ICES | Primary Care Models in Ontario English - page 34

Comparison of Primary Care Models in Ontario by Demographics, Case Mix and Emergency Department Use, 2008/09 to 2009/10
There are some subtle differences between
models that could have affected results. Visit
frequency in capitation-based payment
practices may be lower than in fee-for-
service,
33
resulting in fewer diagnoses and
therefore lower levels of morbidity and
comorbidity on the measures used in this
study. Visit frequency in CHCs could also have
affected these measures, but the direction of
the effect is not known, nor is the effect of
using a single random diagnosis for CHC
visits. Both CHCs and FHTs have
interdisciplinary teams but those teams and
their roles were at a formative stage in FHTs
and well-established in CHCs during the
timeframe of this study.
Encounter data for CHCs were derived from
local electronic records while for other
groups physician billing claims were used.
These differences in data sources may have
introduced differences in measured patterns
of morbidity and comorbidity but the nature
and direction of such effects are not known. It
was possible for an individual to appear in
both CHC data and among patients rostered
in a primary care model, although that
happened rarely. In those cases, the
individual was assigned to the CHC.
25
ICES
The link between ED visits and access to
primary care is mediated by a number of
factors that we were unable to measure.
These include the availability and
appropriateness of local resources, such as
walk-in clinics and urgent care centres,
patient preferences for place of care,
physician practice styles, distances to
facilities, availability of parking or public
transit and hours of operation. It is likely that
these unmeasured factors were responsible
for some of the variation in ED visits we found
across groups.
Finally, many ED visits are not avoidable, even
with the best primary care. The existing
consensus on avoidable ED visits has
identified a very small proportion,
34
consisting
of minor acute infections, but the actual
proportion that is avoidable is not known.
Triage level has been used as a proxy for
avoidable ED visits.
31
It was not used in this
report because coding was substantially
revised during the study period,
35,36
and how
coding changed across urban and rural areas
is not known. Nonetheless, a substantial
proportion of ED visits appear to be linked to
lack of access to timely primary care.
37,38
This work also helped to identify several
provincial data limitations.
Foremost among these is the absence of CHC
encounter data in Ontario’s health databases.
This made it challenging to compare models
as CHC data had to be collected manually
from electronic records, while records for
encounters in other models are collected
routinely as part of physician billing claims.
A second major issue was the lack of
encounter data for nurse practitioners and
other non-physician providers in all models. It
will be very difficult to determine the
contribution of these providers, especially
nurse practitioners, without systematically
collecting data about their activities at the
level of patients and clients.
Although capitation models shadow bill, the
completeness of shadow billing is unknown
and requires study and validation. Finally,
many health numbers in the RPDB have
outdated addresses, making geographic
inferences (such as urban-rural or income
quintiles) subject to misclassification.
1...,24,25,26,27,28,29,30,31,32,33 35,36,37,38,39,40,41
Powered by FlippingBook