09 May Standardizing Medical Care By Patient Complexity Grouping
MedicalResearch Interview with:
Dr. David Cook MD
Professor in the Department of Anesthesiology
Division of Cardiovascular Anesthesiology
Center for the Science of Health Care Delivery
Mayo Clinic College of Medicine
MedicalResearch: What are the main findings of the study?
Dr. Cook: The main finding of the study was that segmentation of a population of surgical patients into groups of higher and lower complexity allowed us to apply a standardized practice, focused factory model to surgical care delivery. A standardized care model improved care process measures such as time on mechanical ventilation or duration of a bladder catheter indwelling. The model reduced resource utilization, decreasing patient time in all care environments (operating room, ICU and on ?the floor?). The care model improved outcomes at 30 days and reduced the costs overall and in every care environment. In addition to the absolute improvements in quality and in cost, the standardized care model reduced variation in all measured variables. That reduction in variation may be even more important than the improved outcomes or reduced costs because we now know it is possible to make the health care experience predictable for these patients. That predictability is critically important to patients and providers, but it also has implications for health care metrics and payment models.
MedicalResearch: Were any of the findings unexpected?
Dr. Cook: In the manuscript we describe a complexity analysis. We used goods and services consumed by each individual patient as an index of their overall complexity of care. Then, we stratified about 2,000 cardiac surgical patients (cared for in 2012) from least to most complex. We found that the complexity index in more than half of the patients was relatively uniform. In contrast the most complex 20 percent were extremely complex, some having a complexity index of 8-10 times our median. This type of analysis points to the percentage of a population for which solution shop care, or conversely, a standardized care model, is most appropriate. Preliminary analysis of complexity in other practices suggests the structure of those practices may be very similar. This has implications for practice design and the distribution of patients to the appropriate type of care model and care environment.
MedicalResearch: What should clinicians and patients take away from your report?
Dr. Cook: Clinicians are often wary of standardization. There are many reasons for that wariness, but one is anxiety that a standardized care model or protocol may be applied inappropriately to those for whom it is not well suited. They are also concerned by decision making that is not at the point of care. Our results show the power of a standardized practice model, when applied appropriately. At a minimum, our work shows that a segment of the surgical population is very complex; for them a comprehensive, protocol-based care model is not appropriate. We were careful not to force patients into a care model for which they were not suited. At the same time, it is evident that standardization may improve care value for more than 60 percent of patients. The other clinician takeaway from our report is that our model was at least as safe as the conventional care model. For patients who received the standardized pathway, there was no increase in adverse outcomes such a respiratory failure, ICU readmission or hospital readmission. In fact, the care model was associated with improved 30-day outcomes. For patients the message is different. That message is that health care can be predictable, and it can be so reliable that, for most patients, we can be confident in telling them what their care expectations will be. I think there is also a cultural message here for patients, and that is that lower- cost care models can deliver excellent care.
MedicalResearch: What recommendations do you have for future research as a result of this study?
Dr. Cook: The implication of this work is that the care model, and potentially the care environment, might be chosen on the basis of the predicted complexity of patients? medical status. We distributed patients to the care model based on a priori (pre-designation) and post hoc methods (physician review for continued suitability at the end of operation). That stratification to the care model (standardized or solution shop) was imperfect. Fortunately, our Mayo Clinic hybrid model allows patients the advantages of having both models there to support them. Ultimately, patients will benefit from the complementary nature of mathematical risk analysis and physician judgment. Some of the most important opportunities for research lie in developing better mechanisms to predict who is best served by a standardized vs. non-standardized care model. We need better tools to discriminate the boundaries of these populations so that we can deliver the right care to the right individual in the right environment. Another area of research will involve developing and evaluating new payment models based on population segmentation and the greater predictability of process, outcomes and costs that can arise from that.