Just went to the conference, Big Data and Analytics in Healthcare Summit 2018, and listened to the CEO of a company called Health Wizz speak. They are using blockchain so that patients canuse a mobile application platform that helps individuals aggregate their health records from multiple sources including Wearables, Electronic Health Records (EHR) Systems and their genome. daks2k3a4ib2z.cloudfront.net/5...

Thirty-five years ago, I envisioned a comprehensive health IT system that is flexible enough to continually adapt to new healthcare knowledge and concepts; data models; value sets; data format, terminology and transport standards; use cases; workflows; and diverse user needs. I imagined it would accommodate the needs of all patients and their extended care teams with highly useful and useable tools.

As the notion evolved over the years and new technologies emerged, I wanted the system to leverage A.I., APIs, cloud-based storage and computing, as well as local computer resources and data stores. It would interoperate with third-party tools that provide additional relevant data from biometric devices, EMR/EHRs, and all other data sources.

It would then provide clinical decision support that continually improves through knowledge-feedback loops among diverse groups of collaborators. These collaborative networks would build, share and refine software models aimed at continually increasing the value of care patients receive through systematic process and outcomes research that cross organizational and geopolitical boundaries. The models would focus on the whole person (biomedical and psychological), SDH, precision medicine, clinical workflows, population health, care coordination, preventive care, treatment and self-maintenance of chronic conditions, and financial matters.

Toward this end, we've developed an innovative software framework that enables apps to be built with the capabilities discussed above. It also utilizes an evolving biopsychosocial ontology that can integrate biomedical/physiological, psychological, and sociocultural data. The apps use common information technologies in novel ways. We're now preparing to expand our network of collaborators and approach investors.

These solutions seem like where we want to go, but I think it will be a long time before we get there. We are all drowning in data, but the trouble is trying to make sense of it. We can collect thousands of data points, but someone has to interpret it. Right now, we don't have validated tools to tell you how the specific foods you eat or the exercises you do will translate into outcomes. In fact, our diet advice is pretty spare and our exercise advice is even less specific. We know it is good to exercise, but we don't know how often, for how long, how intensively or whether these should vary depending on other factors, such as your current health or genetics. Other industries have a big advantage in this regard. Amazon can tell with certainty if you bought something (or just clicked on it). They know an awful lot about our buying habits, but we much less about the state of our health. Without objective outcome measures, we will continue to just make cool graphs.

We need good outcomes research using valid, objective biomedical data, solid psychometrics, and accurate economic data. Part of that research should focus on determining the variety and quantity of data required to measure outcomes reliably. Too much data wastes time and resources. Too little data and you can't have confidence that the analytics will accurately inform clinical decisions. In addition, the analytics themselves should continually evolve along with the data sets in an ongoing knowledge feedback-loop process.

There is no quick fix; doing it right will take time and effort by collaborating parties. As such, we must take a long-term view (a short-term perspective is futile). And we must start somewhere despite the uncertainty and risk.

If we are ever to be successful, we ought to use flexible tools that leverage both old and new technologies and are able to handle daunting complexities, or we're just fooling ourselves. It will take smart, dedicated, cross-disciplinary networks of people to make it happen. I envision this process that continuing to build, evaluate, and evolve vital health-related knowledge for as long as our species exists.

(Repost from Below, given this question is more applicable) Now that we live at the interface of "Big data"--namely having the technology to start to capture and stream more "objective" data sets from a sea of cell phones, sensors, computer vision, IoT, etc--What is everyone's opinion on supplementing subjective feedback from the patient with as much quantifiable data as possible? Or is there a deeper question on having a complete grasp on these various data sets to make sure we are separating "insight" from "dumb data"?

My opinion has always been that with the right technology, moving from mostly subjective data to more quantifiable data is "better" BUT just because we have the data, it doesn't necessarily follow that we have "insight" on it--namely the "where", "when", "why", "how" of the data to generate context and actual meaning from the results. This is at the crux of the weakness of data for the sake of data and now the new trend is forcing A.I. upon it to try and connect the dots to meaning (whole different conversation, but A.I. is only as good as the data+ context used to train it!).

And this problem is compounded if we are asking the wrong questions of patient wellness data, and then using the wrong data sets to answer said wrong questions...

Given that I live in the world of biomechanical analysis, that's where I've seen that objective data + insight delivered by the clinician not only fosters trust and compliance, but also affects the perception of care. (My example is in the next post). But I've also seen "Big Data" overapplied beyond its context. The key to this whole conversation is that the technology we use has to provide the appropriate "insight" to deliver meaningful action toward wellness or preventative health--which obviously answers the following questions: Is the data accurate? Is it repeatable? Is it reproducible? Does it answer the "when, where, why, and how" of whatever my application may be to bolster decision making?

Here is an example of how I frequently see objective data + insight incur a very real change in perception of care:

Imagine a patient (I'll call her Susan) who was scheduled for a hip arthroplasty for degenerative OA, and she undergoes biomechanical evaluation pre-and-post surgery. We were able to use a very accurate and quick computer vision technology to quantify her biomechanics in all planes--The key here being that the technology we chose to employ could calculate all kinematics and kinetics accurately and reproducibly in all planes, to ensure we were comparing "signal-to-signal" instead of "signal-to-noise". Since we had a "baseline" pre-surgery, we could quantify progress post-surgery down to the degree in all planes, and keep tabs on the degradation of non-affected joints/biomechanics as a response to the change incurred by the arthroplasty.

But what was most interesting was what this data did to Susan's perception of care--the conversation went something like this after surgery and after her post-scan:

"Well Susan, how do you think you are doing after surgery? How does your hip feel?"

--"...I don't know if I'm doing very well, it just feels 'funny' to me"

"It may 'feel' funny, but look at how much you've improved--range of motion/balance/control at your hip have improved by xx%, squat depth, mobility, etc have improved by x degrees/inches (insert our explanation of all the quantified improvements, etc)"

Susan listened. Asked questions. Really engaged with the data. Then started to smile when she said "I guess I AM doing pretty well after all!"

No funny business, no subjective scale, she was given a data set that showed real improvement pre-vs-post surgery, it was explained to her in simple terms, and it changed her perspective on the outcome altogether. Now imagine this same objective approach in proactive wellness--where data drives engagement and compliance to get ahead of small problems BEFORE pain, dysfunction, or limitation...

I would take issues with Joel’s suggestion that moving from mostly subjective data to more quantifiable data is “better”. I recognize I may be taking this out of context but want to emphasize that much of the Patient Reported Outcomes data that some may believe is subjective is in fact quantifiable, especially when the appropriate scientific rigor has been used to develop the questionnaire. Take the case of Susan who seems unsure of how she is feeling after hip surgery and is offered a series of measure of motion, balance etc. What if Susan has completed a “subjective” Patient Reported Outcome about functioning prior to her surgery and then again after her surgery. The range of motion may not be as important to Susan as describing her function in a way that is meaningful to her such as how much pain she has with walking, how much pain she has when sleeping at night, can she get on her bicycle, how many stairs can she climb/descend without pain. Sharing her before and after scores on such a “subjective” yet scoring functional measure may well be more meaningful to her than more traditional biometric data.

Likely what we need is a blend of both, traditional biometric data and patient reported outcomes. And it may depend on the health condition that we are helping patients to manage. For instance, managing blood pressure (that typically has few to no symptoms) may always relay on traditional biometric measures. I certainly agree that communicating results to patients in simple terms is critical. Patient reported outcomes is something that patients can use to self-report their physical or mental health and thus empower them to track their own progress over time. Participants may be interested in an example of this from Sweden that emphasizes the combination of traditional biometric measures with patient reported outcomes: youtube.com/watch?time_continu...

Joel make good points. Nevertheless, there are many reasons why I agree with Carolyn that both subjective and objective data are both important. Here's an example: There's a good deal of evidence that psychological stress and depression are associated with poorer biomedical outcomes and care plan compliance, so understanding the whole person (mind-body connection) provides valuable insights about wellness strategies to recommend and referrals to make. Additionally, patient-reported symptoms may be the only data there are, e.g., the existence of certain medication side-effects. Furthermore, the field of quantitative psychology enables subjective data to be quantified, which helps close the gap between objective data from devices and subjective data about one's thoughts and emotions. en.wikipedia.org/wiki/Quantita...

I agree with both Stephen and Carolyn that there will never be an instance where subjective data is not useful--moreover, the goal in my mind would be to capture objective data that directly connects to the things that matter to the patient with regard to a return-to-work/play/function/life and blend that with a comprehensive understanding of the psychological, social, and emotional circumstances of the patient. Likewise, to clarify, by "better" I meant to convey that any available data sets that can be more objective allow us to move toward more reproducible, repeatable, care decisions. But ultimately marrying the objective data sets available with a comprehensive understanding of patient reported data is where the approach comes to life and creates meaningful experiences for the patient.

Another example:
In the biomechanical space, we've witnessed stroke patients be moved to tears when the analysis of simple movements (that directly relate to return to ambulation, activities of daily living, etc) show improvements by a handful of degrees from session to session and when the clinician/therapist explains how these improvements directly correspond to their rehab--which brings me back to the insight point: If the objective data cannot be related to some meaningful goal or context for the patient to help quantify progress or regression toward/away from the goals that matter to them, then that is "dumb data" not insight, and ends up being useless to the patient experience. Connecting the dots on what that data means for the stroke rehab patient's progress, for instance, now delivers meaningful insight with purpose that acts as a force multiplier for the patient and clinician to work together to meet the goals that matter.

Good question...here's my take:

There are many possible reasons why a patient would have a low level of trust, which can carry over to, and adversely affect, the patient-provider relationship. A previously bad experience during an encounter with a provider sometime in the past, or a bad experience with a family member, etc. Cultural and ethnic issues/biases. Concern that PHI will not be adequately protected. Distrust of conventional medicine or the healthcare system in general (e.g., doctors rushed, underpaid, burnt out). Problems communicating. Body language interpretation. Horror stories (even if untrue). A troubled family situation early in life resulting in generalized distrust and difficulty in relationship. Irrational beliefs causing a patient to doubt a provider's motives or competence. Even paranoia due to a serious mental health condition. And so on.

Regardless of the reason, the patient-provider relationship would likely be strengthened if the provider knows that a patient is mistrusting, has an idea of some of the patient's thoughts and feelings in that regard, and has guidelines that recommend ways to interact with that patient to minimize the distrust and build confidence.

In the VA, provider engagement is at the top of most Veteran's concerns. I've seen hundreds of patients upset with having to see a Nurse Practitioner instead of a doctor. Nearly all situations had to do with feelings of confusion and abandonment more than who they see.

Communication builds trust. Coaches talk to boxers as they box in the ring. This distracts the boxer from negativity and keeps them engaged in problem solving instead of giving up. VA "ANNIE App" text messaging is having the same kind of positive impact on patients. ANNIE texts the Veterans and requests the patient enter data specific to each protocol the Veterans are assigned. Also, Annie occasionally sends patients general info for each protocol (weight loss, diabetes, COPD.. ext). This simple solution increased interactions between Veteran and providers and reduced many of the patient concerns Dr. Beller listed above.

How will a life long construction worker or truck driver describe their health problems to a highly educated clinician? One environment wants to hear about your problems and the other spent years telling you not to whine. Rating pain from 1-10 exaggerates this communication gap. A trained athlete might rate a broken arm at 1 but still feels as much pain as the person rating a 10. One doctor might give you pills and a different one might not. Hard to feel comfortable speaking to clinicians when so much is on the line.

Frequently entering data and engaging electronically with healthcare team reduces the stress and uncertainty.

With established patients who have seen me for years, this is not a problem because they have a level of trust with me. However, in today's environment with handoffs, patients often see new providers (not necessarily physicians) who ask these questions. We need to figure out how to explain to the patient that this is to help their overall health, etc. That will remain a challenge.

I think patients want their doctors to be interested in their lifestyle and its impact on their health. Most patients don't want to take medicine, so if we can counsel them on other changes they can make, it builds trust. However, one of the dangers of routinely collecting this information is that patients then expect their doctors to review it. If we ask them to enter the data and then ignore it, they wonder what else we might be ignoring.

Interesting topic--now that we live at the interface of "Big data"--namely having the technology to start to capture and stream more "objective" data sets from a sea of cell phones, sensors, computer vision, IoT, etc--What is everyone's opinion on supplementing subjective feedback from the patient with as much quantifiable data as possible? Or is there a deeper question on having a complete grasp on these various data sets to make sure we are separating "insight" from "dumb data"?

My opinion has always been that with the right technology, moving from mostly subjective data to more quantifiable data is "better" BUT just because we have the data, it doesn't necessarily follow that we have "insight" on it--namely the "where", "when", "why", "how" of the data to generate context and actual meaning from the results. This is at the crux of the weakness of data for the sake of data and now the new trend is forcing A.I. upon it to try and connect the dots to meaning (whole different conversation, but A.I. is only as good as the data+ context used to train it!).

And this problem is compounded if we are asking the wrong questions of patient wellness data, and then using the wrong data sets to answer said wrong questions...

Given that I live in the world of biomechanical analysis, that's where I've seen that objective data + insight delivered by the clinician not only fosters trust and compliance, but also affects the perception of care. (My example is in the next post). But I've also seen "Big Data" overapplied beyond its context. The key to this whole conversation is that the technology we use has to provide the appropriate "insight" to deliver meaningful action toward wellness or preventative health--which obviously answers the following questions: Is the data accurate? Is it repeatable? Is it reproducible? Does it answer the "when, where, why, and how" of whatever my application may be to bolster decision making?

Here is my example of how I frequently see objective data + insight incur a very real change in perception of care:

Imagine a patient (I'll call her Susan) who was scheduled for a hip arthroplasty for degenerative OA, and she undergoes biomechanical evaluation pre-and-post surgery. We were able to use a very accurate and quick computer vision technology to quantify her biomechanics in all planes--The key here being that the technology we chose to employ could calculate all kinematics and kinetics accurately and reproducibly in all planes, to ensure we were comparing "signal-to-signal" instead of "signal-to-noise". Since we had a "baseline" pre-surgery, we could quantify progress post-surgery down to the degree in all planes, and keep tabs on the degradation of non-affected joints/biomechanics as a response to the change incurred by the arthroplasty.

But what was most interesting was what this data did to Susan's perception of care--the conversation went something like this after surgery and after her post-scan:

"Well Susan, how do you think you are doing after surgery? How does your hip feel?"

--"...I don't know if I'm doing very well, it just feels 'funny' to me"

"It may 'feel' funny, but look at how much you've improved--range of motion/balance/control at your hip have improved by xx%, squat depth, mobility, etc have improved by x degrees/inches (insert our explanation of all the quantified improvements, etc)"

Susan listened. Asked questions. Really engaged with the data. Then started to smile when she said "I guess I AM doing pretty well after all!"

No funny business, no subjective scale, she was given a data set that showed real improvement pre-vs-post surgery, it was explained to her in simple terms, and it changed her perspective on the outcome altogether. Now imagine this same objective approach in proactive wellness--where data drives engagement and compliance to get ahead of small problems BEFORE pain, dysfunction, or limitation...

Our clinicians assign their patients self-care tasks. Then the patient does them or doesn't do them or does some of them, which the patient documents, as well as how they are feeling about them and questions/comments they have. Those PGHD results are reviewed by the clinician, typically with the patient, and they make their way into the EMR. Forty percent of our patients report a improvement in care-giver relationship as a result.

The auto industry went through this data transition 20 years ago. Cars are filled with sensors continuously monitoring engine performance.

Like a car, it's the outliers that need attention. Artificial intelligence can monitor patient entered data in a control chart with clinician set limits. Escalating only the outliers for further review.

Hi Joel, I really like your product. Can patients track their objective progress in real time? Any clinical studies yet on this?

Brief summary follows:
Jonathon: Communication builds trust, but patient may have difficultly speaking clinicians. Electronically communication with care team reduces the stress and uncertainty. A.I. can monitor patient entered data in a control chart with clinician set limits, whereby only the outliers received further review.
Sheila: With today's handoffs, patients often see new providers so, to maintain trust, the challenge is how best to explain to the patient that this is to help their overall health.
Michael: While patients want their doctors to be interested in their lifestyle and health impact, it's useful to know when patients want alternatives to common recommendations, so they can be counseled to make other changes, which builds trust. Routinely collecting these data, however, comes with the expectation that providers will not ignore it.
Joel: Quantifiable objective data provides insight by generating context that enables AI to the data meaning that informs action toward proactive wellness. Such objective data drives engagement and compliance to get ahead of small problems before serious symptoms occur.
John: When clinicians assign their patients self-care tasks, the patient would document their (lack of) compliance, attitude about the tasks, and questions/comments they have, which are reviewed by the clinician, typically with the patient. This process strengthens their relationship.

So, we all agree that PGHD is useful in important in building trust and engagement. Objective data (in context analyzed via A.I.) and subjective data can both beneficial depending on how they are used. Beware of the unintended consequence of collecting the data, but not reviewing/using them.

Questions remain about capturing PGSD: How much, what type, how often, from where, when (under what circumstances) and why (i.e., how best to analyze and use them). This appears to be an empirical question research can answer.

Assuming that behavior change would be beneficial to a patient, then the following types of patient-entered data could be helpful:

• Readiness to change since this provides context or framework for which change will be made. These data include the patient's current stage of change (pre-contemplation, contemplation, planning, action, maintenance) and character traits that foster effective change, e.g., resourcefulness, optimism, adventurousness, conscientiousness, passion/drive, adaptability, confidence, tolerance for uncertainty/ambiguity.
• Behaviors to be changed, which should be prioritized to sharpen focus on what needs to be done.
• Previous attempts to change, their results, and if failed/relapsed, why it happened. This history describes what has worked and what hasn't, which is useful knowledge to consider moving forward.
• Method(s)/process(es) for making the changes. This creates a change plan/strategy.
• Knowledge and skills required for the change. This points to required patient education/training.
• Blocks and facilitators of change. The facilitating factors can be used to help overcome the blocks. This includes patients' beliefs/attitudes and their related emotions, behavioral tendencies and coping mechanisms, and SDH.
• Medical conditions, treatments and/or SDH that may be contraindicate the change methods or may contribute adverse unintended consequences.

I doubt that many providers would need all the fine details about the data types above. A brief summary that points to a personalized change strategy and related consideration, however, would likely be helpful.

As a podiatrist, understanding patient preferences where patients have the ability to state there preferences has been an important component in my approach for caring for patients. For example, there are times when patients with vascular disease opt for an amputation rather than limb salvage because the amputation aligns better with their overall goals.
There is a great article by Patricia Brennan (Dir. Nat'l Library of Medicine), et al Improving Health Care by Understanding Patient Preferences: The Role of Computer Technology.
This short abstract succinctly states the value of including patient preferences in making care decisions:
"If nurses, physicians, and health care planners knew more about patients' health-related preferences, care would most likely be cheaper, more effective, and closer to the individuals' desires."

Previous research suggests that patient-entered wellness data can increase preventative care utilization and empower patients to take charge of self-management (Foucher-Urcuyo et al., 2017; Ho & Antonucci, 2017). However, there is also evidence that patient-entered data can have a downside. For example, pain diaries may inhibit recovery from injuries (Ferrari & Louw, 2013; Ferrari, 2015). Other potential pitfalls come from inaccuracies related to data collection devices. To illustrate, recent research has documented that wrist-worn devices such as fitness trackers overestimate the number of calories burned by at least 20% (Shcherbina et al., 2017). Likewise, in a randomized clinical trial, the addition of wearable technology resulted in less weight loss compared with standard behavioral intervention (Jakicic et al., 2016) . With the rapid of development of new devices and software that enable patients to track their health behaviors and lifestyle choices, it is important to keep in mind the potential darkside of patient-entered (or collected) data. Providers should be mindful of the type of information being collected and how it is collected to help patients navigate the changing technological landscape.

Hi Kenzie. What do you think is the reason for the poor outcomes arising from devices? My feeling is that they don't address the underlying problem. Is there more at play here? And if not devices, what other types of data would be more beneficial for instigating behavioural change?

Hi Jonny,
That’s a great question. In general, I think we need more research to understand why wearable and other fitness devices can lead to poorer outcomes. I don’t think we understand how people are using these devices to inform their decisions. People may be putting too much trust in inaccurate data—often eating more calories than they would without the technology. On the other hand, the tracking of health behavior data may lead to a “reward” mentality where individuals reward behaviors such as hitting their 10,000 step goal by consuming more calories (e.g., having dessert or a second helping). Without the knowledge of that goal, individuals may not be inclined to treat themselves. Likewise, tracking specific health behavior data may contribute to individuals overly focusing on one aspect of health and wellbeing—instead of seeing the larger picture.

Other research has suggest that motivational technology can undermine a person’s autonomy by moving internal motivation to external regulation, which makes long-term change less sustainable (Austin & Kwapisz, 2016).

Additionally, I think you’re right about not addressing the underlying problem. A fitness device cannot address issues related to environmental barriers (e.g., unsafe neighborhoods preventing physical activity, long and stressful work hours, or lack of access to fresh fruits and vegetables) that make sustained changes in health behaviors difficult. I’m not sure what kinds of data would be more beneficial for initiating (and sustaining) behavioral change. It’s quite possible that this is not a one size fits all problem. We may need to understand the barriers first before recommending fitness or wellness technology. Some people may benefit from these devices, while for others it may exacerbate existing concerns.

Austin, C. G., & Kwapisz, A. (2017). The road to unintended consequences is paved with motivational apps. Journal of Consumer Affairs, 51(2), 463-477.

In support of Kenzie's hypothesis that wearable device benefits might not be a one size fits all problem, I'd like to propose a strong link between certain personality characteristics (character traits) and benefits from using such devices. Here are three influential traits:
1. Conscientiousness -- Being organized and goal-oriented, driven to do the right thing in the right way.
2. Perseverance -- Not giving up prematurely, able to tolerate failure and frustration without giving up, long-term goal oriented
3. Self-motivation – Having an internal locus of control, self-confidence, self-satisfaction from the effort even if not immediately successful.

A person who lacks characteristics such as these will unlikely gain much benefit from devices, or from any other simple intervention/technique for that matter. It is therefore essential that character traits such as these be assessed and addressed before we can expect fitness/wellness technologies to assist in sustaining positive outcomes for many/most patients. This means using a personalized wellness approach: Match behavior change methods to key patient personality traits (and other influential factors) using appropriate assessment data and outcomes research.

For so many chronic diseases, lifestyle is a major contributor. Every day we see patients who have obesity, diabetes, hypertension, and cardiovascular disease (often in the same patient). Our guidelines all recommend that we start with "lifestyle" interventions, but most physicians skip over those, because they don't feel like they can really affect what their patients will do once they leave the office. Learning about a patient's lifestyle through data they have entered can save physicians time in terms of history collection, but it can also be paired with patient-facing materials that offer counseling and referrals. It also can provide a way to track patients' progress toward goals of healthier eating, more exercise, less stress or better sleep. Right now there is no flow sheet for that in the EMR. We can plot patients' weight, BP and A1C, but we can't plot their diet and sleep alongside in order to show them the correlation.

In terms of decision-making, there may be limited application at present, because we have so few tools to address these issues. As we build out our capabilities with e-coaches and other lifestyle interventions, we can use patient-entered wellness data to intervene with these areas before they become problematic. In particular, we may identify patients with unhealthy lifestyles who are not yet obese, whose blood sugar is normal, and change the trajectory of their health years before it becomes obvious to everyone in the office!

Following up on Michael's good points, I contend that the ability of patient-entered wellness data to assist providers with clinical decision support depends, in part, on what data are collected. Our approach focuses on collecting comprehensive biopsychosocial data. The variety and quantity of the psychological and social data captured, integrated, and analyzed go well beyond other approaches.

If the collected data and analytic models are adequate, then there are many ways the data can be transformed into useful information that builds clinical knowledge. This knowledge can be reflected in clinical guidelines that would recommend what providers can do, for example, to:

• Strengthen the patient-provider relationship by informing providers how best to interact with their patients based on each patient's personality (psychological character traits), beliefs systems, interpersonal relationships, SDH, etc. The chance of patient engagement and positive behavior change increases when the patient-provider connection is a good one.
• Promote positive behavior change and self-maintenance by considering issues in patients' lives that affect their willingness and/or ability follow through with providers' recommendations.
• Engage in shared decision making with their patients.
• Use their awareness of cultural influences on their patients to adjust recommendations.
• Determine appropriate referral sources (e.g., to wellness coaches) based on influential factors.

At Dartmouth Hitchcock Medical Center we have built patient self-report questionnaires into our EMR (Epic) for over 40 different health conditions. In primary care, we have available questionnaires to support annual physical exams and annual wellness visits. These questionnaires include screening for depression, falls risk, activities of daily living, independent activities of daily living, generic measure of overall health (PROMIS-10), exercise habits, eating habits and for women, a screening for breast cancer risk.

Also for pediatrics and obstetrics/gynecology we have implemented screening questionnaires covering similar domains as for primary care but adapted to the specific population.

We have been pilot testing, in primary care, more detailed screening for mental health (EtOH, substance use, depression and anxiety) and for social determinants of health (8 different domains). We are building momentum to scale this across all primary care adult and pediatric practices in our health system.

Patients can complete their questionnaires at the time of the visit in the waiting area on a tablet or if they have signed up for the patient portal they can visit that on the web or using an app on their smart device and complete the questionnaire asynchronous with their visit. Once a patient has completed a questionnaire, the information/results are immediately available for the care team and tools have been designed to make this less work for the care team to reliably and systematically gather and document findings. At a glance, care team members get an overall picture of what is going well and what needs further action. This in turn allows the patient and care team to spend more time of the visit focused on issues that are most important to the patient.

What we would like to see developed in the EMR platform is more patient facing capability and display of results with a user-friendly interface, from the patient perspective.

As we described in our article, we piloted an intervention to collect patient-entered wellness data, including exercise, diet, stress, sleep and other things that doctors often don't have time to collect. We found that physicians were most interested in those things that they knew how to influence. For example, they sent more patients who screened likely to have OSA for sleep tests. They were least likely to take action on diet. Many felt ill-equipped to change their patients eating habits.

The articles are excellent!

About our use case: Our tool, the Personal Health Profiler (PHPro), was developed based on the conception that effective prevention and positive high-value outcomes are most likely when each patient is understood from a sound biopsychosocial (physiological, psychological and social) perspective using patient-entered data.

PHPro considers patients' physical health conditions, risks and SDH as do other tools. What makes it different, however, is the deep level of understanding it provides about the psychological factors that influence patients' health status and behaviors. These psychological factors include patients' attitudes, emotions, behavioral tendencies, social/family relationships, personality traits, coping strategies, change-readiness, self-management capabilities, and more.

I've used PHPro successfully for years in my psychology practice to facilitate shared decision-making, support treatment planning and delivery, and evaluate clinical outcomes and patient satisfaction. It helped me understand how my patients' thought processes, emotional reactions, social relations, built environments, and personal experiences all come together to influence their behaviors. This understanding, in turn, provided insights about how to: (a) interact with my patients for strong patient-provider relationships, (b) determine their readiness to change, (c) be aware of their emotional blocks, (d) identify their knowledge and skill set deficiencies, (e) recognize their maladaptive (irrational/faulty) beliefs, and (f) use their psychological strengths and adaptive abilities to promote positive change.

We've been working to tailor our tool for providers who are not mental health specialists. Our goal is to enable them, in a practical and usable way, to understand the psychological factors affecting their patients' and to know what they can do to promote effective preventive care and to make the right referrals.