2. Prioritize potential projects and choose the projects to begin.
3. Prepare for the project by operationalizing the measures, building support for the project, and developing a business plan.
4. Do an environmental scan to understand the current situation (structure, process, or outcome), the potential barriers, opportunities, and resources for the project.
5. Create a data collection system to provide accurate baseline data and document improvement.
6. Create a data reporting system that will allow clinicians and other stakeholders to see and understand the problem and the improvement.
7. Introduce strategies to change clinician behavior and create the change that will produce improvement.
Evaluating and sustaining a quality improvement program
1. Determine whether the target is changing with ongoing observation, periodic data collection, and interpretation.
2. Modify behavior change strategies to improve, regain, or sustain improvements.
3. Focus on sustaining interdisciplinary leadership and collaboration for the quality improvement program.
4. Develop and sustain support from the hospital leadership.
From Curtis JR, Cook DJ, Wall RJ, et al. Intensive care unit quality improvement: a “how-to” guide for the interdisciplinary team. Crit Care Med 2006;34:211, with permission. Copyright © 2006 by the Society of Critical Care Medicine and Lippincott Williams.
Most clinicians have not been trained in how to develop and implement quality improvement programs. Curtis et al. have published a how-to paper that came out of an outcomes task force of the Society of Critical Care Medicine but has wide applicability to many areas of medicine (Table 5.3). The generalized lack of education and training in quality may explain some of the lethargy that Leape and Berwick allude to in their article “Five Years after To Err Is Human: What Have We Learned?.” They suggest that extraordinary quality improvements have been achieved by implementing discrete strategies in specific environments. If each of our organizations could fractionally achieve similar success, the impact on patient care might be quite remarkable.
P.72 purchasing program sponsored by CMS. In this program, up to 2.0% of a hospital's base-operating diagnosis related group (DRG) payments will be at risk (Table 5.4). CMS calculates the incentive adjustment based on a total performance score achieved by the hospital. The program is structured to allow a continual evolution with measures and weighting changing each year. Because almost $1 billion dollars is at risk, many organizations have highly focused efforts around the measures. As can be seen, CMS is moving from a time when surrogate measures, such as core measures, were the key drivers to outcomes and efficiency metrics (Fig. 5.2 and Table 5.5). The PPACA also required a Hospital-Acquired Condition (HAC) Reduction Program with another potential 1% payment reduction if performance thresholds are not met. Currently, there are eight measures in two domains, but these will likely increase over time (Table 5.6). In addition, many of these measures have applicability for obstetrician-gynecologists. The final major pay-forperformance program is aimed at reducing readmissions. In this program, payments are again reduced (1% in 2013, 2% in 2013, and 3% in 2015 and beyond) if readmission rates exceed the expected level. While the initial disease states being evaluated are acute myocardial infarction, pneumonia, and heart failure, all Medicare discharges are penalized. Reviewing Figure 5.3 demonstrates the significant potential for hospital revenue reductions as all of these programs are phased in.
TABLE 5.4 Amount Available Each Year Controlled by Statute
2013 1.00% of base-operating DRG payments 2014 1.25%
2015 1.50%
2016 1.75%
2017 2.00%
In an early article on the subject of paying for high-quality care, Epstein et al. called for several significant changes to facilitate adoption of pay-for-performance initiatives. These recommendations included an expansion in the scope of the efforts, as well as the amount of incentive available to the clinician or group; the importance of large groups such as CMS becoming involved; an expansion from the current small cadre of clinical indicators;
an improvement in our ability to establish reliable metrics for evaluation of quality; and, finally, a call for investment in electronic infrastructure to facilitate all of the above.
A Cochrane systematic review of the subject of targeted payments to affect outcomes in primary care was unable to reach a conclusion because of limitations in prior study quality and power. Kouides et al. evaluated the use of financial incentives to increase rates of influenza vaccine administration and found that modest incentive
increased an already high baseline administration rate. In another study on immunizations, Fairbrother et al.
evaluated use of cash bonus, enhanced fee for service, and feedback for improving baseline vaccination rates.
They found only the bonus group statistically improved (increased by 25.3%), although they believed much of this effect was due to better documentation versus actual increased administration of vaccine. Roski et al. noted that in incentivized groups, documentation of tobacco use was markedly increased, as was accession to
counseling programs.
Unfortunately, other important clinical end points showed no difference.
FIGURE 5.2 Evolution of value based purchasing from 2013 to 2017.
A more recent investigation evaluated a pay-for-performance initiative by PacifiCare Health Systems. Three quality measures were evaluated: cervical cancer screening, mammography, and HgbA1c. Unfortunately, only in the cervical cancer screening group was there an improvement after the intervention, and in spite of $3.4 million paid in pay-for-performance incentives, only nominal gains were made. Also, physicians with the highest initial baseline performance had the least improvement but were the recipients of the greatest portion of the paid bonus. This trial demonstrates the potential limitation of identifying discrete end points to reward heterogeneous groups of providers, all of whom start from different baseline levels of efficacy. Dudley has suggested that how an incentive is offered (i.e., reward or penalty) and factors such as what percentage of patients in a provider's panel incentives are applicable to may affect the success of these programs. Rosenthal et al., in evaluating multiple trials and multiple industries in which pay for performance has applicability, found little evidence that paying for quality is effective. They noted that even in other industries, results are inconsistent. They raise the concern that incentives may be too low, with limited penetrance within a given provider's or group's practice panel. An additional concern is making the target for incentive so much higher than baseline that the providers feel it is unachievable and therefore do not try. This is consistent with data from Beckman et al., who interviewed providers in a pay-for-performance program. Interestingly, providers reviewed their personal profiles but did not always change their functional behaviors in response to the data.
TABLE 5.5 Outcome Measures
MEASURE DESCRIPTION 2014 2015 2016 (PROPOSED)
Acute myocardial infarction mortality rate X X X
P.73
Heart failure mortality rate X X X
Composite patient safety indicators X X
Central line-associated bloodstream infection (CLABSI) X X
Catheter-associated urinary tract infection X
Surgical site infections X
Bundled payments are a variation on pay for performance where all the providers are paid for the entire care episode and certain outcomes are achieved. In January 2013, CMS announced that a number of organizations had been selected to become part of the Bundled Payments for Care Improvement Initiative. In the program, organizations and providers enter into payment programs that generally include financial and quality performance metrics. For the CMS program, participants could pick from a menu of 48 different episodes of care to bundle.
The hope is that by aligning all providers across the continuum that coordination of care and thus quality and cost will be improved. These types of programs may have had the greatest initial application in the area of cardiovascular surgery, for which outcome metrics of individual surgeons and hospitals have been reported for several years. In addition, there are well-defined clinical algorithms regarding “best practice.” This may have driven the American Heart Association's Reimbursement, Coverage, and Access Policy Development Workgroup to publish a statement on pay for performance. More recently,
the American Hospital Association has evaluated bundled payments and asked the following eight questions:
1. To which conditions should bundled payments be applied?
2. What providers and services should be included in the bundled payment?
3. How can provider accountability be determined?
4. What should be the time frame of a bundled payment?
5. What capabilities are needed for organizations to administer a bundled payment?
6. How should payments be set?
7. How should the bundled payment be risk adjusted?
8. What data are needed to support bundled payment?
FIGURE 5.3 Total hospital revenue at risk from major Medicare quality programs.
TABLE 5.6 Proposed Measures for the HAC Reduction Program
MEASURE DESCRIPTION FY
2015
FY 2016
FY 2017
Domain 1: AHRQ patient safety indicators
Pressure ulcer rate X X X
Foreign object left in the body X X X
Iatrogenic pneumothorax rate X X X
Postoperative physiologic and metabolic derangement rate X X X
Postoperative pulmonary embolus (PE)/deep vein thrombosis (DVT) rate
X X X
Surgery patients with recommended venous thromboembolism prophylaxis ordered
X X X
Accidental puncture and laceration rate X X X
Domain 2: CDC Hospital acquired infection (HAI) measures
Central line-associated bloodstream infection (CLABSI) X X X
P.74
Catheter-associated urinary tract infection X X X
Surgical site infection (SSI): X X
SSI following colon surgery
SSI following abdominal hysterectomy
Methicillin-resistant Staphylococcus aureus (MRSA) X
Clostridium difficile X
It is believed that the results from the demo project will provide clarity around many of these issues, although one could argue that obstetrician-gynecologists have been doing bundled payments for obstetric care for many years.
In addition to the previously noted problems with payfor-performance programs, another limitation is that most programs to date have focused on single disease states or outcomes (such as immunization or Pap test or mammogram compliance). In reality, many patients aged 65 or older have multiple medical problems and treating single disease states in isolation may not lead to overall improved outcomes and enhanced quality. It is not surprising that 89% of Medicare's annual budget is consumed by patients with three or more chronic conditions.
Boyd et al. evaluated current clinical practice guidelines (CPGs), which are often used in pay-forperformance programs, but which also focus on single disease states. In a hypothetical 79-year-old female patient with the common comorbidities of chronic obstructive pulmonary disease, diabetes, osteoporosis, hypertension, and osteoarthritis, they found that following each CPG would lead to 12 prescribed medications and a potential outlay of $406 per month. There were also significant issues with the potential for drugdrug interactions if this “optimal”
regimen was followed. They concluded that pay for performance based on CPGs for single disease states may create “perverse incentives” and actually decrease quality of care and that significant effort should be placed on effectively treating the multiple chronic conditions that many of our patients have.
While these initiatives are only still beginning, there will likely be future pressures both at the state and federal level to achieve an ever-higher level of quality. In addition, pay for performance needs to evolve to incorporate these early findings into rationale paradigms for physicians, which align the needs of our patients, payers, and providers. It may also become quite difficult for groups and practices that do not have access to electronic data to be able to participate in such projects.