Accountable care highlights the value of CLG
Two initiatives to build a more accountable and higher performing health care system, Accountable Care Organizations (ACOs) and Patient Centered Medical Homes (PCMHs), are rapidly being promoted to improve a U.S. health care system that is inefficient and unnecessarily expensive, often uncoordinated, and prone to episodes of inferior quality and error. Succeeding in each initiative requires aligning financial incentives around better outcomes, not increased volume of care; producing more detailed and reliable information; and facilitating patient care coordination across the care system.
These success factors demand new ways to access and examine information. Clinical Looking Glass (CLG), developed at Montefiore Medical Center, offers a breakthrough approach which has led to improvements in outcomes, cost, and care coordination. It empowers clinicians on the ‘front-line’ with robust tools to examine gaps in care and gain new insights into the care process. CLG also offers a patented process for examining delivery of care over time and location thereby creating longitudinal intelligence as it generates valuable analyses of patient care processes and outcomes for use by clinicians. Additionally CLG supports a culture of clinical decision making informed by data and leverages the intellect and creativity of clinicians in improving care.
For over 10 years CLG has been producing the type of information required by ACOs and PCMHs in a way not achieved by any other product and this information has led to repeated successes. By harnessing its analytic power, CLG offers new insights into potentially preventable events and ways to improve clinical outcomes. This information applied in an ACO or PCMH setting can help create a new vision for what health care can become - a better coordinated, results oriented and aligned health care ecosystem.
For example using CLG to reduce health disparities and close gaps in care those served by the South Bronx Health Center (SBHC), a program of the Montefiore Medical Center and the Children's Health Fund which targets a population where almost half the children seen live in poverty and many of the adult patients are uninsured, clinicians have been able to achieve substantial breakthroughs in performance. The following chart demonstrates some of the superior results which have been achieved using CLG – results which even surpass Healthy People 2010 targets.
South Bronx Health Center Performance Using Clinical Looking Glass
||Healthy People 2010 Targets
||SBHC 2010 Performance
||Comparison Group Results
at Montefiore Medical Center
|Immunization rates for children <2
Patients with commercial insurance
|Cervical Cancer screening rates
Patients with commercial insurance
CLG is an “essential tool” for organizations committed to improving care according to SBHC Senior Medical Director, Alan Shapiro M.D. “CLG allows us to systematically look at health outcomes for the entire universe of our patients and not just a sample. We then develop reports that allow us to target non-compliant patients and engage in direct outreach to them. This gives us the opportunity to check in with them, refill medications, order lab tests and perform exams. Patients feel we are really looking out for them and it motivates them to better take care of themselves. In the end we see dramatic changes in our health center’s clinical performance measures which of course translate into better health for the community we serve.”
Empowering the Clinician
Leaders of a PCMH or an ACO cannot anticipate all the places where gaps in care will occur. To find and close these gaps leaders must engage clinicians in continuously identifying and investigating opportunities. This process of local discovery, local problem solving requires an entirely new approach to making information available to physicians, nurses and others involved in the care of patients.
CLG is a tool to support data access which any interested clinicians can be trained to use. It significantly reduces the need for and cost of a cadre of statisticians and programmers to produce desired reports. Furthermore CLG informs the care process at or close to the time care is delivered.
Why is a new approach needed? It has been virtually impossible to get useful patient care reports when the clinician wants them with current IT report generating processes. To get information the clinician has to prepare and submit a request, carefully define the population to include, defend the priority of the request, wait for the data to be run, and, in many cases, when finally getting the initial report, go through the process again … and again to address new questions that are raised by the data.
How can clinicians pressured to serve patients and improve productivity and quality be expected to go through this arduous process? They did not … until CLG was introduced. Now CLG-informed clinicians get actionable insights in minutes. Also they can continuously modify the parameters for patients to be studied, timeframes to be included or excluded, and outcomes to look for and then rerun the study. This ability by CLG users to iteratively refine in real time the cohort to study, key definitions to use and the outcomes to seek dramatically simplifies the effort to get at useful information and more than pays for itself in the usefulness of what is uncovered.
For example in an ACO there is financial value in controlling high blood pressure. The CDC estimates the average lifetime savings of high blood pressure control at almost $1000 per patient. Using CLG to get at blood pressure control improvement opportunities at SBHC, clinicians have achieved a 15% greater control for blood pressure patients than the average for the rest of the organization. Thus CLG has had a positive impact on quality and cost.
CLG has also been used to monitor Hospitalist effectiveness in lowering lengths of stay for inpatient admissions. The study was designed to identify any adverse outcomes from shorter lengths of stay. By turning to CLG, William Southern M.D., the Founding Chief of Hospital Medicine, proved that “using Hospitalists was effective in reducing lengths of stay without adversely impacting readmission rates or mortality”.
In both of these examples, the study was designed and the results generated locally, i.e. in the department, not in IT. Furthermore since the clinical staff has easier and timelier access to the data, they become much more engaged in efforts to improve the care process.
Over 700 clinicians have been trained to design and conduct their own CLG “study” during two 3-hour training sessions. With this training, front-line clinicians and their managers are now empowered with powerful investigatory tools at their fingertips to produce their own clinical reports in minutes. Thus clinicians can better take on their new responsibilities in an ACO or PCMH – to identify gaps in care and improve care coordination for their patients - using CLG’s data analytic platform.
ACO and PCMH organizations now can offer a whole new, accessible approach to using clinical data – one which informs clinical decisions by making patient-specific, clinician-specific, disease-specific and outcome-specific data easily available to any interested clinician. Clinical Looking Glass has provided such an approach since its inception. It has empowered clinicians to ask the right questions and use CLG to get at the answers. This has proved to be a unique and effective success formula!
Creating Longitudinal Intelligence
Responsibility for care coordination over time, i.e. longitudinal care responsibility, requires an advanced form of clinical intelligence. Since each person enters the care process at a different point in time, clinical data must be examined specific to an individual’s experience with the care system and in the aggregate across time periods and groups of clinical providers of care. The capability to properly analyze this data becomes even more important as ACOs and PCMHs reward providers by patient outcomes and less by volume of services delivered. CLG provides far more power than any other product in looking at issues of time and desired outcome. It helps clinicians determine which patient population to analyze for which outcomes by an analytic process which is much more logical than traditional methods as relates to time.
For example for a patient seen first in January, the clinician had a much greater opportunity to improve care within the calendar year than did one seen first in December. CLG addresses this issue of temporal complexity through a patented (US Patent # 7917376) analytic process – where the start date (or index date) is patient-specific. This is essential because each individual enters the care process at a different time so proper analysis requires rolling enrollment. In identifying gaps in care the analytic process must look for results while the individual is under care and exclude periods irrelevant for that person.
CLG-empowered clinicians can more easily track outcomes. For example clinicians may want to intervene differently for groups of patients who have Hemoglobin A1c readings below 7, between 7 - 9, and above 9. Within minutes these clinicians can find out how many and which of their patients fit into each grouping.
In evaluating system performance, clinicians want to know whether traditional patient care is being effective. For example CLG allows the clinician to know whether patients who had HbA1c readings above 9 a year previously and have had at least one visit over the subsequent 90-365 days actually showed an HbA1c reduced to a target of 7 or below.
CLG is especially effective in looking back in order to prioritize efforts going forward. This capability is the essence of longitudinal intelligence. It is most useful in determining when, where and for whom to intervene.
At SBHC the clinicians decided that they wanted to give priority to those patients with diabetes whose results were high on three measures - HbA1c greater than 9, blood pressure greater than 130/80, and LDL-cholesterol above 100. CLG was used to produce lists of patients whose results in these areas were of clinical concern - which led to outreach efforts to get them to come in for care. The outcome: As the following table shows, by the end of 2010, in each of the three targeted areas the percentage of SBHC patients falling within desired ranges far exceeded those of the Montefiore patients with Commercial insurance.
Quality Measures in Treating Patients with Diabetes Mellitus at South Bronx Health Center Compared with Montefiore Patients with Commercial Insurance
|Type of Patient by Quality Measure
||Montefiore Patients with Commercial Insurance
|Diabetes Patients with Blood Pressure <130/80
|Diabetes Patients with LDL <100
|Diabetes Patients with HgA1c >9
(lower % is desirable)
Another example of the benefit of longitudinal intelligence comes from identifying and helping patients with diabetes before their condition progresses to where an admission becomes necessary. The following chart shows the value of good diabetic control on hospital admissions. After one year of poor control for patients with diabetes the hospital admission rate is roughly 1/3 higher than those with good diabetic control, with the gap widening throughout the year. CLG allows the physicians to know where to put their attention. Since the ACO or PCMH participates in the savings, the financial impact for clinicians can be enormous – leading to further reasons for their engagement.
Hospital Admission Rates as Function of Diabetes Control
Clinical Looking Glass has been used to monitor process efficiency, measure quality outcomes and identify priority areas for focused intervention. Assessing gaps and opportunities for improvement require the ability to look at data in a patient-centric way over time and across provider types, not in a calendar determined way. CLG’s process for addressing temporal complexity can lead to change at the patient and program level. In this way CLG is a tool for creating longitudinal intelligence.
Building a Culture of Individualized, Informed Patient Care
Imagine a conversation among clinicians about why the blood pressures of patients with the most primary care visits are the worst controlled, with only 1/6 of all high utilizing patients with high blood pressure showing a BP within the target range a year after their initial visit. Or consider a discussion of how to develop new programs to address asthmatics in three census blocks because it turned out that is where 53% of all patients with a diagnosis of asthma live. Think about how productive it might be for the clinical team to discuss why over 50% of patient blood sugars are still not under control within six months of an initial diagnosis of diabetes. These are all conversations which have been facilitated by CLG.
What will stimulate these types of conversations in a PCMH or ACO? They will be triggered by the intelligence of clinicians and the data that only CLG can create. With that data, conversation can lead to action – program change that promotes improved outcomes.
For example, in still another study using CLG to improve care for people with diabetes, CLG helped identify that 58% of diabetics were not being successfully controlled. This data was used not only for patient identification, but also to target outreach activities. Within months the percentage of uncontrolled diabetics dropped by over 1/3. Enhancing care through better coordination does not just happen. The health care industry is structured vertically - with each specialty having its own clinical guidelines and professionals who are trained to look at care from the perspective of that specialty. Sometimes these vertical perspectives are in conflict.
Care coordination necessitates breaking down these silos and improving care across specialty areas. With information from CLG these inter-specialty conflict zones are more readily identified and more effectively managed.
Consider the case of the geriatrician who was concerned about complications from the use of general anesthesia in certain gall bladder patients. She approached the surgeons who were also concerned – but about the negative impact on patient outcomes if general anesthesia was not used. The surgeons wanted to have general anesthesia become the standard of care. How was this standoff to get resolved – with both specialty groups genuinely concerned about the best interests of patients, but coming to different conclusions? How many meetings? Over what period of time? And toward what end?
The clinicians turned to CLG. They looked at the outcomes for gall bladder surgeries for patients who had general anesthesia and those who did not. Within minutes the study was designed and the data available to these doctors. It showed that there was no difference in outcomes! The surgeons were satisfied – and dropped their demand for general anesthesia.
The opportunities to improve care coordination across the entire spectrum of care require new tools and information – both to help improve the traditional health care system and to reach out to the people it aims to serve. Clinical Looking Glass is a vital tool for supporting a culture in health care which individualizes the patient care process and better informs all participants.
A Peek at the Future – Using CLG to inform and engage patients
ACOs and PCMHs are expected, through program redesign, better information and more effective financial incentives, to encourage clinicians to adopt new ways to coordinate care and improve outcomes. Many clinicians recognize that a number of high frequency and high cost conditions such as cardiac-related and diabetes-related illnesses are behavior dependent and therefore significantly under the control of patients. Thus patients need to be viewed as more than the objects of care – rather they must become fully engaged participants. Usual primary care must evolve to patient-centered care whereby patients become informed decision makers regarding their care. What information will patients need to participate as fully as they are able? Once again the power of CLG can be a breakthrough contributor.
Consider the hypothetical case of a newly identified pre-diabetic. Rather than merely explain the risks of diabetes, his physician might open a CLG study on her mobile tablet which allows her to show this patient what happens to newly diagnosed pre-diabetic patients seen in the practice for the past five years with a diagnosis of pre-diabetes. If a clinician did this exercise it would show that, after two years from the date of diagnosis 11% of patients had a subsequent diagnosis of diabetes and that percentage increased to 17% after three years.
Using real, practice-specific data to inform the individual about patients like him might serve as a “wake-up” call. Moreover it might deepen the doctor-patient conversations about the behaviors which cause diabetes and actions leading to prevention.
In the PCMH or ACO, informing and engaging patients in new ways will be important to success. These conversations might range from actions to reduce the risk of 30-day readmissions to long term ways for promoting behavior change by the chronically ill. Clinical Looking Glass is a unique tool to frame and facilitate these new, critical doctor-patient collaborations in ways that are not available through other products.