Annette Roberts, RN - The Power to Predict

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● cases. And while we've learned a lot at Craco about these cases, we only see about 200 cases per year. So around 20 years ago, Craco formed a strategies division to further leverage and share our tools and extend our collective ability to learn from our MPL experience. And the strategies community is a community of captive and commercial insurers that contribute to the claims data, so it's added to our comparative benchmarking database, which will go into a little more. And it provides a community of learning and sharing of tools of best practice solutions in order to improve patient safety across the country. So, you know, here you'll see Craco is one of his main organization, and then he or some list of contributors to the comparative benchmarking system database. So the comparative database is a shared tool for identifying trans and medical error in patient harm. And there's around 450,000 cases of open and closed cases in our database. We see an average of 8 to 10,000 new cases each year. And this covers over 500 hospitals and physicians practices, and also 195 physicians. The database itself represents about 30% of the claims in the United States. So the benchmarking system is really powerful because it's supported by our coding taxonomy, which is highly governed and overseen by the leaders in our field. It includes hundreds of causation codes that enable the organizations to identify the multifactorial causes of error in harm. And it's updated regularly to ensure that it's responsive to the industry trends. It's reliable and applied by experienced clinicians, many nurses who receive ongoing training and education. And we regularly audit to ensure it's intercoder reliability. So leveraging the materials from medical records and claims, the experts code the taxonomy. And then we translate it into analyzable data, which we'll be presenting today. So I'm excited to be sharing the benchmarking report, the Power to Pre-Defdom, grateful for the team that helped put this together. Kreco has been seeking to understand the drivers of medical liability cases and to be able to translate those insights that we gain from the deep coding and the intelligence. So it helps us to inform us on patient safety and risk management strategies. We know the rate of claims paying more than a million dollars continues to rise. This year we wanted to take that a step further and leverage predictive analytics not only to understand the drivers of these losses, but to derive insights on what you can expect in the future. Predictive analytics have been powering apps that plan our fastest route to work when we were still going into work. I'm providing winning lineups for your fantasy football teams and ensure that your Netflix recommendations are perfectly aligned with your personal viewing preferences. One key differentiator in our ability to do this is that CVS, the CVS database contains both paid and unclaimed unpaid claims. While other medical malpractice and data sets contain only paid claims, by having data on all claims, it allows us to understand what elements make a claim more likely to pay and give us the information to predict that future claims may pay or close without payments. So in our business knowing what's coming can only enhance our longstanding focus on addressing clinical vulnerabilities to prevent patient harm, but it can also inform us on defense strategies and rate settings and more. So about this study, the study explores the relationship between categories of contributing factors, factors that affect the adverse event and the claimant's decision to initiate a case and the odds of a case being closed with an indemnity payment. These factors reflect both provider and patient issues, breakdown in technical skills, clinical judgment, breakdowns in communication, behavior issues, clinical systems, and the environment, teamwork, and others. The team of taxonomy specialists examine these medical records. They look at the expert opinions and other legal documents on every case and identify an average 3.8 factors per case. So with our coding taxonomy, these are all rolled up into tiers of 53 subcategories and 18 categories for analysis at various levels. So as we started to develop this study, our team analyzed 37,000 cases that closed between 2014 and 2018 and we used a multi-stage modeling process. We started with a logistic regression to identify 29 subcategories of contributing factors that can have an effect on whether a case will close with a payment. From there, the team is a random forest analysis to home in on the 11 factors with the greatest importance to the financial outcomes of cases. It's important to note that the model controlled for variables such as the responsible service involved, the clinical severity of a case with injuries, patient age, class type, what the allegation was geographic region and care setting. What we arrived at were these 11 factors of the greatest with the greatest importance for a financial outcome. As you see, listed here on the right, you can see it ranges from the failure to follow a policy and protocols to communication among providers and to technical performance and patient monitoring. Of those, we focused on three subcategories with the highest odds of payment, which were policy protocol not followed or absent, inadequate patient assessment or failures in patient assessment and insufficient documentation. The report delves deeply into those subsets and allows us to present clear opportunities for insurance programs and clinical leaders to develop strategies aimed at reducing risk to patients, providers and insurers. So, our focus today on the study is highlighting the factors when present increase the likelihood of a case closing with an indemnity payment to understand where these issues are most prevalent in which settings and what service lines and how to prevent them from occurring. And when they do, how do we effectively manage these cases when they do hit our providers or insurance companies? So, any questions before I go on? Okay, so first, go to follow an established policy or protocol or failure to have one in place. We analyzed just over 3,000 cases. And cases with this issue have an odds ratio of 2.45, meaning that cases with this issue are 145 percent more likely to close with a payment than cases without this issue. Next was patient assessment failures. We looked at nearly 19,000 cases with an odds ratio of 1.85. And where cases with this issue, there are these cases with patient assessment failure issues are 85 percent more likely to close with a payment versus cases without this issue. And lastly, insufficient documentation or absent documentation, we studied over 6,500 cases. There was an odds ratio of 1.76. And so cases with these issues were 76 percent more likely to close with a payment versus those without. So, of course, the highest odds for payment in these categories, we thought we'd take a deeper look at, just how much will they affect the financial outcomes? So, it's important to note that indemnity payments are expected in anywhere from 47 to 65 percent of cases. As you look at the bar chart with the three study factors listed on the left, relatively to the cases where none of the factors above were in the cases. And you'll see that each one of our study factors are significantly more likely to result in a payment and many in excess of a million dollars. So, as payments are closely linked to the severity of the injury, when a patient suffers, the data can be a directive call to action for patient safety leaders and claimants managers. So, our first key subset, cases where there's a fair to have or follow a policy or protocol. This includes cases where policies where institutions or their institutional or setting specific policies and protocols, which is what we do every single day in hospitals and healthcare settings, either were not followed or based on expert opinions, depositions, and testimonies were not present to guide the caregiver or team in their care or management of the patient. Again, when an issue like this was present, the odds of closing in case with payment increased 145 percent. We know that policies and procedures governing guide much of our care delivery and increasingly so as we continue to standardize care or cross settings in rural groups. While nursing practice is especially policy driven and therefore vulnerable to failures in this regard, you see two charts here on the left and more than half of these cases involve other role groups other than nursing, including services of medicine, surgery, OBGYN, anesthesia. This is reinforced by the case type where the failures are most frequently identified, where cases involving falls and cases involving diagnostic failures are most frequently seen at the top of the list. You'll see in a moment how policy lapses are prevalent in falls and gaps in policies are prevalent in diagnosis cases. I mean, we look at clinical severity of the injuries occurred in these cases where a policy was not followed or one should have been governed or supported clinical care. We see nearly 50 percent of the cases involved a high severity injury or a death. So for example, a 60-year-old male suffered a permanent brain damage due to fall when he was left alone in the bathroom, contrary to the hospital's fall policy. We're a failure to follow reporting protocols resulted in an ED physician unaware of a positive blood culture for a 50-year-old woman who died from sepsis four days after being discharged from the emergency department or is still born delivery, delay attributed to the unoccupied obstetrician, not readily available per hospital policy. So here we stratify the data in the subcategory of policy failure to follow a policy and procedure versus failure to establish one. And we wanted to tease out the difference between these cases. What we see is that while failures to follow a policy and protocol are more prevalent, as you can see here, the need for policy and procedure in the green bar is more costly. And cases that lack appropriate policy and procedures, you know, we'll tease this out further as we go on, but there's different implications in these issues. So starting with cases where policies and procedures were not followed and played a role in the patient's injury, on the left side you can see the care setting that originated in these cases with policy and procedure failures, most prevalent on the patient side in the patient's room. And that tracks closely with the prevalence of nursing involvement in these cases. On the right side, you'll see other concurrent factors. As we said earlier, there's usually more than one factor in these cases, which include communication regarding the patient's condition, monitoring patient's status. Third party involved is around those factors where the handoff didn't happen. Staff training and fall prevention. We know each event includes a constellation of factors and it's easy to imagine a scenario where these factors coalesce in a patient's injury, especially in the inpatient room. In the spiral, you'll also see in the blue and green coloring, the role groups identified in cases where these other factors are also present. I want to point out the visual that risk managers and patients safety leaders consider how to leverage these findings, seeking to understand that environmental and other factors lead to conscious or unconscious policy and protocol lapses, and to make efforts to understand and make improvements in policies must include all role groups. So as we see here, where policies and protocols were needed, we on the left side, you can see the care settings, again, where they originated, most prevalent in the clinicals or office setting, and the less prevalent, but much more costly, where the policies needed in the emergency department. And here again, on the right side, you see the other concurrent factors present in these cases, including the coordination of care with third parties, often the outside practice or health system, on following up with tests or referrals, training and communication. So together, these data link closely to the prevalence of policy gaps and cases involving diagnostic failure, where the presence of standardized processes or protocols directly directing communication and coordination of care would minimize many process gaps in the diagnostic journey. And likewise, in the emergency department, communication gaps, like questions about whose responsibility is it to communicate test results after a patient is discharged, can significantly impact the timeliness of a diagnosis, and we are minimizing those questions through policy and protocols would have had great value. As you can see by the colors in this bar graph, there's significant involvement by physician services in these cases, highlighting the importance of ensuring that any policy or procedure is developed, that is developed considers all the members of the healthcare team. So how do we reduce policy and procedure breakdowns? You want to establish accountability for training and adherence. We need to have constant education on new policies or updates in old policies, to determine if their breaches are due to the policies present in us. And that is relevant staff. And then you want to address failures routinely, address failures to follow if you have event reports follow up on those and see if the event is related to changes needed in a policy or a policy is needed. So focus on the policies that are germane to high risk consequences, engage all impacted disciplines in the development and review of the policy and consider the downstream effects. And then you know, defending these cases, you have to determine whether a formal policy was in place with routine adherence at the time of care. Was there any deviation from the policy that impacted the patient's outcome? And there's an alignment between a formal policy and a standard of care. Next we'll move on to patient assessment failures. So patient assessment failures are cases involving highlights of the multi-complexity of these patients and the processes that impacts caregivers, misunderstanding of patient status either prior to a diagnosis or one's treatment has begun. When the assessment failure has occurred, the study showed that the odds of case closing with payment increased by 85%. And with the average and diminutive payment of nearly 520,000 dollars and nearly 60% of cases involving high severity injuries in this category, the stakes for understanding, preventing and defending these cases are high for both patients and providers. Of course, accurate patient assessment is a foundational aspect of safe patient care, whether during the ambulatory diagnostic process or the management of surgical or medical procedures and care. The assessment failure is most prevalent in our study. How lighted here on the left with their corresponding average and diminutive payments, failure to recognize signs and symptoms, creating a two-narrow diagnostic focus, missed opportunities in testing and consults reflect the high stakes nature of these challenges. These costs are closely linked to high severity events because they frequently lead subsequent missed opportunities to, of course, correct the patient's care. As patients care builds forward from the initial assessment that drives decisions around treatment when an adverse event resulting in high severity injury or death is linked to an assessment information that was missed or misleading, the likelihood of a case closing with payments increases. As such, you see our study illustrated the prominence of patient assessment errors in diagnostic cases and across surgery and medical management as well. So in one case, there was a 65-year-old female whose primary care provider did not recommend cancer screening, became paralyzed or under diagnosed breast cancer and metastasized to her spine. In another case, a 70-year-old male suffered sepsis and paralysis following a total knee arthroplasty after his orthopedist failed to recognize severe indicators for his post-op infection. So we look at both the inpatient and outpatient events and patient safety and mismanagement perspectives. We know that care processes vary drastically across different case care settings. So the nature and outcomes of patient assessment failures in the ambulatory setting differ significantly from that on the inpatient setting. So when we studied those cases through this lens, we could more clearly see that the largest single impact of assessment failures in the ambulatory setting were cases alleging a diagnostic error. And nearly half of those involved a miscancer diagnosis. We also looked at what else was happening, those other contributing factors in these cases, to see what was going on. You know, delayed referrals, a selection and management of treatment, technical performance of a known complication, again the third party involved that handoff and communication regarding a patient's condition. When we see that system that system factors around delayed referrals, communication factors run external providers regarding the conditions were present, our study confirmed that other vulnerabilities in ambulatory care settings can exacerbate these cases. On the inpatient side, while there's slightly more variation in case types, the largest scuba cases for whom patient assessment is challenging are cases involving surgery. And looking at the covers a bit, it's namely management of a surgical patient where assessment factors dominate, where top injuries include sepsis and other infections, perforations, accidental punctures. When we look at the concurrent factors in these cases, they too are complicated by communication among providers and caregivers, awareness, monitoring, and the management of initial complication. So again, you know, how do we look at these to reduce and defend these cases with patient assessment failures? You want to look at what does your history and physical look like? Do you have a concise patient profile? Have a guide for a management or differential diagnosis list, promote smarter testing, review the role of decision support tools, and address the operational environmental factors that may narrow the assessment process. What we're really good at doing is adding more things to click on and to review, but we're not good at taking away the things that we that don't add any value. So when defending, you want to look for the expert reviews and test vulnerabilities that will help gauge clinical judgment related to the patient assessment, whether the clinician had access to adequate information when those decisions were made, and evidence in what happened before and after our key points to the assessment process. So our last key point involves documentation, and it represents a persistent dilemma for healthcare providers. When there are indications that documentation of care or any encounter was inadequate to ensure appropriate care by the next provider or to even guide the patient's decision-making regarding the course of their care, the odds that a case closed with payment increased by 76%. Our study found that nearly 60% of cases with insufficient documentation close with indemnity payment versus only 30% for all of our other medical professional liability cases. And they averaged a payment of $430,000. And we want to pay close attention to the cases where this is identified as an issue because it offers mitigation opportunities for claims managers and education and prevention opportunities for patient safety leaders. So our overall cases where documentation or insufficient documentation was an issue, it comes from a broad range of services and are more likely to derive from the outpatient setting. However, a substantial portion of these 40% of the cases derive from the inpatient setting and most frequently from the patient rooms or the operating room. So both the complex episodic nature of ambulatory care and the team-based nature of inpatient care both rely on upon documentation that enables clarity for all providers about the patient's past, present, and future care plans. When a provider is missing critical information, regarding assessment, test results, or evidence of a patient visit, or a phone call, subsequent management can be compromised, driving diagnostic failures, and errors in surgical medical management. And not only are the payments an average of nearly $400,000, almost half of these involved high severity or death. So a couple examples of this, a 57-year-old male colorectal cancer diagnosis was delayed in part due to confusing documentation over prior colonoscopies. And another example, 71-year-old male- suffered paraplegia following a post-op complication associated with incorrect documentation regarding anticoagulation therapy. So clinical documentation that enables country of care is only one part of the story. When basic standards of documentation aren't met, failing to document and inform consent conversation or failing to document multiple efforts to contact a patient for follow-up, these can be extremely problematic from a defensibility perspective, even when the care and question was unmarred. In fact, 90% cases with breakdowns and documentation standards have a 90% higher odds of closing with payment than cases without these types of failures. So here's another example, a 66-year-old male's allegation of a failure to diagnose colon cancer was complicated because his multiple refusals for recommended digital rectal exam and colonoscopies were not documented. Documentation of informed consent is a dominant defensibility issue and comprise about 42% of the defensibility documentation failures that we see in this group. And not surprising, these mostly involve surgical services, but can involve proceduralist, and so they are available to these issues as well. And from a payment perspective, the dark blue on the right, where informed consent or lack of documented informed consent is involved in a as you'll see here in the OBGYN cases, where there was only 9% of cases, 8 yielded $449,000 in inteminative payments. So you can see the importance of the informed consent documentation. So of course, this is a persistent dilemma for healthcare providers, what should be documented, and for what reason, to support ongoing care or to defend prior care. The adoptions of electronic health records have only complicated things, where once providers were in tors for handwriting issues that has been supplanted by the vulnerabilities of the cut-and-paste mode, user unfriendly templates, and too much buried into many places, and not having that one source of truth. When considering interventions to reduce the risk associated with clinical documentation, these should envision the problem. From the perspective of the subsequent provider, as well as the patient themselves, can they clearly understand their clinical path. And when doing seeing the risk associated with defensibility documentation, motivate vigilant documentation, it doesn't mandate irrelevant targeting. Consider chart audits that can and use case examples to demonstrate about the conundrum and a necessary legal risk that could jeopardize good medicine as a practice. So what now? So you want to start examining your own data, as we said, examining your system-wide opportunities, examining your readiness for change, and take a look at your peers. There are some really high performers and others that your probably risk managers could say could use some help in their documentation or the policy and following policy procedures. So what's your risk to better understand your risk and vulnerabilities? You want to look at those policies and procedures. Are remorse susceptible to policy absences or violations? Where do we see the most significant injuries or losses? Again, the risk management and patient safety team at childrens can help you with that. And our certain services or settings frequently cited for policy violations. It will be interesting to see what are the Department of Surgery's vulnerabilities? To better understand the org's patient assessment vulnerabilities, again, which misdiagnosis are the most problematic? Which categories of tests are most commonly not ordered or delayed? And which specialties or consults or referrals are most commonly delayed? And then what's your risk around documentation? Are we settling cases strictly due to missing documentation? Is documentation a prominent factor for a particular service or setting? And is documentation regarding specific types of medication, procedure, or communications commonly missed or inadequate? And these are things that you can derive and get from your quality and safety department to really take a deep dive into what kind of events are happening in the surgical space. Sorry about that. I'm going to and we'll just end with this last slide. Sorry. You know, we like to say this. It's far better to foresee even without certainty than not to foresee at all. So really having that proactive approach to looking at what your vulnerabilities are, are always key to patient safety and staff safety. So with that, I will take any questions. Thanks a lot. This is Joe Corvall. I'm the Chair of the Anesthesia Department. I'm just wondering, particularly with respect to the protocols and policies, we have so many protocols and policies and I'm sure people inadvertently do not abide by them on a regular basis. In your experience, I can see in a lot of different organizations, there are specific organizational strategies that you've seen that have allowed organizations to have better compliance with specific protocols, particularly those that present risk to patients. You know, I think, I think part of the problem sometimes is that people will put a new policy in place and not remove the old policy. I think really taking an audit of what your policies are is very helpful. I've worked in hospitals where you know, everybody keeps adding and nobody keeps replacing and then you have two different policies that put you at risk. I think that you really want to have a team together to review those. You want to see what's pertinent and what's not. You can have that one source of truth. I think there are tons of policies. There's a policy for just about everything it seems these days. I haven't, I can't say that I've seen a best practice of how to do this, but really good policy management and really including every single person who that policy touches to decide things, because you can't have someone in a policy management office saying, well, this is how we're going to do it without having the clinicians involved. Thanks. I can also send these slides to Katherine and Hester and at the last, next to the last slide, has the link to the CBS report. If you haven't seen it already, the full report. Any other questions? Letting me have an hour of your time to review the report. Please, if there's any other questions that you have, feel free to reach out through Katherine to me and I will gladly answer those.