Creating Evidence-Based Risk Assessments for Individual Patients: The New Consent Forms
August 2010
Can you share some of your research interests and how the individual risk consent software came about?
I am a cardiologist and outcomes researcher, meaning that I am very interested in understanding patient and treatment factors that influence patients’ outcomes. Most of our work, to date, has focused upon patients’ symptoms, function, and quality of life in coronary disease and heart failure. Outcomes researchers, like myself, often develop terrific prediction models to take into account each patient’s unique characteristics, and then to predict, fairly well, their longer-term outcomes. What we have been struck by is what happens after we do this research. We write articles, the articles are published in journals, and then this information just sits on the shelf and never gets used. We became very interested in trying to create a paradigm that could developing a web-based solution, integrated into the flow of care, that could translate the results of our outcomes research to clinical care. It allows individualized estimates of outcomes to be readily available for clinicians and patients so that they can jointly choose the best therapy, and clinicians can tailor the intensity of treatments to patient’s individual risks.
Our first testing of this concept has been in the cath lab. There is a long history of outcomes research in the cath lab; most notably, the American College of Cardiology’s National Cardiovascular Data Registry (NCDR), containing millions of patient records. The NCDR built big, multi-variable prediction models of outcomes from angioplasty, but these are only used to give quarterly reports back to hospitals about how the hospital did and how it was expected to do, given the characteristics of the patients they treated. Our vision was to take that same prediction model, and allow it to be run at the time care is delivered to patients, so that patients and their clinicians could be better informed of the risks and benefits of treatment. Our software offers, prospectively, an individualized patient risk assessment so physicians can decide how best to treat, given the patient’s specific risk profile. We folded that concept into an improved informed-consent document.
Traditionally, informed consent documents are written in legalese and patients don’t read them. A standard consent form is very vague and uninformative. It doesn’t really educate patients, and most importantly, even though clinicians and patients are supposed to talk about the risks, benefits and alternatives to treatment, in general, traditional consent forms offer no support for shared medical decision-making. At our institution, we used the same consent forms for cataract surgery, angioplasty and heart transplantations, with only a blank line to write the type of procedure and the risks associated with it. This was a missed opportunity to engage patients in choosing their treatment, and you are also, as a clinician, unable to tailor the approach to treatment to the patient’s individual risks. In cardiology, there is something called the “risk-treatment paradox” where we tend to most aggressively treat those patients with lower risk, and least aggressively treat those with the highest risk. We thought that if we could transparently display patients’ risks within the consent form, that we could overcome this ‘paradox’ and most aggressively treat those who could benefit most from the more aggressive therapies.
We thus completely redesigned our angioplasty consent forms to make them much simpler to understand (Figure 1). Our consent form is now written at the 8th grade level, as opposed to the 16th grade level, which was where our original consent form was written (Figure 2). The redesigned form is much more educational, with a description of an angiogram, an angioplasty and a stent, and it also has support for bi-directional discussions between doctors and patients to help choose — if the patient ends up needing a stent — a bare-metal or drug-eluting stent. The form helps patients to see the benefit in reducing the risk for coming back to the lab for a repeat procedure, and allows us to connect this benefit to the need for our patients to continue dual anti-platelet therapy for at least a year. Another important element to the consent form is the documentation of patients’ risks for bleeding. There are important therapies, such as bivalirudin, closure devices or a radial approach, that can minimize bleeding, but are most often used in patients at low risk for bleeding and less often in those who could benefit most from such therapies. Bleeding risk is also an important reason to admit an outpatient for observation overnight, since we often don’t know at the end of the procedure whether or not the patient will ultimately have a bleeding complication. Thus, if a patient is at moderate-to-high risk for bleeding, it is probably safer to admit them to the hospital for observation, because many times the bleeding events don’t occur in the lab or even at the cath site.
We implemented the consent form at Saint Luke’s Hospital, starting it as a pilot program for three months, but the nurses liked it so much and thought it was such an improvement that Saint Luke’s has actually sustained it for four years (almost 9,000 patients) and it was at the nurses’ urging that we rolled it out to the other three hospitals in the Saint Luke’s Health System that perform angioplasty. Our experience has been that patients like it and better understand their risks. They are much less anxious about the procedure, and feel more involved in the medical decision-making process. We also found that our clinicians are able to use the data in the consent form to reverse the risk-treatment paradox at our institution and significant lower our rates of bleeding complications.
Each consent form is individually generated based on the individual patient’s risks? Exactly. The data needed to execute the risk models is collected by the nurses just prior to the cath and the form is then printed off so that the doctors and patient can discuss and sign it. We also have a decision aid for physicians with the risks, minus all the explanatory information. This form can be taken to the cath lab and either taped to the monitor or transcribed onto a whiteboard. Even though physicians can enter the data, the elements are fairly straightforward, so it is actually the nurses in our cath lab holding area that complete the forms and print them out. Often our patients in the CCU don’t go through the holding area, so the form may actually be printed up in the CCU, but it could be printed out at the clinic or anywhere else that a computer with web access is available. Patients are given time to read the form, and discuss it with their family and physician. Once the data has been input and run, clinicians have the data upon which to make recommendations and decisions.
How might this tie in with the ACC/AHA/SCAI guidelines for care? Some of the decisions that we are helping to influence aren’t necessarily addressed in the guidelines. However, when the decisions are addressed, the guidelines usually will say something like “for high-risk patients, consider x, y, and z.” But there is no clear, reproducible way to figure out who is at high risk and who is not. Our software provides a transparent, evidence-based way to show risk. It is quite advantageous, because the models come from the American College of Cardiology NCDR, built on hundreds of thousands of patients at hundreds of hospitals. It is essentially the most accurate risk model that exists for these specific conditions and outcomes. It offers a very good supplement to the guidelines, because the models provide very quantitative and accurate estimates for patient outcome risks.
What have you been able to learn about incorporating the form into the flow of care thus far? We have been using the form for some time now without any specific follow up, but have seen a marked reduction in our bleeding rates. Our hospital used to have an elevated risk of bleeding. After angioplasty, it used to be about 5.5%, which was twice the national average. Over the past two years, we reduced our bleeding rate to about 1.3%, or half of the national average. Interestingly, our lowest bleeding rates were observed after we reversed the risk-treatment paradox and began preferentially using bivalirudin in patients at the highest risk for bleeding. While we have used it for years at our institution, we are now rolling it out as part of a research program funded by two grants, one from the PRT-Spina American Heart Association (AHA) Outcomes Research Center Program and the second from the National Heart, Lung and Blood Institute of the NIH. In the AHA grant, we are examining whether clinicians will begin preferentially using drug-eluting stents in those at the highest risk for restenosis and bare-metal stents more often in those at low risk for restenosis. We are also looking at patient compliance. If patients are informed up front of the benefits of a drug-eluting stent, will they be more adherent to taking clopidogrel or two anti-platelet agents at six months, then if they are just traditionally consented? We hope that if we engage the patient in a discussion about benefits and their obligation to stay on dual anti-platelet therapy, that they will be more compliant with that therapy six months later. The NIH grant is testing whether clinicians will use bleeding avoidance therapies more aggressively in high-risk patients when this information is provided up front. This will enable us to explore whether these more informational consent forms will lower the risk of bleeding and create a safer procedure.
What have patient reactions been to the consent form? Very positive. When we first rolled it out, patients read the consent forms more often, which was the first step. They were able to better recall their risks and felt more involved in medical decision-making. The nurses at our hospital wanted to keep the form going because of their observations that patients were asking more questions, were much more involved, and that it was fulfilling the task of what informed consent is supposed to do by engaging patients and providing data so that we can tailor care to individuals. We turned the system on at Integris, in Oklahoma City, about two months ago. They are having an exceptionally positive reaction, and they have sent us some quotes from patients who liked it a great deal, stating that they had never read such an informative consent form.
Could this type of consent ultimately help with reimbursement issues? We began this project because we were interested in creating a tool for evidence-based, safer, more cost-effective care, and for enabling greater patient involvement and shared decision-making. However, there wasn’t much of a business model to support it until Medicare started scrutinizing the decision to admit a patient after elective angioplasty. In the pilot program of RAC audits, the number-one procedure on which Medicare overturned or challenged hospitals was the outpatient who came in for angioplasty and was then admitted. The RAC auditors found that there was seldom documentation to justify that admission and that it probably would have been safe to send the patient home. Because our form prospectively documents each patient’s risks for bleeding, once a hospital decides the threshold at which they think it is unsafe to send patients home, they can use the form to document and support the decision to admit a patient. This is quite important, because the data is based on the best models that exist, models built on hundreds of thousands of patients from hundreds of hospitals. You can say to a RAC auditor, “I executed the very best risk model for this specific patient and I used this objective estimate of the patient’s risks for bleeding, and in my medical judgment and in our hospital’s standard practice, it was safest to admit these patients rather then send them home and run the risk that they would have a bleeding episode at home, which can be catastrophic.” Being able to provide this documentation, while also delivering better, more patient-centered care, offered a very good business case to support disseminating this nationally. There is now a company, Health Outcome Sciences (HOS; www.h-outcomes.com), offering this to hospitals who would like to implement a similar program, with the benefit being that hospitals can not only justify their decisions to admit a patient to RAC auditors, but also decrease their bleeding risks, support better patient adherence to dual anti-platelet therapy after drug-eluting stents, and improve patient satisfaction. Our Press Ganey scores went up in angioplasty patients after we introduced the consent form. We think that it would be to a hospital’s great financial advantage to have a transparent, justifiable basis for admitting rather than discharging outpatients after angioplasty.
It’s interesting that something that might work very well to save money across the board also offers such an opportunity for more personalized, involved care, which is the opposite of where people seem to expect healthcare is headed. That’s why we’re so excited about this; we think that this is a terrific advance. In the long run, I would like all medical decisions to use this evidence-based, personalized approach.
How would this work for a patient who is trying to decide whether to have coronary artery bypass graft surgery (CABG) or percutaneous coronary intervention (PCI)? Our first funding was really focused on building a shared decision-making tool for this exact situation. We built risk models in patients with acute coronary syndromes: what is their quality of life and survival likely to be a year or later if they’re treated with angioplasty versus bypass surgery versus medicines alone? It turns out that there are two challenges in creating such a model. One is that we don’t have a good prediction model, especially for stable coronary disease, as a function of alternative treatments. You can only run data if there is a good outcomes model to deploy. Second, we had a cultural issue. Physicians are very used to having their dialogue with patients and making these decisions, and it would really step out of the routine process of care to execute these models and use these forms as the basis of that discussion. It is absolutely where we want to go, but the key to success is starting simple, and everybody has to get an informed consent anyway.
How much time does using the form add to the process? One very big concern is that cath lab staff and interventional cardiologists are very intolerant of delays in the efficiency of the lab. If this adds ten minutes per patient and you have 9 cases a day, that’s an hour and a half extra delay, which is just too much. This form does not add any time to our cath lab flow. We turn over the rooms just as fast as we did with the old consent forms. The patients are usually in the holding area for quite some time before they go to the cath lab. They read it during that time, and then the doctor comes in and talks to them about it. Sometimes they ask more questions and that process will last a little bit longer, but that’s the essence of what we do as clinicians: educate and address our patients’ questions as we render care, and our doctors felt that it was not a bad thing to have patients more involved and asking more about what we are doing to them.
Could you talk about your partnership with the American College of Cardiology (ACC)? The ACC has, for angioplasty, what we think are the best models that exist today. They are developing models for intracardiac defibrillators (ICDs), and have several terrific registries for which they build outcomes models. When we were thinking about starting the company, we approached the ACC to get a license to use their models. They thought that an individualized consent form was an ideal application, because they are very concerned about patient safety, cost-effectiveness and quality. It would benefit patients greatly, it would help doctors and hospitals, and so for them, it was really a win-win-win.
What are you seeing in the future? In the short term, within cardiology, we would like to have models for ICDs, the third-biggest RAC challenge in the pilot states. That’s a real problem for hospitals, because if they can’t admit those patients afterwards, they lose money on each case, so we would like to provide RAC documentation justification. We would also like to start building models to help regarding the decision of whether to give patients prasugrel or clopidogrel — each has a different influence on ischemic complications and bleeding. Medicare is going to start reporting 30-day readmission rates after angioplasty, and hospitals will want a tool to help identify patients at high risk for readmission. I don’t think that readmissions result from something done by the interventionalist or the procedure, but instead from the fact that we are dealing with high-risk patients. If you could identify patients at greater risk for readmission, then you could have an advanced nurse practitioner or a social worker stop by before the patient leaves to make sure they know who to call if they have a problem, make sure they have their medications, a follow-up appointment, etc. Even though we should do that for everyone, we could at least identify those patients at highest risk to be sure that they have the support they need to avoid unnecessary readmissions. These are patients where you want to be absolutely certain to do a good job transitioning them from the hospital to home. We also want to move more into the shared decision-making between specialties. There are other professional societies that have quality registries, such as the Society for Thoracic Surgeons and the American College of Surgeons, and we would like to be able to use their models to make better consent forms for relevant procedures. Finally, there has been a lot of talk about genetics and clopidogrel resistance or poor clopidogrel metabolism. It is very hard for doctors or patients to understand how to weigh patients’ genetic risk factors amongst all their other clinical conditions. Running these models will be a very intuitive way to incorporate new predictive biomarkers and genes into clinical decision-making.
Dr. John Spertus can be contacted at spertusj@umkc.edu For Further Reading 1. Arnold SV, Decker C, Ahmad H, et al. Converting the informed consent from a perfunctory process to an evidence-based foundation for patient decision making. Circ Cardiovasc Qual Outcomes 2008;1;21–28.
Each consent form is individually generated based on the individual patient’s risks? Exactly. The data needed to execute the risk models is collected by the nurses just prior to the cath and the form is then printed off so that the doctors and patient can discuss and sign it. We also have a decision aid for physicians with the risks, minus all the explanatory information. This form can be taken to the cath lab and either taped to the monitor or transcribed onto a whiteboard. Even though physicians can enter the data, the elements are fairly straightforward, so it is actually the nurses in our cath lab holding area that complete the forms and print them out. Often our patients in the CCU don’t go through the holding area, so the form may actually be printed up in the CCU, but it could be printed out at the clinic or anywhere else that a computer with web access is available. Patients are given time to read the form, and discuss it with their family and physician. Once the data has been input and run, clinicians have the data upon which to make recommendations and decisions.
How might this tie in with the ACC/AHA/SCAI guidelines for care? Some of the decisions that we are helping to influence aren’t necessarily addressed in the guidelines. However, when the decisions are addressed, the guidelines usually will say something like “for high-risk patients, consider x, y, and z.” But there is no clear, reproducible way to figure out who is at high risk and who is not. Our software provides a transparent, evidence-based way to show risk. It is quite advantageous, because the models come from the American College of Cardiology NCDR, built on hundreds of thousands of patients at hundreds of hospitals. It is essentially the most accurate risk model that exists for these specific conditions and outcomes. It offers a very good supplement to the guidelines, because the models provide very quantitative and accurate estimates for patient outcome risks.
What have you been able to learn about incorporating the form into the flow of care thus far? We have been using the form for some time now without any specific follow up, but have seen a marked reduction in our bleeding rates. Our hospital used to have an elevated risk of bleeding. After angioplasty, it used to be about 5.5%, which was twice the national average. Over the past two years, we reduced our bleeding rate to about 1.3%, or half of the national average. Interestingly, our lowest bleeding rates were observed after we reversed the risk-treatment paradox and began preferentially using bivalirudin in patients at the highest risk for bleeding. While we have used it for years at our institution, we are now rolling it out as part of a research program funded by two grants, one from the PRT-Spina American Heart Association (AHA) Outcomes Research Center Program and the second from the National Heart, Lung and Blood Institute of the NIH. In the AHA grant, we are examining whether clinicians will begin preferentially using drug-eluting stents in those at the highest risk for restenosis and bare-metal stents more often in those at low risk for restenosis. We are also looking at patient compliance. If patients are informed up front of the benefits of a drug-eluting stent, will they be more adherent to taking clopidogrel or two anti-platelet agents at six months, then if they are just traditionally consented? We hope that if we engage the patient in a discussion about benefits and their obligation to stay on dual anti-platelet therapy, that they will be more compliant with that therapy six months later. The NIH grant is testing whether clinicians will use bleeding avoidance therapies more aggressively in high-risk patients when this information is provided up front. This will enable us to explore whether these more informational consent forms will lower the risk of bleeding and create a safer procedure.
What have patient reactions been to the consent form? Very positive. When we first rolled it out, patients read the consent forms more often, which was the first step. They were able to better recall their risks and felt more involved in medical decision-making. The nurses at our hospital wanted to keep the form going because of their observations that patients were asking more questions, were much more involved, and that it was fulfilling the task of what informed consent is supposed to do by engaging patients and providing data so that we can tailor care to individuals. We turned the system on at Integris, in Oklahoma City, about two months ago. They are having an exceptionally positive reaction, and they have sent us some quotes from patients who liked it a great deal, stating that they had never read such an informative consent form.
Could this type of consent ultimately help with reimbursement issues? We began this project because we were interested in creating a tool for evidence-based, safer, more cost-effective care, and for enabling greater patient involvement and shared decision-making. However, there wasn’t much of a business model to support it until Medicare started scrutinizing the decision to admit a patient after elective angioplasty. In the pilot program of RAC audits, the number-one procedure on which Medicare overturned or challenged hospitals was the outpatient who came in for angioplasty and was then admitted. The RAC auditors found that there was seldom documentation to justify that admission and that it probably would have been safe to send the patient home. Because our form prospectively documents each patient’s risks for bleeding, once a hospital decides the threshold at which they think it is unsafe to send patients home, they can use the form to document and support the decision to admit a patient. This is quite important, because the data is based on the best models that exist, models built on hundreds of thousands of patients from hundreds of hospitals. You can say to a RAC auditor, “I executed the very best risk model for this specific patient and I used this objective estimate of the patient’s risks for bleeding, and in my medical judgment and in our hospital’s standard practice, it was safest to admit these patients rather then send them home and run the risk that they would have a bleeding episode at home, which can be catastrophic.” Being able to provide this documentation, while also delivering better, more patient-centered care, offered a very good business case to support disseminating this nationally. There is now a company, Health Outcome Sciences (HOS; www.h-outcomes.com), offering this to hospitals who would like to implement a similar program, with the benefit being that hospitals can not only justify their decisions to admit a patient to RAC auditors, but also decrease their bleeding risks, support better patient adherence to dual anti-platelet therapy after drug-eluting stents, and improve patient satisfaction. Our Press Ganey scores went up in angioplasty patients after we introduced the consent form. We think that it would be to a hospital’s great financial advantage to have a transparent, justifiable basis for admitting rather than discharging outpatients after angioplasty.
It’s interesting that something that might work very well to save money across the board also offers such an opportunity for more personalized, involved care, which is the opposite of where people seem to expect healthcare is headed. That’s why we’re so excited about this; we think that this is a terrific advance. In the long run, I would like all medical decisions to use this evidence-based, personalized approach.
How would this work for a patient who is trying to decide whether to have coronary artery bypass graft surgery (CABG) or percutaneous coronary intervention (PCI)? Our first funding was really focused on building a shared decision-making tool for this exact situation. We built risk models in patients with acute coronary syndromes: what is their quality of life and survival likely to be a year or later if they’re treated with angioplasty versus bypass surgery versus medicines alone? It turns out that there are two challenges in creating such a model. One is that we don’t have a good prediction model, especially for stable coronary disease, as a function of alternative treatments. You can only run data if there is a good outcomes model to deploy. Second, we had a cultural issue. Physicians are very used to having their dialogue with patients and making these decisions, and it would really step out of the routine process of care to execute these models and use these forms as the basis of that discussion. It is absolutely where we want to go, but the key to success is starting simple, and everybody has to get an informed consent anyway.
How much time does using the form add to the process? One very big concern is that cath lab staff and interventional cardiologists are very intolerant of delays in the efficiency of the lab. If this adds ten minutes per patient and you have 9 cases a day, that’s an hour and a half extra delay, which is just too much. This form does not add any time to our cath lab flow. We turn over the rooms just as fast as we did with the old consent forms. The patients are usually in the holding area for quite some time before they go to the cath lab. They read it during that time, and then the doctor comes in and talks to them about it. Sometimes they ask more questions and that process will last a little bit longer, but that’s the essence of what we do as clinicians: educate and address our patients’ questions as we render care, and our doctors felt that it was not a bad thing to have patients more involved and asking more about what we are doing to them.
Could you talk about your partnership with the American College of Cardiology (ACC)? The ACC has, for angioplasty, what we think are the best models that exist today. They are developing models for intracardiac defibrillators (ICDs), and have several terrific registries for which they build outcomes models. When we were thinking about starting the company, we approached the ACC to get a license to use their models. They thought that an individualized consent form was an ideal application, because they are very concerned about patient safety, cost-effectiveness and quality. It would benefit patients greatly, it would help doctors and hospitals, and so for them, it was really a win-win-win.
What are you seeing in the future? In the short term, within cardiology, we would like to have models for ICDs, the third-biggest RAC challenge in the pilot states. That’s a real problem for hospitals, because if they can’t admit those patients afterwards, they lose money on each case, so we would like to provide RAC documentation justification. We would also like to start building models to help regarding the decision of whether to give patients prasugrel or clopidogrel — each has a different influence on ischemic complications and bleeding. Medicare is going to start reporting 30-day readmission rates after angioplasty, and hospitals will want a tool to help identify patients at high risk for readmission. I don’t think that readmissions result from something done by the interventionalist or the procedure, but instead from the fact that we are dealing with high-risk patients. If you could identify patients at greater risk for readmission, then you could have an advanced nurse practitioner or a social worker stop by before the patient leaves to make sure they know who to call if they have a problem, make sure they have their medications, a follow-up appointment, etc. Even though we should do that for everyone, we could at least identify those patients at highest risk to be sure that they have the support they need to avoid unnecessary readmissions. These are patients where you want to be absolutely certain to do a good job transitioning them from the hospital to home. We also want to move more into the shared decision-making between specialties. There are other professional societies that have quality registries, such as the Society for Thoracic Surgeons and the American College of Surgeons, and we would like to be able to use their models to make better consent forms for relevant procedures. Finally, there has been a lot of talk about genetics and clopidogrel resistance or poor clopidogrel metabolism. It is very hard for doctors or patients to understand how to weigh patients’ genetic risk factors amongst all their other clinical conditions. Running these models will be a very intuitive way to incorporate new predictive biomarkers and genes into clinical decision-making.
Dr. John Spertus can be contacted at spertusj@umkc.edu For Further Reading 1. Arnold SV, Decker C, Ahmad H, et al. Converting the informed consent from a perfunctory process to an evidence-based foundation for patient decision making. Circ Cardiovasc Qual Outcomes 2008;1;21–28.
NULL
