RocketTools Labs · Hospital network research · October 2026
Shop by procedure, not by hospital
- Off the shelf, the two carriers cost the same. On inpatient hospital cost, UHC is 0.5% more expensive than BCBS. That is a tie.
- Tiering by procedure is where the money is. If every member used the cheapest hospitals for their procedure, cost would fall about 18–20% on either carrier. That is the ceiling.
- With real member behavior, expect 3–5%. Members move only partly toward lower copays. Strong copay gaps and good member support push it to 6–8%. Hospital capacity limits the top end.
- Requiring a narrow network is not much better, unless its hospitals can grow. If members must use Tier 1 hospitals for planned care, cost falls about 12%, but only if those hospitals can take all the extra patients. If each can grow by 50–100%, the saving is 4–7%. Emergencies cannot be moved, and only about half the money (54%) is in planned stays.
- Read the percentages in context. They apply to hospital inpatient facility cost for the modeled procedures, about a third of inpatient hospital spend, which is itself about 17% of total health spend. Per $100M of total plan spend, 3–5% here is about $0.2–0.3M a year, or $0.5–0.9M if the same saving holds across all inpatient care. The simulator works this out for any plan size and enrollment.
- Quality costs some of the saving. Giving quality a 30% weight keeps most of the saving (UHC 3.2% lower). An even 50/50 split leaves about 1%. Ranking only on quality costs more than today.
- The surgeon matters more under UHC. UHC pays per day at many hospitals, and surgeons whose patients go home sooner are associated with lower cost. Steering to those surgeons adds up to about 1.4 points on UHC, and little on BCBS.
Bottom line: the network design, not the carrier, drives the saving. Pick the carrier for other reasons; then tier by procedure, with real copay gaps.
Today's prices on a real mix of hospital stays
- Prices are current (2026). They come from the carriers' own price files and the hospitals' 2026 price files, including per-day rates and the extra payments carriers make on very high bills.
- The patient mix is a stand-in. We have no current claims. So 20,505 real commercial hospital stays from the 2019 Texas hospital discharge file show which stays happen, where, and how long they last. Each stay is priced through both carriers' 2026 contracts. No 2019 price is used.
- Quality comes from CMS Care Compare (safety, outcomes, patient experience, process). Google and Yelp ratings are optional extras.
- Scope: inpatient hospital facility cost in the San Antonio–New Braunfels area, 46 common procedure groups (DRGs), about 31% of commercial inpatient hospital charges. No doctor fees.
A tie overall, decided by length of stay
UHC pays a low amount for the first 2–3 days, then a rate per day. BCBS mostly pays one flat amount per stay. So UHC wins short stays and loses long ones.
| Length of stay | Stays | UHC ÷ BCBS | Result |
|---|---|---|---|
| 1-2 days | 10,604 | 0.89 | UHC 11% cheaper |
| 3-5 days | 7,163 | 0.97 | UHC 3% cheaper |
| 6-10 days | 1,743 | 1.11 | UHC 11% more expensive |
| 11+ days | 995 | 1.20 | UHC 20% more expensive |
Long, complicated stays carry many dollars, which is why the total comes out even. On a narrow set of outpatient services (imaging, ER visits, drug tests, colonoscopy) UHC is about 25% lower at the same hospital.
Tier by procedure: about 18–20% lower if everyone follows
The same hospital is cheap for some procedures and expensive for others. On UHC, the average cost per stay at Resolute Baptist is about half of Methodist Stone Oak's for a vaginal delivery ($4,994 vs $9,574), but about double for major bowel surgery ($65,858 vs $33,939). A network built per procedure uses those gaps.
| Network | Inpatient cost vs BCBS as sold |
|---|---|
| BCBS Blue Choice PPO, as sold | baseline ($414M on this stay mix) |
| UHC Choice Plus PPO, as sold | 0.5% higher |
| UHC, Tier 1 hospitals per procedure | 19.7% lower |
| BCBS, Tier 1 hospitals per procedure | 18.2% lower |
Tier 1 = the cheapest hospitals for that procedure holding at least a quarter of its stays (at least 2 hospitals). Every stay is then priced at the Tier 1 average. This assumes every member goes to Tier 1 and Tier 1 hospitals have room.
With real member behavior: 3–5%, up to about 8%
Members respond to copay gaps, but only partly. The simulator uses response rates from published tiered-network programs (Massachusetts GIC, Blue Cross Blue Shield of Massachusetts, CalPERS). Default design: 3 tiers, inpatient copay $500 at Tier 1 rising to $3,500 at the top tier, and no hospital may grow more than 50%.
| Carrier | Tiers ranked by | Low response | Central | High | Everyone, with capacity cap |
|---|---|---|---|---|---|
| UHC | Price only | 0.3% | 3.8% | 6.9% | 8.4% |
| UHC | Value (30% quality) | 0.2% | 3.2% | 5.7% | 6.7% |
| BCBS | Price only | 0.9% | 4.2% | 7.5% | 8.7% |
| BCBS | Value (30% quality) | 0.8% | 3.9% | 6.6% | 8.1% |
Total cost (plan plus member) for the modeled procedures vs BCBS as sold. Central = 70% of stays can shop, sensitivity 0.04 per $100 of copay; childbirth 85%.
Read this as: both carriers land in the same place once member behavior is real. The capacity cap matters a lot: without it, "everyone" rises to about 17–23% in the simulator, but no hospital can absorb that much new volume overnight. (The simulator splits hospitals into equal-size tiers; section 03 uses a stricter Tier 1, so the two ceilings differ a little.)
Plan cost vs total cost. The table is total cost (plan plus member). The employer's own share falls a little more, because members pay somewhat more in copays: at central response on UHC the plan pays 4.1% less while total cost is 3.2% less, and the average member copay per stay is about $1,620 vs $1,500 today.
Require Tier 1 for planned care: about 4–7%, or 12% with unlimited room
The shopping design above lets members use any hospital and pays them to choose well (copay-shopping designs such as Surest). The other design is a required network: for each procedure, members must use the plan's chosen hospitals, as in a narrow network or a centers-of-excellence program.
- Only planned stays can be moved. A member with sepsis or a stroke goes to the nearest emergency room. Using each stay's admission type, about 65% of stays and 54% of the money are planned. Deliveries count as planned, because the hospital is picked in advance.
- Capacity decides the answer. Tier 1 hospitals must absorb the moved patients. The table shows the saving for different limits on how much one hospital can grow for one procedure.
| Carrier | Required network | Growth +20% | Growth +50% | Growth +100% | Growth No limit |
|---|---|---|---|---|---|
| UHC | Tier 1 only | 1.9% | 4.2% | 5.8% | 11.9% |
| UHC | Tiers 1–2 | 1.5% | 3.1% | 3.3% | 3.3% |
| BCBS | Tier 1 only | 2.9% | 5.4% | 6.8% | 12.2% |
| BCBS | Tiers 1–2 | 2.6% | 4.7% | 5.2% | 5.2% |
Total cost for the modeled procedures vs BCBS as sold. Value blend with 30% quality weight, 3 tiers, 95% of planned stays follow the rule. Patients a full hospital cannot take go to the next-best hospital.
Read this as: a required Tier 1 network beats copay shopping (3–5%) only if its hospitals can grow a lot. A wider network (Tiers 1–2) is easy to staff but saves about the same as shopping. Requiring best-quarter surgeons for planned surgery as well adds more: Tier 1 only with +50% growth goes to 5.6% on UHC and 5.7% on BCBS.
Every point of quality weight has a price
In the default "value blend", each hospital's cost and quality are put on the same scale and mixed by a weight. The simulator has a slider for that weight.
| Weight (UHC, central response) | Total cost vs BCBS as sold |
|---|---|
| 100% cost · 0% quality | 4.0% lower |
| 70% cost · 30% quality | 3.2% lower |
| 50% cost · 50% quality | 1.1% lower |
| 0% cost · 100% quality | 3.0% higher |
In San Antonio, the cheapest hospitals for a procedure are often not the highest rated. A 20–30% quality weight keeps most of the saving while keeping low-quality hospitals out of Tier 1. Hospitals with too little quality data cannot be Tier 1 by default.
Under per-day contracts, the surgeon changes the bill
Inside one hospital and one procedure, some doctors' patients go home sooner. We measured this from the coded doctor IDs in the Texas file, adjusted for procedure, age and how sick the patient was. Doctors are never named in this part.
- Doctors explain about 9% of the differences in length of stay. Hospitals explain about 3%. The patients explain the rest.
- If every surgeon matched the best quarter of their peers, surgical stays would cost 3.8% less on UHC (about $4.0M) but only 1.0% less on BCBS. UHC's per-day contracts are why.
- A bigger UHC amount (about $6.6M) sits with hospital-assigned doctors in medical stays such as sepsis. Members cannot choose those doctors, so this is a contract issue: a flat rate per stay would shift this cost to the hospital.
| Surgeon steering (UHC, value blend 30%) | Total cost vs BCBS as sold |
|---|---|
| 0% of surgical stays | 3.2% lower |
| 50% of surgical stays | 3.9% lower |
| 100% of surgical stays | 4.6% lower |
Named surgeons. Separately, the simulator lists named surgeons from public CMS Medicare data, with their Medicare case counts and their hospitals' tiers. The two doctor sources are never linked.
| Procedure family | Named physicians | Medicare cases |
|---|---|---|
| Heart stents | 69 | 2,283 |
| Heart bypass | 31 | 844 |
| Heart valve | 31 | 832 |
| Bowel surgery | 5 | 82 |
| Weight-loss surgery | 4 | 49 |
| Brain surgery | 2 | 27 |
| Deliveries (OB/GYN) | 104 | – |
Medicare hospital cases, newest year 2020–2024; CMS hides counts under 11. Deliveries have almost no Medicare volume, so OB/GYNs are listed by hospital affiliation only. CMS links doctors to a hospital license, not a building, so a surgeon may not work at every campus shown.
Read before you quote
- Stay mix is from 2019 (the newest free year). Volumes and stay lengths may have shifted. A check using CMS 2026 stay lengths gives the same off-the-shelf answer (UHC ÷ BCBS 0.96–1.04).
- Small rounding differences. The report uses all 20,505 priced stays ($414M on BCBS). The simulator groups hospitals with fewer than 5 stays for a procedure, so its totals use 20,477 stays ($413.5M).
- Contract prices, not paid claims. Implants, carve-outs, denials and coordination of benefits are not modeled. Facility only.
- Baptist's main campuses post no payer rates. They share a license with Baptist Medical Center, so its contract is used for them (about 40% of BCBS-priced stays and 23% of UHC-priced stays).
- Not priced: Guadalupe and Medina inpatient on UHC (paid as a percent of charges), 2,664 stays in total.
- Member response comes from other programs and other markets. The capacity limit is a planning assumption; real room depends on beds, operating rooms and staff.
- Planned vs emergency uses the admission type on each 2019 stay. Some "emergency" admissions could be planned with better care management, so the required-network numbers are on the low side.
- Research study, not a quote or an actuarial certification.