BACKGROUND/RATIONALE:
Toe amputations are often naively viewed as an inconsequential surgical procedure with little functional impact. Instead, a toe amputation is often the inciting event in a cascade of progressive loss of function and quality of life. One in three people undergoing toe amputations due to chronic illness fail to heal the amputation in a timely manner, requiring additional amputation surgery(ies) and one in five die in the following year. Inadequate healing may be related to suboptimal management of comorbid conditions. Veterans with limited social support may be further limited in their ability to self-manage their illness and engage in treatment to facilitate healing and prevent additional amputation. Veterans who undergo repeated amputations lose functional life years during repeated treatments and recoveries.

OBJECTIVE(S):
The aims of this study are to: 1) to characterize amputation trajectories and evaluate trends in subsequent amputation and death in the year after initial toe amputation overall, based on demographic characteristics, selected clinical factors (e.g., vascular status), and geographic location (e.g., Veterans Integrated Service Networks [VISN]), 2) to evaluate the associations between modifiable risk factors and 1-year risk of subsequent amputation or death while controlling for important risk factors that confound these associations, and 3) to acquire knowledge, attitudes, and behaviors related to secondary prevention using semi-structured interviews among patients who have undergone a toe amputation and providers who care for these patients to develop future interventions that improve outcomes in these patients.

METHODS:
We will accomplish these goals by using VA electronic medical record data, and in-depth interviews with providers and patients. Our proposed research - a retrospective cohort study including all VHA patients (~20,000) with diabetes with a first ever toe amputation between FY 2005 and 2016 -- will be the largest study conducted to date, national in scope, and will supplement structured data with unstructured data to ascertain key information about not only the amputation but also other relevant patient characteristics. For Aim 1, frequencies and percentages of the following clinical outcomes will be obtained: subsequent amputation within 1-, 6-, and 12-months; counts of subsequent amputations; the final level of amputation at 12 months; and death. We will formally test for trends across time by including year as a predictor variable in logistic regression models, considering the use of piece-wise regression to allow for non-linear trends. Similar models will be constructed to test for variation across VISN. Key modifiable risk factors for Aim 2 include behavioral factors (e.g., smoking, alcohol misuse), revascularization procedure and type, and adequacy of treatment of mental health conditions. For Aim 3, we will conduct in-depth interviews with providers and patients who have and have not undergone a subsequent amputation, to gain insight into patient and system-level barriers to successful healing, policies or other activities that could improve care for patients who undergo a toe amputation, and patient and provider perspectives on factors that they think would lead to improved patient outcomes.

FINDINGS/RESULTS:
We have no findings to report at this time.

IMPACT:
By conducting a detailed evaluation of temporal and geographic variation, identification of modifiable risk factors, and interviews with patients and providers, we will understand system-level and individual-level variation in outcomes. Our study will lay the groundwork to improve amputation prevention at VA.

---

External Links for this Project

NIH Reporter

Grant Number: I01HX002044-01A2
Link: https://reporter.nih.gov/project-details/9293451

---

PUBLICATIONS:

None at this time.

DRA: Diabetes and Other Endocrine Conditions, Other Conditions
DRE: Epidemiology, Treatment - Observational, Prevention
Keywords: Care Management Tools, Monitoring, Natural Language Processing, Predictive Modeling
MeSH Terms: none