  |
|
|
Update on the Physician Workforce August 2000
A. "ESTIMATES OF PHYSICIANS NEEDED TO SUPPLY
UNDERSERVED AMERICANS ADEQUATELY UNTIL UNIVERSAL COVERAGE" A key feature of this paper is the authors' disaggregation of the
United States into five distinct county types, as follows:
The authors' surveyed experts in the field of physician workforce
planning and analysis for the purpose of eliciting their opinion as
to the number of generalists per 100,000 needed to provide (a) an "adequate"
and (b) a "minimal" level of physician availability by county type.
The responses received from those who responded are shown in Table 1.
These standards differ somewhat from the range of 60 to 80 generalist
physicians per 100,000 recommended by COGME in its Eighth Report
(1996).3
The "expert-based" requirements standards were compared to the corresponding
supply figures for 1995. Supply, as in the case of requirements, was
subdivided by county type; however, for counties classified as metro-core,
the authors provided a further breakdown by poverty versus non-poverty
tract, to reflect the reality that physician supply within large counties
is not necessarily uniformly distributed. The distinction between poverty
and non-poverty tracts was defined by the percentage of households below
the Federal poverty line, with 20% used as the defining threshold. The
relevant supply numbers are shown in Table 1.
Comparing these figures to the "expert-average" requirements shown
in Table 1, the authors conclude that there existed in 1995 a severe
deficit in generalist availability in all but the non-poverty tracts
of metro-core counties. Applying the observed deficit per 100,000 for
each county type to the corresponding population count for that county
type, then summing across county types, they arrived at an estimated
deficit for the Nation of 15,441. Had the COGME high figure of 80 per
100,000 been used in place of the expert averages, the estimated deficit
would have been 41,359; had the low figure of 60 per 100,000 been used,
it would have been 7,676.
The authors then conducted a sensitivity analysis to observe how much
greater (or smaller) these deficits would be were the Nation's health
care insurance coverage or staffing patterns to change in prescribed
ways. This was done by applying the Health Resources and Services Administration/Bureau
of Health Professions' Integrated Requirements Model4 to each of the
scenarios defined earlier to determine how many more (or fewer) generalists
would be needed in the year 2005 compared to 1995.
The scenarios, previously defined by an expert workgroup formed jointly
by COGME and the National Advisory Council on Nursing Education and
Practice (NACNEP),5 are described in Table 2.
The model produced the following percentages which ranged from 10.4%
for Scenario 1 to a high of 22.7% for Scenario 5 with Scenario 6 showing
a decline of 2.2%. Little difference in percentage change is seen between
Scenarios 1, 2, and 3, with Scenarios 4 and 5 running roughly 10% higher.
The 10.4% increase in requirements in the case of Scenario 1, which
was due to population growth and aging, was abstracted from each of
the other percentages shown, thus arriving at an estimate of the incremental
impact associated solely with the changes in insurance coverage and
health care staffing patterns postulated in Scenarios 2 through 6. The
resulting incremental measures of impact are as follows:
Lastly, the authors applied the described methodology on a county-by-county
basis rather than merely by county type, thereby negating the possibility
that surpluses in some counties might cause deficits in others to go
unnoticed. The result of this expanded effort, shown in Appendix B of
their paper, was to increase the deficits associated with Scenario 1
in the following manner:
While not able to quantify precisely the need to increase the National
Health Service Corps (NHSC), the authors noted that COGME could reasonably
recommend relieving some of the deficit with the NHSC, or some combination
of public and private efforts. They ended their paper with the following
conclusions and recommendations:
|
|
