Case Weight

Case Weight as a Relative Cost Weight in Hospital Payment, Costing and Comparison

In activity-based hospital funding, a case weight turns a classified hospital stay into standard units of expected resource use, so that very different admissions can be priced, added up and compared on one scale. The term here has this hospital payment meaning, not that of the sampling weights given to observations in survey statistics. Weights are derived from cost data, become payments in Medicare, in Australia's national weighted activity units (NWAU) and in England's healthcare resource group (HRG) prices, and produce the case mix index. They are also used as unit costs in economic evaluation and in hospital performance comparison.

From cost data to relative weights

A case weight starts with a patient classification that places each inpatient stay in a group of clinically similar, similarly costly cases, such as a diagnosis-related group (DRG). Robert Fetter and colleagues developed DRGs at Yale University, and the United States Medicare programme first used them for hospital payment in 1983.

In the European Observatory study of DRG systems edited by Busse and colleagues, the chapter on hospital payment describes the calculation, in greatly simplified terms, as dividing the average cost of cases in a DRG by the average cost of all cases in the country. That study and the Centers for Medicare & Medicaid Services (CMS) call the result a relative weight, a synonym for case weight in this sense. A weight of one therefore marks a group whose cases cost the same as the average case, as follows.

$$w_g = \frac{\bar{c}_g}{\bar{c}}, \qquad \bar{c} = \frac{\sum_{k} n_k \bar{c}_k}{\sum_{k} n_k}$$

where $w_g$ is the relative weight of group $g$, $\bar{c}_g$ is the mean cost per case in group $g$, $\bar{c}$ is the mean cost per case across all groups, $n_k$ is the number of cases in group $k$ and $\bar{c}_k$ is the mean cost per case in group $k$. Because every weight is scaled by the overall mean, the case-weighted average of the weights equals one in the costing year.

Real systems refine this ratio. The Independent Health and Aged Care Pricing Authority (IHACPA) builds an inlier and outlier cost model for each Australian Refined DRG (AR-DRG), with a mean cost for inlier episodes, a base amount plus a daily rate for long-stay outliers and separate rules for short stays, calibrated so that modelled costs equal actual costs. The cost parameters are divided by a reference cost to give cost weights. The first reference cost was the mean modelled cost of 2009-10 in-scope admitted acute activity, AUD 4,260, and it has since been indexed by unit cost growth that excludes casemix change, so that one NWAU stays a constant measure. Annual movements in inlier price weights are capped at 20% in either direction for groups judged comparable between years.

How a weight becomes a hospital payment

A weight carries no money value until it is multiplied by a conversion rate, usually called a base rate. Adjustments for local costs may follow. The basic payment for a case is as follows.

$$P_i = w_{g(i)} \times B \times a_h$$

where $P_i$ is the payment for case $i$, $w_{g(i)}$ is the weight of the group to which case $i$ is assigned, $B$ is the base rate in money per unit of weight, and $a_h$ is an adjustment factor for hospital $h$, equal to one when no adjustment applies. The Busse study gives a hypothetical example in which a relative weight of 1.95 multiplied by a conversion rate of EUR 2,000 gives a payment of EUR 3,900, and identifies three ways of expressing weights across its 12 European countries, namely relative weights, raw tariffs stated directly in money, and points scores.

CMS describes the Medicare inpatient prospective payment system in these terms. Each DRG has a payment weight based on the average resources used to treat Medicare patients in that group, and the base payment rate, with its labour-related share adjusted by the area wage index, is multiplied by the DRG relative weight. Add-on payments for hospitals serving many low-income patients, for teaching hospitals and for unusually costly outlier cases follow. Case weights thus sit at the centre of provider payment by case, in contrast with fee-for-service or capitation.

National weighted activity units in Australia

Australia names its case weight explicitly. IHACPA states that the average hospital service is worth one NWAU, that more complex services are worth several and simpler ones a fraction, and that the price of a service is its NWAU multiplied by the national efficient price (NEP). A simplified form of the 2026-27 formula for admitted acute care is as follows.

$$NWAU_i = PW_g \times A_{paed} \times (1 + A_{res} + A_{ind} + A_{rt} + A_{dia}) \times (1 + A_{treat}) + A_{icu} \times h_i$$

where $NWAU_i$ is the activity units for episode $i$, $PW_g$ is the price weight of its AR-DRG, $A_{paed}$ is the paediatric adjustment, $A_{res}$ is the residential remoteness adjustment, $A_{ind}$ is the Indigenous adjustment, $A_{rt}$ and $A_{dia}$ are the radiotherapy and dialysis adjustments, $A_{treat}$ is the treatment remoteness adjustment, $A_{icu}$ is the intensive care adjustment per hour and $h_i$ is the hours spent in an eligible intensive care unit. The full formula also deducts private patient amounts and includes shadow-priced (not currently applied) adjustments for hospital-acquired complications and avoidable readmissions. For 2026-27 the NEP is AUD 7,418 per NWAU and acute care is grouped with AR-DRG Version 12.0; IHACPA's examples give a tonsillectomy 0.7417 NWAU, or AUD 5,502, and a minor-complexity hip replacement 3.8021 NWAU, or AUD 28,204.

HRG prices without a published weight in England

England publishes prices in pounds rather than weights. NHS England calculated almost all 2026/27 NHS Payment Scheme prices by updating the previous year's prices for inflation and efficiency, so they rest on 2018/19 cost and activity data, with HRG4+ phase 3 as the basis for prices for many services, including admitted patient care. Draft price relativities are scaled to the 2026/27 prices cost base and adjusted for inflation, a 2.0% efficiency factor and the market forces factor, an index with a minimum of 1.00 that multiplies each unit of activity. Dividing an HRG price by a reference price recovers an implicit weight, and the Busse study classes England among the countries that calculate raw tariffs.

The case mix index as a hospital's average weight

Adding the weights of all cases treated by a hospital gives its casemix, its activity in units of the average case. Dividing by the number of cases gives the hospital's average weight, which summarises its whole caseload in one number.

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

where $CMI_h$ is the case mix index of hospital $h$, $n_{hg}$ is its number of cases in group $g$ and $w_g$ is the weight of group $g$. In the Busse study's chapter on Germany, small rural hospitals typically have indices well below the average of 1, while university hospitals may exceed 1.5.

Group costs as unit costs in economic evaluation

Costing studies often value an admission at the national average cost or price of its group rather than measuring each resource used. NICE's manual (PMG36) calls HRGs a valuable source of information for estimating resource use, notes that the national average unit cost of each HRG is reported through the annual collection of reference costs, and says that using these costs can reduce the need for local micro-costing. It asks analysts to consider all relevant HRGs, since critical care and unbundled items may be costed separately from the core HRG.

The manual adds that HRG data may not be appropriate when the new technology and its comparator fall in the same HRG, or when the mean cost does not reflect the technology's resource use; micro-costing may then be better. The same holds for any case weight. A technology that shortens stays within one group changes no weight, so a model costed only with group prices misses its effect. Group costs are a form of top-down costing, and a payment rate is a payer's price, which suits a budget impact analysis better than a measure of opportunity cost. NHS England states that its prices are calculated on the basis of average costs. Average cost per case can differ markedly from the marginal cost of the cases a technology adds or avoids.

Comparing hospital costs with case weights

Crude cost per case penalises hospitals that treat complex patients. Dividing total cost by casemix gives cost per weighted case. Comparing actual cost with the cost expected at reference prices gives an indirectly standardised ratio, the logic of the standardised mortality ratio, as follows.

$$R_h = \frac{C_h}{\sum_{g} n_{hg} w_g \bar{c}}$$

where $R_h$ is the cost ratio of hospital $h$, $C_h$ is its total actual cost, $n_{hg}$ and $w_g$ are as defined above, and $\bar{c}$ is the reference cost of one unit of weight. A value above 1 means spending above the reference for that casemix. NHS England publishes a National Cost Collection Index, described as a measure of the relative cost difference between NHS providers, and IHACPA notes that jurisdictions use the NEP for benchmarking, for example to compare the cost of a hip replacement in two hospitals. Such ratios are a form of risk adjustment and benchmarking; they compare cost for a given mix of cases, not health produced.

Worked example of deriving weights and paying two hospitals

The figures below are illustrative. A costing sample of 1,000 cases in three groups supplies the mean cost per case.

1. Derive the weights. The overall mean is £4,000,000 / 1,000 = £4,000. The weights are 2,000 / 4,000 = 0.50, 5,000 / 4,000 = 1.25 and 13,000 / 4,000 = 3.25, and their case-weighted average is (600 × 0.50 + 300 × 1.25 + 100 × 3.25) / 1,000 = 1,000 / 1,000 = 1.00.

2. Compute casemix. Hospital X treats 200 cases in A, 50 in B and 10 in C; hospital Y treats 50, 100 and 60.

$$CM_X = 200 \times 0.50 + 50 \times 1.25 + 10 \times 3.25 = 195, \qquad CM_Y = 50 \times 0.50 + 100 \times 1.25 + 60 \times 3.25 = 345$$

where $CM_X$ and $CM_Y$ are the casemix of X and Y in units of the average case. X has 260 cases, so its index is 195 / 260 = 0.75; Y has 210, so its index is 345 / 210 = 1.64.

3. Pay at a base rate of £4,000. X receives 195 × £4,000 = £780,000 and Y receives 345 × £4,000 = £1,380,000, the same as pricing each case at its group mean.

4. Compare actual costs. X spends £858,000 and Y £1,311,000. Crude cost per case is £3,300 for X (858,000 / 260) and £6,243 for Y (1,311,000 / 210), so Y looks nearly twice as costly. Per weighted case the order reverses, with £4,400 for X (858,000 / 195) and £3,800 for Y (1,311,000 / 345). The cost ratios are 858,000 / 780,000 = 1.10 and 1,311,000 / 1,380,000 = 0.95.

5. Interpret. Allowing for casemix, X spends 10% above the reference and runs a £78,000 deficit under case payment, while Y spends 5% below and keeps £69,000. The crude comparison favoured X only because it treats mostly cheap cases. Whether X is truly less efficient still depends on coding, on complexity within groups and on outcomes, none of which the weights capture.

Choices that move the weights

Every weight inherits the choices made in building it. The Busse study notes that DRG payment cannot work without functioning cost accounting, and that erroneous cost information leads to inadequate weights and unintended incentives.

• Costing method and sample. Top-down allocation and patient-level costing can give different group costs, and a few volunteer hospitals may not represent all.
• Scope of costs. IHACPA removes depreciation and other capital costs and direct teaching, training and research costs from its acute model.
• Outliers. Trim points and long-stay rates decide how far extreme cases shape a weight; see outlier and average length of stay.
• Imported weights. Weights adopted from abroad, the Busse study observes, need not reflect national practice patterns unless adapted to national cost data.
• Stabilisation. Caps on annual change keep payments steady but let weights lag behind practice.

Case weights and neighbouring measures

Several nearby terms are easily confused with case weights. The table separates them.

Where case weights mislead

A case weight measures expected average cost, not severity, need or quality, and says nothing about what one patient in the group will cost. Weights from different systems or years are not comparable. One NWAU, one Medicare weight unit and an implicit HRG weight are scaled to different averages.

Coding responds to weights. The Busse study describes up-coding, or DRG creep, in which hospitals add or reorder diagnoses to move cases into higher-paying groups; regular recalculation limits the gain when all hospitals do it but penalises honest coders when only some do, so auditing is still needed. A rising case mix index can therefore reflect coding rather than sicker patients. Payment weights are also policy prices, moved by stabilisation caps and efficiency factors, and in England eligible specialised providers receive top-up payments on top of prices, so an analyst needing resource cost should check what a published weight or price includes.

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 1 (Wiley), 3 (Quentin et al.), 5 (Tan et al.), 6 (Cots et al.) and 14 (Geissler et al.). https://eurohealthobservatory.who.int/publications/m/diagnosis-related-groups-in-europe
• Centers for Medicare & Medicaid Services. Acute Inpatient PPS. Baltimore, MD: CMS; accessed 1 October 2026. https://www.cms.gov/medicare/payment/prospective-payment-systems/acute-inpatient-pps
• Independent Health and Aged Care Pricing Authority. National Efficient Price Determination 2026-27. Sydney: IHACPA; March 2026. Section 2.
• Independent Health and Aged Care Pricing Authority. National Pricing Model 2026-27: Technical Specifications. Sydney: IHACPA; March 2026. Sections 2.1 to 2.3 and 7.3.
• Independent Health and Aged Care Pricing Authority. Understanding the NEP and NEC Determinations 2026-27. Sydney: IHACPA; March 2026. Sections 1.1 and 2.1.
• National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). London: NICE; 2022, last updated 13 October 2026. Sections 4.4.9 and 4.4.10.
• NHS England. 2026/27 NHS Payment Scheme. London: NHS England; published 26 March 2026. Sections 3.3 and 8. https://www.england.nhs.uk/long-read/2026-27-nhs-payment-scheme/
• NHS England. National Cost Collection for the NHS. London: NHS England; accessed 1 October 2026. https://www.england.nhs.uk/costing-in-the-nhs/national-cost-collection/