CORR Insights®: What Are the Migration Patterns for U.S. Primary Total Joint Arthroplasty Patients?

Where Are We Now?

Since 2012, the American Joint Replacement Registry (AJRR) has accumulated hip and knee arthroplasty procedure records from a variety of institutions in the United States [1, 2]. With almost 1.5 million U.S. procedures currently registered, and a 38% growth in the volume over the past year, the AJRR is now the second largest joint replacement registry in the world behind the UK’s National Joint Registry [2, 8].

But registries are only useful if they capture and report standardized clinical and administrative variables within the perioperative environment as well as long-term outcome data following the index procedure like reoperations or revisions. The AJRR is positioned to capture enough data on short-term quality and adverse events associated with primary hip and knee arthroplasty to enable its use as a reliable benchmark for implant tracking, quality initiatives, or pay for performance. However, the U.S. healthcare system presents a formidable challenge to achieving the necessary longitudinal data capture because of the country’s widely disparate provider networks, disconnected data systems, and highly mobile patient population across a massive geography.

Within this context, Etkin and colleagues [5] analyzed Medicare fee-for-service claims and enrollment files spanning 11 years to determine the magnitude and patterns of geographic migration within this large subpopulation of U.S. hip and knee replacement patients. They found that, after 5 years, nearly 10% of patients relocate (migrate) either out of state or somewhere within their state but outside of their original county. At 10 years, roughly 9% of patients will have relocated to another state. Additionally, the authors identified that older age, higher comorbid status, and residence in Western states correlated with a higher likelihood of migration [5].

This study confirms that patient migration represents a major barrier to long-term data capture and implant tracking for either local or national registries. With a better understanding of migration patterns, registries can adjust revision (or other long-term adverse event) rates, as well as prioritize specific geographic regions for capturing long-term followup on migrating patients.

Where Do We Need To Go?

Although the United States contains a highly mobile patient population, data presented by Etkin and colleagues [5] suggest that not all migration has the same impact on effectiveness of the AJRR and raises questions about which subgroups of migrating patients as defined in this study have less clinical impact on the reporting capability of the registry. One question is how to interpret the in-state but out-of-county migration that contributes to roughly half of total migration at each time period studied in the current study [5]. With the evolution toward episode-of-care models and a center-of-excellence orientation, patients frequently bypass local hospitals for joint replacement, traveling longer distances for their procedures [4, 6, 9]. While published data still demonstrate that most patients remain in their local health service area, patients in suburban and rural areas frequently have joint replacement at hospitals 25 miles to 50 miles from their homes [3, 4]. This suggests that in-state migration rates may not be associated with major losses of longitudinal data capture, because patients often travel outside of their local area for their index procedures. Travel distances can complicate the interpretation of in-state migration, which is more pronounced among younger, more active, patients in whom long-term implant surveillance matters the most [6]. Based on this, it would be of interest to exclude the subpopulation of patients migrating to adjoining counties or within 50 miles of their index procedure to account for the probability that meaningful longitudinal data for these patients are less likely to be lost with such a limited relocation. Another subgroup of migrating patients to target for a future study should be those 80-years-of-age and older. Although it would be unexpected to find that this age group has a high risk of migration, it could be reasonably speculated that frailty and end-of-life issues requiring a move to the proximity of a caregiver may be drivers of higher migration in this population. This is supported by life-table data showing 1-year death rates between 5% to 17% and average life expectancy between 7.9 and 3.9 years for men between 80 to 90 years old [7]. If reduced functional levels coupled with much higher mortality in this age group are considered, discounting the impact on loss of longitudinal data capture for this subgroup of migrating patients in the registry would be reasonable. Further studies will be necessary to quantify the age cut-off or a combination of age with other factors that justify discounting the impact of longitudinal data loss in specific patient subgroups in the registry.

How Do We Get There?

Etkin and colleagues [5] have provided a useful analysis of patient migration rates for Medicare hip and knee arthroplasty patients that spans 11 years. Juxtaposing these findings against the relatively short timeframe over which most adverse events and revisions are historically reported supports the case for a parallel strategy in ongoing registry recruitment. One option is to target large institutions in key geographies in the United States based on the dominant migration patterns. These locations may be identified on the basis of either having particularly deficient registry coverage or having a relatively high volume of patients who migrated from other registry institutions where they had index procedures. Meanwhile, it seems that focused recruitment efforts should continue for institutional sites with large volumes that have existing standardized registry data. The AJRR has already incorporated state registry data from California and Michigan, and I anticipate that this will continue. In addition, continued emphasis on growing partnerships with leading outcomes data management vendors, outpatient surgery centers specializing in arthroplasty, and high-quality institutional registries will offer leverage for growing both procedure volume and meaningful procedural reporting within the AJRR. This might be accomplished through offering organizations greater access to the registry for queries, reporting, and benchmarking in return for their time and resources in terms of entering their patients’ results into the registry. With appropriate organizational incentives and technical interfaces to such data repositories, I believe even greater patient volumes with standardized data and relatively complete procedural reporting could be acquired in the future.