Translating Your Medical Tests into Results Team ... - cmsc434-s13

Learn to Speak Doctor: Translating Your Medical Tests into Results Team Members: Joseph Chan, Jeremy Ozymandias Fallick, Emily Aster Jones, Ahmad Rana Section 1: Abstract Medical results often contain large amounts of medical jargon that most people are unable to understand, which unfortunately means that when people read over their results themselves, they have no idea how to interpret them. Previous solutions relied on the doctor interpreting the results for the patient and telling them how to fix whatever the patient suffered from. However, this relies entirely on the advice prescribed by the doctor rather than allowing the patient to take control of their own health, since test results are not put into context or translated into actions. Our solution translates medical results into layman's terms, allows patients to understand the risks of the illnesses they may have, and provides a step-by-step guide for the actions that the person should take to improve their health. We would begin with a lo-fi prototype primarily based on blood tests, but we will also interview physicians to confirm risk analysis and behaviors taken from academic sources and test the prototype on patients. Patients would benefit from better understanding of and engagement in their health, and doctors would benefit from reduced noncompliance. Section 2: Introduction Results from medical tests, whether good or bad, rarely lead to behavioral changes in patients for the better. This is because tests results are difficult to interpret unless you are a physician. However, if patients could interpret the results of their tests, they would have personalized data on their progress, which is a key component in motivating the patient to improve his or her health. In addition, if these results were put in context, comparing them to normal results and providing suggestions for how to improve them, this would translate the results into action (Goetz, 2011). Currently, test results are filled with medical jargon, do not translate results into context, and also do not translate results into actions (NHLBI, 2012). This project is inspired by a medical test redesign by Wired magazine, who asked graphic designers to design what a blood test might look like while consulting physicians for context like risks and suggested behaviors (Leckart, 2010). However, they did not provide an interface to do this yourself, and did not allow you to interact with the data by adding behaviors or moving test values to see the outcome; this is where our app comes in. We help users recognize the risks associated with their illnesses by phrasing their results in an easily understood manner, and we provide a guide for mitigating or avoiding these risks in order to improve their overall health. We determine our app's success by checking usage frequency and number of downloads. A high frequency of use shows that people are utilizing our app, and a high number of downloads means that many people are trying it out. We considered other means of evaluating our app, but they were either inaccurate or impractical. One of these methods involved tracking the patient's health over time, and seeing if he improved. However, there are diseases that may not improve with time or the patient's methods might be ineffective, so this is not an accurate way to judge success of our app. Furthermore, health information should remain confidential, so we will not collect this information for any reason. Another method in consideration was to survey the users to ascertain how useful they thought the app was. Unfortunately, surveying every user is not very practical, especially since this is an app that is meant to be used long-term. This would mean that we would have to survey the users multiple times and many people dislike taking surveys, a fact which would skew our results.

Section 3: Background / Review of Past Work Medical tests have long been an important way for physicians to diagnose conditions, monitor progress, or screen for diseases. According to the most recent status report on laboratory medicine by the CDC, US clinical labs conducted 6.8 billion laboratory tests in 2007 (CDC, 2007). Our project will focus on routine blood tests conducted by the primary care physician, whose primary purpose is to check for risk factors such as hyperglycemia or hypercholesterolemia. The physician can then use these predictors to diagnose the patient's risk of developing conditions such as diabetes or cardiovascular disease. The full blood test is the most common lab test conducted in the United States, which physicians can recommend to be administered once a year in a healthy individual or more frequently in patients with diabetes or other disorders (NHLBI, 2012). The blood test itself provides annotations of whether the values measured are high, low, normal, or critical. However, this information does not indicate what an abnormal value means, nor does it take the other test results in context. For example, a healthy young adult might have an HDL of 39, an LDL of 58, and a total cholesterol of 114. The patient's HDL score would be marked as unhealthy (a good HDL score is 60+), even though a healthy HDL score would be unlikely given that the patient's total cholesterol is far below the normal range (140-200). In order to understand what low or high annotations mean, a patient could consult their physician on each measure. This system is not meant to replace a physician's interpretation of test results. However, given the complexity of blood tests and how often they are administered, it is often difficult for physicians with hundreds of patients to follow-up with each of them to interpret what each measure means and what actions the patient needs to take. Alternatively, a patient could search online for what each value means, a tedious process which again does not take the values in context of each other and does not suggest behaviors. Despite the pervasiveness of clinical tests, these personalized results often do not lead to the patient having a better understanding of their health or taking action to improve their condition. First, waiting for a physician's interpretation can lead to unnecessary stress for the patient. A recent survey of patients found that more than 80 percent said they would rather have online access to such results within three days, no matter how dense the jargon, than wait a week or longer for a doctor's interpretation (Johnson et al, 2012). Second, when medical test results are translated into options for patients to take to improve their health, this opens a dialog between the patient and the physician about the patient's health, helping the patient take responsibility for their own health. Several studies have shown that patients who are informed about their personalized results and their options have better outcomes, and that patients who understand their health and options are more involved in decision-making with their doctors, leading to better compliance with physician instructions (Saxton and Finkelstein, 2011). Our aim is to inform the patient of chronic conditions that can be improved over time, not acute conditions that must be treated immediately. Critical results that require prompt treatment are required by law to be reported to the attending physician immediately. Although this system does not always work, it results in treatment within hours when it is followed correctly (PA Patient Safety Authority, 2009). These patients would be in hospitals for serious conditions, however, and thus would not benefit from our system, which would give interpretations in the context of chronic conditions only. In particular, we are targeting early diagnosis and treatment of heart disease. Heart disease is the greatest cause of death in the United States, killing 600,000 people each year. 26.5 million Americans, or 11.5% of the population, have diagnosed heart disease (CDC, 2013). The best treatment for heart disease is prevention or early diagnosis and treatment. Patients who are informed of their risk factors and encouraged to adopt healthy behaviors like exercise early on are less likely to develop heart disease than those who are not.

Currently, physicians must release medical records to patients within 30 days of their request. However, they can require that the patient first pay for a consultation visit so that they can receive the results with the doctor's interpretation. In 2011, the Department of Health and Human Services announced a proposed rule to require all test results to be directly reported to patients within 30 days, preventing physicians from requiring this consultation. Some doctors are opposed to this rule, concerned that it would lead to patients misinterpreting their results. Their main concerns are information overload from too many numbers out of context and lack of knowledge of the limitations of the procedure, including the false positive rate or the need for follow-up tests. Others have a more positive outlook, suggesting that this will encourage patients to initiate their own dialogue about health and develop a closer relationship with their physicians (Lewis, 2011). While our tool would not take the place of a doctor, it would put the numbers on their blood test in context of healthy and unhealthy ranges, inform them of what each metric means, suggest healthy behaviors, and give disclaimers about when test results can be invalid and what results require follow-up with their doctor. Currently, most health tools on the market aim at recording medical information or logging healthy behaviors. For instance, you can record your family's prescriptions with the Family Drug Guide app (Unbound Medicine, n.d.), or log your workouts and calorie consumption with MyFitnessPal (MyFitnessPal, n.d.). However, these apps but do not address the step in between: translating medical information into healthy behavior. Two applications on the market address blood tests and have similar features to our proposed tool. First is Gazelle, which loads test results into a phone app and allows users to read a description of each test. However, this is only available for patients who receives tests from Quest Diagnostics labs, and the results still do not show the values in context, calculate cardiovascular risk, or suggest behaviors (Quest Diagnostics, n.d.) A second app that resembles ours is Blood Test Guide, which displays a simple guide to common blood test measures (O Clock Software Pvt Ltd, n.d.). However, this does not display your personalized results, does not show what ranges of values are healthy or not, and does not suggest behaviors.

Gazelle by Quest Diagnostics.

Blood Test Guide by O Clock Software.

This project is inspired by a medical test redesign by Wired magazine, who asked graphic designers to design what a blood test might look like while consulting physicians for context like risks and suggested behaviors. This redesign presented data in the context of healthy and unhealthy ranges, gave a summary of what the results meant, and listed a few behaviors at the bottom (Leckart, 2010). However, they did not provide an interface to do this yourself, and did not allow you to

interact with the data by adding behaviors or moving test values to see the outcome. Our goal, like theirs, is to make medical data less complicated, but also to encourage patients to take action by allowing them to generate this layout themselves and to interact with it. Since this article's publication, there have been two projects based on Wired magazine's redesign. First, Stanford Computer Science major Francisco Cai automated this visualization for a graphics course and added features such as modularity and visualization over a time series (Cai, 2011). However, he did not calculate cardiovascular risk, suggest behaviors, or allow the user to visualize how their data would change if they adopted new behaviors. He also did not provide an interface for entering the medical test results, nor did he user test his interface. Finally, the Wired editor who presented the blood test redesign in the TED talk, Thomas Goetz, pursued the Blood test redesign by Wired Magazine. idea as a start-up company called 1 + 1 labs. His prototype, which he presented at the Health 2.0 Conference, does allow patient entry of information, but lacks much of our functionality (1 + 1 Labs, 2011). His project's goal is to produce a 1-page lipid profile, not to produce an interactive display of all blood test results, thus his prototype isn't interactive, doesn't explain what each test result means, and doesn't suggest specific actions outside of "living a healthy lifestyle". In addition, it has been 18 months since Goetz displayed his prototype, but still 1 + 1 labs has produced nothing. Their website only allows you to email them and says nothing about their product, suggesting that they have not pursued this idea (One Plus One Labs, n.d.).

Redesigning Lab Reports by Francisco Cai.

Your Test Results by 1 + 1 Labs.

Section 4: Target Users We envision two main target user groups for our web application: non-genetically disposed cardiovascular patients and healthy patients. When patients with a wide range of ailments receive a blood test report, they do not understand the intricacies of various results and indicators. These users would benefit from a user-centric breakdown of their results. Their results will be tailored to facilitate a more meaningful discussion with their healthcare provider regarding their ailments and potential solutions. Our other group of users are those who are have recently received blood test results from physicals or checkups. These users are not sick, but rather are curious about what their blood test reveals about their bodily health. Our secondary users will be health care providers and community health-care professionals. The ease of use of our application will allow primary users to get understandable data, but health-care professionals can use the application as an aid in educating those at risk of cardiovascular diseases about potential life-style risks. The ability to showcase adverse test results in the context of a person’s overall health is an important part of being an health care professional. Any doctor or nurse that can properly communicate "impacts [patient] satisfaction", creating 'positive outcomes' for said patient (Wilcox, 2013). Our application hopes to provide a comprehensive review of relevant blood test markers in an interactive form. Users will have the ability to see how their health can be impacted by changing the various values in each part of the blood test. This interactivity coupled with analysis will give users the motivation to improve or learn more about their health. The combination of explanations and personalized computed recommendations will allow target users to “see a path forward” in modifying their risky habits or ineffectual health regiments (Leslie, 2011). Section 5: Scenario Walkthrough A patient goes to their doctor, who tells them that they may be at risk for lifestyle diseases and orders a blood test. When the results come back from the lab, they are entirely unintelligible to the patient. The patient goes to our application and enters in the test results. A graphical display comes up, which contextualizes and explains the test results. Each value is shown on a slider with healthy, risky, and unhealthy ranges for the user’s cohort, alongside an explanation of the value’s meaning. At the bottom, there is a listing of the risks for several lifestyle diseases that the test results suggest. Possible lifestyle changes are suggested on the right. By selecting one of these lifestyle changes, the user can see their effect on prospective test results and the risks. The patient can also adjust the risks, which would cause the sliders to display the test values that would be needed to achieve the desired risks. Lifestyle combinations that would help achieve these goals are suggested. By using the application, the user is able to understand the interaction between the test results, their risks for lifestyle diseases, and their lifestyle choices.

Proposed application layout: step by step.

Typical blood test results. Unintelligible to the patient. Proposed application layout. Section 6: References 1 + 1 Labs. (2011). 1 + 1 Labs – Launch!. Health 2.0 Conference, San Francisco 2011. Retrieved from http://www.health2con.com/tv/1-1-labs-launch/. Cai, F. (Dec 2011). Data Visualization: Redesigning Lab Reports. Retrieved from https://graphics.stanford.edu/wikis/cs448b-11-fall/FP-CaiFrancisco Center for Disease Control and Prevention (CDC). (2010). Laboratory Medicine Best Practices Final Report. Laboratory Medicine Best Practices: Developing Systematic Evidence Review and Evaluation Methods for Quality Improvement. Phase 3 Final Technical Report. Retrieved from https://www.futurelabmedicine.org/pdfs/LMBP%20Yr3%20FINAL%20Technical%20Report%20 FINAL.pdf. Center for Disease Control and Prevention. (2013). FastStats: Heart Disease. Retrieved from http://www.cdc.gov/nchs/fastats/heart.htm. Goetz, T. (Jan 2011). It's Time to Redesign Medical Data. TED talk. Retrieved from http://www.ted.com/talks/thomas_goetz_it_s_time_to_redesign_medical_data.html Johnson, A. J., Easterling, D., Nelson, R., Chen, M. Y., Frankel, R. M. (1 Ap 2012). Clinical Simulations to Investigate Patient Preferences. Journal of the American College of Radiology. 9(4); 256-263. DOI: 10.1016/j.jacr.2011.12.023 Leckart, S. (29 Nov 2010). The Blood Test Gets a Makeover. Wired Magazine. Retrieved from http://www.wired.com/magazine/2010/11/ff_bloodwork/all/1 Leslie, B. (Feb 2011). How analytics can get you better medical treatment. Retrieved from http://sloanreview.mit.edu/improvisations/2011/02/03/how-analytics-can-get-you-bettermedical-treatment/ Lewis, N. (14 Sept 2011). Docs Don't Want Lab Results Going to Patients. InformationWeek Healthcare. Retrieved from https://www.informationweek.com/healthcare/policy/docs-dontwant-lab-results-going-to-pati/231601421

MyFitnessPal. (n.d.) MyFitnessPal iPhone App. Retrieved from http://www.myfitnesspal.com/iphone. National Heart, Lung, and Blood Institute (NHLBI). (6 Jan 2012). What Are Blood Tests?. Retrieved from http://www.nhlbi.nih.gov/health/health‐topics/topics/bdt/ O Clock Software Pvt Ltd. (n.d.) Blood Test Guide – Analyze Your Blood Test Report. Retrieved from https://itunes.apple.com/us/app/blood-test-guide-analyze-your/id491681195?mt=8. One Plus One Labs. (n.d.) One Plus One Labs. Retrieved from http://oneplusonelabs.com/ PA Patient Safety Authority. (Sept 2009). Safe Patient Outcomes Occur with Timely, Standardized Communication of Critical Values. Pennsylvania Patient Safety Advisory. 6(3); 93-97. Quest Diagnostics. (n.d.). My Gazelle App. Retrieved from http://mygazelleapp.com. Saxton, J. W., & Finkelstein, M. M. (23 Nov 2011). Patient Engagement to Enhance Patient Outcomes: Test Tracking. MD News. Retrieved from http://www.mdnews.com/news/2011_11/05829_novdec2011_patient-engagement.aspx Unbound Medicine. (n.d.) Family Drug Guide. Retrieved from http://www.unboundmedicine.com/products/family_drug_guide Wilcox, L. (Feb 2013). Promoting engagement with personal health through interactive technology. Presentation delivered in Hornbake Library, HCI Lecture Series, College Park, MD. Section 7: Appendix Original Proposal: Problem background/motivation: Results from medical tests, whether good or bad, rarely lead to behavioral changes in patients for the better. This is because tests results are difficult to interpret unless you are a physician. However, if patients could interpret the results of their tests, they would have personalized data on their progress, which is a key component in motivating the patient to improve his or her health. In addition, if these results were put in context, comparing them to normal results and providing suggestions for how to improve them, this would translate the results into action [1]. Problem statement: We should encourage patients to take control of their health by using their test results as a tool to improve their wellbeing. Providing an interface which shows clearly what their results mean and what the patient should do next will help reduce patient noncompliance with doctor recommendations. Proposed solution: The interface would allow users to enter their test results for a particular test using checkboxes, sliders, and other clickable elements to put the numbers on the paper in context. This data would be presented in a colorful interactive infographic which would explain what each metric means, whether you are in a healthy range for Blood test redesign. Source: Wired magazine [2]. that measure, and predictions for risks such

as heart disease. In addition, there would be a set of checkboxes with behaviors that the patient could select to see how this could change their results; the patient could also move pointers up and down the metric ranges to see what changes they would need to make to change their scores. Differentiation from past solution: Currently, test results are unreadable, don't put the results in context, and don't translate the results into actions [2]. This project is inspired by a medical test redesign by Wired magazine, who asked graphic designers to design what a blood test might look like while consulting physicians for context like risks and suggested behaviors [3]. However, they did not provide an interface to do this yourself, and did not allow you to interact with the data by adding behaviors or moving test values to see the outcome. Who benefits and why: Patients would benefit from better understanding and engagement in their health, and doctors would benefit from reduced noncompliance. Evaluation plan: We would begin with a lo-fi prototype based on current tests, interview physicians to confirm risk analysis and behaviors taken from academic sources, and test the prototype on patients. We could focus on blood tests for the purposes of this class. References 1. Goetz, T. (Jan 2011). It's Time to Redesign Medical Data. TED talk. Retrieved from http://www.ted.com/talks/thomas_goetz_it_s_time_to_redesign_medical_data.html 2. National Heart, Lung, and Blood Institute. (6 Jan 2012). What Are Blood Tests?. Retrieved from http://www.nhlbi.nih.gov/health/health-topics/topics/bdt/. 3. Leckart, S. (29 Nov 2010). The Blood Test Gets a Makeover. Wired Magazine. Retrieved from http://www.wired.com/magazine/2010/11/ff_bloodwork/all/1. Changes from Original Proposal: We have so far stayed with the ideas presented in the original proposal with some refinement. First, we have decided to focus on blood tests since they are the most prevalent medical test. Second, we have chosen Javascript as our platform to keep all data at the client side. Third, we have narrowed down our user base and evaluation plan. Finally, we have drafted and sketched several mockups for the overall design which add more components than the original Wired magazine redesign and consider what would bring a person to use this tool and what would best help them adopt healthy behaviors based on their assessed risks. In particular, we will be examining each test metric separately, explaining what the test means and how the score can be improved, rather than only giving an overall summary of results. We will also be adding behavior checkboxes which, when selected, will change the metric and calculated risks, allowing the patient to visualize how that lifestyle change could affect their results a year from today.