Centralized Data Repositories - What are the Benefits and the Risks?

The following question is posed by Dr. Tracy Monk, a BC physician:

I was wondering about a central question which never seems to have entered the discussion about the approach to E -medical records and ­ who should be the data manager?

The idea of all of us getting stand-alone systems which leave us to be the data managers ­ which means the patient record is lost if computer stolen or goes down plus all security concerns etc ­ doesn¹t seem ideal.

It seems to make more sense to have a centralized, standards based (HL7 or equivalent ) data repository which I would simply connect to when I am with patient. In this model I am not a data manger, not responsible for the security and encryption and firewalls etc. Isn¹t this what the HMOs have? What do they have in UK?

It seems to me that the priority should be to create the centralized data repository ­ esp one that can create a truly cumulative lab report by PHN ­ that would be great ­ sounds like something truly useful to connect to and then could have electronic lab ordering and prevent waste. Then I could have all the useful bits of an electronic record without all the worrisome and irritating parts.

Why would I want to be a data manger? Am I concerned about something that is irrelevant?

Would love to hear your thoughts.

Tracy

If you have ideas about data repositories and would like to comment on this posting, please click on the 'Comments' link below.

Comments

[Morgan]

Tracy,

You raise a very interesting question. With every interesting question there are pros and cons in both directions.

This is definitely not a purely technical question. If it were, it could be argued that data would be better held in a centrally designed system with all the extra security, back up and access bandwidth. This is certainly compared to the local installations that physicians may have in their office that may / may not have properly working back up, redundancy, etc.

Of course with a centralized model a robust and fast connection in each office / place of care provision would be necessary. It would have to be robust and have guaranteed up time, etc. It would still likely be a bit slower than a local solution, but with a good network this isn’t as much of an issue, but when the network is down then the EMR is down. Not everyone has a good connection.

Trust is one of the major non technical challenges. Do you as a physician trust the ministry to house the data? Do all of your patients also Do all other physicians? Do all patients? There is a trust relationship between MD and patient that allows the therapeutic relationship to be effective. Would this be lost by patients knowing that their data being housed by a third party?

Also remember that today we are already the custodians of the data in the paper charts. Although providing “IT support” for paper charts requires an extensive knowledge of how to use tape, staplers and paper clips. Computer maintenance is as much a black box to most physicians as the human body is to an IT expert.

Choice is something else that physicians want in an EMR. With a single centralized solution, this choice would be, obviously, limited to one.

A further question would also be to look at clinical data standards to that elements of the record CAN be appropriately shared. Currently these do not exist in a comprehensive format. If they did and were implemented, then MDs could have choice to use centralized systems or local systems, depending on their need while still being able to appropriately share data.

I think the bigger issue is ensuring open data standards so that there are options for providers to choose various EMR solutions while still being able to move patient data when needed in the delivery of high quality care.

-- Morgan

[Allan Horii]

The EMR I am using stores the data locally on the office server, but also backs up the data hourly to a remote site managed by the EMR vendor. In theory, if my server's data became corrupted or lost, I could still access the information from the remote site. I imagine that it is really a question of "when," rather than "if" I will have an office system failure. Perhaps this arrangement is the best of both worlds.

In this instance, I am still responsible for the security of the data on my local server, since I decide how much to invest in firewalls, encryption, and office security measures. I can only hope that the EMR vendor is a watchful guardian of my off-site backups.

[Morgan]

Allan,

I hope that you have more than hope with your vendor - it should be very explicit in your contract what they do and do not have access to and where they are storing the data.

Morgan

[James Busser]

Well, those who know me can imagine my taking a deep breath... ;-)

At first glance, centralized data can have a lot of appeal for some of the reasons mentioned.

However the following limitations & concerns need to be borne in mind.

If this centralized data represents the only copy of "your" results (as opposed to an "extra" copy) "access" issues arise from its being "offsite" and outside your control:

- slower performance compared to a local server
- inability to access the data during internet service/connection problems
- inability to access the data if the central server is down or if there are queuing problems
- inability to control the timing of any server service interruptions for maintenance or upgrades
- dependence on an additional layer of access policies and procedures potentially putting your access at risk e.g. vulnerability to be "locked out" of during disputes, also unclear access policies e.g. if you move or retire and are "removed" from the list of authorized people permitted access, how do you get access for questions/issues that arise after?

Even if it is only an "extra" copy, and you retain electronic "originals":

- where does the data reside? Is it somewhere over which neither you, nor government has physical custodianship as is a concern with MSP data which I understand is in Texas, thus subject to foreign breach of our privacy stipulations

- how attractive a target is it to being hacked? One giant box whose location cannot be kept hidden is much more attractive to be hacked than a bunch of smaller systems, more value for identity theft etc.

- does your patient **want** their data in a giant repository? How would they ever (later) get it removed?

It is much easier to restrict what Joe Bloggs, Miscellaneous doctor (or health care worker) can access, if that data does not even *reside* in the giant repository. Patients would prefer their doctors to manage their information than to have government do it. Admittedly patients DO get upset by doctors who do not pass the information along to those who need it.

I agree that having data in a repository can bypass the need to depend on other doctors for the information, for with a repository you may not even need to know who else is looking after the patient (though you should want to!). Nowadays, patients sometimes go to walk-in clinics expressly *not* wanting news of the visit to go back to the family doctor. Kind of hard to do, once the data is in the repository!

So poor communication among doctors ought not be the basis to *require* patients' data to be put in a repository. Let us not forget we COULD have a model in which patients carry or own a copy of their data or a summary thereof.

I would rather a model in which the custodians of information (family doctors and specialists) can "store and forward" information to each other. I like the idea of a patient-specific "keyring" that permits identification of one or multiple doctors in a chronic care role, according them ongoing access (with provision for pieces to be "held back" at a patient's request). Over and above this, the referral process would add consultants until they discharge care to one of the other physicians. Doctors on the keyring (and doctors within their on-call groups) would have authority to access and share each other's information on a patient except where vetoed by the patient.

Admission of a patient into a recognized facility (Emerg, Treatment facility) could activate access **from that facility** to doctors' EMR information, until discharge.

So in closing I think
- it would be foolish for doctors to relinquish ownership of a sole/primary copy of results into a repository
- for performance purposes, doctors should want to have the data locally
- if doctors opt for an .asp model, conceding the performance downsides for competing benefits, they must still assert custodianship responsibilities (due diligence).

And I think if we do support/tolerate repositories, they should serve an archive (non-primary) function, and should be subject to the types of control and restrictions I offered above.

[tracy]

Many great thoughts - this is great!

I still think the repository model offers me more choice - inlcuding the option of not havinga "paperless" office if I don't want to. We should simply use a terminal/ device of our choosing to access patient data from a centralized, secure, redundant data service that stores data from whatever source - labs, hospitals, outside imaging, ECG's, important and relevant consults. The data should be accessible by any authorized and triple-validated user from wherever they happen to be, not just from a single PC. All access should be logged and auditable - by patients as well as government. The provider must be trusted, but independent of government (much as a bank is).

The other nice thing about this model is that it doesn’t press me to have to have a “paperless” office – if that’s not what I want or I don’t feel ready for it. I would have access to all the useful lab data and cumulative reports – but it wouldn’t be essential that I chart my patient’s erectile dysfunction or quantity of vaginal discharge in electronic form. Just how much of a rush are we in to make that info electronic anyway? and how big an improvement in patient care results if we do? Maybe some things are actually best left in a manila folder. . . (legibility remains a prob – but in a longitudinal continuity of care model – I’m the only one reading the note – not same as hospital – and even I, seldom re-read the note. Could dictate a note if I want to . . .) All the labs, x-rays, pharmanet and consults would be organized and searchable in cumulative form in the repository.

I agree that I need to retain custodianship and part of how I would do that is not turning some of the patient's most private info into electronic data - ever.

It’s the notion of the paperless office, the costs purely on MD shoulders and data management and security issues that are focal points of discomfort and resistance – not the technology and its potential usefulness. (“Paperless office” is far from paperless – have to pay to scan in the paper that comes in, then pay to shred it – then pay to print it again to send out if needed as hardly anyone is connected up anyway – repository model makes all that unnecessary.)

I think it’s really important to flesh out this debate before we go too much further down one road. I'm so HAPPY to see it happening!! Would love to hear from more people with EMRs about what they see as the strengths and weaknesses of their systems and how the repository may or may not address them.

[tracy again]

Further thoughts - maybe differnt data should be treated differently and stored differently
1) re: patient carrying own info - smart card would be nice. Would be esp great if it could contain the 10 things an ER doc or ambulance crew should know in first 20 min of treating my patient - code status, CO2 retainer, allergies, dementia etc
2) When it comes to labs and pharmanet- that info is already in giant potentially hackable boxes - it would be nice if it was in a more useful cumulative form in one location. I'd really like to stop all the wasteful lab ordering - means that we need true cumulative reports across province by PHN, and electronic ordering. I think it also means Meditech has to come up with a way to translate all its document storage and focus more on clincial applications and HL7 compliance.
3) Maybe printed material like consults should be held only in patient ownership, doc office or hosptial system.
Maybe deeply personal matters belong only in manila folders or within in-office systems for those who already have them.

I'm no techie - maybe these ideas make no sense. It also seems really imporant for the public to understand these issues and participate in the debate. EMR is seen as a panacea (it's not), and many seem to think electronic continuity can replace longituduanl human continuity (I don't think it comes close) and I think everyone has unrealisitic expectations about what an EMR can do. We have to focus on which bits of it are actually doing something useful. Seems to me that data crunching is the forte - so let's focus on at least doing that properly and centralizing the labs and drugs for a start. Disturbingly - clincial applications do not seem to be a priority for national or provincial IT strategies.

[tracy last thought]

One last thought:
Maybe CMA and/or College of Physicians should be the trusted authority/custodian of the centralized database.
We are poised on the brink of huge health care delivery re-dseign.
At the very least, the fact that we have a choice needs to be clear, and there should be some debate before we go any further down the road of stand-alone EMRs. Clinicians and patients need to understand the choices and their implications, and particpate in the debate.

[Peter]

Centralised data as mentioned above is two edged. So you can have all the IT experts needed in one site to maintain services. Service guarentees, up time and such. I am not so worried about many technical issues raised above. I am more interested in personal and business issues... what happens if they don't perform... or stop performing... or are bought out. Access to your data is then very much at risk.

[Alan Brookstone]

Just published on the web: Doctors fear the computerization of medical records could threaten patient confidentiality. This issues was debated at a BMA conference.

A survey released by the BMA on Thursday shows members of the public recognise the potential value of the new system - but many are seriously concerned about its security.

Dr Richard Vautrey, a GP from Leeds and lead for IT on the BMA's GP committee, said: "Doctors have a fundamental duty to protect the confidentiality of the information that patients tell them.

"And many doctors remained concerned that we seem to be sleepwalking into a system where a significant amount of personal and medical information is held on a national database without either us or patients fully understanding the long-term implications."

Survey Results

75% of patients would not mind their health information being held on a central database
75% had concerns about the security of information
81% were worried about accessibility by people other than the healthcare professionals providing their care
93% said the public should be fully consulted about the proposals before they are finalised.

Link: BBC NEWS | Health | Confidentiality fear over records.

[James Busser (reply)]

Some of my assertions attracted commentary prompting me to support or clarify some of what I had written. Having gone to the trouble, I may as well post it FWIW even though I understand that the commentary has, in the meantime, been removed.

I wrote "Patients would prefer their doctors to manage their information than to have government do it." in reply to which was posted [SAYS WHO. DOCUMENT THIS OR I DON'T BELIEVE IT. IF YOU CAN DOCUMENT IT I'LL ASK WHETHER THOSE POLLED WERE GIVEN THE CURRENT CONTEXT FOR THEIR CONSIDERATION.]

reference: Angus Reid Health Care Information Survey (April 1999) in which patients believed that physicians should have primary responsibility for holding and controlling access to the complete medical record. However, they may not have been shown this blog as a pre-requisite to responding. >:-)

I wrote: "it would be foolish for doctors to relinquish ownership of a sole/primary copy of results into a repository" in reply to which was posted [ THIS IS PATIENT INFORMATION. THE QUESTION IS HOW THE DATA SHOULD BE USED AND STORED, NOT OWNED.]

The patient does own the *information*, and the right to ascertain how it has been used, as well as any use beyond the purpose for which it was collected (e.g. research or use by insurers).

However, the doctor owns the record of the information (or the "copy"). I believe this provision exists in law, and provides for protection of doctors. Their ability to to later justify actions that they took (or did not take) in providing care would be seriously undermined if any party could demand removal or alteration of information on which the doctor's actions had been based. Note, this does not preclude *auditable* improvements to the record of the information.

[Ron Joe]

To all,

Very good thoughts here.

I’d like to throw another idea out into the arena, which I feel is important to take into account. I agree that the treating physician should own their copy of the patient data and thought processes documented on each patient encounter. As professionals were are accountable for our actions based on our “owned” data set. Whether it is on paper or digital form should bear little relevance. In the digital arena, although some of the barriers to sharing data have been removed, this should not automatically translate to data sharing without first agreeing upon the scope of each professionals responsibility.

For instance, I am a primary care physician. What if a critical result from a lab comes in after hours? As the College of Physicians would reiterate, we need to be contactable 24/7 for these critical results and act on them accordingly. If we share our data do we share our responsibility? Would a nurse practitioner (who has access to the lab data) at a computer terminal elsewhere act on the critical result by calling the patient into the ER to treat them? What if something untoward happens to the patient overnight. Who would be liable? The family physician who had his cell phone off, or the nurse practitioner or other provider who was aware of the critical result but did not act on it?

At what point does access or available access to data in a repository translate into responsibility?

I have a patient in front of me with a medial complaint. Would central repositories and other data sources need to be accessed and then digested before I treat this patient? What if there is a lot of information, say 1000 pages? How would we as physicians find the extra time to digest this extra information?

The walk-in clinic model has been successful “business wise” as they treat patients based on a reduced data set. What if the standard of care is such that it is requirement to download the said 1000 pages and go through it all before seeing the patient?

[tracy]

Spent the weekend having fascinating discussions with several friends from US who work in IT, one who works for UN, one in banking, both were astonished that BC would be going the route of stand-alone EMRs for medical information. Both noted the enormous advantages of centralization of a subset of critical info held in a secure standards-based data repository. Both noted that info held in this way is dramatically more secure than that held in the c drive of a GP’s office - and that centralized info is also far more useful to patients in that it can be more easily shared with triple validated user health providers. In Boston, rigorous audit of access to all patient records has ensured security of centralized patient info for over 10 years.

As per Alan’s British survey results - Patients are OK with their data being centralized and want it to be usable, accessible and secure – all arguably best achieved with central repository, especially for the most useful subset of info which is labs, drugs and summary. Any street level thief can walk away with a doc’s c drive – it is not necessarily secure in my office for a multiplicity of reasons – my incompetence and lack of training at data management being at least part of the problem. With the centralized repository model – we can have economies of scale by centralizing expertise of data management. Patients think that a key advantage in making certain elements of their record electronic, is to improve doctor to doctor communication of their key data – but - if data is in stand alone EMR in doc office – the computer has to be turned on and connected all the time for patient data to be accessible elsewhere – this is not practical

Many of the concerns about central repositories in the discussion above – are simply arguments for the central not being the only copy of the data. It seems reasonable that if docs want to have an EMR at their end of the respository, replication can keep local records up to date. As my IT friends have pointed out, lots of systems work centrally without any of the concerns above being “showstoppers” and centralized systems are working extremely well in many industries with highly secure data. It is easy to restrict access to any computer system and governments of many countries have been doing it for years.

Everyone felt that it would be great for patients to carry or own a copy of their data and that this should certainly be a component of a centralized system.

Everyone felt strongly that the information belongs to patients and that neither government not doctors should “own” these data. The question is how different types of patient data should be used and stored, not owned. It was agreed that docs should still assert custodianship responsibilities as long as it was about due diligence, not ownership, nor did it seem reasonable to place docs in the position of being data base managers – much better to centralize this expertise..

Consider the numbers: 6,000 MDs in BC - Stand-alone EMR costs $5,000 – 10,000 per doc per year in ongoing expenses just to maintain (ie proper backup, training and support etc) – that’s $30,000,000 - $60,000,000 for the province annually. At the end of 10years we will have spent 600 million in BC alone, to have a system of 6,000 stand-alone data repositories.

The most useful subset is the labs, drugs and summary (perhaps patients should carry their own summary in a smart card). We need to consider that there are different types of info in the patient record – (data points like labs, vs narrative info like chart notes – which we seldom re-read or review in old records) – and it may be more appropriate to store some information differently from other info depending on the delicacy of its content and utility for other care providers. We need to focus on where the VALUE IS ADDED in the technology. The centralized repository of a subset of clinically useful data is the most useful bit of the technology. We would all love to have a nice searchable cumulative lab summary for our patients and save on filing costs and be better organized. We need not take on the headaches of data base management with stand-alone EMR to achieve this. A centralized, expert-managed data repository can do this for us. This is the priority – stand-alone EMR is not.

We ALL need the computer hardware and the connectivity – supplying it to us, so that we can better link to one another as well as regional and web-based resources and repositories, would be very useful.

I think our medical leadership needs to exert political pressure to ensure the creation of a secure and accessible and clinically useful central repository which contains an agreed upon subset of patient info. The CMA should take the lead in creating informed public discussion about the options, and should have a vision for what we want the system to look like in 10 years. CMA and BCMA must remain objective advocates for patients and clinicians, educating us all about the strengths and weaknesses of different approaches to systems development as we stand together on the threshold of a brave new age.

[yep tracy again]

Re: If we share our data do we share our responsibility?

I think the answer is no - this question highlights the importance of the fact that patients actually NEED a family doctor – they need one person who is the responsible nodal point to whom results go and who is responsible for coordinating their multi-disciplinary care. Nurse practitioners etc have an important role to play – but responsibility for action must remain on the shoulders of the treating MD.

IT systems development must reflect the importance of continuity of care. One patient and one doctor traveling the road of health and illness together through time makes for wiser and more fulfilled doctors and patients who use the system more cost-effectively over time – the evidence supports this. We cannot allow longitudinal relational continuity of care to be designed out of the system. It is the jewel in the crown and the key to the sustainability of the system. IT systems must reflect this central truth.

Re: What if there is a lot of information, say 1000 pages?
If we focus on where the value is added in the technology and only centralize labs, drugs and summary – could not be 1,000 pages. We get the “killer ap” of organized lab data, without the headaches of data base management at our end – unless we want it.

[Ron Joe]

Hi Tracey,

You have very good points. I’ve been following the US system to a certain extent too. I agree that the Canadian model should emphasize the family practitioner as the nodal point and have general custody of the patients’ treatment timeline.

However, there are certain barriers to implementing a centralized data repository system here. The Medicare fee structure in Canada is completely different to that of the US. From my last review, FP’s in the U.S. Get paid twice as much more per visit. There is also reimbursement for coding and granulating a visit by charging higher value codes. A complete granulation of review of systems and physical exam for each visit takes a lot of time. For doing all those pesky check boxes dropdowns, pick lists: diagnosis codes, medication codes, etc, they get a higher fee (Approx $75US).

Medtuity is a good program if you wish to granulate and feed to a central repository here is a link to a flash demo:

Link: DocumentingAnEncounter.

This link demonstrates a headache consultation, but my typical patient may come in with 2 – 3 problems and you’d need to granulate those too to send to the central repository. Of course this takes a lot of time. Many FP’s are retiring and leaving their offices without replacement because of low reimbursement. How would this extra work be reimbursed aside from the costs procurement and maintenance of computer equipment and software?

[Tracy]

Thanks Ron

I agree completely that the work involved in longitudinal continuity of care needs to be supported and better remunerated and that the profession is hemorrhaging as a result of failure to do so at many levels.

I am spending more and more time each year filling in forms and spending hours on voicemail with pharmacies and community nurses and all the members of the interdiscinplinary team. This important work has become more complicated, and consumes more and more time yet it is the stuff that keeps patients out of hospital and is a considerable service to the health region for whom we carry it out – as well as the patients. Because the walk-ins are not doing this work, it is an ever larger burden on those of us who value the longitudinal continuity model. Classic example would be the patient that I note over time to be dementing slowly. Because I KNOW the patient – I notice this and begin to liase with all the home supports and pharmacy to blisterpack meds and family members and make placement plans – years before the patient ends up in crisis and hospitalized for 8 months. There is eneromous and unrecognized value in all this – I find it all very fulfilling and wouldn’t have it any other way – but my deep sense of fulfillment with the work is not going to make it any easier for me to find a new partner when mine retires next year, because we are not attracting enough new docs to this wonderful work that is the key to the sustainability of the health care system.

However, when it comes to documenting patient encounters – this seems like the lowest priority and lowest value in e-technology applications to health care. The patient narrative is really just our meeting notes – how much time do we ever spend reviewing these when we summarize an old chart? How often do we re-read our own narrative note? Consider the nature of the info in these notes: colour and quantity of patient’s vaginal discharge, specifics of their erectile dysfunction, marital conflicts etc. Exactly how much of a rush are we in to make some of this info electronic anyway?. And exactly how has patient care been improved when we do? Maybe some info actually is best kept in a manila folder.

In terms of value added – the narrative note documenting my encounter seems like the absolute least value. In the US – administrative micromanagement has mandated that the encounter be documented and it is on the basis of this type of micromanagement that the US has the highest administrative overhead of any health care system in the world – why would we rush to implement aspects of this costly micromanagement if nothing is gained in patient care?. Maybe we should just take the best and most useful bits of the technology as a starting point – all of which could be available in a repository, without me necessarily adding anything to the respository at the time of the patient encounter – I could dictate a note that stays in my office or continue to chart on paper while having a computer in the examining room to access labs and drugs from the repository and useful decision support tools at point of care.

I think the repository should contain only the info where value is added with the technology– labs, drugs and summary. Forget my narrative note - it is not a priority in terms of value added. Why on earth are we focusing on the least useful, most cumbersome aspect of the technology which contains the most delicate of patient content – as the first step in the adoption of very useful technology? Pressing docs to get stand alone EMRs mandates all the most difficult aspects of adoption of the technology – while saddling docs with all the costs and headaches of database management. I’m no technie – but the whole thing makes no sense to me. If I was to enter any data into the repository on my encounter – it would be vaccine status, maybe weight and BP – the things that are actual data points – plus summary info. (Summary info would only be the 10 or so key things ER doc or ambulance crew need to know about my patient in first 20 minutes of treatment – eg no code, CO2 retainer, dementia etc.)

Ron – as you’ve noted: “A complete granulation of review of systems and physical exam for each visit takes a lot of time. . . doing all those pesky check boxes dropdowns, pick lists: diagnosis codes, medication codes, etc” Everyone I know who has an EMR notes that it has slowed them down – but the one thing everyone loves is nice organized lab data on retrieval. So why are we in such a rush to adopt the aspects of the technology that will decrease our efficiency instead of just adopting the aspects that will increase our efficiency. I am mystified that we are in such a rush to adopt aspects of the technology that were developed specifically to serve expensive bureaucratic micromanagement of MDs in the US. We are in an ideal position to learn from the experience of other nations and adopt only the parts of the technology that will be cost-effective, efficient and improve patient care. It seems to me that stand-alone EMR is not the right first step. I am astonished that our medical association has been so smitten with this approach that it has gone into the business of vending a particular brand of EMR. Clearly more debate is needed before we travel any further down the path of stand-alone EMR and our association must poll its membership on the question of whether or not it has put itself into a conflict of interest position by becoming an EMR vendor.

There is huge potential value in electronic info in health care – but just as there is value in MRI – we don’t choose to use the technology to do total body screening scans – it isn’t an effective, efficient use of the technology – employing the technology in this way is expensive and doesn’t get patients any better cared for. We should be using the same judgment and common sense in our use of e-technology. We should be sure at all times to focus our energies on what we do best – clinical assessment and treatment. Passive clicking on boxes and drop downs seems unlikely to improve our clinical skills and will certainly waste our time. A fabulous searchable cumulative database of labs and drugs would be a great asset. Our use of technology must make us wiser not just wired.

[Alan Brookstone]

There is a debate currently taking place in the UK on the privacy issue and implied consent - how far does implied consent extend?

In an article just published in Health Data Management, this is discussed in more detail -

Link: Health Data Management | England Debates Access vs. Privacy (July 13, 2005).

There is also a link to a downloadable Word document that is referenced in the article and which makes interesting reading relating to the issues of consent.

[tracy]

Very interesting stuff Alan. What is not clear to me is if those who are pro-stand-alone systems are suggesting that this is somehow a better way to protect patient confidentiality? This seems like a false pretext and seems to be based on an economically flawed sort of protectionism on the part of our medical community. Stand-alone systems seem like the worst of all possible worlds. It saddles us with burdens which are not our expertise and decreases our productivity. It is basic economics that economic growth occurs best when we allow trained experts to do the things which they do best.

It seems that we are engaged in an economically self-destructive sort of protectionism. The following excerpts from Andrew Wheelan’s Book Naked Economics seem pertinent
(also fun vacation reading . . .)

Governments, like individuals, will get themselves in serious trouble if they consistently overspend on things that do not raise future productivity.

The fallacious logic of protectionism is alluring . . . strategies such as “self-sufficiency” and “state leadership” were hallmarks of the postcolonial regimes, such as India . . . India has been one of the world’s great economic underachievers for decades.

Economics tells us that companies shielded from competition do not grow stronger; they grow fat and lazy. . . The preponderance of evidence suggests that open economies grow faster than closed economies

People in Seattle should not grow their own rice. Instead they should build airplanes (Boeing), write software (Microsoft), and sell books (Amazon.com) – and leave the rice growing to farmers in Thailand

Basically - productivity is what makes economic growth – specialization and engaging in what we do best, makes us productive. Things only improve economically if we are able to focus on what we do best. There are ways that technology can assist us to increase our productivity and ways that it can slow us down – economists apparently actively debate whether or not computers and other kinds of information technology have made Americans more productive. (Undoubtedly, in some ways they have, and in some ways they haven’t) And here we MDs go – knowing in advance – that that the e- documenting of the encounter is the aspect of the technology which decreases our productivity, and that the burdens of database management are costly and time consuming – all ready to take on this on. We seem to be doing it under the false pretext of being the guardians of patient security – instead of realizing that we may be the incipient authors of our own economic demise.

Once any info is electronic – it is potentially much less secure. If some subset of info is made electronic – it is safer in a repository managed by experts – much like the repository the feds have just built for our census data. It is highly secure, all access is logged and audited. In the case of medical info – arguably only treating MD and patients themselves should have access to all levels of the subset of info that is in repository and should be the only ones – to forward or permit access to parts of the repository. Hospital access could be similar to current standards or those used in Boston. It seems to me that these kinds of safeguards could be built into a system and that implied consent is a dangerous assumption.

The debate in the UK seems like all the more reason to pause before we go any further. I want to use the aspects of the technology that will improve patient care without compromising their confidentialty and will make me a wiser and more productive doc. Stand-alone EMR just doesn't seem like the route to any of this. If we continue down this route - in 10 years we will have spent 600 million to create 6,000 out-of-date data farms - all the software etc will need replacing etc by then - what is the point?

By all means we should be pressing to get funded for the hardware and connectivity - but perhaps we should be focussing our political efforts on ensuring that national and provincial IT efforts are going in to creating useful things for us to connect to - instead of all the Ministry tracking applications that Infoways money is pouring into. None of it will improve patient care and we are standing by and letting the money be spent in this way.

I have no idea if I’m right about any of this – I just know that we need the debate. If there are others reading this – I would love to hear from them.

[Scot Mountain]

I've just caught up on this thread, and I wanted to comment on Dr. Monk's posting from July 14th regarding what information is most effectively collected in an office based EMR. I have spent over a year working in an EMR based environment, capturing everything including clinical notes, labs, and physical exam findings, including the data points such as BP, weights, etc.

While initially the EMR slowed me down, in fact I found that eventually I was just as fast capturing notes on the computer as I was writing them by hand. This was largely due to the fact that I started out by being able to type a little bit, and after a few months of capturing all my notes this way, I was able to type very quickly. The big advantage of capturing the clinical notes was that they were LEGIBLE, and for some of my partners, that proved to be a monumental leap. Other major advantages to capturing clinical notes on the computer included the fact that they could be cut and pasted into referral letters, and the fact that they were filed by diagnosis, so that I could pull up only the notes dealing with a certain issue (eg. all the notes relating to anxiety), without having to wade through the whole chart. In addition, assigning of the diagnosis to the note could automatically update the face sheet, so that a diagnosis became part of the patient summary.

Having said that, I fully agree with Dr. Monk that the clinical note should be relatively low on the priority of useful information to collect. By far more useful, as she points out, is anything with discrete data points (labs, BP's, weights, growth parameters for kids), or patient summary/face sheet information. It is extremely useful to be able to efficiently study trends in any of this data, and this adds a whole new element to the ability to provide effective care. For anyone considering a staged implementation of an EMR system, I wholeheartedly agree with Dr. Monk's proposal that this information should be collected first. Eventually, I think there is also something to gain by capturing clinical notes, but this is definitely more difficult and more time consuming, with less initial return.

I have recently returned to work in a large hospital based environment, with a very extensive networked computer system which has taken exactly this approach. Lab data and anything dictated to the system is available on the computer, and is available system wide (with appropriate access safeguards), while patient notes are still kept in paper charts. While it is a bit frustrating having to move from one to the other to get the whole clinical picture, it is clearly a reasonable start to moving towards a comprehensive electronic medical record, as evidenced by the fact that it has been adopted by an institution that has hundreds of users every day.

[tracy]

Hi again - thought this was an interesting point on the confidentiality issue from BMJ

Derrick C Attfield,

Anonymous clinical data repository technology negates the need for consent from research subjects

Information confidentiality is clearly a function of the ability of those analyzing and using the data to identify particular individuals from identifing demographic information. With the design and implementation of eHR's that separates identifying data completely from clinical data the paradox is finally resolved. One way to achieve this separation is to endow the patient with the only identifying access key - electronic or biometric/biological

I would add that in this way, info in a central repository could be accessible for research purposes without any connected identifying info. Patient would give consent for ongoing access to their primary MD, and it would be up to patient and their doc to grant access or forward info in the repository. eg - referral to a specialist would grant access to patient's repository data. In no way shape or form should government be the custodian of the repository. Perhaps CMA should be guardian in collaboration with a global proivider of secure data farms like those used by banks etc. And again - it should only contain a subset of less persoanl info than is contained in narratives. I can have an EMR over top at my end, but I don't need to.

The enormous costs of our shouldering the burden of data base managment really seems tantamount to all the economic pitfalls of "protectionism" as outlined above. As stated above, it makes no sense for people trained to build jet engines to grow rice, they should work for Boeing. Likewise, I do not want to be a data farmer, as Bones always said to Kirk "Dammit Jim I'm a doctor not a ______"

Why does my assocation want to make me a data farmer? While a select number of gifted GP's (like most of you reading these pages), may have the skills to run their own EMRs,and the time and patientce to do so - i know that I do not. As the experts who are provinding advice at this time of sweeping policy change - I plead with all of you to ensure that the broader view will be represented, and to ensure that the BASIC needs of the patients are met first.

We need to remember that all of this is really very low priority in the vast scheme of things. Last month - one of my patients was unable to get a timely appendectomy. We are supposed to be advocates for our patients. What could be more basic than an appendectomy? Every dollar spent on all this IT is one less dollar to staff an OR. Primary care is in crisis, the system cannot deliver the basics of timely treatment for the seiously ill or ensure that people have a family doc, and our assocaitons are obsessing about costly IT

I believe that we need the IT - but we have to be very careful to use only the parts that will actually be helpful and will not bankrupt either docs or the system as a whole. Again - I believe that the repository model with only the selct modules of drugs, labs and summary is the most cost-effective, appropriate use of resources that ensures greater protection of patient data and has the flexibility to allow anonymous data access for research purposes, and idetified and highly regulated access for clinical purposes.

[tracy]

Association’s Double Standard on Repositories:

I have been very surprised by the fact that while our association has taken a stand against data repositories, it has put the use of a repository (the web-based CDM toolkit) into its proposed new contract for GP’s. It seems to me that there are good repositories and bad repositories.

A good repository , would be like those used by banks and credit card companies, or probably like the new national census data repository. These are maintained by expert staff and are as close to unhackable as can be achieved. I would presume that while these are virtually impossible for unauthorized users to access, they are easily used by those who are authorized and don’t interfere with rapid workflow in banking and credit. A good repository medically should be held by a trusted third party guardian, like the College.

A bad repository, would be easy to hack, not held by a third party guardian, and difficult for intended users to use seamlessly. The CDM toolkit appears to be all these things, and yet the association, while opposed to repositories on what seems like flawed principals, has put use of this repository into the proposed new contract for GP’s.

Here are some thoughts of GP’s about using the CDM toolkit:

Re: Privacy concerns using the toolkit
Some doctors have checked with both the CMPA and the College and have been sent letters warning them against the use of the toolkit.

On using the toolkit:
Much effort to get everyone on as a group, set up all those digital certificates, usernames, passwords etc etc. After all that I tried entering data on a few of my patients – very time consuming, frustrating because it demanded exact dates (including the day) for lab results, date of diagnosis etc, when an approx. to a month or quarter would be fine, that made too many columns (on for each date of a lab test for instance) etc etc. Then your password is blocked every 6wks, have to make and remember a new one. and you can’t reuse an old one…then final straw I learned that we can’t ever get the data back

I am profoundly confused by our association’s position. If anyone can enlighten me it would be much appreciated.

[Alan Brookstone]

Tracy, you bring up some important points. This discussion on data repositories is long overdue and is certainly only beginning to scratch the surface of issues that must be debated in greater detail.

For those not from British Columbia, the CDM toolkit is a web-based data entry tool that allows physicians to track certain chronic diseases according to specified criteria and then view and manipulate the data in order to evaluate individual performance vs. groups of peers or physicians on a province wide basis. The toolkit also provides a mechanism to print lists to facilitate patient recall.)

My personal impression of the CDM toolkit (as a physician user) is that it is an intermediate solution for physicians who are pre-EMR or who elect not to implement an EMR in their practice. Once physicians have EMR systems that allow them to manage and track information within their practice for chronic disease, the need will be lessened for a standalone web-based CDM tool.

If the CDM Toolkit is to remain viable and useful as an adjunctive tool to EMR, there needs to be a mechanism to port information (that has been manually inputted into the toolkit by physicians or their staff) into an EMR so that the data is not 'wasted'. This is the interoperability piece that needs to occur.

The confidentiality of patient information stored in the toolkit is of paramount importance. This issue has to be addressed and the data management needs to be protected by air-tight policies in terms of who is able to access that information.

One of the features that I do like about the Toolkit is the ability to view my patient population in relation to others in my community and across the province. This is only possible because of the fact that we have a repository. So, what we need to do is get to a point where we have the best of both worlds - security and privacy of information, interoperability with EMR systems in physician offices and the ability to manipulate data (where appropriate) to analyse and improve patient outcomes.

My experience with technology, particularly in healthcare, is that it is a strange beast. Sometimes the best direction forwards is to try an imperfect solution that does certain aspects of the job effectively. Over time the application is refined to the point that by version 3 it is a much more effective solution and by version 5 becomes indispensable.

Although I know this is not an answer to many of the questions you have posed, I think it is very important that we begin and continually move forward despite the unknowns.

As Lewis Caroll said, "Where shall I begin, please your Majesty?' he asked. 'Begin at the beginning,' the King said, gravely, 'and go on till you come to the end: then stop."

[Alan Brookstone]

This is an article that I wrote for Canadian Healthcare Technology magazine - Sept 2005. Some more added discussion around the different groups that govern data elements and who should have overall responsibility for management of the information.

Link: Alan_Brookstone_Health_data_Sept_2005.pdf (pdf document).