Introduction

Health services and health systems research are developing as true interdisciplinary fields of inquiry where ideas from many different perspectives are combined into conceptual frameworks to address health problems. One of the key challenges to these developments is difficulty in resolving misunderstandings of shared terminology. And one of the most problematic words that I will assert here is never understood in a useful and shared manner is ‘efficiency’. This paper will document some of the attempts to define efficiency in health care and illustrate circular reasoning and other problems that arise in these definitions, especially in trying to understand health care innovation. The driving motivation of the paper is to try to point a way forward to finding approaches to defining efficiency that will promote innovation and advancements in health services and health systems.

Information technology advances, in particular, have been slow to be adopted in health systems as compared to the rest of our technology-connected lives. Many of the advances that one would think would be seen as advancing innovation have raised questions in policy circles as to whether they actually add value. And questions can be raised as to whether the health system actually is demanding these advances. But the key question is what the definition of value one is using is and how it relates to other words, such as efficiency. Recent thinking in defining access to health care can also be shown to assist in forming some interdisciplinary conceptualizations that may be able to help move the field forward. Nevertheless, the overarching sense of the present state of thinking is a shocking level of misunderstanding and unclear use of language that will not be remedied easily.

Most of this paper will be a tour through many of these definitions and difficulties in seeing clarity in existing research. Especially in the United States, free market ideas have suggested that optimum levels of efficiency and innovation will be driven by the marketplace, and thus there is no need to define these terms specifically. This view suggests that the market determines value; however, information asymmetries between patients and providers make it impossible for these markets to work in the usual market efficiency sense. Therefore, though the rest of the world has completely accepted the need for public health care systems and regulatory structures that need to define and promote optimum levels of efficiency and innovation, we primarily will use the unique American lens as an example. But every community and every population in the world faces these same challenges in dynamic health systems within their own complex contexts. This relatively short essay will illustrate and relate those complexities as clearly and concisely as possible, so that health services and health systems researchers around the world can use these ideas as a springboard to research that will lead to better outcomes and healthier populations. At the end, important conclusions can be drawn that will help to see where key recent advances have been made, but where much more work is needed. The key underlying difficulty is determining the role of quality of care in these definitions.

Definitions and background

In circumstances in which health care service or delivery quality is deemed by policymakers or other decision-makers as irrelevant, or is easily measured, the issues raised in this paper will not be important. But, in the United States, almost all relevant stakeholders working together through the National Quality Forum (National Quality Forum, 2012) have endorsed over 1000 quality measures, and therefore efficiency in health care must take some account of quality. Other countries have similar arrays of metrics, most being driven and organized by their public health care systems. For example, in the U.K., the National Health Service has over 200 Indicators for Quality Improvement (National Health Service, 2012), which is focused on measures of clinical effectiveness in areas such as acute care, long-term care, and planned care. Once we have an accounting for quality, conceptually we seek value, and value is simply defined as quality divided by cost. And then once we have value, the question is access: Which people or populations have access to this value? And, finally, everything is dynamic and capabilities are not static, so how do we account for, motivate, and measure innovation over time relative to efficiency?

Quality and value

The pervading definition of quality in health care for decades has been to use the structure, process, and outcome formulation (Donabedian, 1966), though in recent years ‘patient experience’ has been added to this list, and even cost, though this potential addition will need more discussion below (https://www.talkingquality.ahrq.gov/content/create/organize/type.aspx). Since value has been defined as quality/cost, cost cannot be seen as a separate component of quality very easily. Efficiency then could be a concept related to cost (saving or minimizing costs), one related to both cost and quality (some sense of optimization), or as equivalent to or directly related to overall value. Unpacking these concepts, and doing so in a manner that does not just speak to economists but all researchers in health systems or health services, will turn out to be much more difficult than it might seem at first.

Another problem is whether we need to ‘price’ quality or not. Cost is naturally expressed in a currency form that can be aggregated across classes of costly activities. But if we really are going to have a value equation expressed in a currency format, then we also need to be able to price quality out. There are at least three concerns that could be or have been raised about such pricing; most of these issues are outside of the scope of this paper, but need to be mentioned for completeness. First, one has to consider questions of composite quality measurement (Shwartz et al, 2009) that remain relatively poorly understood. Composite quality measures can be reflective as they aggregate different measures all assessing the same underlying latent construct, or they can be formative as they aggregate different measures each assessing different (possibly quite uncorrelated) attributes of a multidimensional construct. Resulting cost/quality trade-offs can move in many directions at once and lead to difficulties in measuring a single price of quality, with different problems when quality is determined formatively vs reflectively. Second, cost-effectiveness analysis as opposed to cost-benefit analysis, which prices all relevant measures, has arisen in health care precisely for these concerns, with a long history of discussion (e.g. Shepard & Thompson, 1979; Torrance et al, 1981). Converting effectiveness into currency is not necessary to make resource allocation comparisons, and direct monetary valuation of human life can be avoided in cost-effectiveness analysis. Third, we cannot trade quality across time the way we can borrow and lend currencies. This problem actually is part of a much deeper problem regarding the nature of time preferences and discounting, where we would want to express all value in net present value. But assumptions regarding using real interest rates for discounting, which may make sense for monetary costs, may not make sense for quality and for other aggregating health care decisions such as purchasing health insurance (Bradford & Burgess, 2011). Individual preferences may vary, but the most important interpersonal allocation issue is equity.

Economists tend to draw a titanium wall between efficiency and equity, though in the context of this quality/value discussion one actually can question whether the clear distinction is possible. Standard economic theory tends to suggest that efficiency is about the size of the pie and equity is about how the pie is divided up. And, moreover, that the pie generally can be made ‘bigger’ if one does not worry about how it will be divided. But another issue, which I will not discuss too much, is that even if we get ‘efficiency’, this does not imply equitable allocation of those services to those who value them the most. This issue is highly related to the concept of access, a term that one seldom hears in domains of economics other than health economics, but is a given in most health discussions.

Access

Access to health care generally is considered a societal right and has roots in the ideas of horizontal and vertical equity (Culyer, 1995). Horizontal equity is defined as providing equal health care to people who objectively have the same characteristics in some respect; sometimes this is noted as having the same need. Vertical equity is defined as treating people differently who objectively are different in some respect; sometimes this is noted as having different need. Challenges in determining both horizontal and vertical equity also will be generally outside of the scope of this review; however, issues in determining differences between them will be related to some issues in defining efficiency, as noted later.

What does access to care mean in a socially networked interactive world? Though increased interactive technological networking has the potential to improve what we have termed digital access, this also could create greater access disparities for some patient populations who are not electronically socially networked (Fortney et al, 2011). We also posit that five domains of access (geographical, temporal, financial, cultural, and digital) each need to be separated into actual and perceived access as determined by objective vs subjective criteria. And separations between actual and perceived access also can be sources of the types of quality differences that make efficiency and value so difficult to define and measure. The person-centered care movement, for example, is an attempt to make sure that the people receiving care and their perceptions of their own access are important and should not be overlooked by a societal or system approach to decision-making. And the more information technology innovations that have applications to high-quality health care are developed, the greater the risks of these actual/perceived wedges developing that may create vertical/horizontal equity challenges as well as efficiency problems. Access to care in most societies in the world leads to the development of national health care systems, such as the British National Health Service, which work to guarantee particular levels of access to all citizens.

Innovation

Innovation is the art of the possible, what can we do to improve the health of populations and to provide health care to citizens of societies. In particular, the expansions we have made in information technology around the world are instrumental in work to expand what is possible in health care and health systems. While it might seem that innovation is orthogonally separate from efficiency, as with equity/access, this separation is not going to be possible. Of course, innovation can be in processes needed to improve efficiency, using knowledge management in ‘the utilization of information technologies to help organizations better identify, develop, access and apply the skills and experiences of their employees to augment business processes and drive innovation’ (Kudyba, 2003). Or, it can be in innovating new treatments or products. But either way, the definition of efficiency will be essential to understanding how these terms relate.

A tour through existing definitions of efficiency in health care

The assertion of this paper is that we do not really know what we mean when we refer to efficiency in health care, numerous definitions have been promulgated; however, none are satisfying or adequate except in limiting reductive cases. And, the key operative factor in making these definitions difficult is the role of quality.

Input- and output-oriented definitions of efficiency

Here are two of the common economic definitions of efficiency, rewritten to highlight the role of quality:

  • Maximize health care outputs produced from a fixed set of health care inputs and input quality, holding health care output quality constant.

  • Minimize health care inputs (related to cost minimization) producing a fixed set of health care outputs where input/output quality also are fixed.

The central problem with these definitions is: What do we mean by quality and what does it really mean to hold all this quality fixed? A particularly attractive definition for thinking about this is from the Institute of Medicine, which defines quality as multidimensional and characterized by care that is safe, timely, effective, efficient, equitable, and patient-centered (recently, this last term of patient-centered has begun to be viewed as too oriented toward sickness as opposed to wellness, so in a spirit of loss aversion is turning toward person-centered). A similar typology is used by the Australian National Health Performance Framework (Australian Institute of Health and Welfare, 2009) where health system performance is measured by effectiveness, safety, responsiveness, continuity of care, accessibility, and efficiency and sustainability. These approaches implicitly encompass the Donabedian definition that also highlights the multidimensionality of quality. And once quality is multidimensional, holding it fixed is a major concern. Can we even hold quality fixed in any practical manner other than a thought experiment as we actually change other things? Is quality an input, an output, or most likely both in the sense of these definitions? Is quality actually a separate input or output, or does it adjust other factors? Complex health care and health services delivery context obviously is needed to approach these questions, but one could provide examples of any of these situations. The best approach in a conceptual framework for understanding how to treat a particular quality measure or measures is to dig into the data-generating processes where clinical and non-clinical employees interact with patients to generate measured outcomes. Actions of individual people may or may not be the most important of these factors, and many of the data-generating processes are characterized by deep uncertainties, and therefore large samples are needed to sort out population effects. In the sense of these two definitions, pointing toward one or the other of the bulleted options depends most on what is fixed in the model or change being considered. We can think of the first definition as output-oriented (expand outputs for given inputs) and the second as input-oriented (shrink inputs for fixed outputs), and the orientation clearly matters depending on how it relates to quality (inputs or outputs?).

This overview paper is not going to provide a detailed conceptual framework or a review of the empirical methodology literature for all of these possibilities. Instead, we need to discuss the issues in defining quality for efficiency carefully, but we can do so in a context-free fashion. The result is intended to be general and provides a way of thinking about all populations and societies, from the poorest and most resource challenged to the most technologically advanced systems. All of these systems have constraints that limit what they can do. Harkening back to the composite quality discussion in the background, one issue is whether we want to consider quality measures individually and independently or whether we need to construct a composite measure of quality. And then composite measures of quality can be formative (formed from different dimensions of quality with some type of weighting – this issue of weighting is itself quite controversial – Shwartz et al, 2009) or reflective (latent manifested from correlated measures). How we treat statistical correlations among the variables is then the key concern. These concerns impact quality that in many cases may be operating along a different dimension, orthogonal to the input/output-based production process measurement that standard efficiency analysis does.

Note that one solution to this challenge in the input/output-based efficiency definitions above, which is contrary to a standard economic dictum that above all one seeks efficiency first and then deals with access/equity/quality, is that perhaps we should hold efficiency fixed and focus our efforts foremost on increasing quality. This is extremely important in the U.S. dialogue since it underlies the Patient Protection and Affordable Care Act development and is inherent in the value movement that puts quality in the numerator of a quality/cost equation. But there is a problem here, too, which we will come back to below, in that there is a circular reasoning problem. But, first, let us introduce some other popular efficiency in health care definitions, staying with the United States systems as the primary examples. Across the world, health care systems have been struggling with the same problems for some time, including the World Health Organization (Evans et al, 2000).

Definitions of efficiency in health care promulgated by major U.S. organizations

Medicare Payment Advisory Commission (MedPAC): ‘Using fewer inputs to get the same or better outcomes, efficiency combines concepts of resource use and quality’ (Agency for Healthcare Research and Quality (AHRQ), 2008). Note that this definition is closer to the ‘value’ definition, combining cost and quality.

National Quality Forum (NQF): ‘A measurement construct of cost of care or resource utilization associated with a specified level of quality of care’ (NQF, 2009). This definition incorporates the ‘fixed quality’ concept described above.

AQA Alliance (formerly Ambulatory Care Quality Alliance): ‘A measure of the relationship of the cost of care associated with a specific level of performance measured with respect to the other five IOM aims of quality’ (AHRQ, 2008). And here is one of the circular definitions, since note above that one of the five IOM aims is ‘efficiency’. Therefore, in this definition, efficiency is subsumed under quality.

RAND and AHRQ: This blue-ribbon panel report contains a typology of efficiency measures and its Figure 1 illustrates another approach to organizing a definition of efficiency in health care (AHRQ, 2008). This model has Society (including Providers, Intermediaries, and Individuals) sitting on top of a relational model between Outputs and Inputs. The Outputs in the typology are Health Services and Health Outcomes (which in the Donabedian sense is one domain of quality). The Inputs in the typology are Physical and Financial. This entire report is well worth reading, as it covers an immense amount of ground in greater depth than this paper can. The typology focuses on the different perspectives that elements of Society bring to considerations of inputs, outputs, and efficiency. The approach in this paper agrees that perspectives of different stakeholders are vitally important. But none of these typology definitions succeed in resolving the quality challenge that this paper takes on. It does focus on one key quality dimension of patient outcomes, and we will come back to this perspective after a digression into what perhaps is the most popular public view of inefficiency in health care, especially in the United States: waste and fraud.

Waste and fraud as definitions of inefficiency

In the United States, in the last 4 years, President Obama has used ‘waste and fraud’ as a key metaphor in speeches discussing ways of reducing inefficiency and promoting efficiency in health care. Steadily through these years, the number of hits from a Google search on the three terms: ‘Obama’+‘health care’+‘waste’ has been growing to well over 100 million. However, if you add ‘fraud’ to that search it drops by more than half to 43.1 million hits on 13 February 2012, mostly since there are so many hits on ‘Health Care Reform as a Waste’. He made a major speech in Missouri on 10 March 2010 to announce this initiative and has maintained http://www.whitehouse.gov/health-care-meeting/proposal/whatsnew/waste-fraud-abuse as an active high-level White House web page.

However, can waste and fraud (defining inefficiency in the negative) be a useful way to come at this question? The Obama administration's implementation of the Patient Protection and Affordable Care Act health reforms to date focuses on relatively small initiatives aimed at community health centers, durable medical equipment, and relatively marginal increases in funds to seek out waste and abuse in the system. This is not a comprehensive approach to efficiency, and it is not clear where the incentives and motivations for quality would come from out of these initiatives. In fact, there are concerns that unintended consequences could arise from aggressive prosecution from data mining methods that have been developed to seek out potential fraud and abuse from large aggregated claims files. This moves in the opposite direction to a person-centered focus that gives direction to a way forward that focuses more on each individual person and the demand side of access and needs.

Person-centered focus on final outcomes

Another approach to defining efficiency, which is a more specific way of taking the RAND/AHRQ patient outcome approach forward, is to define efficiency as being concerned with the relationship between resource inputs (perhaps thought of as costs in the form of labor, capital, equipment, and supplies) and either intermediate outputs (such as numbers of patients treated or access measures of care) or final health outcomes (working from lives saved to life years gained to quality-adjusted life years or QALYs) with the ideal focus on the final health outcomes (Palmer & Torgerson, 1999). This brings a focus back to the economic evaluation of treatments and quality that is subsumed in the QALY structure or elsewhere in the health outcome structure. The issues raised in this essay regarding measuring and determining relevant quality measures are still concerns not addressed by Palmer & Torgerson though. But their approach is related to the claim above that perhaps we should flip around what we call ‘fixed’ and try to find solutions that are focused on increasing quality. Inefficiency in their terminology is then said to exist when resources could be reallocated in such a manner that we could maximize health outcomes (Palmer & Torgerson, 1999). This idea has evolved in the United Kingdom in the work of the National Institute of Health and Clinical Effectiveness and the understanding there of the idea of Comparative Effectiveness (Barnett et al, 2009). This essay is staying focused on the idea of efficiency, but effectiveness is also challenging to define.

But then, the three measures of standard efficiency in economics (technical efficiency, productive efficiency, and allocative efficiency) are defined by Palmer & Torgerson as follows:

Technical Efficiency: In technical efficiency, we want to be using each of the productive resources to their best and take maximum advantage in maximizing health outcomes.

Productive Efficiency: In productive efficiency, we want to be choosing the combinations of productive resources that achieve the maximum health benefits (that are quality adjusted ideally) for a given cost (fixing cost).

Allocative Efficiency: In allocative efficiency, we want to seek the right mixture of health care programs and health systems that maximize the health of society overall.

The implication of these measures is that allocative efficiency is the highest goal and subsumes the others. In that way, unlike in standard economics where quality can be traded and competed for among providers, we may not want to begin with technical efficiency and then move up the ladder toward allocative efficiency. But this frames the questions directly and simply so we can consider what comes next.

Discussion and seeking a way forward

The intent of this essay is to be hypothesis generating, helping to improve conceptual frames of important problems in the study of health systems across disciplinary frameworks. Though my original disciplinary training is in economics, and some of the language here comes out of that tradition, the attempt is to make the key issues more transparent to all disciplines that touch on health systems research and health services research. The overriding emphasis is the definition and role of quality in defining efficiency and considering paths to innovations and quality improvement. Underlying this has been a focus on trying to understand what people are calling ‘the value equation’ (or quality/cost) and the challenges in trying to define efficiency in the value environment. The approach to relating efficiency to quality and life outcomes that Palmer & Torgerson have concisely described (Palmer & Torgerson, 1999) offers more grounding than the other efficiency definitions we have considered here for incorporating quality comprehensively. But it has not accomplished that task.

To seek a way forward requires a divergence between macro-system approaches and micro-system approaches, and there is much to be gained especially in seeking more micro-work (Hollingsworth & Street, 2006). Understanding country-level performance has generated some interesting work as well, though, both generally (beginning with Mahlberg & Obersteiner, 2001; Despotis, 2005) and specifically to health care, especially with the World Health Organization Health System Indicators (beginning with Hollingsworth & Wildman, 2003; Lauer et al, 2004).

However, health systems are composed of people too, and while it may seem obvious that a person-centered focus to determining efficiency and judging innovative productive improvements would be clearly seen as superior, the people inside health systems do not always see it that way (Reinhardt, 2012). Knowledge management approaches also have been seen as ways to organize information-focused service industries to identify, develop, and communicate employee capabilities in order to drive innovation and efficiency (Kudyba, 2003). The multi-disciplinary nature of health systems studies has much promise to yield more transparent communication and system solutions. Efficiency and innovation are perhaps the most important goals of a health system and quality of care measurement is the biggest challenge in that effort. The first step toward achieving these goals is establishing clear communication that is not stuck in the language of particular intellectual disciplines, which we do not yet have.