3062 — Small Area Analysis of Psychiatric Admissions in Northern New England

Author List:
Watts BV (White River Junction VAMC)
Weeks WB (White River Junction VAMC)

Objectives:
There is evidence that psychiatric admission is subject to considerable variation which may indicate inconsistent admission criteria or inefficient resource utilization. The development of hospital service areas using small area analysis is useful in examining geographic variation in medical and surgical care among various hospitals. However, the techniques of small area analysis have not been applied to psychiatric admission rates.

Methods:
We used patient ZIP Codes from psychiatric admissions to hospitals in Maine, New Hampshire, and Vermont to develop psychiatric specific hospital service areas (PHSA) within the three states. We used census data from the three states to calculate an admission rate for each hospital service area. We then compared these admission rates with the number of psychiatric inpatient hospital beds per population for each area.

Results:
We identified 25 PHSAs within the three states; the mean localization index for these PHSA was 0.69, which is comparable to those found in HSAs developed for medical and surgical services (0.45-0.87). Each PSHA had an average population of 156,742 (range 34,674-485,423), 116 hospitals located within it (range 1-8), and 736 psychiatric beds within those hospitals (range 6-124). Across these 25 PHSAs, the rate of psychiatric admissions varied from 2.4 to 13.4 admissions per 10,000 residents. Admission rates were strongly correlated with number of psychiatric hospital beds (r=0.72, p<.0001).

Implications:
We were able to use psychiatric discharge data to develop psychiatric hospital service areas that have similar robustness to other hospital service areas. Similar to findings from medical and surgical admissions, supplier induced demand may partially explain the variation in admission rates.

Impacts:
As part of CARES, VA is exploring expanding the availability of psychiatric beds through community contracting. Policymakers should consider that admission criteria may vary in different geographic regions, and that the very private hospitals likely to have additional, vacant beds for contracting may admit patients at higher rates.