EPORT SUMMARY
HOSPITAL CASE MIX COSTING
Some hospitals are more efficient than others. In terms of common sense, that stands to reason. In terms of dollars and cents, it has cost Manitoba taxpayers millions. In 1991/92 alone, it is estimated that if all Manitoba hospitals had treated patients with the same cost-efficiency as its most efficient hospital, Manitoba might have saved as much as 100 million dollars on in-hospital costs - perhaps more.
Manitoba 1991/92This is one of the findings of the Manitoba Centre for Health Policy (MCHP) using an approach called Hospital Case Mix Costing. We used Hospital Case Mix Costing in response to a request from Manitoba Health, who asked us to develop a means of assessing the cost efficiency of Manitoba hospitals. At a time of great fiscal restraint, this need was never more keenly felt.
Assessing efficiency was not easy; it was not just a matter of dividing a hospital's expenditures by the number of cases it treated in a year. Some conditions are more costly to treat than others; some hospitals had more of those costly cases than others. So, Hospital Case Mix Costing, as the name suggests, took into consideration, and compensated for, the case mix - different patients with different medical conditions - particular to each hospital. Using Hospital Case Mix Costing, here is what we found.
Findings: 1991/92Since Winnipeg and Brandon hospitals - two teaching hospitals and six urban community hospitals - served 65% of the patients and accounted for 76% of the provincial expenditures for inpatient care (fig. 1), variations in efficiency among these institutions were particularly important. The results indicated that:
Two hospitals (Grace and Concordia) appeared to be very efficient relative to the other urban community hospitals. Compared to Concordia General Hospital, for example, the others ranged from 4% to 35% more expensive. The two teaching hospitals (St. Boniface and Health Sciences Centre), although similar to each other in costs, spent close to 55% more than Grace General Hospital to provide care to patients. As one would expect, most of the serious, resource-intensive cases were at the teaching hospitals, but they also had a considerable portion of low acuity (severity), low resource-use cases-particularly pediatric and obstetric admissions. This suggested that Manitoba teaching hospitals functioned not only as tertiary care facilities, but also as large community hospitals. Comparing hospitals across the province, there were other noteworthy findings:Northern hospitals were higher cost institutions than other hospitals, among the most expensive in every hospital type. Reasons for this difference may have been higher wages, shipping costs and, in many cases, low occupancy rates. The cost of care was particularly high at the northern isolated institutions; however, these units accounted for less than 1% of provincial hospital expenditures. Among the major rural hospitals, three (Portage la Prairie, Steinbach and Swan River) appeared to be very efficient relative to the others. Compared to them, costs at the other (non-northern) major rural hospitals were 21% to 38% higher. The intermediate rural hospitals appeared to be the most efficient group of hospitals, having the lowest overall average cost. Within that group, three (Beausejour, Neepawa and Souris) appeared to be more efficient than the rest, whose costs ranged from 11% to 27% higher. Regardless of hospital type, where occupancy rates were higher, cost-efficiency was higher. The most cost-efficient hospitals were no more likely than other hospitals to have their patients readmitted within thirty days of discharge. Recent FindingsIt is important to bear in mind that these analyses were conducted on 1991/92 data. Since then, there have been major changes in Manitoba, with the teaching hospitals in particular experiencing bed closures and budget constraints. If the decrease in cases was small relative to the dollars cut at the teaching hospitals, then their relative efficiency should improve markedly. Replication of these analyses are ongoing; results from urban hospitals using 1993/94 data (fig. 2) show there has been little change.
ConclusionsHospital case mix costing did allow for a fair comparison of hospital costliness irrespective of their mix of patients. It revealed a wide variation in hospital efficiency (fig. 3), not just between types of hospitals, but between hospitals of the same type: for example, some urban hospitals were more efficient than some major rural hospitals; some urban hospitals were also more efficient than other urban hospitals. Our analysis, based on 1991/92 figures, suggested that substantial savings - as high as 20% of the $582 million provincial inpatient budget-could have been achieved by increasing the cost efficiency of the less efficient hospitals.
Of that 20%, over half would have come from the two teaching hospitals alone. We estimate that if they functioned at the level of efficiency of the average urban community hospital, the province would have saved almost 11% of the total inpatient budget. This does not suggest they were the least cost-efficient hospitals; other hospitals rated lower in efficiency. But because they are so large - treating 35% of all patients in Manitoba - and consumed such a large portion (46%) of the inpatient hospital budget, improvements in their efficiency would have had the greatest impact. By comparison, although northern isolated hospitals were very expensive and there were large variations across the small rural and multi-use facilities, the estimated potential savings at these facilities combined were low since they treated such a small percentage of the patients in Manitoba.
It should be noted that our estimates have considered the fact that teaching hospitals were just that - teaching hospitals. We excluded direct teaching costs such as salaries for interns and residents. But even if we had allowed for additional teaching costs (12% according to the U.S. reimbursement system), the Manitoba teaching institutions still would have been considerably more costly than the average urban community hospitals.
It is also worth noting that for the non-teaching institutions, our estimates are not based on the assumption that all hospitals became as efficient as the most efficient hospital in the province. Rather, they become as efficient as hospitals of the same type-such as the lesser vs more efficient major rural hospitals.
One of the key influences on hospital efficiency was occupancy rates. Simply put, an empty bed reduced cost efficiency. This had a greater impact on rural hospitals than on urban hospitals since, for the most part, they had a relatively higher percentage of empty beds.
Another major influence on cost was length of stay. When patients stayed longer than the average for their conditions, it lowered a hospital's cost-efficiency. It follows that two of the most cost-efficient urban hospitals were also identified in a previous study as hospitals that discharged patients more efficiently. That is, their patients had shorter stays compared to other hospitals, yet were no more likely to be readmitted to hospital within 30 days of discharge (a common measure of quality of care). This suggests that higher cost efficiency did not produce lower quality care.
Thus far, Hospital Case Mix Costing is proving to be a highly effective means of assessing hospital efficiency. Ultimately, if analyses in subsequent years continue to support these initial assessments, such data could provide important information with which to adjust global funding.
However, as effective as case mix costing seems for measuring hospital cost efficiency, we do not recommend adopting it as a sole funding mechanism. Experience in the U.S. and in Canada indicates that such a system could be manipulated. We do recommend, however, that Hospital Case Mix Costing be used as part of a hospital's "report card", perhaps including other indicators such as efficiency of discharge and community needs. That way, efficient hospitals could be rewarded for their efforts. More importantly, all hospitals would be encouraged to improve. If not, what will be the cost?
Summary written by RJ Currie, based on the report: Hospital Case Mix Costing Project 1991/92: by Marian Shanahan, Noralou Roos and Marni Brownell, with Michael Lloyd as external consultant.
