Multi-location practices are told to build a page per location per service. The arithmetic gets large quickly: six locations and eight treatments is forty-eight pages, most of which differ by two nouns. It is worth knowing that Google has written down what it thinks of that.
The policy, in its own words
Google's spam policies define doorway abuse as sites or pages created to rank for specific, similar search queries, which lead users to intermediate pages that are not as useful as the final destination. The listed examples are unusually direct.
One is having multiple domain names or pages targeted at specific regions or cities that funnel users to one page. Another is creating substantially similar pages that are closer to search results than a clearly defined, browseable hierarchy. A third is generating pages to funnel visitors into the actual usable portion of a site.
Read those against a typical location-page set and the overlap is uncomfortable. If every city page ends with the same enquiry form pointing at the same central intake, the page is an intermediate step that funnels to one destination.
What makes a location page legitimate
The test is whether the location is genuinely a different thing to buy. Sometimes it is, and substantially so.
Different clinicians, named, with their own credentials. A different service list, because not every site offers every treatment. Different pricing, where pricing varies. Different hours, parking, accessibility, languages spoken. A different phone number that reaches that front desk. Photographs of that building rather than the stock interior used on all six.
A page carrying four or five of those is not an intermediate step. It is the destination for someone deciding whether to go to that clinic, and it answers questions the central page cannot.
A page whose only variable content is the city name in the heading, the meta description and a sentence about serving the surrounding area is the case the policy describes.
What to do with a set you already have
Do not mass-delete. Thinning a site to satisfy a policy reading is its own mistake, and some of those pages are receiving real traffic and real calls.
Rank them by whether they could pass the test with work. Pages for real, staffed locations usually can: they need the clinicians, the actual service list and the local details added, which is a content problem with a known fix. Pages for service areas where there is no premises and no staff usually cannot, because there is nothing true to add.
For that second group the question is whether the location genuinely serves those patients, in which case the honest structure is one page describing the catchment, or whether the page exists only because a tool generated it.
The part that is not about policy
Even setting the policy aside, forty-eight near-identical pages is forty-eight pages to keep current. Prices change, clinicians leave, a treatment is discontinued. The realistic maintenance outcome is that most of them go stale, and a stale page with a departed clinician's name on it is worse for a prospective patient than no page.
That maintenance argument usually decides it faster than the spam policy does.
What this means for an operator
Take one location page and list what on it is true only of that location. If the answer is the city name, merge it. If it is clinicians, services, pricing and hours, keep it and put those higher up the page.