Welcome to the Abstract Variation and Outcome Measures in Lumbar Spondylolisthesis Patient Registries and the Need for Harmonization. On behalf of my co-authors and the multi-stakeholder group that have been coordinating around this program and our funding agency, the Agency for Healthcare Research and Quality, I thank you for listening. The Outcome Measures Framework is originally a model developed as a common conceptual way for classifying the range of outcomes that are relevant to patients and providers across most conditions. It was developed using a multi-stakeholder-driven process that included, by this point, over 600 different stakeholders, incorporating many different rounds of reviews and revisions and use and application across multiple different types of disease entities and treatments. The model itself is generic and seeks to understand what outcomes are relevant to clinicians and patients and can be standardly measured. This model basically breaks down into three components, characteristics, treatments, and outcomes. If we look at the outcomes on the right, we see the five core domains that have been common across all the conditions that we have studied to date. They include survival, which is whether a patient survives and also whether or not the survival is related to the disease or condition or its treatment. Clinical response, which differs from condition to condition, meaning it's different in cancer than depression than asthma than surgical procedures. However, it reflects how a clinician and a patient views whether they are improving or worsening and often is described or monitored within a condition in a consistent way. Events of interest are the things we generally count, like complications, adverse events, and other coded information. Patient-reported outcomes can have an impact on clinical response, as it does in lumbar procedures, but it is a domain in and unto itself. And then resource utilization or cost would be the fifth of those domains. As we look to the left, we see ways that we have to think about risk-adjusting those outcomes. For treatment, it's the type of treatment and the intent of treatment that might vary. And characteristics are also critically important. Characteristics that the patient brings to their condition prior to developing the condition, other comorbidities, age, exposures, and so forth. The characteristics of the disease itself, its severity, and so on. And where the treatment is taking place and who is performing the treatment, because the experience of the provider and so forth, as we know, can have a direct impact on patient outcomes. In this program, we had defined a lumbar spondylolisthesis working group whose goals were to develop standardized library of outcome measures in lumbar spondylolisthesis. First, it would identify a minimum set of outcome measures that should be collected in all future registries. Second, agree on definitions for each measure in the minimum set. And third, translate text definitions into standardized terminologies for use in electronic health records and other data collection systems. The goal ultimately is to answer the following questions. Is it feasible to reach consensus on which outcomes to measure? To harmonize those measure definitions? To identify the barriers to harmonization and whether there are outcome domains where new measures might be needed? And then ultimately, how should harmonized measures be disseminated? Methodologically, it was a six-step process. We collected outcome measures from different registries and other relevant efforts, including those by the NIH. We categorized the measures using the Outcome Measures Framework and built a proposed minimum measure set. We then endeavored to harmonize definitions for measures in the minimum set, to identify key characteristics to support the risk adjustment I described earlier, and then to produce a final standardized library. Participating registries and other data collection efforts were vast. We mapped a number of different registries and similar programs that were generous in the use of their measures, including the American Academy of Physical Medicine and Rehabilitation, the AOS Spine North American Research Network, programs from both payers and industry, as well as the efforts of the North American Spine Society and the NeuroPoint Alliance QOD program. Participating registries and others represented a number of different purposes and populations. It's important to consider that when seeing the measures that were actually used. They ranged from quality improvement, maintenance of certification, to safety monitoring, to compare effectiveness, and so forth. When we actually compared the outcome measures that we collected to the Outcome Measures Framework standard template version, we could see how those measures fall into those categories of those five domains I described earlier. The greatest number were categorized as clinical response. The second were those that were patient reported. The use of interest was also common, and not surprisingly, those around survival and resource utilization were a bit less common. The challenges in harmonizing these measures to the level of clear, consistent definitions demonstrated by this example. Step one in our process for harmonization was to identify existing definitions and prepare comparisons for workgroup discussions. We in the workgroup participated by reviewing examples like this across all of the different measures we were looking at. In this example, on the left, we have a postoperative complication definition from the QOD registry. On the right, we have one from the NAS spine registry. While similar, there are subtle differences that can impact how these measures are actually rolled out. The one on the right divides the complications into intervention-related and those that are systemic. The one on the left doesn't. In other areas, there are subtle differences in the terminology. So on the left, we see urinary tract infection. On the right, we see urinary complications. Similar, but not the same. So these are the types of things that the workgroup had to work with to achieve harmonization. The other steps in the process were to create harmonized definitions with workgroup feedback, post those definitions for anonymous editing in a virtual platform, and then repeat steps two and three until the group arrived at a consensus definition. This is just an example of the list that we are generating for final publication of characteristics that need to be considered for risk adjustment and treatment types. None of these will be a surprise to anybody in the field. At this point, I'd like to provide several examples of the measures or the common harmonized measures that the workgroup completed, one of which was in the area of clinical response, which utilized both patient-reported and other measures. One of the key areas is pain interference. So pain interference was found to be critically important, and it was recommended that the PROMIS domain would be utilized for it. Also, in clinical response, it was felt that morphine-equivalent pain medication, meaning collecting the type, duration, and dosage, and at specific time intervals would have great value in the collection for these measures across most efforts. In the domain of events of interest, there were a number of adverse events associated with surgical and non-surgical treatments, some of which are shown here and will be familiar to many of you. In some cases, there were measures that were identified which were not felt to be necessary in the common minimal measure set but could be used for supplemental data collection, as shown here, the work productivity being measured by WPAI that could have value in certain efforts. In summary, the Outcome Measures Framework is a common model for developing outcome measures that are relevant to patient and clinicians across multiple use cases in research and in clinical practice. This effort applied that model to lumbar spondylolisthesis. As we demonstrated, significant variation exists between outcome measures in lumbar spondylolisthesis that require harmonization in order to develop standardized measures for research and practice. As the examples demonstrated, even measures that are similar in domain may differ subtly or significantly in definition and time frames. A multi-stakeholder iterative approach to developing consensus was successful in achieving harmonization of outcome measures for lumbar spondylolisthesis in this study. Thank you.

harmonization

outcome measures

lumbar spondylolisthesis

patient registries

Outcome Measures Framework