To describe Selayang Hospital’s IT issues as “cancerous” is apt, and I applaud the health minister for realising that, like cancer, the problem is incredibly complex.
Having seen the general sentiment regarding this, please allow me to further elaborate on the complexities of the issue – in the hopes that interested spectators will be able to understand the magnitude of the problem.
Oncogenes
Some patients are born with mutations that predispose them to cancer. Selayang Hospital is no different. With all due respect to the pioneers who spearheaded the THIS project back in the late 90s, the seeds for collapse were already planted at the time. The two major problems were:
1. Non-strict implementation of a relational database. In essence, although the clinical notes appeared structured on the front-end, it was a total mess in the database.
In addition, it was stored in a massive BLOB. Figuring out this mess significantly complicates data migration into any new system, although we finally did manage to (kind of) do it at the end.
2. Lack of obsolescence management. A patient may eventually die of cancer, but preparation of their death makes things easier after they pass.
Attempts to replace Cerner began about 10 years ago, but faced challenges at every step of the way due to the fact that Cerner was expected to live forever.
Since this version of Cerner could only run on Windows XP, every other software acquired by Selayang Hospital had to be backward compatible with Windows XP, further amplifying the obsolescence problem.
Collapsed At Home
Sometimes, cancers can result in symptoms which may bring the patient to hospital. Selayang Hospital had an episode of that in 2014, which resulted in the activation of its business continuity plan.
That was my first experience of “CPR”, so to speak, and I watched in horror as the “plan” was for doctors to write their notes in Notepad (yay!), store them in a shared network drive (where everybody had read/write access), and copy-paste those notes back into Cerner after it came back online.
A perfectly fine approach for one or two hours of downtime. Not that good when the downtime lasted a month like in 2014, resulting in the requirement to copy-paste thousands of clinical notes (i.e. the number of patients seen in Selayang Hospital for that entire month).
Cue the original version of BHIS, which I presented nervously as a young, naive, recently-gazetted radiologist to the hospital director – who essentially patted me on the head, said “Good Job”, and promptly ignored it – deciding to keep to the tried-and-“tested” (ahem, “failed”) shared network drive method.
Chemotherapy (The Wrong One)
Most cancers have effective chemotherapeutic agents. I’m not an oncologist, dammit, but I’m pretty sure that using the wrong first-line agent would significantly complicate management.
We had placed our hopes on an upgraded version of Cerner — but in a totally unexpected twist, we watched in dismay as the tender was awarded to another HIS company.
Needless to say, the next five years trying to implement that other HIS was an exercise in futility – but we managed to (kind of) solve the data migration problem and successfully backed up everything (including the BLOB) into Speedminer (a data warehouse provider) before Cerner kicked the bucket.
Remember how Spock transferred his consciousness into McCoy at the end of Star Trek II: The Wrath of Khan before dying? Just like that. No magic blood or reincarnated Tribbles.
CPR CPR CPR
In reality, the terminally ill cancer patient has breathed their last breath, last rites have been performed, and everybody has made their peace – it’s time to move on. But in our timeline – Cerner’s death took everything from us.
Everything, that is, except all the data which we had already backed up in Speedminer. Working with the Speedminer engineers (bless their souls!) in the midst of the pandemic, we managed to create an interface for the doctors to access all the old clinical notes previously entered into Cerner.
As for current notes — some resorted to using the shared network drive (hence, the complaints about notes being deleted), although most finally (and somewhat begrudgingly) resorted to BHIS, which was serving 30,000 hits a day without issue when I left.
What’s Left?
To be best of my knowledge, here’s a summary of what Selayang Hospital has left which, when you think about it on a macro scale, isn’t that bad:
- Clinical notes, radiology reports, and lab results from the dawn of time (1999) until January 2021 in Speedminer.
- Clinical notes from February 2021 until January 2023 in BHIS (but only for doctors who used BHIS instead of the shared network drive).
- Radiology reports from February 2021 until present in the Backup Radiology Reporting System and the still-functional RIS.
- All the radiology images, stored in the still-functional PACS.
- Fully functional lab analysers, but without data connection.
- Fully functional network infrastructure, albeit mostly 100Mbps.
The glaring omissions are lab results from January 2021, which may no longer be retrievable or easily migrated.
The subtler fault is the fact that the clinical notes are no longer unified, i.e. one patient may have data scattered under different identifiers throughout the multiple databases.
What’s Next?
Personally, I would summarise the solution into three broad areas:
- Replace all the hardware so that Selayang Hospital can move away from EOL (end-of-life) software like Windows XP and Windows Server 2003 – and follow proper enterprise-level IT management norms.
- Implement a new HIS – one with the willingness and expertise to do the complex additional work needed to clean up this mess, i.e making sure that data from all the data sources above are easily available to the doctors.
- Change management – the new HIS, whatever it may be, will face massive pushback from the overworked, under-appreciated, and long-suffering staff in Selayang Hospital. Confidence in this new HIS needs to be earned from the users (and by extension, the patients).
He’s (Only Half Brain) Dead, Jim!
The moment we see something like “CPR x 3” in a patient’s clinical notes – it usually means that if the patient survives, there will be some sort of neurological sequelae.
Selayang Hospital will never be the THIS hospital it was in 1999.
Too many mistakes have been made, too many corners have been cut, and the damage is permanent. It will not be possible to retrieve the lost data.
Even when data is not lost, it will not be possible to unify every detail of each patient into a single streamlined database.
The realistic aim would be to bring Selayang Hospital back to a stable, functional state and accept that not everything will be perfect, but enough to ensure good patient care, maintain data privacy and integrity, and facilitate, instead of burdening, the most important resource of all; the health care workers currently stuck in this seemingly endless nightmare.
“That Radiologist” is the radiologist who developed the Backup Health Information System (BHIS) at Selayang Hospital and has since resigned from the public health service.
- This is the personal opinion of the writer or publication and does not necessarily represent the views of CodeBlue.
