Processes and Outcomes in Acute Myocardial Infarction
How Technical Should We Get?
Article first published online: 12 JAN 2002
Journal of General Internal Medicine
Volume 16, Issue 12, pages 868–869, December 2001
How to Cite
Kiefe, C. I. (2001), Processes and Outcomes in Acute Myocardial Infarction. Journal of General Internal Medicine, 16: 868–869. doi: 10.1046/j.1525-1497.2001.11016.x
- Issue published online: 12 JAN 2002
- Article first published online: 12 JAN 2002
As physicians, we strive to do the best for our patients. The meaning of “the best,” however, is as elusive as it is important. Best care means, for example, administering β-blockers on discharge to all acute myocardial infarction (AMI) patients without a contraindication, an intervention that has been shown to reduce mortality by 23% to 40%.1 Best care also means treating all our patients with respect and compassion, even empathy. Their importance notwithstanding, the consequences of practicing medicine according to these humanistic standards of excellence are not easily quantified. Both β-blockers and respect might be deemed processes of care. Whereas the former may be described as “technical,” the latter is usually not so described. Can we measure their relative, and absolute, importance? By what yardstick?
In this issue of JGIM, Fremont et al. use observational data to seek associations between patient perceptions of the quality of care received during hospitalization for AMI and self-reported health status 1 year later.2 Patient perceptions of quality were measured using the standardized Picker questionnaire, which captures “nontechnical” processes of care such as respect for patient preferences and emotional support. Fremont et al. found that patients with higher Picker scores have better subsequent health status; furthermore, their data suggest that Picker scores reflecting ambulatory encounters during the period between discharge and 1-year follow-up may mediate this association. Do the associations observed in this interesting observational study imply causation? If so, what inferences might we draw regarding the relative importance of “nontechnical” aspects of care in defining and improving quality?
Clearly, drawing etiologic inferences from observational data is fraught with hazards. There are many ways in which both patient perceptions of better acute care and better self-reported health status 1 year later might be consequences of an underlying phenomenon causing both, rather than the former causing the latter. For example, one might attribute both more satisfactory acute care and better 1-year outcomes to higher income, or to more optimistic character traits; both of these are known to be associated with lower morbidity or mortality (better outcomes) and could plausibly be associated with perceptions of better processes. Most importantly, it is reasonable to suspect that better technical care would also cause both patient perceptions of better care and better outcomes; were this the case, then better outcomes might even be entirely attributable to better technical care rather than to higher scores on the Picker scales. Neither patient income, nor optimism, nor technical quality of care could be measured with the data available for the Fremont study.
The instruments used to assess outcomes in the Fremont article are not universally accepted. For example, why use the epidemiologically but not clinically oriented Rose questionnaire, rather than validated disease-specific instruments such as the Quality of Life after Myocardial Infarction questionnaire or the Seattle Angina Questionnaire?3 Using different instruments might have led to different conclusions. The authors acknowledge limitations, but nonetheless suggest a “potential impact of patients' experiences with care on their outcomes.”
The hypothesis that care perceived to be better leads to better self-reported outcomes can hardly be tested against the gold standard of clinical evidence, the randomized controlled trial: imagine randomly allocating patients to a “worse care” group… Hence, adjustment for potential confounders becomes imperative when seeking a causal link between patient perceptions of care and subsequent functional status. Fremont et al. could not do so, thereby limiting the strength of the conclusions. However, it is rare to find longitudinal data, even observational, that link patient perceptions of care to long-term outcomes. Therefore, we need to pay attention to the message suggested by the Fremont study; if confirmed, it will have far-reaching implications.
The concept that better processes of care lead to better outcomes underlies modern-day health care improvement efforts.4 By better processes, the medical community tends to mean medical interventions that have been shown by the highest standards of scientific evidence to prolong life or decrease suffering. Administering β-blockers to patients who have undergone an AMI is such a process. Yet many studies have shown that large proportions of patients eligible for this proven intervention, in some states more than 40% to 50%, do not actually receive it.5 Improvement efforts have been targeting adherence to such technical aspects of care because their efficacy has been demonstrated, generally in highly selected patients recruited into a randomized controlled trial. Consequently, their effectiveness in improving outcomes such as mortality in an unselected general patient population requires only a modest leap of faith. Increasing adherence to proven technical processes of care is accepted as improving care in an evidence-based fashion.
The “outcomes movement” has succeeded in convincing the medical community that not only “hard” physiologic measures, such as mortality or reinfarction, are relevant outcomes, but that we should also measure and emphasize more patient-centered outcomes, such as functional status and health-related quality of life.6 The acceptance of patient perceptions of care, such as those captured through the Picker questionnaire, as patient-centered outcomes, worthy of being improved for their own sake, has been slower than the acceptance of, say, functional status as an outcome worth measuring. This is partly due to the fact that the link between better health (the ultimate goal) and patient perceptions of care is more tenuous than the link between better health and functional status. Nonetheless, the importance of measuring and improving patient perceptions of quality of care as outcomes in their own right is becoming increasingly accepted. This is driven to some extent by the need of health care systems to market themselves, and the understanding that public perception of high quality is an important marketing tool.7 Still, it is not generally accepted that patient perceptions of quality of care should themselves constitute actionable processes that cause better outcomes in the future. The article by Fremont et al. leads us in that direction.
In a world of limited resources, not all processes can be improved. Do “nontechnical” processes deserve the same degree of attention as technical processes on the part of those who seek to improve a health care system? Is the dichotomization artificial? We all want patients to know that they have received good care; it is a goal unto itself. But does its attainment, as suggested by Fremont et al., also result in better health as defined by morbidity, mortality, or health-related quality of life? At the current time, we simply do not know the answer. Until we do, we will have to continue to press for technical improvement, while considering excellence in nontechnical processes as more of an ethical than an evidence-based imperative. As physicians, it is our obligation to attend to both. —
- 4Evaluating the quality of medical care. Milbank Q. 1996;166–203..
- 7Measuring health care quality comparing and contrasting the medical and the marketing approaches. Am J Med Qual. 2001In press.& .