25x5 Pitch Event

• Products and tools that alter clinician workflow in ways that reduce the burden of documentation
  • Better and more affordable voice recognition integration into workflow
  • Vendor agnostic AI/heuristics to expedite the way clinicians document
• Separate clinician notes for care process & completion of the extra record keeping required for billing by AI
• Vendors to integrate capture of patient generated health data into clinician workflow automatically, so that clinicians do not have to re-enter this data again
• Other solutions to prevent data re-entry
• Eliminate any work for clinicians to document what patient data have been reviewed

• Core HIT enablers that unblock the full HIT vendor community from contributing to reduce documentation burden by 75%
• Eliminate the need for manual entry of patient data from external sources (e.g., devices, labs, etc.) into documentation by providing true interoperability and availability of this data into clinician workflows
• Vendor agnostic support for use of multiple devices for information collection and documentation
• Interoperability interfaces and user interfaces designed to support clinician workflow in a way that reduces documentation burden
• Eliminating barriers to entry (technical AND business model) for accessing the ecosystem to integrate into EHR workflows
• Create capability to decompose CCDA for granular reuse of data points beyond certification requirements
• Establish standards with regards to documentation for compliance

• Develop a metric that automatically grades notes on length/efficiency/redundancy; can utilize AI algorithm and schedules training; automated benchmarking/analysis of documentation trend, individual & department level
• Methodology/tools to support EHR vendors improving their responsiveness to user suggestions/complaints; amplify the voice of the users; improved transparency with user feedback/problems

• Vendors can propose ideas that fall outside of the above priorities if they reduce documentation burden

Within our project, a decision box for voice recognition, digital scribe, and dictation software will be produced.

A time-out (Screen time-out) key of two minutes for every 20-minute interaction with the screen to enable physicians to stop, look away from their screen, and reconnect with the patient while allowing a digital scribe to take notes of the encounter with the assistance of the voice recorder.

Place them within the appropriate categories within the electronic health record (EHR) while allowing for a replay of the script, and editing if corrections are needed at the end of the encounter.

This project will satisfy the provider's need to pause and increase valuable time to respond to the patient while giving their full attention and reducing the stress brought on by click fatigue and prolonged screen time; during this added process, patients will feel validated in their interactions which will increase satisfaction between physician and patient, leading to improved care quality and reduced error which will yield better revenue for the organization.

On Interoperability:

For instance, A  Standard Clinical summary document for a patient with a diagnosis should have an ICD-10 coding, depending on the EHR (Electronic Health Records) Vendor which can be generated in SNOMED-CT mapping to FHIR Resource Category for Diagnosis as well as the Prescribed medication dosing and formulation code in Rx Norm while noting any drug or medication interactions, allergies and the list of recently consumed medications, both over the counter, substances, and herbal or alternative medication listed.

This will promote cost reduction for repeated tests at the PCP in events of return, medication reconciliation, and treatment updates while preventing re-admission since the information required is provided.

This is in hope that policies will be made to support Certified Electronic Health care Vendors using similar Ontology standards to conform to a common HL7 SMART (Specific Measurable Achievable Relevant Time) on FHIR integration database Standard for improved Clinical Document Architecture (CDA) and Continuity of Care documents (CCD) and other Protected Health Information documents that will be available at health information exchange (HIE) request endpoints.

We have a robust governance structure at our health system which includes several bedside nurses. We currently document everything for patients which includes normal values in assessments. I am leading a project to move away from that practice and move towards documenting by exception by using the WDL method. However, there are several ways WDL (Within Defined Limits) can be implemented. We believe we have designed a good structure to help nurses reduce unnecessary documentation and support nurses only documenting on the required or needed assessment items. To start all assessment rows start collapsed under each main body system.

If exceptions are noted or selected, then nurses will select the specific assessments needed. Also, instead of documenting WDL on each assessment, we found that nurses feel obligated to continue documenting on assessments that have returned to normal. Instead, we will use the term "Return to WDL" so that nurses do not continue to document on once abnormal assessments. This project is in the design phase and we are engaging nurses from several inpatient areas to change the documentation practice in order to decrease the burden of documentation.

Despite medical device enhancements, the documentation burden on caregivers in 2022 remains high. This creates a barrier to positive patient-to-caregiver relationships, negatively impacts the experience of the patient and contributes to the burnout of the caregiver.  Interoperability solutions that decrease the time spent documenting while supporting nurses’ clinical decision making, are solutions that will improve the patient experience, drive outcomes, and reduce clinician turnover.

Many devices in the clinical environment and connected to the patient are collecting valuable and insightful information but are limited in their ability to communicate that relevant information directly to the EMR.  Through vendor agnostic interfaces from Optical Character Readers to Serial Interfaces to Server Interfaces, our health IT company is automating the collection and documentation of vitals directly to the EMR to drive more time at the bedside, more time spent delivering care and less time spent in “data entry” mode.

Once data at the point-of-care is collected, it can be further automated by applying clinical decision support algorithms that support nurses in both their collaboration and care delivery decision making.  Application of these algorithms will provide early recognition of deterioration and contextual information around alerts for response prioritization.  In the drive to reduce the documentation burden we should be driven to support care delivery through the application of algorithms and present results through visualization tools that give caregivers actionable insights and reduce their cognitive workload across disparate clinical data sources.  

Through collection, analysis, and visualization, we propose to offer an automated, end-to-end, connected care and clinical collaboration solution that reduces the documentation burden, improves the accuracy of the data record, reduces the latency of available information, improves the patient experience, improves the clinician experiences, reduces care delivery costs, and most importantly, directly and positively impacts patient outcomes.

Create foundation-built hyperlinks that can be added to notes with appropriate verbiage (ie "The Problem List" [Medication list, Allergy list, Social History, Family History] was personally reviewed and updated as of [now].).  Then once those foundation-built hyperlinks exist, create a metric to see what percentage of the user/system note templates include those hyperlink smartlinks.   

This will allow the user to edit the Problem List, Med list, etc in the respective activity within the EMR, and then still take credit for doing that documentation work within the note without adding all of the details to the note itself.  Part of the metric would be to also include what percent of notes have other smartlinks which have high character counts.

Currently, providers generally don't use Flowsheets; however, there is a lot of required provider documentation that currently gets dumped into various progress notes or provider notes and are difficult to retrieve the information.  We could create flowsheet rows (Epic lingo) that can be automatically populated from orders which can satisfy the mandated documentation so that providers can don't have to document twice and so that their required documentation is easier for others to find.  

For example, to transfer a patient, due to Emergency Medical Treatment and Labor Act (EMTALA), providers have a litany of things to document.  We have a form that fills in flowsheet rows.  If we can do that instead of also documenting within the Provider Note, then the information is still discoverable within the flowsheet and is not bloating the ED Provider Note.

Clinical documentation serves several distinct purposes including providing accurate records of clinical decision-making and management plans, facilitating communication between providers, and directing billing and reimbursement.  

To fulfill these purposes while decreasing documentation burden, we propose performing a documentation needs assessment from end users and reviewing the latest CMS billing requirements to guide modification of current note templates.  Then, applying these criteria, note templates would be standardized with maximized use of Epic SmartLinks to automatically pull in relevant data from the chart, thus eliminating redundancy.  We also propose building an artificial intelligence (AI) algorithm based on QNote, a validated instrument for scoring documentation quality, to grade note quality from which individual note-writers would receive practical feedback on a quarterly basis while groups of physicians’ compiled scores would be trended to gauge overall improvement.  

Impact would be measured pre- and post-intervention.  Objectively, change would be quantified by collecting template utilization rates, QNote scores, billing audits, Epic Signal data measuring time required for charting, and an objective burnout score such as the Maslach Burnout Inventory.  Subjectively, survey data would be gathered from end users evaluating their satisfaction with the new notes, both their own and those of consultants -- perceived time required to complete documentation, and degree of burnout due to documentation.  

Note templates would be modified and tailored to each department’s needs, and could be introduced in stages to improve each iteration, and scalable to be shared between institutions for nationwide impact.  We aim for at least 75% of the chosen pilot department to adopt the new note templates and have statistically significant improvements in post-intervention QNote note documentation scores and provider satisfaction and burnout metrics.

Increased clarity is needed to determine what documentation elements are required. The lack of clarity and the paucity of guiding principles has potentiated documentation burden. Documentation by design is a tool created through collaboration with the Office of Burden Reduction & Health Informatics. After establishing a set of guiding principles, a check-list type tool will be created. The tool outlines the data element to be documented, associated elements, the frequency required, and regulatory and/or safety and quality reasons for the documentation. The tool will serve as a resource for regulatory agencies, professional associations, and individual healthcare organizations to make clear standards and ensure smart decision making with regards to electronic health record (EHR) documentation. A symbol trademarked 'D' will be used in written standards to display when a data element requires EHR documentation.

Rapid titration of life-saving medications is a vital aspect of critical care management during emergent situations. Documentation of these medication dose changes are also important to providing safe patient care. However, clinicians are faced with the challenge of trying to care for the patient with the additional burden of real time documentation of medication titrations, hence the nurse must balance patient care and documentation. Our audits revealed that challenges with compliance in documenting real-time medication dose changes during these situations were not meeting regulatory compliance.

A survey was sent out to critical care nurses on M11 highlighted this dilemma. Eight-six (86%) of nurses surveyed expressed interest in learning an alternative method for rapid titration documentation. Ninety-one percent (91%) of nurses felt that documentation takes them away from providing adequate patient care and ninety-three percent (93%) felt that documentation during drip titration is burdensome. A new standard method for documentation called “block charting” was introduced in 2020 to help nurses in these situations. Block charting allows the bedside nurse to meet documentation requirements and continue to provide patient care without major disruption, allowing the nurse to spend more time at the bedside caring for the patient (The Joint Commission, 2021). Block charting gives the nurses a 4 hour grace period, during emergency situations for documentation. During this time the nurse is only required to document Start and end times for the block charting episode. The complete and accurate order for each medication for titration. The start and end rate/dose during the block charting episode, and the maximum rate/dose. This allows the nurse to focus on providing safe and efficient care to their patient first.

Clinical documentation in pressure injuries is often inaccurate and sometimes missed leading to poor interventions, failure in early detections, and the worst unnecessary litigations. Healthcare organizations worldwide have been devising ways to set standards for accurate and timely documentation of pressure injuries. This often leads to more burdens for clinicians, including nurses. Bedsores occur when mobility limitations exist, but these are preventable with early detection and proper documentation. Most organizations in the US use standardized tools like the Braden Scale to categorize and measure bedsore severity; however, the method of visually classifying is still done by describing using the naked eye. This leads to gaps and inaccuracies because not everyone has the same level of knowledge and experience when it comes to clinical assessments. But what if we can harness the power of Artificial Intelligence (AI), image recognition, and Light Detection And Ranging (LiDAR) technology to solve these gaps?

LiDAR technology is used in the car industry today. It utilizes a pulsed laser to accurately and constantly measure distances to a given target or area. Newly developed smartphones already have LiDARs. We can combine AI, image recognition, and LiDAR to seamlessly capture skin breakdowns, automatically providing an accurate description, location, size, depth, and category. Imagine doing an initial skin assessment and seeing a wound or skin discoloration that you want to document and report immediately; all you need to do is capture the skin using this powerful device that is integrated into electronic health records and instantly transmits complete documentation to the physicians without writing a lengthy note.

The availability of these technologies cements this idea's feasibility. Clinicians will have more time for their patients rather than the computers. This approach can solve the documentation burden and significantly reduce unnecessary medical care costs and delays in treatment or prevention.

Thrive in Clinic is an advanced team-based EHR training program for ambulatory clinics. Our intent is to raise EHR proficiency for every staff member, standardize workflows, optimize Epic build, and ultimately improve patient experience. We are data driven, we report with transparency, and we aim to simultaneously reduce burnout and rekindle the joy of healthcare. We believe documentation burden encompasses not only time spent on documentation but the overall user experience, perceived ease of use, and EHR satisfaction.

We have found that engagement in efficiency and optimization training is greatest when facilitated by like-roles and when the entire clinic is included. Our training team gets to know each clinic prior to our training program through questionnaires and observation. The curriculum is developed based on our discovery period with each clinic and also includes our core, standardized educational topics. Training is provided through one-to-one and group sessions over a two-week period.

Effectiveness of our program is measured primarily with survey responses gathered by a third party.   Group teamwork, cooperation, level of burnout, factors contributing to burnout, ease of use, ability to prevent mistakes, and accessibility to data are some of the measures that we track.  We have seen a 21 percent increase in the Net EHR Experience Score with the clinics that we have worked with thus far.  We have also seen improvements in communication, teamwork and a reduction in burnout.  Quantitative measures of documentation time and efficiency tools are also possible through data curated by our EHR vendor.

Our program started as a small pilot, and we have grown to a fully-funded department.  We believe our training model is scalable and it is imperative to reduce documentation burnout for all clinicians and staff.

It is common for patients to fill out many forms when seeking care as a new or returning patient. These forms contain information that are essential in delivering high quality care. However, the process of gathering and translating that information into actionable care could be burdensome: the patient must complete a form, several members of the healthcare team must review and re-transcribe the from paper into the electronic health record (EHR), and then use that information to provide care. Manual processes take time and effort, delays care, and increase transcription error risk.

To reduce burden on our clinicians, convert all patient facing paper forms into an electronic form. The form can be completed by the patient via the patient portal, through a tablet in clinic or any patient handheld mobile device.  

After the patient submits the form, responses can be shared across other forms as needed and is integrated directly into the EHR. The clinicians will open their clinical note and responses will automatically populate their note.  The clinician will review, clarify, and verify with the patient and make changes as needed. This will be a bi-directional update between patient portal and the EHR. From the same note, referrals/orders (such as social work, chaplaincy, nutrition) or any actionable items are automatically generated, reducing end-to-end steps in the process.

In the pilot of voice recognition technology (VRT) for nurses, I’ve learned that there is still much work to do to fit VRT into nurses’ structured flowsheet documentation workflow.  It is difficult to dictate on a flowsheet, where nurses spend most of their time in documentation.  There are limited available commands and it is challenging to insert the dictated data in the right place on the flowsheet.  These barriers can cause VRT adoption challenges for nurses.  

My proposal is to create a software that will allow nurses to simply dictate in a free-flowing narrative format and leverage natural language processing (NLP) to parse out the needed data.  Nurses can either dictate at the bedside in the narrative format using a mobile device or at a desktop instead of using the keyboard and mouse.  The nurse will review the NLP generated suggestions of what can be extracted into flowsheet (structured) format.  Clinical decision support can be built-in to include recommendations on any gaps or missing information required for a specific assessment or topic.  Nurses can review and address these recommendations in this same screen before data is saved into the electronic medical record (EMR).  I believe this will greatly simplify the use of VRT for nurses and increase adoption, leading to a solution in reducing documentation burden for nurses.  It is different from the nascent ambient technology, where it can pose challenges in capturing conversations between patient and nurse in semi-private hospital rooms. 

S: My goal is to create better machine learning (ML)/Artificial Intelligence (AI) to reduce the amount of time providers spend charting by 50%. This will be achieved by using Augmented Reality (AR) glasses that will utilize Automatic Speech Recognition (ASR). During a provider’s visit with a patient the glasses will use ASR and display blocks(paragraph/sentences) of speech that provider has spoken. The provider then will be able to select which blocks that are pertinent to charting.

A personal ML algorithm trained by the providers previous notes and voice will rewrite the blocks in proper medical format. The provider will then be able to finish by speaking the rest of the note and submit it.  

Another ML algorithm will be utilized to recognize when a provider prescribes medication. The provider will be able to use previous spoken words to create and digitally sign prescriptions. The prescription then will be sent to the pharmacy reducing time.

M: The decrease in time will be measured by timing patient visits and charting before implementation and then measured after implementation. The average time per (patient visit + charting time) from pre and post implementation will be used to calculate a percentage for the decrease in time.

A: The goal will be achievable, but will require a team with many different roles. These roles will include ML/AI experts, AR glasses vendor, EHR vendor, providers, IT infrastructure, etc.

R: This goal will help work towards the 75% decrease in charting burden set by the AMIA.

T: This project will take approximately 2 years. The first year will be spent creating the ML algorithms and training the AI. The second year will be focused on bringing in the AR glasses and introducing them into the workflow.

Many hospitals and clinics have seen the number of patient portal messages they receive triple over the last few years, requiring providers to spend more time outside of work answering these messages. Since the COVID-19 pandemic further accelerated this growth, patients have become more accustomed to Telehealth services and expect text based communication with their providers.
Clinicians, nurses, and medical assistants are tasked with assessing messages over a far range of topics from scheduling requests, medication refills, lab test questions, new untreated symptoms, and clinical follow up questions. However, patient portal messages are unsorted, untriaged, and lack a systematic way to view and document the relevant clinical information used in message responses. The increase in messages has not only increased the time it takes for a patient to receive a response but also increased the burden on care teams.

Our platform analyzes both the patient’s message and patient’s clinical information pulled from any major electronic health record (EHR) vendor in real-time. Utilizing industry-leading deep learning algorithms that incorporate millions of clinical data points, the platform sorts the messages into the broad topics and identifies the chief complaint utilizing natural language processing. In conjunction with the chief complaint and current problem list, we synthesize an automated note detailing pertinent problems mapped to relevant risk factors, past medical history, family history, laboratory trends, imaging, and previous assessments and plans pulled from the patient’s record. By automatically integrating patient clinical data into care team workflows when replying back to patients, the note serves as documentation for clinical information reviewed and complaints addressed.
 
We have already partnered with hospitals across the US to develop this solution with the aim of reducing burden on their care teams and automating manual documentation/sorting of the increasing number of patient portal messages.

Patient reported outcomes (PROs) are personal health information collected directly from the individual for assessment without interpretation by a clinician. Patient reported outcome measures (PROMs) are typically electronic based patient questionnaires tools developed to measure the patient’s reported health status. At our health system are developed to be integrated into clinical documentation to safeguard hospital based clinical standards, promote more meaningful clinical-patient interactions, and minimize clinical documentation burden.

The modern landscape of healthcare shifts patient care from the clinical setting to the home environment which necessitates creative solutions for effortless and intuitive integration of PROMs into clinical documentation. At our health system, PROM tools are integrated into our Connected Care Monitoring (CCM) Program, a remote monitoring initiative that aids in minimizing disruptions in care as patients navigate from the inpatient, home, and outpatient environments. The surgical arm of CCM is the Recovery Tracker, a post discharge 10-day electronic symptom reporting tool received by post-surgical patients through their portal messaging system (PSM). Patients discharged with electronic home devices can upload device information through the iHealth APP and the health data is integrated into the patient questionnaire. As these daily questionnaires are submitted, PSM alerts are generated when a predetermined threshold is met, notifying the clinician of the potential need for intervention. Alerts are categorized to yellow and red depending on severity and thresholds indicated allowing for easy clinical triage. All clinician-patient PSM communications and patient reported information is imported directly into the clinical documentation form. As technology continues to remove barriers to communication it will also increase the documentation burden of clinicians. Clinical documentation should be clear, concise, and intuitive. It should prepopulate with essential patient information and auto-document clinical interventions including prepopulated responses and ad hoc follow-up questionnaires to reassess interventions.

Current documentation paradigms are modeled after paper charts & encounter based care.  Though care is provided on a continuum and by teams instead of by individuals, daily progress notes on the inpatient side & per visit notes in ambulatory, require clinicians to duplicate entire sections of their patient care plans for each encounter, often by multiple team members. It is not surprising that research shows about 50.1% of EHR text is directly duplicated from a prior note (pending publication in JAMA Network Open, Steinkamp, J. Et al).  Additionally, current EHR models do not have the concept of a patient-centric to do list.  I.e. what needs to get done for the patient, why, by whom, and what is the status?  Instead, most systems rely on messaging, sticky notes, email and texting to keep track of who is doing what.  This is inherently duplicative and disorganized, leading to inefficiencies, communication failures, and important information being lost in these other communication methods, never making it into the patient’s record.

While care delivery models have changed, our documentation structure has not. We propose a different type of written communication and documentation paradigm that supports team-based care, iterative care planning, and cross care setting collaboration, while still meeting documentation requirements of encounter based charting.

What does that look like?

Building off the concept of the problem oriented medical record (PROM), a large quaternary academic center created a team collaboration platform with a wiki-based care planning tool.  Organized around a patient’s problems or diagnoses, the wiki contains semi-structured elements including problem names, plans (assessment of the problem, differential, updates, relevant data), action items to carry out that plan, as well as a patient-level one-liner summary statement.  The semi-structured elements enable clinicians to flag any item as being important anticipatory guidance for all care team members to  be aware of, or being related to transitions in care and more.   Tasks can be identified as being primary team tasks, on call tasks, they can be dated, timed, assigned etc.  This enables the wiki to be sorted and viewed based on different workflows such as TCM visits, post-discharge follow ups, or on the inpatient side rounding, handoff and more. Typically these different workflows are supposed by paper lists, sticky notes, in basket messages, email and more.  Instead of relying on these piecemeal modes of organization, this wiki model incorporates the plan and tasks within the care plan, while still allowing clinical teams and staff to view “just the tasks” of what needs to get done, without having to keep track of these separately.

Think of this as a living, breathing, wiki-plan of care... or progress note that anyone can see, anyone can edit, anyone can contribute to as a working document. At any time, the plan can be reviewed, captured and saved in the EHR as an encounter note (for inpatient OR outpatient). Meanwhile, the wiki captures work done between outpatient encounters and between inpatient daily notes, so that when the next note is written, it already incorporates all the change sin the plan since the last note, instead of requiring the clinician to start from an already outdated plan and have tore-write all the updates that took place since then - pulling form email, in-basket, text, paper and more.  At any time, once the plan is updated, a snapshot of it can be brought into the EHR progress note, leveraging already existing progress note templates, and eliminating double documentation.

This wiki-style documentation workflow has been in use on the inpatient side at a large quaternary academic center, across specialties, disciplines and facilities.  It is now also used in post-acute care settings, home care and value based care, with the same results of time saved, improved team collaboration, less errors and improved diagnosis capture. The application used for these sites is device agnostic, EHR agnostic, and care setting agnostic. It is intuitive and easily learned: most clinicians report feeling comfortable using it within 1 shift after watching a brief 10m video. Pilot implementation of this work has been published in ACI.