The “problem(s)” with quality improvement in health care

Health care has always had a quality issue. For much of our history, health care was provided by practitioners whose basis for their craft was spiritual (or quasi-spiritual) or unsubstantiated theory, or by charlatans peddling cure-all concoctions. * As such care often had little or no effect on the patient's disease or ailment, one would be justified in considering that care to be of low quality. Some would argue that little has changed. Variations in the care patients received, first observed by Glover2, 3 in the early part of the last century and notably by Wennberg and colleagues 40 years later (eg, Wennberg and Gittelsohn4), raised concern amongst health-care stakeholders that resources were being used inefficiently and/or some patients were not receiving the best care available †—both of which are considered by many as issues of quality in health care. Variations in care can still be seen today despite increasing knowledge of which therapies are effective for which health conditions and considerable attention to the organization and distribution of health care resources. It is difficult to see how quality care would result in patients with similar needs receiving different care (especially where effective therapies are known and available and the resources exist to provide those therapies). However, quality care does not end with patients obtaining effective therapies. Quality has many dimensions, and whilst progress has been made on identifying what they are and how to address them, there is still much that can be done to improve the quality of health care. Over the past two decades, we have seen a growth of interest in health care quality and quality improvement (QI). That interest was not limited to providers and scientists—public interest has also grown in response to several concerns, including medical error, access (and wait times), and value for money. How we tackle quality issues depends on how we define quality. For example, one might define quality health care in terms of the extent to which patients receive proven therapies. Thus, QI is a matter of ensuring that patients obtain effective care. The much praised report by the Institute of Medicine (IoM), To Err is Human, framed quality in terms of medical error, implying that quality care is safe care.7 In that case, QI would focus on reducing avoidable adverse events, such as avoiding administering the wrong (and potentially harmful) medication or dose. In their next report, Crossing the Quality Chasm, the meaning of quality was expanded to include six domains: quality health care is care that is safe, effective, patient-centred, timely, efficient, and equitable.8 The National Health Service defines quality along similar terms as the IoM.9 Formal definitions of health care quality have also been proposed in the literature. For example, the IoM defines quality as the “degree to which health services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge” [10; p.4]. Campbell et al11 suggest quality of care for individuals is “whether individuals can access the health structures and processes of care which they need and whether the care received is effective” (p.1614). Batalden and Davidoff12 provide perhaps the most ambitious definition: quality is “the combined and unceasing efforts of everyone—health-care professionals, patients and their families, researchers, payers, planners and educators—to make the changes that will lead to better patient outcomes (health), better system performance (care) and better professional development (learning)” (p.2). These definitions share much in common. However, they also all struggle in that it is not clear exactly how to operationalize them such that quality can be measured, which limits our ability to identify when quality is poor or has subsequently been improved. There are several considerations required to operationalize a definition of health care quality so that it can be used to identify a QI. One must first identify what to measure and how. ‡ For example, one might consider a therapy effective if its relative benefit (compared with available alternative therapies or placebo) has been demonstrated in a high-quality randomized controlled trial (RCT). If so, then quality is a function of the proportion of patients obtaining that therapy when it is indicated (as suggested above). Regardless of the chosen metric, one must use a method of determining when it is achieved that is both reliable and valid. Scope is also a consideration. Should we measure quality at the level of a population or the individual? There is no guarantee that what is measured at the population level translates to or is meaningful for the individual patient.13 Consider a scenario where quality is defined according to patient satisfaction measured using a continuous scale. It is possible that an intervention can induce an increase in the average score on the satisfaction scale (indicating a QI) but also result in a reduction in the proportion of patients who meet a defined threshold for “satisfaction” on that scale (indicating a deterioration in quality). Scope also includes the breadth of what is being measured. Improving quality in one part of the care chain or aspect of care might reduce quality elsewhere for the individual patient or for other patients. Thus, measurement of quality in one clinic or one service (or even one aspect of care within a clinic) may not be a sufficient indicator of the impact an intervention has on QI (ie, what is measured may lack sufficient breadth to see the whole impact of the intervention). Finally, one must consider how one indicator of quality behaves relative to others. For example, observed improvements in the proportion of patients obtaining effective care as determined by which therapy had fewer events in an RCT does not guarantee an improvement in patient-centred care or may even reduce it (perhaps some patients do not want that therapy for reasons of cost, potential side effects, etc). Different indicators of quality must be integrated, and how to do so is not a trivial task, especially when the indicators are not aligned (ie, when one shows a positive response and another a negative response). The issues I raise here might be solved through consensus of health care stakeholders. § However consensus is achieved, the process must be valid given the health care goals, less it be useful for those interested in health care quality. We cannot know if we have improved quality if we cannot identify it. In this issue of the Journal of Evaluation in Clinical Practice, Djulbegovic et al14 suggest an additional integration problem affecting the science (and ultimately, the achievement) of health care QI. Currently, health care quality and QI is a focus of multiple scientific and technical initiatives that cross several academic disciplines. Djulbegovic et al14 suggest that the poor progress in QI is in large part due to the fact that the different initiatives examining quality are doing so in silos. Health System Science (HSS) is proposed as a means to unify the community so that QI activities can be optimized. The purported advantage of HSS is that it provides a common basis to align QI efforts—one that is rooted in evidence-based medicine (EBM) and decision science (thought to be a common feature amongst the relevant initiatives). In a related paper, Mondoux and Shojania15 challenge the examination by Djulbegovic and colleagues, suggesting that QI problems persist even in contexts where there is good evidence of what is effective and/or where decision-making is satisfactory, and in doing so, question the extent to which EBM and decision science are indeed the basis for QI. Despite this debate as to the role and usefulness of EBM and decision science in QI, both groups share much in common. Notably, both are committed to the view that QI interventions should be grounded in high-quality scientific evidence. Likewise, both recognize the importance of context and the fact that one size often does not fit all. ¶ The debate between Djulbegovic et al14, 16 and Mondoux and Shojania15 raises some important questions. First, is it the case that different initiatives that may represent different disciplines can be unified? A phenomenon, in this case, quality, can be examined through the lens of any of several disciplines. However, what is examined (and how it is measured) and how what is measured is interpreted is bound by the theoretical commitments of each discipline. # In some cases, the theories that ground each discipline may be incommensurable, precluding a unification. Furthermore, each discipline may have a distinct epistemic culture, and thus may not agree on what constitutes evidence (or how to weigh the evidence). For example, an engineer concerned with wait times in the emergency department might highly value mechanical explanations of patient flow (perhaps based on time studies), whereas a sociologist interested in the same issue might highly value information on stakeholder experiences derived from qualitative method. Putting aside these concerns, let us assume for a moment that the different disciplines can be unified and can share a common notion of evidence. Is it the case that such a common notion of evidence can be, is, or should be that proposed by EBM? The answers to the questions I raised here will determine the extent to which HSS can be the way forward for QI. One implication of grounding QI in EBM is the privileged position it gives to clinical practice guidelines (CPGs) as a benchmark of quality health care. The advantage of CPGs is that they are often developed with more than just the trial derived therapeutic effect size in mind. The Grades of Recommendation, Assessment, Development and Evaluation (GRADE) framework is the current standard for developing CPGs.18, 19 Most of the domains of quality identified by the IoM are incorporated in some way into the GRADE framework's components for determining a clinical practice recommendation and its strength; safety and effectiveness (through “certainty of evidence”), patient-centredness (through “values and preferences”), efficiency (through economic considerations), and more recently, equitable care.20 Thus, CPGs developed using the GRADE framework might constitute a suitable benchmark for quality (and thus, adherence to such CPGs might indicate quality). However, the appropriateness of GRADE-derived CPGs for indicating quality depends on how well each component addresses each quality domain, which is a function of both the structure of framework and judgement of its users. ** Other frameworks might also be developed. We should also recognize that quality is a value-laden construct, and thus the scientific merits of any tool to deal with issues of quality is only one of perhaps many considerations. Regardless of which method one uses, it is important that the indicators of quality that are derived and implemented in practice reflect what is best for each patient, given her or his health care needs and circumstance. Until we have good reason to believe that has been achieved, we may continue to have a health care quality issue. I thank both Benjamin Djulbegovic and Shawn Mondoux for their insightful comments on an earlier draft of this editorial. I appreciate their (and their co-authors') willingness to participate in an important discussion on QI.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.