CMI and DRGs in Labor & Delivery: What a Delivery Is Actually Worth in FY 2027

On October 1, 2026, the FY 2027 inpatient payment rates take effect. CMS published the final rule on August 4, 2026 with a 2.3% payment update for hospitals that met quality reporting and EHR requirements. Buried in Table 5 of that rule is something most obstetric services will never look at and every one of them should: the relative weights for the twelve MS-DRGs that classify a delivery.

Those weights moved this year, and they did not move in the same direction. Cesarean delivery weights came down. Vaginal delivery weights went up. And the largest single financial variable in an obstetric admission is still not whether the patient delivered vaginally or by cesarean. It is whether anyone documented that she was sick.

What Case Mix Index Is, and Why L&D Gets Misread

Case mix index is the average DRG relative weight across a set of discharges. Sum the relative weight of every case, divide by the number of cases, and you have a single number meant to express how resource-intensive your patient population is.

Two things about CMI matter enormously in obstetrics and are almost always confused.

First, the CMI most people quote does not contain a single delivery. The CMI that appears in CMS impact files and in most benchmarking products is a Medicare CMI, calculated on Medicare discharges. Obstetric patients are not Medicare patients. A hospital that delivers 4,000 babies a year and a hospital with no maternity service at all can have identical Medicare CMIs. If your health system evaluates your unit against a Medicare CMI, it is evaluating you against a number your unit does not appear in.

Second, the all-payer CMI that does contain deliveries is structurally depressed by them. Obstetric DRG weights are among the lowest in the system. Add a large volume of them to a hospital’s all-payer denominator and the average falls, not because the care is less complex but because the classification system prices childbirth low. This is a persistent problem for high-volume maternity hospitals, which look like low-acuity institutions on a metric that was never designed to describe them.

The FY 2027 Weight Ladder

Twelve MS-DRGs cover delivery. Six of them account for the overwhelming majority of cases: three tiers of vaginal delivery without sterilization, and three tiers of cesarean without sterilization. Within each pair, the only thing separating the tiers is whether a secondary diagnosis on the claim qualifies as a complication or comorbidity (CC) or a major complication or comorbidity (MCC).

Here is what those tiers are worth under the FY 2027 final rule. The dollar column applies the FY 2027 national operating standardized amount of $6,848.98, which is the unadjusted base before any wage index, teaching, or disproportionate share adjustment. It is a yardstick for comparing tiers, not a prediction of your payment.

MS-DRG	Description	FY27 weight	At base rate	vs. base tier
807	Vaginal delivery, no sterilization, without CC/MCC	0.6802	$4,659	base
806	Vaginal delivery, no sterilization, with CC	0.7889	$5,403	+$744 (+16%)
805	Vaginal delivery, no sterilization, with MCC	1.1301	$7,740	+$3,081 (+66%)
788	Cesarean, no sterilization, without CC/MCC	0.9127	$6,251	base
787	Cesarean, no sterilization, with CC	1.0869	$7,444	+$1,193 (+19%)
786	Cesarean, no sterilization, with MCC	1.6369	$11,211	+$4,960 (+79%)

Read the ladder rather than the rows. An uncomplicated cesarean carries 0.2325 more weight than an uncomplicated vaginal delivery, a difference of 34% or about $1,592 at the base rate. That is the number everyone assumes is the story. But moving one vaginal delivery from the no-CC tier to the MCC tier is worth 0.4499 weight, nearly twice the cesarean differential. And an uncomplicated vaginal delivery to a cesarean with an MCC is a 2.41-fold spread.

The mode of delivery is a modest lever. The severity tier is the large one.

What Changed for FY 2027

CMS recalibrates these weights annually from claims and cost report data. This year the obstetric recalibration split cleanly by delivery mode.

MS-DRG	FY 2026	FY 2027	Change
805 Vaginal with MCC	1.0792	1.1301	+4.7%
806 Vaginal with CC	0.7540	0.7889	+4.6%
807 Vaginal without CC/MCC	0.6742	0.6802	+0.9%
786 Cesarean with MCC	1.6495	1.6369	-0.8%
787 Cesarean with CC	1.1168	1.0869	-2.7%
788 Cesarean without CC/MCC	0.9588	0.9127	-4.8%
783 Cesarean with sterilization, with MCC	2.4551	2.2475	-8.5%
785 Cesarean with sterilization, without CC/MCC	0.9578	0.8769	-8.4%

Every vaginal delivery tier without sterilization went up. Every cesarean tier went down, and the uncomplicated cesarean fell the most of the three at 4.8%. Cesarean with sterilization dropped more than 8%.

One quirk worth knowing: for FY 2027, a vaginal delivery with sterilization and no CC/MCC (DRG 798, weight 0.9246) now carries more weight than a cesarean with sterilization and no CC/MCC (DRG 785, weight 0.8769). The classification system is not a clean ranking of surgical intensity, and reasoning about it from clinical first principles will mislead you. Look up the weight.

Which Obstetric Diagnoses Actually Move the Tier

This is where most conversations about obstetric documentation go wrong, because clinical severity and grouper severity are not the same list. The following reflects Appendix C of the FY 2027 MS-DRG Definitions Manual, Version 44, the official grouper logic effective October 1, 2026.

Major CC (moves 807 to 805, or 788 to 786)

• Chorioamnionitis (O41.12-) and infection of the amniotic sac
• Severe preeclampsia, second or third trimester (O14.12, O14.13)
• HELLP syndrome, second or third trimester (O14.22, O14.23)
• Eclampsia (O15.02, O15.03, O15.1, O15.2)
• Pre-existing hypertension with preeclampsia (O11.1, O11.2, O11.3)
• Puerperal sepsis (O85) and other infection during labor (O75.3)
• Shock during or following labor and delivery (O75.1)
• Uterine rupture during labor (O71.02, O71.1)
• Peripartum cardiomyopathy (O90.3)
• Amniotic fluid embolism with specified encounter (O88.111, O88.112)
• Acute kidney failure, specified type (N17.0, N17.1, N17.2)
• Acute pulmonary edema (J81.0), ARDS (J80), DIC (D65)
• Sepsis and severe sepsis (A41.9, R65.20, R65.21)

CC (moves 807 to 806, or 788 to 787)

• Postpartum hemorrhage (O72.0 third-stage, O72.1 other immediate, O72.2 delayed)
• Acute posthemorrhagic anemia (D62)
• Third and fourth degree perineal laceration (O70.2-, O70.3)
• Mild to moderate preeclampsia, second or third trimester (O14.02, O14.03)
• Body mass index 40.0 and over (Z68.41 through Z68.45)
• Acute kidney failure, unspecified (N17.9)
• Placenta previa, depending on the specific code (O44.-)

Not on the list at all

These are the ones that surprise people, and every one of them is verifiable in Appendix C:

• Gestational diabetes (O24.41-). Clinically consequential, classified as neither CC nor MCC.
• Obesity complicating pregnancy (O99.21-). Not a CC. The BMI code is.
• Placenta accreta (O43.21-). Not a CC, despite being among the most resource-intensive conditions in obstetrics.
• First and second degree perineal lacerations (O70.0, O70.1).
• Poor fetal growth (O36.5-), post-term pregnancy (O48.0), abnormal forces of labor (O62).
• Severe preeclampsia, unspecified trimester (O14.10). The condition is an MCC. The unspecified-trimester version of it is nothing.
• Anemia, unspecified (D64.9). Acute posthemorrhagic anemia is a CC. Plain anemia is not.

Three Documentation Pairings That Decide the Tier

Read that last section again and a pattern appears. In three common situations, the same patient with the same physiology lands in a different DRG depending entirely on how the note is written.

Pairing 1: Trimester specificity on preeclampsia

“Severe preeclampsia” without a trimester documents to O14.10, which is not a CC or an MCC. “Severe preeclampsia, third trimester” documents to O14.13, an MCC worth 0.4499 additional weight on a vaginal delivery. The clinical content is identical. One word separates them.

Pairing 2: BMI value, not the word obesity

“Obesity complicating pregnancy” is not a CC. A documented BMI of 40.0 or above codes to Z68.41 or higher, which is. The BMI is already in the chart as a calculated value. It has to appear in the clinician’s assessment to be codeable.

Pairing 3: Acute blood loss anemia, quantified

Postpartum hemorrhage (O72.1) is a CC, and acute posthemorrhagic anemia (D62) is a separate CC. Both require documentation that goes beyond a hemoglobin value on a lab flowsheet: a quantified blood loss and a clinician statement naming the diagnosis. Visual estimation of blood loss reliably understates it, which is one reason quantitative measurement is now standard practice for clinical reasons and a coding consequence follows.

What This Adds Up To

Consider a unit with 2,000 annual deliveries at the national cesarean rate of 32.4%, so roughly 648 cesareans and 1,352 vaginal deliveries. Hold the clinical care constant and vary only what gets documented and coded.

Severity captured	Total weight	Obstetric CMI	vs. no capture
None (every case in the base tier)	1,511.1	0.7555	base
10% CC, 2% MCC	1,558.6	0.7793	+47.5
20% CC, 5% MCC	1,616.9	0.8085	+105.8
30% CC, 8% MCC	1,675.2	0.8376	+164.1

The gap between capturing nothing and capturing a 20% CC and 5% MCC distribution is 105.8 relative weight points across the year. At the FY 2027 national base rate that is roughly $725,000, or about $362 per delivery. At a lower Medicaid base rate it is proportionally less, and at a commercial rate it can be considerably more.

The arithmetic above is transparent so you can redo it with your own numbers: multiply your case counts by the FY 2027 weights in the first table, sum, and divide by total deliveries. The capture percentages are illustrative rather than benchmarks. Your actual distribution is the point of the exercise.

Why This Is a Documentation Problem, Not a Coding Problem

Coders can only assign what the record supports. In obstetrics the record has a specific structural weakness: the clinical information that establishes severity is generated during the labor course, in flowsheets and nursing documentation, while the narrative that coders abstract from is written afterward, often quickly, and often without reference to the details that accumulated over the preceding eighteen hours.

The blood pressure trend that justified the severe preeclampsia diagnosis is in the vitals. The fever and fetal tachycardia that support chorioamnionitis are in the flowsheet. The quantified blood loss is in the delivery record. None of that becomes a codeable diagnosis until a clinician states the diagnosis in an assessment. The gap between what happened and what the claim says is a documentation timing problem, and it is the single largest source of understated obstetric severity.

How Birth Model Closes That Gap Automatically

This is precisely the problem Birth Model was built to solve. The severity is already in your data. It is sitting in the flowsheets, the vitals, the medication administrations, and the delivery record, and it is being lost somewhere between the labor course and the signed note. Birth Model reads that data continuously and turns it into documentation while the patient is still in front of you.

It watches the labor course, not the discharge summary

Birth Model’s FHIR integration pulls the full labor course in real time: vital signs, cervical exams, medications, interventions, and delivery events. That means the platform is evaluating the same clinical picture a CDI specialist would evaluate in a retrospective audit, except it is doing it during the admission, when the clinician can still act on it.

The three pairings described above are exactly the kind of pattern this surfaces. A rising blood pressure trend with proteinuria is visible in the data long before anyone writes an assessment. Maternal fever with fetal tachycardia is visible in the flowsheet. Quantified blood loss above threshold is recorded at delivery. Each of those is a severity signal that today depends on somebody remembering to name it in prose hours later.

Diagnotes drafts the note with the specificity that matters

When the delivery is complete, Birth Model’s Diagnotes feature generates a draft delivery note automatically from the clinical data it has been collecting throughout the labor course. The draft includes the elements that determine DRG assignment and that are most often omitted under time pressure: the trimester on a hypertensive diagnosis, the documented BMI, the quantified blood loss, the laceration degree, the infection findings.

Diagnotes then suggests the supporting ICD-10 codes alongside the CPT codes, so that the complicating diagnoses which establish the CC or MCC are surfaced rather than left to memory. A physician who would have written “severe preeclampsia” at two in the morning is handed a draft that already says “severe preeclampsia, third trimester,” because the platform knows the gestational age.

The clinician stays the author

Nothing is submitted on a clinician’s behalf. Diagnotes produces a draft; the provider reviews it, edits anything that is wrong, and signs. That is the workflow answer and it is also the compliance answer. The platform surfaces what the data supports, and a human being attests to it. Documentation that takes 30 minutes or more to build from scratch is typically reviewed and signed in under five.

What that is worth

Return to the arithmetic above. The difference between capturing no severity and capturing a realistic distribution was roughly 106 relative weight points a year on 2,000 deliveries, or about $362 per delivery at the national base rate. That gap is not created by care that did not happen. It is created by care that happened and was never written down with enough specificity to survive the grouper.

Automating the capture converts a recurring documentation burden into a recurring accuracy improvement, and it does so without asking an exhausted clinician at the end of a shift to remember a CC/MCC list.

Two Cautions

A condition reported as the principal diagnosis cannot also serve as the CC or MCC. Appendix C attaches a principal-diagnosis exclusion list to nearly every obstetric code. If severe preeclampsia is the reason for admission and is sequenced as the principal diagnosis, it will not additionally function as the MCC. Sequencing rules are the coder’s domain, and this is one of several reasons that adding words to a note does not mechanically produce a higher-weighted DRG.

The goal is an accurate record, not a heavier one. Documenting conditions the patient does not have is fraud, and documenting conditions she does have in order to reach a tier is the same act with better intentions. What makes this worth pursuing is that obstetric severity is genuinely and systematically understated. The patient who bled 1,200 mL, required a transfusion, and left with a hemoglobin of 7.1 was sick. If the claim says otherwise, the claim is wrong, and correcting it is not optimization. It is accuracy.

There is a quality dimension too. The same coded severity data feeds risk adjustment for severe maternal morbidity comparisons and public quality reporting. A unit that systematically understates the acuity of its population will have its expected complication rate set too low, and will then be measured against that expectation. Understating severity does not only cost money. It makes good care look worse than it is.

Where to Start

• Pull your own distribution. Twelve months of delivery discharges, grouped across DRGs 805, 806, 807, 786, 787, and 788. The share sitting in 807 and 788 is your opportunity, and your starting point.
• Calculate an obstetric CMI and trend it monthly. Stop using a house-wide or Medicare CMI to describe the unit.
• Audit twenty charts against the tier. Take twenty deliveries that grouped to 807 or 788 and have a clinician and a coder read them together, asking one question: was this patient actually uncomplicated? The disagreements are your curriculum.
• Fix the three pairings. Trimester on every hypertensive diagnosis, BMI in the assessment, quantified blood loss and acute blood loss anemia named explicitly. These three account for a large share of recoverable severity and require no new technology.
• Check your contracts. Determine which of your payers use MS-DRG weights, which use APR-DRGs, and which pay a flat maternity case rate where none of this affects payment. It still affects risk adjustment and benchmarking everywhere.
• Stop relying on memory at the end of a shift. Every item above is a recurring task that a platform reading the labor course can do continuously. Birth Model captures the clinical detail that establishes severity as it happens and drafts the note with the specificity the grouper requires, which is what turns a one-time audit into a permanent change in your distribution.

The FY 2027 weights take effect October 1, 2026. Cesarean deliveries are worth slightly less this year and vaginal deliveries slightly more, but the ladder inside each pair is unchanged in shape and steeper than the difference between the pairs. That has been true every year, and it will be true next year, and it is the part that is actually within your control.

Sources

1. CMS, FY 2027 IPPS Final Rule (CMS-1849-F). Relative weights and geometric mean length of stay from Table 5; national operating standardized amounts from Tables 1A through 1E. Published August 4, 2026, effective October 1, 2026.
2. CMS, FY 2026 IPPS Final Rule (CMS-1833-F), Table 5, for the prior-year weight comparison.
3. CMS, ICD-10 MS-DRG Definitions Manual, Version 44, Appendix C (list of CC and MCC codes and principal-diagnosis exclusions), effective FY 2027.
4. NCHS, Births in the United States, 2024, Data Brief No. 535, July 2025. Cesarean delivery rate 32.4%, primary cesarean rate 22.9%, private insurance 51.8% and Medicaid 40.2% as source of payment for delivery.
5. MACPAC, Medicaid Inpatient Hospital Services Payment Policy, on DRG-based payment as the predominant Medicaid inpatient method and the use of APR-DRGs for obstetric and neonatal populations.
6. Peterson-KFF Health System Tracker, Health costs associated with pregnancy, childbirth, and infant care, for total cost differentials between cesarean and vaginal birth.