Case Mix Index

Case Mix Index in Hospital Payment and Benchmarking

A hospital with a case mix index of 1.5 treats cases whose diagnosis-related group (DRG) weights average 1.5 on the payer's schedule, so its average discharge is expected to use half as much again as a case of weight 1.0. Payers, regulators and analysts use the index to compare hospital workloads, to set budgets and eligibility, and to put cost per case on a like-for-like footing. The index is built from discharge weights and read against a reference point. The Medicare programme publishes and uses it in the FY 2027 rules, European systems use it for budgets and benchmarking, and a rising index needs careful reading. An illustrative worked example separates a change in patient mix from a change in the weight schedule.

From discharge weights to one hospital number

The index starts from the relative weight of each group, which expresses the expected resource use of a case in that group against an average case. How those weights are derived from cost data is covered on the case weight page. The case mix index takes the weights as given and averages them over a hospital's discharges in a period.

$$CMI_h = \frac{\sum_{g} n_{hg} w_g}{\sum_{g} n_{hg}}$$

where $CMI_h$ is the case mix index of hospital $h$, $n_{hg}$ is the number of its discharges assigned to group $g$, and $w_g$ is the relative weight of group $g$ on the schedule in use. The numerator is the hospital's casemix, its activity in weighted cases. In the European Observatory study of DRG systems edited by Busse and colleagues, the casemix is the sum of the weights of all DRGs a hospital produced and the index is that casemix divided by the number of discharges. The Centers for Medicare & Medicaid Services (CMS) describes a hospital's index as its average DRG relative weight, calculated by summing the DRG weights for all Medicare discharges and dividing by the number of discharges, and calculates it on both transfer-adjusted and unadjusted cases.

The index is a ratio without units, and every element of it is a choice, including the payer population (Medicare only or all payers), the grouper version, the weight year, the period and the treatment of transfers. Two indices computed on different bases describe different things even for the same hospital.

Why the index tracks payment so closely

Under case-based payment each discharge earns roughly its weight multiplied by a base rate, so total DRG payments before adjustments move with discharges multiplied by the index. CMS states the link directly in its FY 2027 final rule, saying that because the DRG weight determines the prospective payment for each case, any percentage increase in the index corresponds to an equal percentage increase in hospital payments. That link is what makes the index both useful to payers and sensitive to coding, and it places the index at the centre of case-based payment and activity-based funding.

What an index of 1.0 does and does not mean

A weight of 1.0 marks the average case in the costing data from which weights were first scaled. Systems differ in whether they keep the national average at 1.0. In Portugal, Mateus reports that the budget model is recalibrated after changes in cost weights or base rates to keep the national index at one.

Medicare does not re-centre each year. In the FY 2027 rule, CMS normalised the recalibrated MS-DRG weights by a factor of 1.945875 (as corrected in September 2026) so that the average case weight after recalibration equalled the average before it, which keeps recalibration by itself from raising or lowering total payments, as section 1886(d)(4)(C)(iii) of the Social Security Act requires. The median index of all urban hospitals in FY 2025 data, which CMS uses in the test below, was 1.778. A Medicare index of 1.0 is therefore well below a typical urban hospital and does not mark the Medicare average.

How Medicare publishes and uses the index in FY 2027

The FY 2027 inpatient prospective payment system (IPPS) final rule, published on 4 August 2026, lists hospital indices in Table 2, posted on the CMS website and titled "Case-Mix Index and Wage Index Table by CCN" (the CMS certification number). The rule uses the index in at least three places.

Rural referral centre eligibility. Under 42 CFR 412.96, a rural hospital that does not meet the bed-size route must meet two mandatory criteria, a minimum index and a minimum number of discharges, plus at least one of three optional criteria. For initial status in cost reporting periods beginning on or after 1 October 2026, a rural hospital with fewer than 275 beds needs an FY 2025 index of at least 1.778 (the national median for all urban hospitals) or the median for urban hospitals without approved teaching programmes in its census region. CMS specifies the index that is not transfer-adjusted, computed on all Medicare discharges subject to MS-DRG payment.

The capital update framework. CMS separates three reasons the index can change. The first is real change in the mix and resource needs of Medicare patients, the second is documentation and coding changes that move cases into higher-weighted DRGs without higher resource needs, and the third is reclassification and recalibration that is not budget neutral. For FY 2027 it projected a 0.5% total increase in the index and a 0.5% real increase, so the net adjustment for case mix change was 0.0 percentage point.

Long-term care hospital weights. For long-term care hospitals, CMS sets weights iteratively using hospital-specific indices, starting from an index of 1.0 for each hospital, and applies a normalisation factor of 1.27345 so that recalibration neither raises nor lowers the average index.

Budgets and benchmarking outside the United States

The Busse study shows the index in two further roles. In countries that allocate hospital budgets by DRG, Quentin and colleagues report that the casemix and the index are usually the determining factors for distributing the budget. Portugal states this as a formula, in which a hospital's inpatient budget is its index from two years earlier, multiplied by its predicted equivalent discharges and by a group base rate, and 2011 budgets rested on 2009 indices.

The index is also a standard benchmarking indicator. Cots and colleagues describe the index and length of stay per DRG as the basic indicators of Spanish hospital scorecards, and Kahur and colleagues report that the Estonian Health Insurance Fund has given hospitals regular comparisons including the index since 2008. In Germany, Geissler and colleagues note that small rural hospitals typically have indices well below 1, while university hospitals may exceed 1.5.

Reading a change in the index

A rising index has more than one possible source, and the policy response depends on which applies. CMS's three reasons cover most of them, but the definition of the inpatient population matters as well. Mateus reports that the Portuguese index, computed on fixed 1998 weights, rose from 0.99 in 1997 to 1.13 in 2006 while treatment of patients in ambulatory settings grew, a trend Mateus reads as a sign that inpatient cases had become more complex. Moving simple cases to day surgery raises the inpatient index without any patient becoming sicker.

Coding intensity is the hardest source to separate. Fuller recording of complications moves cases into higher-weighted severity levels, so the index and payment rise while the patients stay the same. The mechanism, the evidence on its size and payers' offsets are covered on the DRG creep page. For analysis of the index itself, the main defences are to hold the grouper and weights fixed across years, to compare recorded severity with independent clinical data, and to examine the distribution of cases across groups rather than the average alone.

Worked example separating patient mix from reweighting

The figures below are illustrative. A hospital treats four groups, A to D. In year 2 the payer recalibrates its weights; the hospital also treats fewer simple cases and more complex ones.

1. Year 1 index. The group products are 400 × 0.60 = 240, 300 × 1.00 = 300, 200 × 1.80 = 360 and 100 × 3.50 = 350. The casemix is 240 + 300 + 360 + 350 = 1,250 weighted cases, so the index is 1,250 / 1,000 = 1.250.

2. Year 2 on the old weights. The casemix is 180 + 300 + 450 + 420 = 1,350, so the index is 1,350 / 970 = 1.392. The change in patient mix alone raises the index by 1.392 / 1.250 = 1.113, or 11.3%.

3. Year 2 on the new weights. The casemix is 186 + 300 + 437.5 + 444 = 1,367.5, so the published index is 1,367.5 / 970 = 1.410. Reweighting adds 1.410 / 1.392 = 1.013, or 1.3%, for this hospital even if, nationally, normalisation leaves the average unchanged. The total rise is 1.410 / 1.250 = 1.128, or 12.8%, and 1.113 × 1.013 = 1.128.

4. Cost per weighted case. Total cost is USD 7,500,000 in year 1 and USD 7,760,000 in year 2. Cost per discharge rises from 7,500,000 / 1,000 = USD 7,500 to 7,760,000 / 970 = USD 8,000, up 6.7%. On fixed weights, cost per weighted case falls from 7,500,000 / 1,250 = USD 6,000 to 7,760,000 / 1,350 = USD 5,748, down 4.2%. Using the published year 2 index gives 7,760,000 / 1,367.5 = USD 5,675, down 5.4%.

5. Interpretation. The crude figure suggests rising costs, but allowing for case mix the hospital spends less per unit of expected resource use. About 1.2 percentage points of the 5.4% fall come from the new weights rather than from the hospital, which is why trend analysis should hold the schedule fixed. If part of the rise from 1.250 to 1.392 reflected fuller coding rather than sicker patients, even the 4.2% fall overstates the improvement.

Uses and limits in economic analysis and decision making

The index summarises expected resource use, so it suits questions about payment, budgets and workload. In hospital cost comparisons and efficiency studies, dividing cost by casemix gives cost per weighted case, and the index can enter data envelopment analysis or stochastic frontier analysis as an output adjustment or covariate. Such uses inherit the weights' focus on average cost, so a hospital with a high index may simply be paid more without producing more health per unit of resource.

The index is a poor tool for outcome comparisons. It reflects the cost structure of DRGs, not each patient's risk of death or complication, so dividing a mortality rate by the index is not a valid adjustment. Outcome comparisons need patient-level risk adjustment or indirect standardisation such as the standardised mortality ratio. Nor is the index a measure of quality, despite its use in hospital scorecards. In economic evaluation, a hospital's index rarely enters directly; group prices or costs value individual events, and the index matters mainly when hospital-level costs are transferred between settings with different case mix.

Common misreadings

Several errors recur in reports that cite the index. Most come from reading an average of payment weights as if it were a direct measure of patients.

• Treating it as severity. The index measures expected cost relative to a schedule. Two hospitals with the same index can treat very different patients, and a single average hides the case complexity within groups.
• Assuming 1.0 is average. On Medicare weights the median urban hospital is well above 1.0; on recalibrated European schedules 1.0 may be the national mean.
• Comparing across schedules. Indices from different grouper versions, weight years or countries are not on one scale.
• Reading every rise as sicker patients. Coding, reclassification and changes in what counts as an inpatient case all move the index.
• Ignoring small numbers. In a small hospital a handful of very high-weight cases, such as transplants, can shift the average noticeably from one year to the next.

Sources

• Busse R, Geissler A, Quentin W, Wiley M, editors. Diagnosis-Related Groups in Europe: Moving towards transparency, efficiency and quality in hospitals. Maidenhead: Open University Press, European Observatory on Health Systems and Policies Series; 2011. Chapters 3 (Quentin et al.), 14 (Geissler et al.), 17 (Kahur et al.), 21 (Mateus) and 22 (Cots et al.). https://eurohealthobservatory.who.int/publications/m/diagnosis-related-groups-in-europe
• Centers for Medicare & Medicaid Services. Case Mix Index (Acute Inpatient PPS files for download). Baltimore, MD: CMS; accessed 6 October 2026. https://www.cms.gov/medicare/medicare-fee-for-service-payment/acuteinpatientpps/acute-inpatient-files-for-download-items/cms022630
• Centers for Medicare & Medicaid Services. Medicare Program; Hospital Inpatient Prospective Payment Systems for Acute Care Hospitals and the Long-Term Care Hospital Prospective Payment System and Policy Changes and Fiscal Year (FY) 2027 Rates; final rule. Federal Register 2026;91(148):49570-50461, 4 August 2026. FR Doc 2026-15833. Sections on MS-DRG recalibration, rural referral centres, MS-LTC-DRG weights and the capital update framework.
• Centers for Medicare & Medicaid Services. FY 2027 IPPS and LTCH PPS rates; correction. Federal Register 2026;91(187):61328-61343, 29 September 2026. FR Doc 2026-19946.