Does service "leak" to the fee-for-service sector under an alternative funding method? Experience at a Canadian Academic Health Center.

ABSTRACT

This study was designed to determine whether shifting specialists from fee-for-service practice to annual block funding at an academic health center in Ontario, Canada, would be accompanied by shifts in some of the services provided in the center to specialists who remained in the fee-for-service sector. A data set of the procedural workload was collected for four specialties (Cardiology/Cardiothoracic Surgery, Obstetrics/Gynecology, Orthopaedics, and Urology). Catchment areas were defined for the study center, the four other academic health centers in the province, and for nearby hospitals in the study center's region, using patient postal codes from provincial health insurance data. Patient volumes and patterns of service provision were compared for each provider specialty for 2.5 years before and after the change in funding. The study found no clear trend in migration of care to other centers following initiation of a non-fee-for-service payment plan at an academic health center.

**********

In 1994 Queen's University and its hospital partners formed the Southeastern Ontario Academic Medical Organization (SEAMO) in order to initiate the Alternative Funding Plan (AFP). Under the new plan, negotiated with the Ontario Ministry of Health and Long-Term Care, the academic health center was to receive a fixed annual sum for five years. Clinical faculty would be remunerated for all medical care from this fixed sum that had been based largely on previous fee-for-service billing levels. Under the AFP, as under any fixed payment system, the Ministry's virtual monopsony on the purchase of medically-necessary physicians' services meant that the volume of clinical activity would be de-linked from faculty members' incomes. A review of planning documents (Queen's Health Policy, 1996) revealed the program was intended to alter financial incentives so as to eliminate marginally-necessary care, thus allowing faculty additional time for research, educational activities, or enhanced patient care.

Did the new payment system achieve its intended clinical impact? Our evaluation studies have demonstrated that at SEAMO there was, indeed, a decline in the volume of clinical activity between 1992 and 1996. In the case of outpatient appointments, however, the trend began well before the AFP and continued to decrease at its previously-established rate throughout the study period (Shortt, 2002). We have previously shown that the decrease in inpatient services was consistent with changes that were associated with hospital restructuring initiatives seen throughout the province (Stanton & Shortt, 2003). However, documenting a decrease in the volume of clinical activity alone would establish only that some care was no longer provided at the study center. It would not address the important issue of whether this represented the complete elimination of care or simply transferal of care elsewhere to the fee-forservice sector.

STUDY OBJECTIVE

This study was designed to determine whether AFP implementation was followed by "leakage" of patients from SEAMO to the four other academic health centers in Ontario or to the eight nearby regional hospitals. Specifically, the study questions were as follows.

Was there a change in the proportion of patients from the SEAMO catchment area who received health services:

* at SEAMO after the implementation of the AFP?

* at the other academic health centers or regional hospitals area after the implementation of the AFP?

METHODS

1. Study Design

This retrospective study, based on five years of administrative data, used a pre-post design that included 2.5 years before and after the AFP was implemented. The data captured all of the procedural workload for four clinical specialties in SEAMO, the other Ontario academic health centers, and the regional hospitals in SEAMO's catchment area. For this study, only non-restrictive acute care hospitals affiliated with medical school's care were considered when defining academic health centers.

2. Procedure Data

A literature review identified procedures that were both relatively common and had a high discretionary component in the specialties of urology, obstetrics and gynecology, and orthopedics (Gibson, 1997). Since change was deemed most likely to be seen for discretionary procedures, analysis was confined to these three specialties, with cardiothoracic surgery added as a control. The study procedures (and ICD-9 codes) were: prostatectomy (72.1-72.4, 72.52, 72.59), non-obstetrical dilatation and curettage (81.09), total hip replacement (93.51, 93.59), total knee replacement (93.41, 93.47) and permanent pacemaker insertion (49.72, 49.73).

Data from the Canadian Institute of Health Information's Discharge Abstract Database was obtained from Ontario Ministry of Health and Long-term Care. It included every Inpatient and Day Surgery discharge record involving a procedure that was performed by one of the study specialties in an acute care institution affiliated with an academic health center or at one of the regional hospitals in the SEAMO region between January 1992 and December 1996 (inclusive).

3. Defining Catchment Areas

Catchment areas were defined by the geographic areas represented by the Forward Sorting Area contained in the postal code of patients' residences. A Forward Sorting Area was assigned to a center's (academic health center or regional partner) catchment area for each specialty if the center provided over 60 percent of the care to the residents in that area who received hospital specialty care between June 1992 and June 1994. If less than 60 percent of the residents' care was provided in any one center, the care area was defined as "Mixed". Leakage was defined by the discharge of residents of SEAMO's care area from either another center's hospitals or regional hospitals.

4. Analysis

As stated above, the care areas associated with each academic health center were defined using bivariate analysis. Aggregate measures (proportions, averages, or totals) were used to minimize the threat of biases resulting solely from the difference in sizes of the centers. Paired t-tests were used to determine if there had been changes in practice patterns within SEAMO and whether these changes differed from changes elsewhere.

RESULTS

Overall, there was a statistically significant decrease in the volume of obstetric/gynecology patients and urology patients in SEAMO during the study period. This was accompanied by a general upwards trend in the severity of the patients - a trend that was significant for obstetrics/gynecology and orthopedic patients. The only significant difference in the proportion of patients from outside the SEAMO area who received care from SEAMO specialists was in urology. When the difference between changes in SEAMO and those in other academic health centers was examined the single point of marked difference was found to be in the change of the proportion of patients from the SEAMO region for all services. These differences, while significant, accounted for less then a 1 percent change in SEAMO area patients and was of a downward direction in urology. Furthermore, the only statistically significant shift between SEAMO and regional hospitals was a decrease in the proportion of SEAMO catchment area urology patients seen in regional centers. These results are shown in Exhibits 1-3.

DISCUSSION

The study found no clear trend in migration of care between sites following initiation of a non-fee-for-service payment plan at an academic health center. Several factors may have contributed to this finding.

It is possible that care in other specialties, rather than those chosen for this study, was shifted out of SEAMO. However, the selected specialties are known to encompass some relatively discretionary procedures, a characteristic which should render them more sensitive to change than specialties in which procedures tend to have more definitive indications.

Though some of the specialties' procedures were relatively common, some nearby regional hospitals may have lacked the specialist expertise to do them, thereby preventing a shift in service provision. The same would not apply to other academic health centers. However, the nearest of these is approximately 100 miles from SEAMO which may have discouraged patients from accepting a referral to such centers.

Established patterns of clinical service provision may be relatively inflexible, such that change may require more than the 2.5 years of post-AFP follow-up used for this study. While the time lag between provision of financial incentives and practice change has not been studied in the Canadian context, it appears to occur rapidly in the United States (Nguyen & Derrick, 1997; Mitchell, Wedig & Cromwell, 1992; Christenson, 1992).

It is possible that clinicians experienced no alteration in financial incentives under the AFP since under the previous fee-for-service system their personal income was capped by a ceiling agreement with the center. However, under the previous system, in contrast to the AFP, there were incentives to maximize income: to cover practice expenses beyond the ceiling, to justify a future increase to the ceiling, or to gain stature in a department by virtue of the amount of funds contributed.

Finally, it is possible that the literature on the volume-enhancing consequences of fee-for-service, and the converse for prepaid systems, is specific to the American context where most of the research originates. For example, two Canadian studies dealing with the response to alterations in medical fees (Labelle, Hurley & Rice, 1990; Hurley & Labelle, 1995) failed to report an income-generating shift in behavior resembling those seen in the United States. Similarly, unlike their American counterparts, Ontario physicians participating in HMO-like practice settings did not reduce the hospitalization rates of their patients, despite financial incentives to do so (Hutchinson, Birch, Hurley, Lomas & Stratford-Devai, 1996). Therefore, the assumption that the introduction of the AFP might lead to a shift of service provision to the fee-for-service sector may not be valid in the context of the Canadian system of publicly-funded, universal health care.

While various factors may have influenced the findings of this study, it is reasonable to conclude that there was no clear pattern of migration of care to other centers following the initiation of a non-fee-for-service payment plan at an Ontario academic health center. This has implications for public policy since the Ontario government is committed to replicating the new funding system at all academic health centers in the province (Provincial Working Group on Alternative Funding Plans, 2002).

EXHIBIT 1

DIFFERENCES WITHIN SEAMO

                               OVERALL   INPATIENT DAY SURGERY
                               Pre-AFP   Post-AFP  Pre-AFP
Cardiothoracic
  Volume                         5,393     5,709     2,811
  % from "outside" care area       36%       35%       34%
  Severity                        3.15       3.2       4.9
Obstetric/Gynecology (1)
  Volume                        11,641   10,175 *    6,818
  % from "outside" care area       25%       26%       24%
  Severity                        3.43    4.72 *      4.97
Orthopedic
  Volume                        13,230    13,153     7,953
  % from "outside" care area       35%       33%       39%
  Severity                         2.9    3.11 *      3.78
Urology
  Volume                         8,862   7,460 *     2,478
  % from "outside" care area       44%     38% *       83%
  Severity                        3.25      3.16      7.35

                               Post-AFP  Pre-AFP   Post-AFP
Cardiothoracic
  Volume                         2,779     2,582     2,930
  % from "outside" care area       35%       38%       35%
  Severity                      5.37 *      1.28    1.14 *
Obstetric/Gynecology (1)
  Volume                       5,823 *     4,753   4,346 *
  % from "outside" care area       24%       27%       27%
  Severity                      7.42 *      1.25    1.16 *
Orthopedic
  Volume                          8337     5,277     4,816
  % from "outside" care area       38%       27%      26%
  Severity                      4.07 *      1.56    1.45 *
Urology
  Volume                       1,922 *     6,384   5,538 *
  % from "outside" care area       77%       29%     25% *
  Severity                        7.82      1.66      1.55

(1) Less than 1 percent of the Obstetric/Gynecology records did not
have a valid Inpatient/Day Surgery indicator. This resulted in the
exclusion of 70 pre-AFP and 6 post-AFP records when categorized.

* Significant difference at the p < = 0.05 level.

EXHIBIT 2
SHIFT BETWEEN SEAMO AND OTHER ARCS

Specialty                           Other AHC         SEWO
                                  Pre       Post      Pre
Cardiology
Volume                           66,286     77,611    5,393
Pts from outside area               25%        24%      36%
% Pts from within SEAMO area         1%         1%
Severity                           3.64       3.73     3.15

Obstetrics & Gynecology

Volume                          199,032    183,854   11,641
% Pts from outside area              9%         8%      25%
Pts from within SEAMO area         0.1%       0.1%
Severity                           2.51       2.69     3.43

Orthopaedics

Volume                          113,088    119,098   13,230
% Pts from outside area             14%        13%      35%
% Pts from within SEAMO area       0.2%       0.2%
Severity                           3.23       3.29      2.9

Urology

Volume                          100,519     97,987    8,862
Pts from outside area               16%        13%      44%
Pts from within SEAMO area         0.2%       0.1%
Severity                           3.42       3.43     3.25

Specialty                                  t-test
                                 Post     (p value)
Cardiology
Volume                            5,709        0.48
Pts from outside area               35%        0.15
% Pts from within SEAMO area                   0.01
Severity                           3.20        0.21

Obstetrics & Gynecology

Volume                           10,175        0.44
% Pts from outside area             26%        0.81
Pts from within SEAMO area                     0.00
Severity                           4.72        0.41

Orthopaedics

Volume                           13,153        0.51
% Pts from outside area             33%        0.06
% Pts from within SEAMO area                   0.04
Severity                           3.11        0.34

Urology

Volume                            7,460        0.18
Pts from outside area               38%        0.20
Pts from within SEAMO area                     0.00
Severity                           3.16        0.56

EXHIBIT 3

THE BASIC DATA: THE VARIABLE COST SCENARIO

Occupational   Annual Wage    Ratio   Hours    Rate    Total
Category       Expense                                 Expense

A              $47,000        0.10     5,500   22.84   $125,600.96
B               38,000        0.05     2,750   18.27     50,240.38
C               31,500        0.40    22,000   15.14    333,173.08
D               29,500        0.25    13,750   14.18    195,012.02
E               29,000        0.50    27,500   13.94    383,413.46

REFERENCES

Christenson, S. (1992). Volume response to exogenous changes in Medicare's payment policies. Health Services Research, 27: 65-79.

Gibson, L. (1997). High volume, high cost and discretionary procedures identified in the literature and South Eastern Ontario Health Science Centre, (Technical Report). Kingston, ON: Queen's Health Policy Research Unit.

Hurley. J., & Labelle, R. (1995). Relative fees and the utilization of physicians' services in Canada. Health Economics, 4:419-38.

Hutchinson, B., Birch, S., Hurley, J, Lomas, J., & Stratford-Devai, F. (1996). Do physician-payment mechanisms affect hospital utilization? A study of health services organizations in Ontario. Canadian Medical Association Journal, 154: 653-61.

Labelle, R., Hurley, J., & Rice, T. (1990). Financial incentives and medical practice: evidence from Ontario on the effect of changes in physician fees on medical care utilization, (Working Paper) Hamilton, ON: McMaster Centre for Health Economics and Policy Analysis.

Mitchell, J.B., Wedig, G., & Cromwell, J. (1989). The Medicare physician fee freeze: what really happened? Health Affairs, 8: 21. ON: McMaster University Centre for Health Economics and Policy Analysis.

Nguyen, N.X., & Derrick, F.W. (1997). Physician behavioral response to a Medicare price reduction. Health Services Research, 32: 283-98.

Provincial Working Group on Alternative Funding Plans. (2002). Alternative funding plans for academic health science centers, Toronto, ON: Ontario Ministry of Health and Long-term Care.

Queen's Health Policy. (1996). Alternative funding plan evaluation, documents model, (Technical Report). Kingston, ON: Queen's Health Policy Research Unit.

Shortt, S.(2002). Paying doctors: impact of a change in remuneration method at a Canadian academic health centre. Canadian Journal of Program Evaluation, 17 (1): 73-96.

Stanton, S., & Shortt, S. (2003). The influence of payment method on patterns of physician practice: Experience at a Canadian academic health centre. Research in Healthcare Financial Management, 8:43-58.

Research for this study was made possible by a grant from the Canadian Health Services Research Foundation, and by salary support from by the South Eastern Ontario Academic Medical Organization.

Address for correspondence: Samuel E.D. Shortt, Centre for Health Services and Policy Research, Queen's University, Kingston, Ontario, Canada K71 3N6 [e-mail]

Samuel E.D. Shortt Queen's University (Canada)

Sally Stanton Kingston General Hospital (Canada)