Measuring access to health care has been traditionally done with waiting times, that is, how long a patient waits between requesting an appointment and actually seeing their provider. The Veterans Health Administration (VHA) acknowledged the limitations of this metric and aimed to better grasp what timely access to quality care really looks like for Veterans with the passage of the MISSION Act of 2018. In particular, Section 401 of the law required VHA to develop a measure of underservedness that would holistically assess access at every Department of Veterans Affairs Medical Center (VAMC).
New Research
In partnership with VHA’s Office of Integrated Veteran Care (formally, the Office of Veterans Access to Care), the Partnered Evidence‐based Policy Resource Center (PEPReC) responded to the MISSION Act’s mandate to measure and mitigate underservedness at each VAMC. PEPReC outlines this new evidence-based approach to modeling access to care in a commentary published in Health Services Research. In it, PEPReC explains how to identify underserved VAMCs by using the econometric principles of supply and demand.
PEPReC defines underservedness as an imbalance between the supply of VHA care and the expected Veteran demand for VHA care. In the first year after the law’s passage, PEPReC developed a statistical model to measure underservedness in primary care. (PEPReC has similar methodology ready for implementation in specialty care as well but those models have not yet gone live.)
Model and Variables
The model is centered around new patient waiting times, but takes a step further by considering the various factors that can impact waiting times. These variables and their relationships to waiting times are explained by being assigned numerical weights. Some increase waiting times while others decrease waiting times; some impact waiting times a lot and others not so much. A larger weight indicates more influence over waiting times.
The model includes 21 variables to comprehensively capture VHA supply of care and Veteran demand for VHA care. For example, on the supply side, PEPReC includes clinic capacity and clinic efficiency, measures of staffing and productivity, respectively. On the demand side, PEPReC includes the household median income in the area surrounding the VAMC, the percent of Veterans who also have private insurance, Veterans’ Nosos risk scores (a measure of how sick they are), and the percent of Veterans who are 65 or older.
Scores and Facility Rankings
The model estimates the relationship between the variables listed above and raw waiting time data and creates an underserved score for each VAMC. The higher a VAMC’s score, the more underserved they are. In other words, the higher a VAMC’s score, the more they may struggle to provide timely access to quality care for their Veterans.
After running the model and producing underserved scores each year, VAMCs are ranked relative to each other.
Impact
The underserved scores for all VAMCs are shared with the Office of Integrated Veteran Care. From there, the VAMCs with the highest scores are notified of their underserved status and are required to submit action plans explaining how they plan to mitigate underservedness in the coming year. The most underserved VAMCs and their action plans are shared with Congress via an annual congressionally mandated report.
Each year, the model is refined and the underserved scores are recalculated.
Importantly, this novel approach to measuring access to care can be used beyond the MISSION Act. The model and scores allow VHA to systematically identify and address imbalances in the supply of and demand for VHA care through evidence‐based policy making and equitable resource allocation. For example, VHA can used the scores to guide mental health clinic operations forecasting, budget forecasting, and local clinic management.
The approach can also be used by other health systems, too, helping Veterans and non-Veterans alike access the quality care they need.
Understanding and managing access to care for any population is tricky, and the MISSION 401 models are just one way to tackle that complexity.
