CPOE and Order Sets in the ED, and Cost Effective Care
A recent study by Terrence Adam, MD, PhD et al published in the AMIA Annual Symposium Proceedings Archive on the Implementation of Computerized Provider Order Entry in the Emergency Department: Impact on Ordering Patterns in Patients with Chest Pain revealed a perhaps unexpected impact on the cost of care: CPOE seems to increase the use of some diagnostic tests, though not others. The overall impact of CPOE on the cost of care, patient safety, and outcomes has been studied to an extent, though much of this research is funded by vendors of electronic health information systems (EHIS) and is less than transparent. It is likely that negative impacts on the cost of care mostly go unreported, as there is a lot riding on the success of EHIS for providers, payers, governments, and vendors.
The Leapfrog Group has addressed this issue, and the First Consulting Group and the AHA published a cost-benefit analysis of CPOE in 2003. David Classen, MD et al gave a pretty good summary of the state of the Evaluation and Certification of Computerized Provider Order Entry Systems in 2007; but many such studies focus on the costs of implementation, the costs of having physicians perform these oder-entry functions, and the cost of the overall HIS. None-the-less, these studies often tout the potential savings from fewer medication errors and adverse drug events, from substitution for less expensive medications, from revealing hospital charges for various tests and studies to the ordering physician, from the reduction in unnecessary repeat testing, from the increased use of clinical pathways, and the indirect potential economic benefits of a reduction in morbidity and mortality. The Leapfrog Group predicted a $549 M a year savings nationally from CPOE; and a study at Wishard Memorial Hospital found a 12.7% reduction in charges that would amount to more than $3 M annually (tens of Billions nationwide). Billions and billions: this is Carl Sagan territory here.
A very recent review of 22 articles related to CPOE in the ED, published in the Annals of EM by Andrew Georgiou, PhD et al, found that CPOE “was associated with an increase in time spent on computers (up to 16.2% for nurses and 11.3% for physicians), with no significant change in time spent on patient care. Computerized provider order entry with decision support systems was related to significant decreases in prescribing errors (ranging from 17 to 201 errors per 100 orders), potential adverse drug events (0.9 per 100 orders), and prescribing of excessive dosages (31% decrease for a targeted set of renal disease medications).” Two of the common components of CPOE that may directly impact the cost of care include the use of standardized order sets, and decision support functionality. In a paper entitled Physician Perspective on Computerized Order-sets with Embedded Guideline Information in a Commercial Emergency Department Information System, Asaro et al reported that “guideline-consistent CPOE order-sets alone failed to improve adherence to a clinical practice guideline (for acute coronary syndrome) in our ED. The primary contributions to this failure appear to be cultural and organizational, but the lack of decision support functionality is a likely contributor”. Payne, et all discussed the Preparation and Use of Preconstructed Orders, Order Sets, and Order Menus in a Computerized Provider Order Entry System that involved the development of an amazing 667 order dialogs, 5,982 pre-configured (quick) orders, and 513 order sets organized in 703 order menus for particular contexts, such as admission for a particular diagnosis. Unfortunately, other than the AMIA study mentioned in the beginning of this post, there just isn’t a lot of pre- and post-implementation evidence base to indicate the overall cost-effectiveness of CPOE and order sets in ED practice. In part this is because of the issue of defining what cost-effectiveness is, and in part it is because there are so many different variables involved (hospital type and commitment, physician integration, financial incentives, etc) and so many different EHIS vendors.
Those whose job it is to implement EHIS and CPOE in the ED and the hospital need to strike a fine balance between the costs of these products, their impact on physician productivity, the complexity and integration of clinical decision support; the need for constant updating, and the impact of order sets on the cost of care. Certainly, functions that promote patient management efficiency and safety can have an impact on reducing length of stay, expensive complications and errors; but like squeezing a balloon, CPOE can save money one way and increase costs in another. When the provider is a capitated hospital, ACO, or other integrated, at-risk health care organization; ignoring the less apparent downsides of CPOE and order sets in order to achieve a quantifiable upside can be disastrous. In striking this balance, it appears to me that there are a few requisites that the literature or common sense appears to support:
1. Order sets should be easily amended on the fly to eliminate unnecessary components from the set
2. Charges or costs for each test, procedure, medication, and order set total should be posted alongside the check box
3. Order sets for particular clinical conditions should include lesser and more comprehensive (costly), or risk-stratified, alternatives
4. Simplified clinical decision support algorithms, and pretest scoring tables and calculators (PECARN, Wells, CCHR, etc) should be incorporated, in a non-mandated way, into CPOE, focusing particularly on more expensive and often overused services
5. Cost effective care strategies, like those in Choosing Wisely, should be easily accessible through the HIS and CPOE
6. Nursing use of order sets prior to physician (or NPP) evaluation should be restricted to less expensive or time-sensitive ‘always ordered for these signs and symptoms’ sets, and easily cancelled if needed
7. Order sets should be broken down into subsets – for example, not every major trauma case needs blood products or liver panels or pelvis x-rays.
8. Only a limited number of tests truly need justification to be ordered and reimbursed – don’t flood the provider with mandatory justification indicators for every order entry option
9. Physician scripts for discussions with patients should be readily accessible through the HIS.
10. Alerts for potential duplication of previously performed tests or procedures should be provided if possible
11. Avoid complicated clinical algorithms in decision support, as these will be ignored in any case
All of these should be weighed against the hassle factor and physician productivity constraints they may impose. This all assumes, of course, that the claims that CPOE and order sets and clinical decision support improves outcomes and reduces errors, complications, and costs, are valid and proven. Dr. Kevin Klauer has a nice video presentation on CPOE that questions these claims; and the Physician Order Entry Team (POET) at Oregon Health & Science University has an excellent website that addresses many of the still-debated issues and unintended consequences of CPOE. Like Dr. Klauer, I am not yet convinced that the alleged benefits of CPOE or order sets, especially in reducing the cost of care, are achievable or real. If, like many of the ‘advances’ in medicine, CPOE and order sets are thrust upon providers in the ED or in the hospital; following the above provisions should facilitate a reduction in the cost of care, or mitigate some of the unintended consequences. This is especially important if the hospital or providers are financially at risk for these costs.
