A Hospital's Medical Records Were Entered Through Its Meal Provider — The Ransomware Attack on Osaka General Medical Center (2022)

Hello, I am Kurashiki Bouzu (kurashiki-bouzu), born in 1971. I am a former funeral director, and I now live in an old house in Kurashiki, Okayama, working together with my AI butler, Moka.
This time, I am digging up an incident from a little while ago. A large hospital’s electronic medical records were stopped, but the entry point that was targeted was not the hospital itself. The places that think “this has nothing to do with us” are exactly the places that can become an entry point. That is the kind of story this is.
The result first
The result first:
- In the early hours of October 31, 2022, a ransomware attack was confirmed at Osaka General Medical Center (865 beds). Most of the servers, including those for electronic medical records, were encrypted and became unusable (according to the center’s investigation committee report)
- The entry point was the system of the company that delivered meals to the hospital. Attackers are believed to have entered through a weakness in remote-maintenance equipment (VPN equipment) placed at that company, passed through the hospital’s meal-service server, and spread inside the hospital (same report)
- As reasons for the spread, the report cited equipment that had not been updated, passwords that were the same on every server, administrator privileges given to everyone, and the fact that the boundary of “who is responsible for protecting what” had not been decided between the hospital and its contractor
- It took 43 days for the electronic medical records to return, and 73 days to restore the care system as a whole. The number of inpatients in November 2022 was about one third of the same month the previous year (same report)
- The investigation found no sign that patient information had leaked outside. Backup tapes stored far away were safe, and that helped the recovery
What happened — the entry point was the meal company
On the morning of October 31, 2022, the center noticed a failure in its electronic medical records, and at around 8:30 a.m. a ransomware infection was confirmed. Because how far the infection had spread was not known, the network connected to the electronic medical records was stopped, and care was switched to paper records. That night, the incident was announced at a press conference.
At a press conference one week later, on November 7, an assessment of the entry point emerged: it was highly likely that the attackers had come in not through the hospital itself, but through the system of the business entrusted with meal services.
The later investigation report organized that route into seven stages. First, the attackers entered through a weakness in the maintenance VPN equipment placed at the meal company, and stole the company’s IDs and passwords. Next, using the stolen information, they entered the hospital’s meal-service server, used it as a foothold, and gathered the keys (passwords) to other servers inside the hospital. They then reached core systems such as the electronic medical records, and set off the ransomware.
In terms of a house, it is like this. The lock on your own front door was solid. But the key to the warehouse of the delivery company that comes and goes every day was broken, a whole bunch of duplicate keys was taken from there, and the attackers passed through your storeroom into the house itself. The hospital’s front door was not broken; the attackers came in from “a place it was connected to.”
Why did it spread this far?
The report lists several reasons for the spread. None of them is an unusual story.
- Equipment had not been updated: The VPN equipment that became the entry point had not received its fix (a firmware update), and leaks of IDs and passwords had also been confirmed
- The passwords were all the same: The hospital’s servers and terminals all shared the same password. Once one was taken, every server could be entered
- Everyone had administrator privileges: Because of that, the attackers were able to delete the antivirus systems
- The electronic medical record server had no antivirus protection: Other servers had it, but this one central machine alone was missing it
- “Who does it” had not been decided: Whether the hospital or the contractor was responsible for updating the VPN equipment and managing its weaknesses — including in the contract — was not clear. The report summed this up with the words “a lack of IT governance”
The report also raised, as background, the assumption that “a medical institution is safe because it is a closed network.” Even if it seemed closed, it was connected to the outside through the meal company and maintenance lines.
Impact on care — recovery took 73 days
The impact appeared in care itself.
- Most of the core servers, including those for electronic medical records, were encrypted, and traces of unauthorized access were found on about 2,200 terminals inside the hospital. All servers and terminals were initialized and reinstalled
- It took 43 days until the electronic medical records ran again (December 12, 2022). Full recovery, including the systems of each department, was on January 11, 2023 — day 73
- Care results for November 2022 fell, compared with the same month the previous year, to 558 new admissions (33.3%) and 465 first visits (17.9%). Highly urgent surgery and treatment continued, but accepting new outpatients remained stopped for a long time
- Ministry of Health, Labour and Welfare materials state that, at that point, there was “no impact on patients’ lives or similar”
The money side was not small either. The report wrote that investigation and recovery costs were expected to exceed several hundred million yen, and losses from restricting care to exceed more than one billion yen (still being examined closely at the time of the report).
What helped the recovery
It is not all a dark story. What helped the recovery is also written clearly in the report.
One was the backup tapes stored far away. The backup equipment placed inside the hospital was encrypted, but the tapes stored at a remote location were safe, and records up to October 27 could be read out. From November 10, 20 terminals that could view those records were prepared, and emergency admissions were gradually resumed.
The other was the preparation for disasters. There was no plan assuming a cyberattack, but this hospital, as a disaster base hospital, had prepared how to run meetings to keep its operations going (its BCP). At noon on the day of the incident, that meeting was held, and decisions were made on how to continue care with paper records. That, on the day the system stopped, “how to keep things running with paper and people” had been decided was one reason care was not stopped completely.
As for patient information, the investigation of communication records and other evidence found no sign of a leak outside, and the report concluded that the possibility was extremely low.
For people running a small company — entry points also lie in the places you are connected to
The lessons of this incident are not only for hospitals. They apply just as they are to small companies and shops.
- The entry points of the places you are connected to are also your entry points. Meals, cleaning, accounting, website maintenance. Please write down, once, through what kind of system each place you entrust work to is connected to you
- Put “who does it” down on paper. Whether updates to equipment and fixes for weaknesses are done by the place you entrusted or by you. Places where the boundary is vague are left unattended the longest. Checking with a single word when you make the contract makes a difference
- Use a different password for each machine. Keep things in a shape where, even if one leaks, it does not spread to the others. Settings that limit the number of login attempts, and not giving administrator privileges to everyone, are also basics that stop the spread
- Review remote connections left connected all the time. Close connections opened for maintenance when they are not in use. That alone removes one entry point
- Keep another backup in a distant place. And test whether you can restore from it. What saved this hospital was a tape far away, and a way of working on paper
To close
The heart of this incident is not “even a large hospital can be targeted.” The entry point was in a place that looks far from the main work — the meal company. And the reasons it spread were things that can happen in any workplace: not updating, reusing passwords, and not deciding the boundaries.
Put the other way around, the way to protect is not special either. Count the places you are connected to, decide the boundaries, and separate the keys. The record of a large hospital from three years ago can be read, just as it is, as a checklist for a small workplace today.
Sources: Osaka Prefectural Hospital Organization, Osaka General Medical Center, Information Security Incident Investigation Committee, “Investigation Report” and summary (March 28, 2023). Ministry of Health, Labour and Welfare, “Report on initial response to a cybersecurity incident case” (December 15, 2022). Ministry of Health, Labour and Welfare, “On strengthening cybersecurity measures at medical institutions and similar (alert)” (November 10, 2022).
Entrance (other languages)