|
 |
 |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
Printer Friendly Version (PDF)
This synthesis focuses on the reduction of errors and improvement of health care quality using Evidence-based Hospital Referral (EHR) strategies. A substantial body of peer-reviewed literature published in top-tier journals confirms that the more experience a clinician or hospital has caring for patients with select high-risk conditions or performing high-risk surgeries--like heart surgery--the more likely patient care will be provided appropriately and without error. Evidence consistently shows that such high-volume providers tend to deliver superior quality of care to patients, as reflected in patients’ lower mortality rates following complex procedures. The volume-outcome relationship is the likely result of a complicated causal chain of events linking high-volume to better patient outcomes. Some people believe that experience (i.e., “practice makes perfect”) accounts for the volume-outcome effect. Others contend that hospitals achieve a high volume of patient referrals because they have a reputation for outstanding quality. It is more likely, however, that the volume-outcome relationship is due to system-wide factors. Physician skill, well-organized clinical teams, use of practice guidelines, availability of technologically sophisticated equipment, and the skills and abilities of auxiliary medical personnel combine to produce positive results. Despite the fact that we do not understand fully the mechanisms underlying the volume-outcome relationship, volume is a proxy measure of the quality of care for specific procedures. Health care purchasers, public and private, can use volume as a criterion in physician and hospital selection and contract negotiation. The Leapfrog Group calls the use of volume or outcome data to refer patients to particular physicians or hospitals evidence-based hospital referral. There is strong evidence that if patients in need of specific high-risk surgeries or treatments in the United States selected high-volume hospitals for their care, more than 4,000 lives could be saved annually. However, this quality-improving, error-reducing, patient safety practice is not widespread. Purchaser Tips (top) The Leapfrog Group consists of more than 110 large purchasers and coalitions that work to “mobilize employer purchasing power to trigger breakthroughs in the safety and the overall value of health care to American consumers”. The Leapfrog Group is promoting evidence-based hospital referrals for a select set of high-risk surgeries and neonatal conditions. Leapfrog member purchasers will employ various mechanisms to guide patients needing care in these areas to hospitals and clinical teams that treat a comparatively high volume of patients needing a given procedure or, when a valid outcome measurement system exists, that demonstrate superior outcomes. The Leapfrog
Group has identified favorable volume thresholds for seven treatments,
as shown in Table 1.
In addition to enrollee education, purchasers can promote EHR through benefit design and health plan and provider contracts. Currently, the incentive structure in closed-network plans discourages external referrals. A purchaser could contract with health plans to refer patients who need certain complex procedures to high-volume providers (or those that have demonstrated superior outcomes). Purchasers could also create financial incentives to reward plans that refer enrollees to high-volume hospitals (Dudley et al., 2000a). Synthesis of Research1 (top) The proportion of patients who do not survive a specific medical procedure can be used as a direct or “outcome” measure of the quality of care for that procedure. Applying this type of measurement in comparisons across hospitals can be helpful to patients trying to select a hospital for care. However, such comparisons can create controversy in the medical community because it is difficult to account for differences in the severity of illness of the patients different hospitals treat. Removing the potentially confounding effect of health status differences in patient populations requires risk adjustment, which statistically removes health status differences among the patient populations for which mortality rates are calculated. Otherwise, differences in mortality rates may reflect differences in health or sickness prior to treatment, rather than differences in the quality of care patients received. Unfortunately, there is no generally agreed upon risk-adjustment methodology. Only a few states (New York, Pennsylvania, and California) have risk-adjusted, statewide mortality reporting systems for select, complex procedures, such as coronary artery bypass graph surgery (Lara, 2001). When risk-adjusted outcome data are not available, the number of times a doctor or hospital has performed a procedure, administered a treatment or provided a certain type of care can be a proxy for measuring quality. For certain procedures, volume is a “predictor” of outcomes because higher volume is positively associated with better patient outcomes. The research literature reveals the existence of statistically significant associations between high physician and hospital volume and better patient outcomes for a number of treatments and procedures. In the 2000 Institute of Medicine (IOM) workshop summary on interpreting the volume-outcome relationship, Halm, Lee, and Chassin examined 137 research articles on the relationship between hospital volume, physician volume or both, and patient outcomes for eight selected conditions and procedures.2 Based on their review of 88 studies that met their criteria for inclusion3, the researchers found that 77 percent of the studies reported statistically significant relationships between either higher physician volume or higher hospital volume and better outcomes. The remaining 23% of studies found evidence of these relationships, but the results were not statistically significant. None of the studies documented a statistically significant association between higher volume and poorer outcomes (Halm et al., 2000; Halm et al., 2002). The nature of the causal relationship between volume and outcomes is not fully understood, although there are several plausible explanations. Figure 1 portrays the relationship between hospital volume and mortality rates measured in studies that received the highest possible quality rating across ten dimensions specified by Halm et al. in their review for the IOM. For each of three surgical procedures4, high-volume hospitals have substantially lower mortality rates than low-volume hospitals. Figure 2 shows that the same relationship exists, though to a somewhat lesser degree, for high-volume physicians and mortality rates for these surgical procedures.
Figure 2
Studies that
examine the effects of both hospital and physician volume suggest that
hospital volume and physician volume have a combined impact on patient
mortality that is greater than the impact of either alone. In short, studies
suggest that high-volume physicians working in high-volume hospitals have
the lowest patient mortality rates overall. Why does high volume produce better outcomes? Studies that examine the effects of physician and hospital volume on patient outcomes provide insight about the actual mechanisms at work. The traditional explanation is “practice makes perfect” (Luft et al., 1987). Although it is likely that specific cognitive and physical skills are improved with increased experience, leading to better outcomes, physician skill is just one component of a system of care delivery that affects the quality of care and patient outcomes. The fact that hospital volume is related to patient outcomes independent of physician volume indicates that high-quality care is the result of both individual effort and organizational structure and processes (Halm et al., 2000; Halm et al., 2002). An alternative explanation builds on the observation that physicians are inclined to refer patients to hospitals and clinical teams with a demonstrably superior track record for specific procedures. In other words, the best hospitals and physicians are high-volume providers because they have a reputation for high quality (Luft et al., 1987). Empirical evidence does not support this theory, however. One study found that even when pertinent data regarding provider volume and outcomes are available, physicians do not use it to make referrals (Schneider and Epstein, 1996). Without guidelines (and aligned incentives) directing physicians to refer patients for high-risk, elective surgery to high-volume providers, it is unlikely that selective referral explains the variation in mortality rates between low- and high-volume physicians or hospitals (Halm et al., 2000; Halm et al., 2002). The best explanation for the volume-outcome relationship may lie in the “systems perspective” described and endorsed in recent IOM publications on patient safety and quality improvement (Kahn et al., 1999 and Institute of Medicine, 2001). The systems perspective suggests that a number of factors - physician skill, well organized clinical teams, the use of practice guidelines, and the ability of ancillary personnel and various units within the hospital to work together - jointly produce the positive association between volume and patient outcomes. For example, while the cardiologist’s experience and competence is essential for minimizing in-hospital mortality rates for patients with acute myocardial infarction, so are the skills of the nurses who support the physician. For patients in an intensive care unit, the expertise and organization of the ICU staff also affect patient mortality rates (see Intensivist Staffing in Intensive Care Units (ICUs)). The organization of the surgical team and related hospital systems contribute to the efficacy of clinical decision-making and the way processes of care are carried out. Utilization of standardized protocols and advanced technological support tools, such as computer-based systems for drug prescribing and ordering, also positively affect patient outcomes (see Computerized Physician Order Entry (CPOE)). Potentially avoidable deaths Researchers
have estimated that 602 lives could have been saved in California in 1997
if EHR had been instituted and patients needing high-risk elective surgery
had been channeled to high-volume hospitals (Dudley et al., 2000b)11.
Although this is only a projection, the study provides a reasonable estimation
of the impact selective referral could have on patient safety. If the
number of potentially avoidable
deaths in California is extrapolated to the entire United States,
thousands of lives could be saved annually by implementing EHR nationwide. What are the costs of EHR? In assessing the costs of EHR, one begins by asking “to whom”? Referring patients to high-volume hospitals (HVH) necessarily means referring them “away” from low- or lower-volume hospitals (LVH). Thus, in terms of market share, there are winners and losers. However, from the purchaser’s perspective, while the evidence is scant, some research indicates that length of stay (LOS) is significantly shorter for patients undergoing procedures in HVH (Phillips et al., 1995). (Depending on the nature of the payment agreement negotiated in purchasers’ or plans’ contracts with hospitals, shorter LOS can result in significant cost savings.) The same study found that costs for AMI patients were considerably greater at LVH than at HVH. The Leapfrog
Group also commissioned researchers at Dartmouth Medical School to analyze
the economic implications of implementing The Leapfrog Group safety standards.
Birkmeyer et al. (2000b) calculated average hospital profits (based on
a cross-section of New England hospitals) for each EHR procedure or condition
included in the estimates of lives saved. Table 2 summarizes the findings.
Where many
see opportunities to improve patient safety and practice value-based purchasing,
others see obstacles. Most states collect condition-specific hospital
volume data. But, with few exceptions, the health care system is not using
this information to refer patients to high-volume physicians and hospitals. There are other concerns about EHR. For instance, some physicians and patients worry about the disruption or discontinuity of care when patients receive care outside of their usual source. For some providers, EHR may mean loss of control and income. Some patients would rather see a familiar physician or go to a nearby hospital, even if evidence suggests that they stand a better chance of a good outcome elsewhere. Finally, the impact of selective referral on LVH can lead to regionalization of care. That is, if fewer and fewer patients go to LVH for certain procedures, the clinicians at these facilities will have less and less experience providing those procedures over time. For example, a hospital may have a lower chance of saving a patient in need of an emergency angioplasty if its clinical team hasn’t done one in six months or a year (Birkmeyer, 2000a). This problem is likely to be particularly acute in rural areas. The reader is referred to Epstein (2002) for a detailed discussion of this and related issues concerning the volume-outcome relationship. Areas for Future Research (top) The wealth of evidence that HVH have lower risk-adjusted mortality rates than LVH provides strong support for adopting EHR for certain medical conditions and procedures. However, there are several areas for future research on this topic (as identified in the IOM Workshop Report (Institute of Medicine, 2000)).
Glossary (top) Case mix: The grouping of patients handled by a practitioner or hospital according to such factors as age, sex, diagnosis, treatment, and severity of illness. Comorbidity: An additional condition or conditions that a patient has at the same time as their primary condition or an illness or injury before a patient’s hospitalization that may extend the patient’s hospital stay. Error: the failure of a planned action to be completed as intended (i.e., an error of execution) or the use of a wrong plan to achieve an aim (i.e., an error of planning). Evidence-Based Hospital Referral (EHR): Ensuring that patients with high-risk conditions are treated at hospitals (and by physicians) with characteristics shown to be associated with better outcomes. Purchasers, health plans, hospitals, practitioners, and patients can use available information to select hospitals and physicians with the best track records. Hospital volume: The number of specific procedures performed by a hospital over a defined period of time (usually one year). The Leapfrog Group: Composed of more than 110 public and private organizations that provide health care benefits, The Leapfrog Group (www.leapfroggroup.org) works with medical experts throughout the U.S. to identify problems and propose solutions that it believes will improve hospital systems that could break down and harm patients. Representing more than 32 million health care consumers in all 50 states, Leapfrog provides important information and solutions for consumers and health care providers. Medical error: is the “failure of a planned action to be completed as intended (i.e. error of execution) or the use of a wrong plan to achieve an aim (i.e. error of planning)” in the course of managing a patient’s medical condition (Kohn, 1999). The Pacific Business Group on Health (PBGH): A non-profit coalition of more than 40 employers based in San Francisco that works to improve the quality of health care while moderating costs. PBGH supports value-based purchasing and works closely with payers, providers, and researchers in quality assessment and improvement efforts. Physician volume: The number of specific procedures performed by a physician over a defined period of time (usually one year). Potentially avoidable deaths: Deaths that potentially could be avoided if patients are treated at a high-volume hospital instead of a low-volume hospital. A quantification of deaths attributable to low-volume hospitals or physicians, calculated by subtracting the number of expected deaths (based on risk-adjustment) from the observed number of deaths. Preventable adverse events: are preventable injuries due to the management of a patient’s medical condition rather than to the patient’s underlying health condition. Regionalization: The process of concentrating high-risk surgical procedures in high-volume hospitals through the implementation of selective referral. Risk adjustment: Risk-adjustment techniques are used to increase the accuracy of mortality rate comparisons. The elements of case mix, patient selection, the severity of the condition, and the presence of comorbidity are calculated and mortality rates are adjusted accordingly. Larger hospitals that treat older, sicker patients who have multiple health problems may have a higher mortality rate for major operations relative to other area hospitals, despite the fact that they are high-volume. By risk-adjusting the mortality rates, the effect of volume can be isolated and identified. Selective referral: The referral of patients needing certain high-risk procedures or with high-risk conditions to hospitals and practitioners that have a reputation for or have data that reflect high-quality patient care. Severity of Illness: The seriousness of a patient’s condition. Severity of illness may differ among patients diagnosed with the same condition. Researchers have created “severity indices” or “scores” to quantify the severity of illness. The methods correlate the “seriousness” of a disease in a particular patient with the statistically “expected” outcome (e.g., survival, morbidity, etc.). Related Links (top) Related Topics (top)
Key Experts (top) Although agreeing to be listed, these experts are not necessarily endorsing the ideas presented in this synthesis. Also, the experts have not explicitly agreed to respond to inquiries about the topic. The list is meant as a reference point for those interested in learning more about this topic. R. Adams
Dudley, MD, MBA John D. Birkmeyer,
MD Ethan A.
Halm, MD, MPH Harold S.
Luft, Ph.D Arnold Milstein,
MD Ciaran S.
Phibbs, Ph.D. Works Cited (top)
Additional References (top)
Acknowledgements (top) This research was funded by the Robert Wood Johnson Foundation, under its National Health Care Purchasing Institute. Junette Williams, Laura McDaniel and Mark Chow provided research assistance. Suzanne Delbanco provided expert reviews and AcademyHealth colleagues provided editorial assistance. Although the analysis and conclusions are solely my own and do not necessarily reflect the views of the foundation or AcademyHealth colleagues, I am deeply indebted to them for their support and assistance, which made this work possible. Citation Guidance (top) Lee, J. "Evidence-Based
Hospital Referral (EHR)," Research Synthesis, Endnotes
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|
|
|
|
|
 |
