have improved and widely implemented. The accuracy of diagnostics allows for precise treatments. The same can be said for trauma. Through the use of imaging techniques and knowledge of anatomy, surgeons can quickly learn the details of a trauma, and reconstruct these damages. Advances are being made in this field, reducing complications and expenses.

This analysis looks at policies that ACOs can adopt to achieve their goals to reduce cost when caring for patients with genetically driven maladies while improving their quality of life through the utility of a precision medicine program.

2.    Background: The Shift towards Big Data

The completion of The Human Genome Project in 2003 set the stage for new forms of diagnostics to be developed. The human genome has approximately 3 billion nucleotides, which are the building blocks that act as the blueprint for our existence, and are inherited from parents through equal shares. The human genome contains about 20 thousand genes that encode proteins. Figure 1 shows an illustration of how the genome is organized, where genes are interspaced by non-coding and regulatory regions (Shumate et al., 2020). Thanks to developments of massive parallel sequencing technologies, it’s now possible to sequence the entire human genome quicker and cheaper than ever before (Tucker, Marra, & Friedman, 2009). This area of diagnostics is developing quickly, but due to the volume of data generated, it adds additional layers of complexities that must be dealt with. Many genetic diseases are complex and can stem from combinations of many genomic aberrations. Various combinations could give rise to a range of symptoms within the same disease. An example is Autism Spectrum Disorder (ASD), which has over 59 million genetic aberrations identified that, in different combinations, can cause this disease (Turner et al., 2017). Symptoms range from mild to severe and affect communication, emotion, reasoning, and spatial skills (Carpenter, Soorya, & Halpern, 2009). Depending on the symptoms and combination of mutations, specific treatment options can be provided which directly impact outcomes. It is nearly impossible for any specialist to comprehend this volume of data in order to develop optimal treatment plans. The aid of bioinformatic approaches, including artificial intelligence (AI), will have to play a vital role.

Epilepsy, like ASD, is influenced by many genetic variants. A team at the Children’s Hospital of Philadelphia recently developed a novel AI algorithm which allowed them to mine electronic medical records (EMRs) for outcomes in patients with epilepsy and to link them to their genetic finds. A cohort of 658 patients with 62,104 combined encounters during a 7-year period made it possible for this algorithm to successfully identify specific mutations that give rise to specific outcomes (Ganesan et al., 2020). By knowing these future outcomes, treatment options can be planned ahead and the process can be streamlined, with the result of iteratively improving outcomes.

Imaging techniques are also taking on new roles. Imaging qualities have improved and AI algorithms are being developed to interpret these. AI algorithms are quicker than radiologists, and have the potential to be more accurate. In a study where AI algorithms were performed to detect breast cancer, it was found that the false positive rate dropped by an average of 3.5%, and false negative rate dropped an average of 6.1% (McKinney et al., 2020). Likewise, some imaging techniques in combination with AI can identify structural differences that are driven by genetic mutations. Many mutations are somatic and can only be detected locally within the human body, requiring a biopsy sample in order to detect it. Mutations that drive structural differences can potentially be detected through the utilization of imaging. A recent study looked at gliomas and through the use of magnetic resonance imaging, it was possible to predict mutations in the gene encoding isocitrate dehydrogenase (Suh, Kim, Jung, Choi, & Kim, 2018). This gene is involved in key pathways within the central nervous system, and when mutated cause or enhance gliomas to thrive. Through this non-invasive identification of a group of mutations treatments with specific inhibitors can be prescribed (Huang et al., 2019).

It is evident that the digital era has allowed for increased medical data accumulation, and EMRs are the assigned database (Kalf & Meinecke, n.d.). Besides the data that is entered by physicians during a patient’s visit, other data as described above will make up the vast majority of the data stored. Data types vary in complexity and each require its own structure for storage. Unique algorithms make it possible to read and interpret these very different types of data, and through AI it will be possible to create links in new ways that are developing quickly. The different disciplines collectively form precision medicine.

3.    Landscape

Presented here is an overview of stakeholders affected, including the intended client (ACOs), physicians, IT departments, and patients. For each, concerns and key issues will be discussed.

3.1 Accountable Care Organizations

The implementation of the Affordable Care Act of 2010 gave rise to many new forms of healthcare delivery models. ACOs are the most recognized model (Matulis & Lloyd, 2018). In general, ACOs focus on providing improved value of care that is patient centered while reducing cost. Reimbursement is typically in the form of a lump sum in order to take care of specific patients for an extended period of time, typically over a year (Beasley, 2015). This lump sum, provided under a shared savings contract, is often calculated based on the historical incurred expenses incurred on annual fee-for-service payments that include various health services. This is referred to as the baseline. When implementing value-based care, the ACO aims to reduce cost compared to this baseline, and the agreement includes how these savings must be distributed. Part of this distribution is revenue for the ACO. If losses are incurred, these will be partially recuperated from the payer, thus risk is shared (Bailit, & Hughes, 2011).

3.2 Care Providers

Here, the term “care providers” is used as a collective term that includes all physicians, clinicians, nurses, and counselors, involved in the delivery of care. Over the years there have been numerous publications describing physician burnout, of which symptoms can be grouped into three distinct categories; lack of accomplishment, exhaustion, and depersonalization. Their lack of accomplishment can be attributed to the fact that each care provider is not always part of the entire diagnosis and treatment plan, and therefore do not observe the total benefit patients might endure. This is especially true when dealing with patients who have complex diseases, where patients are under the care of various specialists. Additionally, the traditional fee-for-service model incentivizes high volumes of patients to be seen, which decreases the allotted time a care provider can spend with their patient. This results in shortcuts and often treating symptoms rather than root causes. The suboptimal time spent with patients, combined with administrative burden, result in depersonalization and exhaustion (Katzen, Dodelzon, Michaels, & Drotman, 2020).

Some entities have started to vertically integrate services to deal with complex diseases. One such entity is the University of Texas MD Anderson Cancer Center (MDACC). When thinking of care providers as individual processes, it makes sense to streamline these processes by aligning them. MDACC continuously analyzes the flow of patients through the system and identify commonalities between a group of patients who utilize a specific pattern of specialists. These specialty groups are then joined into one unit to optimize the process. For example, those patients with head & neck cancers have unique needs. Their symptoms tend to be very visual, often expressed by altered speech and dental issues. This group of patients often requires surgical care, alcohol and/or smoking cessation, and endocrinological care due to the frequent involvement of the thyroid or parathyroid glands. MDACC employs their providers, removing the incentive to see large volumes of patients. By combining these specialties within one unit, patient flow is optimized. Care providers have improved communication, and are able to observe outcomes and how these were tied into their personal contributions. This vertical integration improved overall outcomes as measured by 5-year mortality rates, improved patient satisfaction, and decreased physician burnout (Porter & Jain, 2018).

3.3 Laboratories

Many diagnoses are made based on a patient’s history, and laboratories performed supportive tests. As science has evolved, there is more evidence-based diagnoses, where laboratories play a central role. With the ability to accurately identify bacteria in combination with knowledge about their specific sensitivities to antibiotics, the correct treatment can be prescribed. This is a form of precision medicine (Gaynes, 2017). These types of samples are mostly obtained through non- or minimally invasive techniques and testing is done within the microbiology department of the laboratory. Samples are either cultured to identify the bacteria and to assess sensitivity to a drug, or through molecular testing.

Diseases with genetic components require a broader range of samples. Samples used to detect inherited diseases can be collected through a non- or minimally invasive swab or blood collection, and tested directly in the genomic laboratory. Specimens for diseases stemming from somatic events require invasive strategies, such as surgeries or interventional radiology to perform biopsies. These biopsies require various disciplines within the laboratory to preprocess before actual tests can be performed, namely anatomical pathology, histology, cytology, flowcytometry, cytogenetics and genomics laboratories. Figure 2 shows the flow of specimens through the various laboratory departments. The anatomic pathology lab processes the actual biopsy, after which it is sent to the histology or cytology laboratories. These departments perform tissue staining so that malignant cells within a tissue can be identified at the microscopic level. The flowcytometry department can identify malignant cells as well, but this is typically done on liquids such as blood and aspirates. These departments will then send a fraction of the obtained sample to subsequent laboratories. Cytogenetics is capable to detect large genomic alterations within the cells, whereas the genomics laboratory detects aberrations at the nucleotide level. These specimens frequently move back and forth between various sections, and sometimes the tissue obtained is not enough, thus more specimen is requested from the histology or flowcytometry laboratories.

Typically, laboratory tests performed on these types of samples are done in a combination of in-house laboratories and sent out to reference labs. Due to the invasive nature and associated cost of how these samples are collected, they are only obtained once, and resampling is kept to a minimum. By shuttling samples between in- and outside labs, the specimens are exhausted quicker and some tests cannot be performed.

Recent advances in genomic testing have focused on detecting somatic genetic alterations in blood and aspirates, and focuses on the premise that localized cells ‘leak’ genetic material due to apoptosis and other events that break down cells. These minute amounts of genetic material circulate through the body until fully degraded, and can be detected using sensitive tests like next-generation sequencing on circulating tumor DNA, obtained from liquid biopsies (Verma et al., 2020).

Raw data requires tremendous computer processing power. Data is sent to in-house servers or a Health Insurance Portability and Accountability Act of 1996 (HIPAA) compliant cloud for processing. This processed data is then returned and needs to be further interpreted by molecular pathologists. Interpretation is cumbersome due to the vast amount of data obtained, and there is a certain amount of scientific upkeep because of daily advances made in this field. Data obtained from the various tests performed collectively make up a diagnosis and inform care providers of possible treatment options. This data is obtained from in-house and reference laboratories, and is received at different time intervals that are sometime spaced weeks apart. This adds yet another layer of complexity.

3.4 Imaging Departments

Imaging departments utilize various devices, ranging from mobile devices to large stationary built-in devices, and play a critical role in the initial diagnoses for cancers and other abnormalities. The devices include sonograms, X-ray, magnetic resonance imaging, positron emission tomography, and computed tomography devices.

Images are captured in digital formats, and due to their high resolution, require vast amounts of storage space. On- or off-site radiologists interpret these images after which a diagnosis is made. Currently, there is an evolution within the use of artificial intelligence algorithms interpreting these images. The AI algorithms are trained with previously identified images, thus over time the ability of the AI improves. Because a computer is analyzing the images, results are obtained quicker than when radiologists can interpret these. As any test, there are always false-positives and false-negatives. The AI algorithms have proven to obtain equivalent accuracies as radiologists, with the potential of improved accuracy as these models develop further (Wong et al., 2020).

In addition to obtaining images for diagnostic purposes, these departments are also involved in obtaining biopsies. Fine needle aspirates and needle core biopsies require guidance from the imaging departments to accurately locate the malignant tissue previously identified, and guide the needle to this location. These procedures reduce complications for patients when compared to traditional biopsies obtained through surgeries and are cheaper to perform.

Other advances allow for imaging departments to play new roles. One such development allows for biopsies to be obtained from tissue without the need of surgeries or aspirates. Microscopic air bubbles can be injected and popped at the malignant tissue site using radiofrequencies. This allows genetic material to leak into the peripheral blood, which can then be collected through a regular blood draw (Pacia et al., 2020).

3.5 Information Technology Departments

Information technology (IT) departments comprise of many different teams, including security, networking, and clinical informatics. The biggest foci for IT are to ensure compliance with HIPAA by protecting personal health information (PHI), securing the network from breaches, and maintaining connectivity between departments. Data breaches can come in many forms, such as external attacks by hackers, who either want to obtain information, or block critical features within the network, in order to obtain a ransom. Breaches can also come from within, through malware (un)consciously delivered by service personnel or employees.

In order for a network to run properly it requires constant maintenance, updating, and redundancies must be built in, in order to be able to process an onslaught of data input and retrieval. Connectivity between computers across various locations and equipment must be ensured for seamless operations.

The clinical informatics team’s main focus is to maintain the EMR, and to build out modules in order for care providers to work more efficiently. A problem they face is that deposited data comprises out of many formats, including voice recordings, free text, standardized text, numerical values, images, and videos. Additional complexities arise because data is collected over time and is attached to specific events. Some data belong to the visit, and other data belong to a specific sample of that patient. Sometimes there are multiple data entries for a specific sample, or a sample can be split into multiple sources. Collectively, these data must be presented in an organized way in order for care providers to optimally use it and redundancies reduced (A. Barroso, personal communication, November 2, 2020).

The Health Information Technology for Economic and Clinical Health Act of 2009 (HITECH Act) was introduced to provide a framework and to incentivize adoption of EMRs with the goal of improved patient care. This act led to the establishment of the Office of the National Coordinator for Health Information Technology (ONC). ONC developed the Health IT Playbook, which describes IT roles, and with regulatory updates in mind aims to shape new roles. As value-based care has become a bigger focus, and reimbursement models are increasingly favoring better outcomes, ONC incorporates more strategic advice into its playbook. Emphasis on analytics is one such advice, but is still focused on billing, and quality improvement as part of patient safety. However, it lacks guidance on implementation of analysis to improve outcomes.

Transferring patient data should be easier with the implementation of EMRs. This transfer is not limited to ensuring continuing care, where data is transferred between various care organizations, and also between care organizations and insurance companies, health departments, and commercial entities. Legislation through HIPAA inhibits the transfer of data with the aim to protect patient’s privacy. This inhibition creates difficulties and increased administrative burden for care providers, especially when dealing with entities that are not covered (HHS Office of the Secretary & Office for Civil Rights, 2013). Additionally, EMRs are expensive to maintain due to technological barriers; institutions either have limited capital or are reluctant to invest in EMRs, and the local labor market has a limited supply of skilled IT workers (Catanzaro & Kain, 2019).

3.6 Patients

In terms of precision medicine, patient empowerment is key. The primary objective involves education, creating awareness, and ensuring access to care. The value must be apparent and infrastructures created (Pritchard et al., 2017). Creating awareness and educating patients is difficult, especially because this service is not directly reimbursed, nor is its value easily recognized. Entities like Cleveland Clinic and Mayo Clinic have developed programs that create awareness and educate patients. However, its reach is limited.

Most patients are only aware of the existence of precision medicine within the oncology space, but not in other areas such as cardiovascular diseases and muscle disorders. Patients are afraid to obtain this knowledge. It can lead to concerns of losing control over their health management and guilt towards the possibility of passing on an inheritable disease (Frost et al., 2018). Patients are also scared that this information might impact them in other discriminatory ways including loss of employment and loss of insurance.

Patients involvement in their care plan’s decision-making has increased (Fowler, Levin, & Sepucha, 2011). Technologic advances, such as wearables, improve on this engagement. For example, Fitbit recently received FDA approval to use its new smartwatch to measure electrocardiograms (Wetsman, 2020). This data should also find its way into the EMR.

As patient data is being transferred between a network of entities that are involved in delivering care and precision medicine required to establish proper diagnosis, consent forms are needed in order to comply with HIPAA. Currently, these consent forms are not written with the patient in mind. In order to empower patients these forms must increase in clarity.

4.    Health Policy Options

From the aforementioned sections, it is apparent that precision medicine is complex and consists of a breadth of data spanning further then what is analyzed in the laboratory. ACOs focus on outcomes by reducing cost. The adoption of EMR systems, which allowed for improved communications between its various entities, resulted in improved continuing care. New policies must be adopted in order to leverage EMRs further to reduce cost of care whilst improving its quality. These options include: 1) Integration of clinical informatics, 2) Implementation of commercial clinical informatics pipeline, 3) Outsourcing all laboratory services, and 4) Maintain EMR.

4.1 Option 1: Integration of Clinical Informatics

With the introduction of EMRs healthcare systems have been collecting data in formats that allow them to communicate better, and implement a wide variety of analytics. Most analytics applied on EMR data has been focused on billing and quality assurance metrics. With healthcare cost rising, and a pressure to improve on outcomes, specifically for ACOs, this data can be utilized in ways to reliably decipher outcomes through means of precision medicine. However, this means that IT will have to play an integral role into clinical diagnostics.

The volume of healthcare data is increasing drastically, going from histories obtained at that one doctor’s visit and specimen analysis to continuous data acquisitions obtained through wearable devices, such as Fitbit (Wetsman, 2020). The total acquired patient data, together with genomic analytes, can be analyzed utilizing AI and has recently been proven to decipher anticipated outcomes for some complex diseases, such as epilepsy (Ganesan et al., 2020) and many other diseases, including rare diseases (James, Phadke, Wong, & Chowdhury, 2020). By knowing patient specific outcomes, a care plan can be optimized, including its supply chain, and therefore reduce cost and improve outcomes.

In order to reap the benefits from big data, acquired through EMRs and laboratory information systems, a proper clinical informatics pipeline must be developed. These pipelines must remain dynamic, and capable of growing accordingly. A HIPAA compliant cloud-based infrastructure allows for scalability without requiring in-house technical maintenance. A cloud-based infrastructure can be private, and allows for secure data storage as well as dedicated state-of-the-art computing power, needed to analyze big data (Kadri, 2020). By having all data organized in one location, future proposed upgrades which will likely be mandated in order to comply with HIPAA and reduce possible breaches, such as blockchain, can be easily implemented (Catanzaro & Kain, 2019).

IT should invest in hiring bioinformaticians. They execute a hybrid discipline between computer science and biology and specialize in data analytics from a biological perspective utilizing computational programming, and are capable of bridging and marrying the two disciplines seamlessly (James, Phadke, Wong, & Chowdhury, 2020).

A flexible database structure needs to be developed which allows for new formats to be integrated. The current EMR system and modules can deposit data through its Health-Level 7 (HL7) language exchange, and other data formats stored in their native languages. HL7 was introduced as a standard language in order to make EMRs interoperable (Health Level 7 International, n.d.). At the time this language proved to be very useful, but as data volumes and variety increased, this language has proven to have some barriers.

For example, HL7 is not capable of harboring image files, nor genomics data. Instead, these files are stored separately from the other data transmitted through the HL7 message, and HL7 merely points to the location of an image file. When data are aggregated into one well organized database clinical analytics can be applied more effectively. Images can be linked to mutations, resulting in reduced needs to perform expensive biopsies (Huang et al., 2019). This reduction also decreases associated risks and complications, and improve on patient satisfaction. Other EMR deposited data can also shed light on otherwise missed outcomes, especially in diseases that comprise of many genetic variations (Carpenter, Soorya, & Halpern, 2009).

This proposed option is costly, as it requires hiring specialty personnel, and a makeover of existing IT infrastructure. Estimated cost can be in the range of 3 to 9 million dollars (Catanzaro, & Kain, 2019). The HITECH Act incentivizes developments of this sort; thus, grants might be obtained. This is a long-term solution, where integrating clinical informatics will lead to iterative improvements and provide a competitive edge in the market place where patients are becoming more aware and consume healthcare based on cost and better outcomes (Shrank, 2017).

4.2 Option 2: Implementation of Commercial Clinical Informatics Pipeline

Because many care providing entities observe IT departments and to some extend laboratory services as ancillary services, investments are poor. The variety of commercial services who offer quick answers are vast, hence the incentive to not invest is high, but shortsighted. Implementation of EMR systems allowed for improved communication, and also the ability to aggregate data. This wealth of data can be utilized to advance quality of care, but can only be done through proper analytics. The need for bioinformaticians is rising, but the labor market is small (Dranove, Forman, Goldfarb, & Greenstein, 2014).

An available solution is to implement off-the-shelf clinical informatics modules, which touch on a variety of issues faced when dealing with big data; storage, interoperability, user interface, database structure, and analytics software (Kadri, 2020). These modules communicate to each other through the use of HL7. Originally, EMRs focused on optimizing billing, and generation of reports. To change an EMR is complex, very costly, and requires retraining of all personnel.

In order to keep up with advances new modules can be purchased and implemented fairly easily, because these modules communicate through HL7, and typically require small improvements on existing IT infrastructure. Big data can be analyzed with greater efficiency and accuracy, and customized user interfaces aid in the usability for each department. However, implementing off-the-shelf products will in the long run become difficult to adapt to future changes which come through unanticipated clinical advances, and is subject to upgrades dictated by companies. Data will remain segregated, because each module requires its own database structure. Inevitably, IT departments will segregate into many teams, each specializing in a group of modules (Kadri, 2020).

4.3 Option 3: Outsourcing All Genomic Laboratory Services

Due to speedy advances, where new clinical impacts of genes are being discovered daily, the clinical genomic test menu is expanding rapidly. As a result, care providers are continuously seeking to test for new analytes with the hope to find other drug targets. ACOs focus is on patient care, not on diagnostics. Therefore, diagnostics can potentially be outsourced. Outsourcing these non-urgent tests to other specialty laboratories provide the benefit that these companies are set up to implement changes into their test menu. The cost and time required to validate new tests will be saved, and other costs associated with capital equipment purchases, maintenance, and other overhead as well. The labor market for specialists in genomic testing is expanding, but at present is still small (Dranove, Forman, Goldfarb, & Greenstein, 2014).

The downside of outsourcing services related to genomic testing is the lack of sample optimization, resulting in repeat surgical or interventional radiology procedures to obtain biopsies. Data provided by outsourced laboratories are in text format, and only list mutations found and omitting data that is considered irrelevant to the condition of the patient at the time, and all the negative data points. This loss of data is devastating for ACOs, because future discoveries of clinical significance for genomic sequences cannot be tied into previously tested specimens, and retesting might be necessary (Kadri, 2020). Additionally, outsourced services are on average 5 times more expensive when compared to performing these tests in-house (Prevention Genetics, n.d.).

4.4 Option 4: Maintain EMR

Average maintenance costs for EMRs at urban hospitals are about $1 million annually, and require approximately 70,000 hours of labor to maintain (Dranove, Forman, Goldfarb, & Greenstein, 2014). ACOs have adopted EMRs and utilize them across all their entities to enable data uniformity and enhancing communication. This results in improved continuing care models, which are especially needed for those patients with complex diseases, such as diabetes, cardiovascular diseases, cancer, or other diseases which require extended care.

By maintaining the EMR as is, cost will change at a minimal pace and it will allow for the ACO to focus on patient care. However, commercial companies who provide EMRs will continue to update their systems which will have to be implemented at the end-user’s side. The EMR systems will evolve with the market, and thus will provide improved billing and quality analysis. Reports, required by various accrediting and government organization, can also be generated. Most IT personnel don’t need to be specialists. Instead, they can be trained specifically to utilize and set up features in the existing EMR systems. These personnel can be promoted from within the system, and these members have an intricate knowledge of the various departments, which can be leveraged into the EMR’s design. Unfortunately, promoting personnel from within the system, and with a lack of true IT knowledge, leads to labor waste. These members often do not possess the required knowledge to implement optimal ways for new module implementation. It requires knowledge of user interface design, user behaviors, coding, and the ability to recognize cross-departmental dependencies. Therefore, newly implemented models need rework, which is another type of waste (Catanzaro & Kain, 2019). Additionally, data deposited will remain fragmented without proper linkage to boost each other’s value.

5.    Discussion

Traditionally, precision medicine is seen as a laboratory diagnostic which would dictate treatment protocols, and is especially geared to pharmaceuticals. Genomic data is vast and requires unique ways to be analyzed. From the millions of data points only those that are clinically relevant are reported. Maintaining the EMR as is, and incorporating updates as advised by the providing company, ACOs can focus on providing care. As the providing companies offer new modules, these can be seamlessly incorporated by existing personnel, and therefore associated cost kept at a minimum. Basic genomic data obtained in-house can easily be deposited into these EMRs, and some modules might be able to deal with advanced genomic data. Advanced genomic data is analyzed in the lab with standardized software, after which an abstract of clinically relevant data can be deposited in text or scanned format. Treatment plans can be developed according to best practices and standard of care as established by various accreditation and governing agencies. However, maintaining EMRs as is will only allow to measure quality solely on outcomes, resulting in minor process optimizations to be implemented in order to reduce cost.

Outsourcing all genomic laboratory services allow to reduce cost. These services are mostly non-urgent and therefore can be centralized. By outsourcing all genomic services, it also reduces the need to deal with complex data and associated infrastructures. As test menus grow, there is no need for further investment by the ACO, but still allowing to test for newly associated clinically relevant analytes. Though set up and maintenance costs will be avoided, outsourced tests are more costly. Additionally, biopsied material will be utilized in a suboptimal manner, thus tests need to be prioritized in order to avoid performing more biopsies.

Implementing a commercial clinical informatics pipeline can be done utilizing existing IT infrastructures. Modules, either provided by the EMR vendor of choice, or others, can be set up easily. These tend to be plug-and-play, and through HL7 can communicate with the existing EMR. Each type of data can be analyzed more efficiently and databases will store these complete data sets, thus can be accessed in the future. Because these databases are segregated it will remain difficult to perform integrated data analytics.

Integration of clinical informatics offers a complex solution, but it will pave the way of the future as observed in the literature. The genomic test menu is expanding rapidly, and the volume of data deposited in EMRs is vast but segregated. Integration allows for this data to come together and remain dynamic, and HIPAA compliant cloud-based storage and computing allows for scalability. Because of this, other forms of analytics can be performed and geared towards deciphering patient specific outcomes correlated to their complete genomic makeup. As a result, care plans and its supply chain can be optimized, reducing the need of unnecessary or adverse treatments otherwise not noted, and implement treatments geared towards the prevention of future more expensive treatments.

6.    Recommendation

ACOs would benefit most by integration of clinical informatics. It allows for new ways for an entire care team to interact with each other, and each care provider’s contribution becomes more visible. ACOs are largely vertically integrated, and the integration of clinical informatics enhances this. Bioinformaticians will bridge the disconnect between IT and care providers. By utilizing a well-organized database structure all data can be correlated and analyzed. HIPAA compliant cloud-based storage and computing provide scalability without the need for hardware maintenance or hardware update implementation. Future regulatory requirements to secure PHI into blockchain will be easier to implement. The continuing care plans for those patients with complex diseases, including diabetes, cardiovascular disease, and cancer, can be iteratively optimized through the added analytical capabilities provided by an integrated clinical analytics pipeline. Therefore, it will reduce the cost of care and improve outcomes. Further reduction of costs can be observed as the supply chain of care delivery is optimized, and physician burnout reduced. These superior outcomes and acquired analytical agility will position the ACO more competitive in the market place, and patients will be more satisfied.

7.    Conclusion

ACOs are always in pursuit to reduce costs and improve outcomes, which is incentivized by their reimbursement models. The preceding analysis shows that IT and genomic laboratory services are not properly integrated, due to their vast and complex data structures, hence many benefits cannot be reaped. The aforementioned options each provide some specific advantages for ACOs, and are mainly beneficial from an immediate cost perspective with limited benefits pertaining to the value for patients. Only the integration of clinical informatics will provide adequate improvements to outcomes and reduce cost of care. Additional benefits include that it will set up an infrastructure that is dynamic, scalable, and allows for future scientific advances to easily be incorporated. The entirety of data deposited will always be accessible and new correlations can be deciphered.

7.  References

Bailit, M., & Hughes, C. (2011). Key design elements of shared-savings payment arrangements.

Issue Brief (Commonwealth Fund), 20, 1–16.

Beasley, D. (2015). Alternative payment models: a changing landscape. MLO: Medical

Laboratory Observer, 47(8), 20–23.

Carpenter, L. A., Soorya, L., & Halpern, D. (2009). Asperger’s Syndrome and High-

Functioning Autism. Pediatric Annals, 38(1), 30-35. doi:10.3928/00904481-20090101-

01

Catanzaro, Z. L., & Kain, R. (2019). Patients as Peers: Blockchain Based EHR and Medical

Information Commons Models for HITECH Act Compliance. Nova Law Review, 44(3), 289–336

Dranove, D., Forman, C., Goldfarb, A., & Greenstein, S. (2014). The Trillion Dollar

Conundrum: Complementarities and Health Information Technology. American

Economic Journal: Economic Policy, 6(4), 239-270. doi:10.2307/43189410

Fowler, Jr., F. J., Levin, C. A., & Sepucha, K. R. (2011). Informing and Involving Patients To

Improve the Quality of Medical Decisions. Health Affairs, 30(4), 699–706.

Frost, C. J., Andrulis, I. L., Buys, S. S., Hopper, J. L., John, E. M., Terry, M. B., . . . Daly, M. B.

(2018). Assessing patient readiness for personalized genomic medicine. Journal of

Community Genetics, 10(1), 109-120. doi:10.1007/s12687-018-0365-5

Gaynes, R. (2017). The Discovery of Penicillin - New Insights After More Than 75 Years of

Clinical Use. Emerging Infectious Diseases, 23(5), 849-853. doi:10.3201/eid2305.161556

Ganesan, S., Galer, P. D., Helbig, K. L., Mckeown, S. E., O’Brien, M., Gonzalez, A. K., . . .

Helbig, I. (2020). A longitudinal footprint of genetic epilepsies using automated

electronic medical record interpretation. Genetics in Medicine. DOI:10.1038/s41436-020-

0923-1

Golden, A. (2015). What Is an Accountable Care Organization? The Journal for Nurse

Practitioners, 11(8), 843–844.

Health Insurance Portability and Accountability Act, 42 U.S.C. § 13001 (2009)

Health Level Seven International. (n.d.). About HL7. Retrieved November 09, 2020, from

https://www.hl7.org/about/index.cfm

HHS Office of the Secretary, & Office for Civil Rights. (2013, July 26). Summary of the HIPAA

Privacy Rule. Retrieved February 25, 2020, from https://www.hhs.gov/hipaa/for-professionals/privacy/laws-regulations/index.html

Huang, J., Yu, J., Tu, L., Huang, N., Li, H., & Luo, Y. (2019). Isocitrate Dehydrogenase

Mutations in Glioma: From Basic Discovery to Therapeutics Development. Frontiers in

Oncology, 9. doi:10.3389/fonc.2019.00506

Innovation Models. (n.d.). Retrieved October 23, 2020, from

https://innovation.cms.gov/innovation-models

James, K. N., Phadke, S., Wong, T. C., & Chowdhury, S. (2020). Artificial Intelligence in the

Genetic Diagnosis of Rare Disease. Advances in Molecular Pathology, 3, 143-155. doi:10.1016/j.yamp.2020.07.013

Kadri, S. (2020). Building Infrastructure and Workflows for Clinical Bioinformatics Pipelines.

Advances in Molecular Pathology, 3, 157-167. doi:10.1016/j.yamp.2020.07.014

Kalf, R., & Meinecke, A. (n.d.). Healthcare databases: Focus on electronic health records.

Retrieved October 29, 2020, from https://rwe-navigator.eu/use-real-world-

evidence/sources-of-real-world-data/healthcare-databases-with-a-focus-on-electronic-

health-records/

Katzen, J., Dodelzon, K., Michaels, A., & Drotman, M. (2020). Lessons learned: A single

academic department's unique approach to preventing physician burnout. Clinical Imaging, 67, 58-61. doi:10.1016/j.clinimag.2020.05.025

Matulis, R., & Lloyd, J. (2018, February). The History, Evolution, and Future of Medicaid

Accountable Care Organizations. Retrieved October 28, 2020, from

https://www.chcs.org/resource/history-evolution-future-medicaid-accountable-care-

organizations/

McCandless, S. E., Brunger, J. W., & Cassidy, S. B. (2004). The Burden of Genetic Disease

on Inpatient Care in a Children’s Hospital. The American Journal of Human Genetics,

74(1), 121-127. doi:10.1086/381053

McKinney, S. M., Sieniek, M., Godbole, V., Godwin, J., Antropova, N., Ashrafian, H., . . .

Shetty, S. (2020). International evaluation of an AI system for breast cancer screening.

Nature, 577(7788), 89-94. doi:10.1038/s41586-019-1799-6

Miller, K. E., Hoyt, R., Rust, S., Doerschuk, R., Huang, Y., & Lin, S. M. (2020). The Financial

Impact of Genetic Diseases in a Pediatric Accountable Care Organization. Frontiers in Public Health, 8. doi:10.3389/fpubh.2020.00058

Pacia, C. P., Zhu, L., Yang, Y., Yue, Y., Nazeri, A., Gach, H. M., . . . Chen, H. (2020).

Feasibility and safety of focused ultrasound-enabled liquid biopsy in the brain of a

porcine model. Scientific Reports, 10(1). doi:10.1038/s41598-020-64440-3

Porter, M. E. & Jain, S. H., (2018). The University of Texas MD Anderson Cancer Center:

Interdisciplinary Cancer Care. Harvard Busines Review. 9-708-487.

Prevention Genetics. (n.d.). Tests by Category. Retrieved November 09, 2020, from

https://www.preventiongenetics.com/ClinicalTesting/TestCategory/category

Pritchard, D. E., Moeckel, F., Villa, M. S., Housman, L. T., Mccarty, C. A., & Mcleod, H. L.

(2017). Strategies for integrating personalized medicine into healthcare practice.

Personalized Medicine, 14(2), 141-152. doi:10.2217/pme-2016-0064

Shrank, W. H. (2017). Primary Care Practice Transformation and the Rise of Consumerism.

Journal of General Internal Medicine, 32(4), 387–391. doi: 10.1007/s11606-016-3946-1

Shumate, A., Zimin, A. V., Sherman, R. M., Puiu, D., Wagner, J. M., Olson, N. D., . . . Salzberg,

S. L. (2020). Assembly and annotation of an Ashkenazi human reference genome.

Genome Biology, 21(1). doi:10.1186/s13059-020-02047-7

Suh, C. H., Kim, H. S., Jung, S. C., Choi, C. G., & Kim, S. J. (2018). Imaging prediction of

isocitrate dehydrogenase (IDH) mutation in patients with glioma: A systemic review and

meta-analysis. European Radiology, 29(2), 745-758. doi:10.1007/s00330-018-5608-7

The Human Genome Project. (n.d.). Retrieved October 27, 2020, from

https://www.genome.gov/human-genome-project

The Office of the National Coordinator for Health Information Technology. (2020, March 11).

Health IT Playbook. Retrieved November 03, 2020, from

https://www.healthit.gov/playbook/population-public-health/

Tucker, T., Marra, M., & Friedman, J. M. (2009). Massively Parallel Sequencing: The Next Big

The thing in Genetic Medicine. The American Journal of Human Genetics, 85(2), 142-

154. doi:10.1016/j.ajhg.2009.06.022

Turner, T. N., Coe, B. P., Dickel, D. E., Hoekzema, K., Nelson, B. J., Zody, M. C., . . . Eichler,

E. E. (2017). Genomic Patterns of De Novo Mutation in Simplex Autism. Cell, 171(3).

doi:10.1016/j.cell.2017.08.047

Verma, S., Moore, M. W., Ringler, R., Ghosal, A., Horvath, K., Naef, T., . . . Gunn, S. (2020).

Analytical performance evaluation of a commercial next generation sequencing liquid

biopsy platform using plasma ctDNA, reference standards, and synthetic serial dilution

samples derived from normal plasma. BMC Cancer, 20(1). doi:10.1186/s12885-020-

07445-5

Wetsman, N. (2020, September 14). Fitbit’s Sense smartwatch gets FDA clearance for the EKG

app. Retrieved September 20, 2020, from

https://www.theverge.com/2020/9/14/21436090/fitbit-sense-ekg-heart-fda-clearance-

apple-Samsung-Withings

Wong, D. J., Gandomkar, Z., Wu, W., Zhang, G., Gao, W., He, X., . . . Reed, W. (2020).

Artificial intelligence and convolution neural networks assessing mammographic images:

A narrative literature review. Journal of Medical Radiation Sciences, 67(2), 134-142.

doi:10.1002/jmrs.385

8.    Appendix

Table 1

Disease categories and their associated average annual costs per capita

Note. Adapted from “The Burden of Genetic Disease on Inpatient Care in a Children’s Hospital,” by S.E. McCandless, J.W., Brunger, & S.B. Cassidy, 2004, The American Journal of Human Genetics, 74(1) & “The Financial Impact of Genetic Diseases in a Pediatric Accountable Care Organization,” by K.E. Miller, R., Hoyt, S., Rust, R., Doerschuk, Y., Huang, & S.M. Lin, 2020, Frontiers in Public Health, 8

Figure 1. Illustration depicting the organization of the human genome. The helical DNA is comprised of nucleotides that together form stretches of genes interspersed with non-coding and regulatory regions. DNA is folded into chromosomes to reduce the space it occupies in order to fit within the cell’s nuclei. Copyright 2019 by ttsz | Getty Images/iStockphoto. Reprint with permission.

Figure 2. Flow chart of biopsy specimens through various laboratory departments.