Boston · Urban form
Half of the Boston area's hospitals stand within 7.5 kilometers of downtown by road
A 2026 study found that, of four measures of a city's shape, distance to the center and to jobs relate most to how far people drive to work. Rebuilt from public data for 3,681 hexagons around Boston, the measures place 22 of the area's 36 mapped hospitals within 10 kilometers of downtown, in surroundings more densely settled than 95% of the region. They describe where the hospitals stand, not how their staff travel.
· Census, EPA, OpenStreetMap; Medicare cost reports
A study published this year in Environmental Research Letters asked how the shape of a city relates to how far people drive to work. Its authors used about 10 million trips drawn from phone records and GPS data in six cities, Boston among them, and found that across the cities, distance to the city center and to jobs mattered more than population density or the street grid.
The trip records are not public, but the study's code is. Delgam ran that code on public data for the Boston area: the 2,796 square kilometers that the European Commission counts as Boston's functional urban area, cut into 3,681 hexagons of about 0.76 square kilometers each. Every hexagon gets the study's four measures. The 36 hospitals in the area that file Medicare cost reports, and that could be placed from the addresses they file, are the dots.
Four measures of urban form across the Boston area, and its hospitals
Each hexagon about 0.76 km². Darker: closer to downtown or to jobs, more people, more street intersections. Dots: 36 hospitals; point at one, or tap it, for its values.
To the center
To 5% of jobs
People per km²
Intersections per km²
Source: Delgam, running the study’s code on U.S. Census population (2009–2013), EPA jobs (2017) and OpenStreetMap streets (October 2026); study area: the European Commission’s functional urban area for Boston.
Close in, and crowded
Half of the 36 hospitals stand within 7.4 kilometers of downtown by road, and 22 within 10 kilometers. Five are more than 20 kilometers out; the farthest, Beverly Hospital in Beverly, is 28.7 kilometers away. The median hospital is 3.0 kilometers from the nearest 5 percent of the region's jobs, against 6.9 for the median hexagon.
Within 2 kilometers of the median hospital live about 4,376 people per square kilometer, more than in 95% of the region's hexagons, and its streets are more closely knit than those of 91%. Several hospitals sit so close together that they share a hexagon: New England Baptist Hospital, Brigham and Women's Hospital, Dana-Farber Cancer Institute and Children's Hospital Corporation in one; Massachusetts General Hospital and Massachusetts Eye and Ear Infirmary; Steward St. Elizabeth's Medical Center and Franciscan Hospital for Children.
The 36 hospitals and what surrounds them
Distances from the hospital's own hexagon; people and intersections averaged over the hexagons within 2 km. Nearest to downtown first. Bars scale to the largest value in each column.
| Hospital | To downtownkm by road | To 5% of jobskm | Peopleper km², within 2 km | Intersectionsper km², within 2 km |
|---|---|---|---|---|
| Tufts Medical CenterBoston · general, 364 beds | 0.0 | 0.4 | 7,301 | 104 |
| The Shriners Hospital for ChildrenBoston · children's | 1.2 | 0.5 | 7,408 | 114 |
| Massachusetts General HospitalBoston · general, 997 beds | 2.7 | 0.8 | 7,124 | 109 |
| Massachusetts Eye and Ear InfirmaryBoston · general, 41 beds | 2.7 | 0.8 | 6,748 | 106 |
| Dr. Solomon Carter Fuller Mental Health CenterBoston · psychiatric, 60 beds | 2.8 | 1.2 | 9,054 | 115 |
| Boston Medical CenterBoston · general, 442 beds | 3.0 | 1.4 | 8,756 | 115 |
| Spaulding Rehabilitation HospitalCharlestown · rehabilitation, 132 beds | 4.1 | 2.1 | 5,713 | 73 |
| Spaulding Hospital CambridgeCambridge · long-term care, 180 beds | 4.7 | 1.5 | 7,787 | 113 |
| New England Baptist HospitalBoston · general, 75 beds | 4.9 | 0.9 | 7,734 | 99 |
| Brigham and Women's HospitalBoston · general, 826 beds | 4.9 | 0.9 | 8,750 | 100 |
| Dana-Farber Cancer InstituteBoston · general, 30 beds | 4.9 | 0.9 | 9,028 | 95 |
| Children's Hospital CorporationBoston · children's, 485 beds | 4.9 | 0.9 | 8,672 | 96 |
| Beth Israel Deaconess Medical CenterBoston · general, 705 beds | 5.0 | 0.9 | 8,948 | 89 |
| Human Resource InstituteBoston · psychiatric, 66 beds | 5.7 | 1.8 | 7,384 | 91 |
| Mount Auburn HospitalCambridge · general, 189 beds | 6.6 | 2.0 | 4,101 | 76 |
| Steward St. Elizabeth's Medical CenterBoston · general, 244 beds | 7.3 | 2.5 | 5,594 | 82 |
| Franciscan Hospital for ChildrenBrighton · children's, 112 beds | 7.3 | 2.5 | 5,848 | 82 |
| The Arbour HospitalJamaica Plains · psychiatric, 138 beds | 7.3 | 2.6 | 4,519 | 75 |
| Lemuel Shattuck HospitalJamaica Plain · long-term care, 248 beds | 7.5 | 3.4 | 4,233 | 67 |
| Cambridge Health AllianceMalden · general, 225 beds | 8.5 | 3.7 | 5,371 | 100 |
| Faulkner HospitalJamaica Plain · general, 147 beds | 9.4 | 3.9 | 2,817 | 60 |
| Steward Carney HospitalDorchester · general, 106 beds | 9.6 | 4.2 | 4,195 | 72 |
| Hebrew Rehabilitation CenterRoslindale · long-term care, 667 beds | 10.2 | 4.3 | 2,996 | 68 |
| BIDMC-Milton HospitalMilton · general, 102 beds | 12.3 | 5.1 | 1,198 | 38 |
| McLean HospitalBelmont · psychiatric, 309 beds | 13.0 | 4.0 | 1,541 | 46 |
| Winchester HospitalWinchester · general, 182 beds | 15.2 | 4.6 | 1,302 | 38 |
| Newton Wellesley HospitalNewton · general, 216 beds | 16.3 | 3.9 | 1,202 | 51 |
| Encompass Health Rehabilitation HospitalWoburn · rehabilitation, 179 beds | 16.9 | 4.6 | 1,211 | 32 |
| Walden Behavioral CareWaltham · psychiatric, 51 beds | 17.4 | 3.5 | 1,753 | 35 |
| Beth Israel Deaconess Hospital-NeedhamNeedham · general, 58 beds | 18.5 | 5.3 | 1,192 | 44 |
| Encompass Health Rehabilitation HospitalBraintree · rehabilitation, 187 beds | 19.1 | 4.9 | 693 | 25 |
| Pembroke HospitalWestwood · psychiatric, 120 beds | 20.8 | 6.4 | 831 | 24 |
| Steward Norwood HospitalNorwood · general | 21.5 | 5.9 | 1,435 | 49 |
| South Shore HospitalSouth Weymouth · general, 374 beds | 22.8 | 6.1 | 1,106 | 25 |
| North Shore Medical CenterSalem · general, 268 beds | 23.1 | 5.3 | 2,575 | 66 |
| Beverly HospitalBeverly · general, 299 beds | 28.7 | 6.2 | 1,393 | 47 |
Source: Delgam, running the study’s code on U.S. Census population (2009–2013), EPA jobs (2017) and OpenStreetMap streets (October 2026); study area: the European Commission’s functional urban area for Boston. Hospitals: latest Medicare cost report; points from the Census Geocoder.
Does the rebuild match the study?
The study published averages for its 608 Boston neighborhoods, built from the trip records. Its neighborhoods are not public, so the code was also run on the 2,242 census block groups that touch the study area. The results land close to the study's, on every measure:
The study's Boston figures and this rebuild
Average, with the middle half of values in parentheses, in the units the study used (see the notes below).
| Measure | The study, 608 neighborhoods | Rebuild, 2,242 block groups |
|---|---|---|
| Distance to downtown by road, km | 15.5 (7.2–21.5) | 16.6 (8.2–22.7) |
| Distance to the nearest 5% of jobs, km | 5.3 (2.8–6.2) | 5.5 (3.4–6.4) |
| People per km² | 4,256 (915–6,072) | 4,391 (897–6,019) |
| Intersections per km² | 70 (31–102) | 70 (28–99) |
| Income class | 2.9 (1.8–4.0) | 3.0 (2.0–4.0) |
Source: Wagner et al. (2026), supplementary table S8; Delgam. Income class: the code sorts block groups into income classes 0 to 6; the study reports 1 to 6.
One finding about the code came out of the check. It holds two versions of the jobs measure. The second counts the nearest jobs until it reaches 5 percent of the region's total, as the study describes, and matches the study: 5.5 kilometers against 5.3. The first stops only when it meets a single place holding 5 percent of all jobs, which none does, so it measures the distance to every job: 20.8 kilometers. The maps use the second.
What these measures do not say
They do not say how far anyone drives. The study's model predicts the average car commute of people who live in a neighborhood; its fitted models and trip records are not published, so no commute distance appears here. In Boston, the average across the study's neighborhoods was 7.56 kilometers per morning car trip, from call detail records collected over six weeks in 2010.
They describe a hospital's neighbors, not its staff. Whether a site works for the nurses, technicians and support staff who keep it running depends on where they live and how they travel, which none of these data record.
How these figures were made
- The study's own code computed every measure, xml4uf at commit e5b447b; nothing was rewritten. It ran on hexagons of the H3 grid (resolution 8) rather than the study's neighborhoods, which come from its private trip records.
- Distance to downtown is by road, on the OpenStreetMap drive network, from each hexagon's center to the authors' point for downtown Boston (42.354° N, 71.063° W), which they took from Google Maps and printed in their code's example. Distance to jobs runs in straight lines between hexagon centers and block groups. Population and intersections are counted per hexagon and divided by its area.
- Units. The code measures every American city in a map projection centered far to the west (UTM zone 14N), which in Boston stretches lengths by about 6.6 percent and areas by about 13.6 percent. The maps and the hospital table are converted back to true kilometers. The check against the study keeps the code's units, as the study did; in true units its averages would be about 6 percent shorter and 14 percent denser.
- The inputs are those the study used, rebuilt from public files: population and household income from the American Community Survey, 2009–2013; jobs in 2017 from the EPA's Smart Location Database; streets from OpenStreetMap as of October 2026, which is newer than the study's download. Straight-line distance to downtown is in the data as well: 15.1 kilometers on average over the block groups, against 16.6 by road.
- Hospitals: the 98 Massachusetts hospitals with a Medicare cost report were placed with the Census Geocoder from the addresses they file. 88 were found, 52 of them outside the study area. 10 could not be placed(two because the address matched more than one place): MelroseWakefield Healthcare, Emerson Hospital, Milford Regional Medical Center, Nashoba Valley Hospital, Lahey Clinic Hospital, Chestnut Hill Benevolent Association, Bournewood Hospital, Haverhill Pavilion, Aurora Westborough BHS and Hospital for Behavioral Medicine. None was placed by hand.
- Street data © OpenStreetMap contributors, available under the Open Database License; the hexagon data behind the maps, urban_form_cells.json, is available under the same license.
Sources
- Wagner, F., Nachtigall, F., Milojevic-Dupont, N., Franken, L., Koch, N., Runge, J., Pereira, R. H. M., Gonzalez, M. C. and Creutzig, F. (2026). Refining urban typologies: causal insights into urban form, car commuting, and related CO2 emissions. Environmental Research Letters 21, 104031.. License: CC BY 4.0.
- xml4uf, the paper's code (commit e5b447b), with its ufo-map library. License: MIT.
- European Commission, Joint Research Centre: GHS functional urban areas R2019A (2015), the study area. License: European Commission reuse notice.
- U.S. Census Bureau: American Community Survey 2009-2013 five-year estimates (population, household income) and 2013 block group boundaries. License: Public domain.
- U.S. Environmental Protection Agency: Smart Location Database version 3, total employment 2017. License: CC0.
- OpenStreetMap contributors, via Geofabrik's Massachusetts extract (data to Oct. 23, 2026). License: ODbL 1.0.
- U.S. Census Bureau Geocoder (hospital addresses from Medicare cost reports). License: Public domain.
| Measure | Median hexagon | Breaks between shades |
|---|---|---|
| Distance to the center by road | 25.5 km | 16, 22, 29, 36 |
| Distance to the nearest 5% of jobs | 6.9 km | 4.8, 6, 8.1, 12 |
| People per square kilometer | 399 per km² | 160, 310, 560, 1200 |
| Street intersections per square kilometer | 14 per km² | 1.3, 9.2, 20, 41 |
Other stories built on the same cost reports: stories.