Electronic medical records – a disappointing mirage for clinicians and research
Nathan Lawrentschuk
Abstract
Throughout Australia and New Zealand we are in a phase of rollout of large electronic medical record (EMR) systems. This should be the best time for research in the past century. Data should be able to be collected prospectively and, with improvements to clinician workflow to incorporate collection, research should be booming [1]. It is not. Why have EMRs failed and why are they still failing? Predominantly because EMRs are originally designed for "Billing and Management" strategies first, patients second, health workers third …with research not even placing. How does this occur? Well, firstly most of the larger systems are from the USA (e.g., EPIC [Epic Systems Corporation, USA] and Cerner [Oracle Corporation, USA]). Hence the focus is on billings. That is what they were designed for – systems focused on finance and management. It is easier to work out how many lignocaine jelly tubes were used in a day (due to the need to record for billing) as opposed to knowing the number of catheters that were difficult to insert. What would be better for patients and clinicians? Both could be recorded but the emphasis should be on clinical care and not billing. Clearly, the order is incorrect. One has to wonder: do the designers of these EMRs ever actually follow a doctor around for a day and work out the bottle necks and difficulties in navigating an EMR? Most people agree the best designed systems are done by clinicians and not IT aficionados. However, to commercialise EMRs they of course require high IT functionality – but then the actual purpose is forgotten – they should be focused on clinical care and research, not billing and management structures. Certainly, the latter needs to be there but they should not be the major priority – clinical care and research need to be squared off and done first. This would make software work for clinicians, and not against them. Also, when large EMR companies present to hospitals the local management are dazzled by graphs and measurable clinical outputs. The clinicians are certainly not consulted in any meaningful way. It always seems like going to a car yard for EMR salespeople – the best car model with sunroof, electronic windows and turbo is test driven with the opportunity to upgrade to leather upholstery and even further safety features. However, what is delivered is the more than slightly used base model with smaller engine and no features. The prospects of an upgrade that were touted as "easy" are limited, generally futile and extremely expensive. Wouldn't it be refreshing if hospitals agreed to pay the same amount for systems but insisted that the emphasis must be switched with features A–D that are clinician and research focused? The conglomerate EMR corporations will then have to change their systems – but instead the dog takes us for a walk. The EMRs for private practice are no better. They try to be everything to everyone and really end up being a master of none. They are the Steven Bradbury of software – they win because all the others have fallen over – not because they are necessarily the best. It is not good enough that we praise being able to access EMR remotely or that we can join multiple hospitals (of course they need the same system – yet another huge problem known as incompatibility that no one wants to tackle). These are all base features that are really not features – like telling us a car should have wheels – and so the narrative goes. What should be baseline is now passed off as an added extra. Integration with common research databases like REDCap (Vanderbilt University, USA) and Prostate Cancer Outcomes Registries [2, 3] should be already plugged into the EMR and not a huge, unrealistic and expensive "build". Every mouse click should be a piece of prospective data that makes a clinician's job easier not harder. That is what will improve workflow and collect prospective data. Digital images saved in clinical context not randomly [4]. Advanced scanning technology exists, so why are we reading PDFs where we cannot cut and paste any data (e.g., results) to correspondence that was not done in the hospital? Too hard we are told. Not possible. No. With prostate-specific membrane antigen (PSMA) PET-CT being funded for staging and biochemical recurrence from July 1st in 2022 in Australia [5], our EMR should be able to tell us how many were positive, negative or otherwise. We should know answers prospectively if the EMR were built correctly. Before we go and spend more money on EMR, let us at least understand what the priorities should be and only agree to "the best model car" – for clinicians and research. Nathan Lawrentschuk was a co-author on the pro-PSMA study.