Thursday, July 19, 2007


Scanning continues to be an issue in EMR offices. Electronic lab results have eliminated a great deal of scanning, but DI, consult notes, and various other non-electronic data from the outside (such as forms that patients bring in to be filled) still need to be integrated into the EMR.

In my office, we keep no paper data for patients; as a result, there are 10 to 15 things that need to be scanned in daily. Here is my scanning process:

I have stacked in-boxes beside my scanner. When something comes in, it gets put into one of three boxes: correspondence, DI or lab. Mail comes in the morning. My scanning tech then scans paper from each box to a folder on the PC's desktop called "files to upload"; everything is scanned into pdf format (that gives the clearest picture). She'll do one in-box at a time, and name the file using the patient's last name: example, smith.pdf. When all the in-box has been scanned to the folder, she'll upload all the files to the EMR, and then attach each one to the proper patient. I order things via the EMR, so if there is a consult/DI request pending, that gets matched to the file that was uploaded; the consult/DI then gets taken off my outstanding list (that's how I know a letter was received).

Once an in-box is uploaded and filed, the tech deletes all the files from the folder on the PC desktop. The physical paper gets moved to the lowest in-box, which we call "pending shred". All uploaded files now appear in my Practice summary page as "unfiled", and awaiting my review. At the end of the day, if I'm happy, I take the paper and put it in the "to shred" outbox.

If a file was attached to the wrong patient, I can see that when I load the file; I can take the physical paper from the "pending shred", and put it to re-scan. Alternatively, I can save the file to my desktop, and reattach to the correct patient; I can also print the file, and put it back in the in-box. There is a high level of redundancy to avoid misfiling. We can probably get rid of the "pending shred" box.

The scanning tech will add extra information when attaching the file to each area. For example, there are drop down lists in the Lab area, indicating whether this is an ECG, a histopath report, etc. She can also put in extra comments, such as "insurance form". When there is a "match" with a pending request, the specialist's name or the name of the DI facility automatically appears in the information area. It is a good idea to work with your scanning tech to make sure that the scan is done correctly, and that extra information that is helpful to you is added; things that are done at the front save you time.

We scan everything to pdf; we tried different file formats (jpg, tiff), and pdf was by far the best. I thought about whether to scan to OCR (optical character recognition) so that text was recognized from the outset, and decided against this. The reason for that was that I needed an exact copy of the form, since the original is shredded; OCR is not 100% accurate. If I need to copy part of a note to put into comments (such as the last paragraph in a consult letter), I will use OCR on my tablet. This leaves the original form unchanged. I bought Adobe Pro; I click Document, OCR. The same thing can be done using MS Document Imaging, which comes with MS Office. To use Imaging, load the document, then Print to MS Document Imaging. Use OCR (the little eye) on the resulting file, then copy and paste to comments.

Comments will give you a very quick overview of everything when you load an area in the chart.

Here is an example of a DI area:

Signed Off DI Reports

Date Collected Date Signed Off DI Facility

Dec 30, 2006 Jan 4, 2007 North York Diagno
comments : XR left ankle normal

May 15, 2006 May 18, 2006 Unknown
comments : XR right knee There is slight prominence of the tibia1 spines and there are osteophytes on the patella. The changes are in keeping with early osteoarthritis.

I know that all this seems like a complicated process, but it works, and now it is not even time consuming. It just takes some time to get used to, and to make sure that it goes smoothly. Working with your staff and your scanning tech really helps.

I think it may be helpful for a regulatory college like our College of Physicians and Surgeons to have a look at scanning. We need some ground rules on whether it is OK or not to OCR from the outset, or whether it is preferable to save an exact copy of the incoming material. Perhaps even some rules about acceptable file formats. There seems to be a lot of confusion about the right thing to do, and we need to have some guidance here.



Anonymous said...

Oh wretched scanning. When you have an EMR
Medical Records = Scanning

Anyway, one point you should recognize about OCR is that you can scan a multipage image that retains it as an image and then does the OCR and has it available in the background. Essentially maintaining the original document, but having the advantage of the OCR.

Is OCR needed? I think pretty rarely right now, but a full text search engine of all the scans might be interesting one day. Very useful in a document management company I worked for that focused on lawyers. Not sure it would be that useful in a medical practice.

Now I'm babbling.

I love the pending shred box. We can't bring ourselves to shred the scanned papers either. Too many What ifs!!

Dr. David M. Kaplan said...

Michelle: We scan to MTIFF format (multipage tiff) in a compressed form. The file size is as small as PDF and is VERY clear.

Moreover, the Advanced Document Manager in Nightingale reads this format and it is extremely easy and fast for my secretary to file for the four MDs in our office.

The format opens in MS Office document automatically which allows for direct OCR of portions of the notes like the final paragraph of a consult note.

After four years (and one week) of dealing with this issue in the EMR, I have come to the conclusion that this is the best solution in Windows.

As you know, our Hospital has agreed to interface with nightingale this fiscal year and this will eliminate a lot of the scanning of CT, MRI, other hospital DI and Hospital consult notes.

Michelle Greiver said...

Thank you for the comment on OCR retaining the original; I did not know that. The text searching is not that likely to be useful; there is a lot of filler in our reports. The important information is often in the last paragraph(diagnosis, management suggestions and meds changes). We need to figure out how to put that in some form that is compatible with EMRs.

I'm really looking forward to having a hospital interface; at least that will remove some of the scanning.


Parimi said...

Some interesting issues to talk about:
How do you approach HIPAA when a third party(scanning service, scanning tech) is involved in scanning MRs?
Dr.Greiver: How did you approach non A4 size sheets like US images on photo paper?
I think there should be a greater reason when to shred your hard copies; not when you feel good. I am just kidding.
Thanks for your input on OCRs.I agree with the comments.
I incline on PDF rather than TIFF files. I cannot imagine TIFF file occupying the same space as PDF.


Michelle Greiver said...

Thanks, Dr Parimi

Our situation in Canada is a bit different than yours. In Ontario, we have the Health Information Protection Act ( As health information custodians, we can have a third party scan medical records, provided they abide by the PHIPA regulations.

Scanning non standard pages should be in your scanner's manual. We sometimes scan legal sized documents. My scanner won't accomodate very large sized pages, but I have yet to receive one.

As far as destroying the original records, here is what our College's policy is:
"When a physician scans his or her paper records to convert them to electronic form, the original paper records may be destroyed in accordance with the principles set out in this policy."


Dane Meuler said...

We have noticed that the process of scanning that includes scanning to a network directory and then attaching to the file is time consuming and prone to mistakes.

As a result we created a scanner interface that allows the user to see the image and verify quality it on the touchscreen of the scanner and then indexing is done using the database of the EMR to validate that the patient ID or name is accurate prior to comitting the scanned document to the EMR's document repository.

Do you believe that this method of scanning and indexing would be beneficial in your case?