Ha! A timely and critical question indeed. Suggest the current September 2013 issue of Journal of AHIMA (American Health Information Management Association) focused on "Healthcare's Data Revolution: How Data Is Changing the Industry and Reshaping HIM's Roles" on the web at:

http://journal.ahima.org/

(Note: portions of the contents available only by subscription).

Nearly every article in this issue is related to your question.

regarding the subtopic "evidence based" - a question remains: for how many 'services' can/do we have strong evidence? only a few guidelines have a S3 level, and more will never reach this level, b/c of the complexity of health care services.

From an implementation perspective we already know, that RCT results cannot be transferred 1:1 into action; thus we already know that a decision made by a healthcare professional including a participatory approach, leads to an evidence-informed (or evidence oriented) decision. This does influence the question: "what do you want to measure" while using your clinical data? performance quality? patient related quality of life? health related outcomes?

In which areas will/should improvement happen? for example

do you want to know something about the feasibility of an innovation (while gathering clinical data) or

about the degree of implementation (not in your clinical data, but important to judge where only change happened or real improvement can be seen) or

is there a requirement to deliver clinical data as part of an external quality assurance system and you want to decrease the burden to gather or optimizing your routine data (than see also lit-tipp from James Bowmann)?

Thankyou all for these very helpful thoughts. My real focus is very practical - I am impatient to see some changes at the clinical coalface, so some of the refs from the AHIMA journal provided by James are close to what I am looking for. I am working on the premise that if we get more knowledge and INTEREST in the use of our local data, we might help transform the culture and thus affect higher level evidence-based policy development eventually. I think we need to work on that conduit between the real world clinical data analysis, especially of genuine clinically relevant outcomes and the higher policy/monitoring work. So I use this quote in my work sign-off: "The time has come to shift from evidence-based practice to practice-based evidence' - Dr Christopher A. Longhurst. I think if we engage everyday clinical practitioners including medical, nursing, allied health in making use of the data that they have, to see what they are doing and what is working when they make changes, we have a better chance of driving collection of relevant clinical outcomes, improving information management practices including data quality - this then all cascades on to recognition for the need for clinical registries and standardisation, etc. But I think we need to first engage the community at a practical level to make use of the available data and this is what I am trying to develop an approach for. But this is not a simple thing. I think it requires access to locally relevant data, but also applied analytical expertise to make use of the data at service level now - clinicians want something that will make a difference to their practice outcomes now, I think....? So I'm looking for how to frame up these barriers and find affordable, practical solutions...

One of the barriers you may well strike is the EMR traditionally has not been designed to provide "evidence". The provision of simple tools that can easily mine the data is what is required. Now the quick caveat; I speak only from the radiotherapy field; whilst we have always had those people able to crawl through a database and extract the appropriate data it has been somewhat of a specialised process and often the effort greatly exceeded the resources. It is only a relatively recent development that we are provided with tools to be able to easily "mine" the data. Your point about the need to standardise the way in which we record the data. Reporting information has some meaning to the reporter however when we mine "a brisk reaction" how do we quantify that to useful information. We want the simple tools that can help us to complete Deming's Cycle in our practice. It is not a simple task upon which you are embarking and as you have identified there is need for broader education and engagement of all involved, but there is also a need to break some of the "walls" that may exist between different data repositories. Good luck on the changes you are seeking to implement.

There is also, as I understand it, an issue with legal responsibility. Any expert whose decisions are underpinned by the use of an expert system are required to understand the expert system and take responsibility for their decisions. At the coal face, the clinician must be able to justify a course of action to the patient, their family and their superiors / professional body, especially in the case of adverse outcomes. It doesn't matter how accurate a decision support tool is, if it seems like a black box to the clinician, it is not fit for purpose.

From our experience in primary care/general practice, the care pattern that you choose to initiate data driven quality assessment and the scope for improvement for this care pattern is important. You're likely to get more engagement for patient safety issues, and those that can identify individual patients with actual opportunities for improvement (low false positive rate). These are likely to be those with less complex decision making (as perceived by the clinicians assessed) attached to them. For example, when developing an intervention to improve prescribing in primary care, we got good buy in from GP practices for reducing high-risk prescribing of NSAIDs and antiplatelets as a QI topic (but less so for reducing underuse of preventer inhalers for asthma or for increasing uptake of anticoagulants in atrial fibrillation). You can find the list of indicators we finally used as outcome measures for a trial evaluating the intervention here:

http://www.implementationscience.com/content/7/1/24/abstract

I think the first thing is to think about the purpose . What kind of information is used to make decisions at all levels of the organization? .

The information system should be able to answer all questions and point to possible questions that require extensive analysis of the data.

If users need to research clinical information in order to conduct clinical trials , systematic reviews or meta- analysis , that is an option. But it's another if the systems must provide data for epidemiologic study of medical records , clinical information that is basic to be very well systematized to be retrieved .

The other type of information to be taken into account is the information derived from the many data, which can be obtained by analyzing the behavior of the variables in the system. That is fundamental to organizational decisions .

What is needed to solve these questions ? Experience . Experience of use , analysis of the real needs of those who use the system to retrieve information every day. Whether statisticians , clinicians, librarians, administrators ... everyone should contribute in the process of management information to build a solid foundation in the design of systems information retrieval .

What is an information system for? to solve an information problem ? .

Key to the design is to connect the actual need with practical resolution .

Central to the health area is the search for evidence for practice and decision -making. There are few who specialize in this type of information. International agencies and organizations to practice and evidence-based research , are those who seek and propose to apply and recommend standards in the development of different types of research.

It is crucial in the design of information systems consider EBM terminology and research methodology.

The best evidence and decision is obtained from the best research , and that is of direct benefit to the patient. But without a good information system , decision makers can do little .

Greetings

The first basic issue in answering this question is what is the quality of the clinical data? Clinical data is often incomplete, recorded in many different ways, classified differently in different clinical units, and often very difficult to translate into the form wanted for use in quality improvement.

Another key issue is paying attention to potential political sensitivities. If the data can be used to indicate that one unit or hospital is performing better than another (which you probably would want to know if you are aiming at quality improvement), you can be fairly sure that these results can be misused by management, policy-makers, clinicians or politicians, leading to a massive quagmire of no progress as people argue about the meaning and use of the data. Therefore it is useful to spell out early in the process exactly what one is trying to achieve and how, then see if it is possible to get all parties on side with your work.

Victoria, which clinical data ... would be my first question ... depending on what my intention is.

I agree that the political impact on which data when and how to use, can not be ignored

HL7 is worth considering, as it is based on standards and interoperability http://en.wikipedia.org/wiki/Health_Level_7

I think the starting point is for managers to understand the importance of information for planning. If they understand that they cannot plan without critical information, they will search for it be it from a paper base or an advanced IT system. People often say we have always done it this way and they expect what worked ten years ago to work today despite the change of conditions e.g more patients etc. Understanding that a lack of information to base one's planning on leads to misplanning will compel people to make use of clinical data especially during this evidence based era. If one claims an improvement, then results should show it. Are the patient outcomes better? How were they before and how much resources have been used to produce this result? Is it efficient etc?

Susan - this is a great question. One of the issues regarding quality is that client satisfaction is not ... yes NOT part of CQI or TQM ... at least not as an outcome variable. Health outcomes, however are. For example, over the past 15 years all of the nation there has been the revamping of hospital foyers to look and feel like a great entrance to a resort. While this may result in an indirect way to influence the culture and climate of the organization and hence quality but was predominantly done because of the perceived need to increase customer satisfaction. However, the important outcome should be some global measure of 'health' outcome ranking done by an independent agency - but this means you will have to do your analysis while avoiding an ecological bias.

All very salient points to keep in mind. I agree, the political consequences of performance and monitoring related analyses are a huge issue to manage. And tying the purpose of an infrastructure improvement to specific clinical topics is difficult, but important. I am quite interested in patient reported outcomes as an alternative to the patient satisfaction/patient experience metrics. Quality of life tools seem to be a necessary tool for comparative clinical effectiveness studies and I gather they are quite different to patient satisfaction scores...

Susan - thanks for the opportunity to respond to your comment. The issue is a practical one not political - some accreditation boards do not take patient satisfaction as part of the quality scores. However, in my experience in medical research where patients and product usage interact to mediate compliance; if patients don't like the product or service compliance goes away. So it does depend on how much the end consumer likes the product if they have a say in the timing of the product administration. So you can have a great product but if folks have a natural aversion they simply will eventually go to the service they like better even if the clinical outcome is not quite as good. This is one of reason why the hospitals remodeled the rooms and foyers. With that said there are pivotal areas that indicate poor quality in health services delivery. For instance, iatrogenic illness, post procedure infections like clostridium difficile and staphylococcus aureus (whether MRSA, VISA, etc.), etc. However, some increases in quality measures may have assignable causes like large service areas increasing wait times between home visitations, which could be a target for managerial tampering of the quality process if the manager is not aware of the service providers catchment parameters.

It seems that "quality" invovlves three major domains: (1) access (timely appointments, waiting times, etc), (2) traditional medical outcomes, and (3) patient satisfaction. If at all possible, avoid "process measures" and focus on outcomes: access outcomes, medical outcomes, and customer satisfaction outcomes. Patients are not always in a position to "vote with their feet" (go to another clinician or hospital) - their health insurance may limit the choice of providers, or in urgernt/emergency situation, they might have no choice but to go to whoever is open or available or closest. Yes, "patient satisfaction" does not correlate directly with "medical outcomes" (a surgeon with an attitude might still produce the best valve replacement outcomes), but since the patient is a paying customer (directly or indirectly), and is often forced to go to a limited group of providers, then the medical community should be held accountable for patient satisfaction as well as access to care and medical outcomes.

Donald, it sounds like you are right in the thick of performance measuring services. It reminds me of an interesting finding in the UK, where the NHS produces a public hospital scorecard. The interesting thing was that people didn't go to the highest quality hospital, they still went to their nearest . there seem to be many hidden factors that influence patient choice other than quality, so relying on patients choosing the best hospital (thru patient satisfaction) to increase quality might not be too effective. Use of patient satisfaction does not seem to be as popular in Australia, but I wonder what is a better way to use patient input to drive quality.

Susan and James - I think the merging of these last 3 posts, yours, James, and mine moves us a bit closer to understanding a measure for CQI. I agree with James that patients are not always in a position to choose but if not then that may be a barrier to service as clients / patients stop seeking those preventative services. You could see patient satisfaction as a contributor to preventative services more than those services that are limited by insurance plans, medically underserved areas, poor public transportation, drug use, crime, socioeconomic status, etc. Considering that the U.S. Affordable Care Act is covering the uninsured, there should be less need for charitable service by U.S. hospitals which would require a greater partnership between allopathic providers and local public health agencies on preventive services. The types of metrics that focus on hospital and public health (preventative) services may add to the next generation of taxonomy for health care service quality. This would result in a strategic rather than incremental improvement in the U.S. health care system.

Donald, incorporating public preventative health In performance monitoring would be great. What outcomes metrics could one measure for this rather than process measures? The time frame for preventative services outcomes would be very long for most things?

James are patient reported outcomes such as QoLs an alternative to patient satisfaction? It's not quite the same but it does incorporate the patient input component.

James, I think that is a really good point - the WHO notes that patient satisfaction/ experience is very important in health system evaluation. I guess, especially so where patients have less choice in their healthcare provider. So perhaps an overemphasis on hospital foyers, etc, could almost be viewed as confounding or 'gaming' in this area of measurement. Or perhaps the different dimensions of patient satisfaction need to be appropriately weighted (eg surroundings x1, food x3, staff interactions x10) with some standardisation of the measurement process between organisations, otherwise I would think it could be quite meaningless and misleading. That would then be better using the benchmarking of these results for QI rather than expecting patients to use their feet & market pressure, which they do not really have the option of.

Most of the key issues that need to be addressed by a program to improve use of clinical data for health service management and clinical quality improvement are discussed in following sites:

www.health.vic.gov.au/qualitycouncil/.../vqc_guide_to_using_data.pdf

www.nas.edu/21st/health/health.html‎