Imaging and PACS in a Small Clinic: What You Need

A clinic that buys its first imaging device — an ultrasound, an X-ray unit, a 3D scanner — quickly discovers that the device is the simple part. The hard question is where the images will live, who will see them, and how they will reach the patient’s record. This is where acronyms that sound heavier than they are enter the conversation.

Three concepts, briefly

DICOM is the standard in which imaging devices store images. A DICOM file is not just an image: it carries with it information about the patient, the study, the device and the settings. Because of this, an image produced on one device can be opened in another system without losing context.

PACS is a system that stores the images and serves them for viewing. In a small clinic this can be a single server; in a hospital it is a large system with a long-term archive.

A viewer is the software you actually look at the images in. There are basic viewers in the browser, and there are dedicated tools for measurements and processing. Usually the basic viewer is enough for routine viewing, and the advanced viewer is needed by whoever reads the studies.

Does a small clinic need PACS

Not always. Three common scenarios:

The situation	What usually fits
One device, few studies, viewing mainly at the clinic	Orderly storage with backup, and a link from the record. A full PACS is usually overkill
Two or three devices, several clinicians, a need to view from outside the clinic too	Central storage that speaks DICOM, with a browser viewer and permissions
High volume, external reading, sharing with other parties	A real PACS, including an archive and a retention policy

The expensive mistake is not buying too little, but storing images in a way you cannot get out of. A folder on a computer in the exam room, in the device manufacturer’s proprietary format, is a trap: the images exist, but they are not part of the record, they are not truly backed up, and they cannot be transferred.

What really matters: the link to the record

The clinician’s daily problem is not the quality of the viewer but retrieval. When a patient comes for follow-up, the clinician wants to see what exists — a clinical image, a video, an imaging study, a scanned document — without having to remember which system each thing is in.

The right solution is not necessarily to move everything to one place. You can leave the imaging files in the system that suits them, and present in the patient’s record a unified list that points to the items across all sources. When you click an item, the viewer suited to its type opens. From the user’s standpoint there is one place; from the storage standpoint, each file type sits where it belongs.

What this requires in practice

• Consistent identification of the patient across systems, otherwise the linkage will break. A national ID alone is not enough in the case of tourists or infants.
• Linking the study to the visit, and not just to the patient, so it is possible to know in the context of which treatment it was performed.
• Permissions: who may see imaging, and who may only know that it exists.
• Access logging: opening a study is logged, like any viewing of medical information.

Storage, backup and cost

Imaging and video files are orders of magnitude larger than text. That changes the cost calculation: storage grows over time and does not shrink, and so it is worth setting a policy in advance. Three questions are enough: how much space accumulates in a typical month, for how long the material needs to stay available for immediate viewing, and what happens afterward — a cheaper archive or deletion.

As for backup, the practical rule is that a backup that has not been tested is not a backup. A sample restore of a single study, once a quarter, gives a real answer to the question of whether the material is protected.

One question for the device vendor, before purchase: does the device export standard DICOM to a network address that I define, without dedicated software from the manufacturer. An evasive answer here is a warning sign that will cost money two years from now.

What you can skip

A small clinic usually does not need dedicated reading workstations, monitors approved for reading, complex routing of studies between readers, or a two-tier archive. These are high-volume needs. AI components for image analysis are also a separate decision, one worth making after the basic infrastructure works and not before.

Sharing with external parties

A clinic that refers a patient to another physician or to an external reader runs into a practical question: how to send a study. Three common ways, each with a price. A CD or a portable drive works everywhere, but it creates an uncontrolled copy that leaves the clinic. Sending by email is convenient and problematic, because it involves identifiable medical information over an insecure channel. A temporary link from within the system, which expires after a set period and logs who opened it, is the preferred solution when it exists.

With each of these ways it is worth verifying three points: that the patient knows the material was sent and to whom, that the sending is documented in the record, and that the file sent includes the required identifying information — a study with no name and date is worthless to the recipient.

The quality of the order, not just the quality of the image

A point that gets forgotten: a good imaging study starts with a good order. A clear clinical indication, an explicitly marked body side, and a link to the visit in which the study was requested. When these are missing, the reader works in a vacuum and the answer received is less useful. A system that requires an indication to be filled in before the order is sent saves a round of questions and answers on every other study.

A short checklist

1. The device exports standard DICOM to a destination I define.
2. The images reach the patient’s record, linked to the patient and the visit.
3. There is a viewer that works in the browser, with no installation on every computer.
4. There are permissions, and there is a log of every opening.
5. There is a backup, and at least one study restore has been performed successfully.
6. I know how much space accumulates per month, and what the retention policy is.
7. I know how I will get all the images out if I decide to switch systems.

A clinic that answers yes to these seven lines can focus on medicine. A clinic that answers no to two or three of them will discover them at exactly the least convenient moment: when a patient asks for a copy, when the computer in the exam room stops working, or when switching vendors.

Back to the blog