Why the quality measures used in health care are deeply flawed

Nobody stands up to argue against quality and value in health care. You might as well argue against motherhood, or puppies. Yet many physicians are inherently skeptical of definitions of “quality” that are imposed from above, whether by outside evaluators like The Joint Commission, or (worse) by the government.

There’s good reason for skepticism. Some of the “evidence” behind “evidence-based medicine” has turned out to be flawed, tainted by financial conflict of interest, or outright fraudulent. Any experienced physician knows that there are fads in health care just as there are in fashion, and today’s evidence-based medicine may be tomorrow’s malpractice. Let’s take a closer look at what’s really going on in the world of quality metrics, and why it matters if payments to you and your hospital are increasingly linked to how you score.

Surgical site infections

The financial toll of surgical site infections (SSIs) is huge, estimated in the U.S. at more than $10 billion a year. A recent retrospective review from the Veterans Affairs Surgical Quality Improvement Program showed that the majority of SSIs are diagnosed only after hospital discharge, and that 57% will require hospital readmission within 30 days.  The Centers for Medicare & Medicaid Services (CMS) stopped paying for care related to SSIs in 2008 by designating them as “never events,” or non-reimbursable serious hospital-acquired conditions. Now SSIs are part of a long list of hospital-acquired conditions that can result in reduced CMS payments to hospitals, and will bring further reduction in payments over the next several years with the implementation of “value-based purchasing.” More than 1400 hospitals will see their Medicare payments cut by as much as 1.25% this year — a margin that could spell financial disaster for hospitals already struggling.

You may already be among the more than 50% of anesthesiologists who have been reporting performance metrics to the Physician Quality Reporting System (PQRS), which is administered by CMS. When the system started in 2007, CMS offered a bonus payment of 1.5% for successful participation, but that soon shrank to 0.5% and will be discontinued after 2014. Starting in 2015, CMS will impose a 1.5% payment reduction for physicians who do not participate in PQRS, and will push the pay cut to 2% in 2016.

If you participate in PQRS reporting, you know that two of the measures that anesthesiologists report are directly aimed at SSI prevention: perioperative temperature management, and antibiotic timing. PQRS measure #193 specifies that the patient must receive “active warming” or have a temperature above 36C recorded within 30 minutes before or 15 minutes after anesthesia end time. Measure #30 specifies that prophylactic parenteral antibiotics must be administered within one hour before skin incision. Compliance with these two measures isn’t hard to achieve, though no one seems to question the cost to the American healthcare system of all those forced-air warming blankets and machines, or ask why giving antibiotics 61 minutes instead of 59 minutes before skin incision is an automatic “fail.”

But have CMS threats and PQRS compliance done any good? A just-published editorial in Anesthesiology concluded: “Despite early efficacy literature establishing the value of specific antibiotic timing and active warming, repeated large database analyses have not observed robust effectiveness across hundreds of hospitals.”   Simply put, as many of us have noticed in our own hospitals, SSI rates have remained about the same.

What about skin preparation?

Did your hospital, like so many, abruptly switch from povidone-iodine antiseptic solution to ChloraPrep®? If so, a 2010 study from the New England Journal of Medicine, “Chlorhexidine-Alcohol versus Povidone-Iodine for Surgical Site Antisepsis,” was probably cited as proof that ChloraPrep would be more effective. No doubt, the staff was warned of the increased risk of operating room fires with ChloraPrep, since it contains 70% isopropyl alcohol in addition to chlorhexidine. But the future benefit in reducing SSIs would outweigh the risk of fire, or so the story went.

Imagine our surprise when the U.S. Department of Justice announced in January that ChloraPrep’s manufacturer, CareFusion Corp., would pay the government a $40.1 million settlement to resolve allegations that the company violated the False Claims Act by paying kickbacks to boost sales of ChloraPrep and promoting it for uses that aren’t FDA-approved.

Who received kickbacks? According to the DOJ’s press release, the complaint alleged that “CareFusion paid $11.6 million in kickbacks to Dr. Charles Denham while Denham served as the co-chair of the Safe Practices Committee at the National Quality Forum, a non-profit organization that reviews, endorses, and recommends standardized health care performance measures and practices.” Another physician with close ties to CareFusion, Dr. Rabih Darouiche, was the lead investigator on the NEJM article, and one co-author, Cynthia Crosby, was an employee of Cardinal Health, the parent company of CareFusion.

The Leapfrog Group, launched by the Business Roundtable in 2000, claims that its hospital survey is “the gold standard for comparing hospitals’ performance on the national standards of safety, quality and efficiency.” On January 30, Leapfrog announced that it accepted the resignation of Dr. Denham, who had served as chair of Leapfrog’s Safe Practices Committee since 2006, amid concerns that Dr. Denham had failed to reveal his “potentially compromising relationship with CareFusion.” At the same time, Leapfrog said it would undertake “a thorough scientific review of its full slate of endorsed safe practices.”

Meanwhile, the Washington State Surgical Care and Outcomes Assessment Program was busy between 2011 and 2012 conducting a prospective cohort analysis comparing commonly used skin antiseptic agents for patients undergoing clean-contaminated surgery (including colorectal cases).  Among the 7,699 patients (a much larger group than the 849 in the NEJM study), the researchers compared chlorhexidine, chlorhexidine in isopropyl alcohol, povidone-iodine, and iodine-povacrylex in isopropyl alcohol. Their conclusion: “For clean-contaminated surgical cases, this large-scale cohort study did not demonstrate superiority of any commonly used skin antiseptic agent in reducing the risk of SSI, nor did it find any unique effect of isopropyl alcohol. These results do not support the use of more expensive skin preparation agents.”

In other words, we don’t yet have the “magic bullet” that will reduce the rate of SSIs to zero. Simplistic efforts to link SSI rates to “quality” measures such as active warming, antibiotic timing, or using one type of skin antiseptic, fail to take into account obvious risk factors such as diabetes, socioeconomic status, obesity, and poor general health. Looking at the larger issue of postoperative complications and 30-day hospital readmission rates, a recent editorial in JAMA Surgery concluded, “Accounting for these differences may be a critical prerequisite before penalizing hospitals that disproportionately and admirably care for populations at higher risk for readmission.”

The saga of tight glucose control

The Surgical Care Improvement Project (SCIP) is responsible for defining many of the quality measures that your hospital probably tracks. These are subsequently endorsed by the National Quality Forum (where Dr. Denham formerly headed the Safe Practices Committee) and included in The Joint Commission’s Specifications Manual for hospital quality core measures. One measure, known as SCIP-INF-4, tracks the percentage of “cardiac surgery patients with controlled postoperative blood glucose.” To meet this quality target, a cardiac surgery patient must have a blood glucose less than or equal to 180 mg/dl “in the timeframe of 18 to 24 hours after anesthesia end time.”

The rationale for this quality measure is that hyperglycemia has been associated with increased in-hospital morbidity and mortality for multiple medical and surgical conditions. The SCIP measure has been revised several times, and the blood glucose target hasn’t always been 180. Starting around 2001, there was an international push toward much tighter control of blood glucose levels in ICU patients, aiming for a target of 80-110 mg/dl. The evidence basis for tight control was a single-center study that associated intensive insulin therapy with all sorts of improved outcomes, from fewer infections to less time on the ventilator and a lower incidence of acute renal failure.

The only problem was that those results couldn’t be replicated, and in fact more patients on tight glucose control ended up dead. In a landmark 2009 project now known as NICE-SUGAR, a multi-center, multi-national group of 42 hospitals including the Mayo Clinic conducted a randomized study of 6100 patients. Compared with conventional management and glucose level maintained at less than 180, the patients on intensive insulin therapy with glucose levels kept between 81 and 108 were shown to have more hypoglycemia, higher mortality, and no difference in morbidity or length of stay. Suddenly tight glucose control was no longer the rage.

In March 2014, the Journal of Thoracic and Cardiovascular Surgery published an article with the straightforward title, “Surgical Care Improvement Project measure for postoperative glucose control should not be used as a measure of quality after cardiac surgery. The authors looked at outcomes from 1703 patients, 90% of whom achieved postoperative 6 a.m. blood glucose levels of less than 200. No significant differences were observed in major morbidity, mortality, or resource usage among the propensity-matched cohorts.   Concluded the authors, “These data suggest that this metric might not be a valid measure of postoperative surgical quality.”

Fraud, stroke, and beta-blockade

SCIP has defined beta-blocker use in non-cardiac surgery patients as another core measure of hospital quality. SCIP-CARD-2 tracks the percentage of “surgery patients on beta-blocker therapy prior to arrival who received a beta-blocker during the perioperative period.” The medication must be taken, SCIP says, either the day before or the day of surgery, and on either postoperative day 1 or 2.

Is there a controlled study that demonstrates conclusively that the SCIP beta-blocker measure makes surgical patients healthier? Or that they recover better from surgery? No, there isn’t. Does the SCIP measure track whether the patient’s blood pressure or heart rate are maintained within a normal range? No, it doesn’t. All it says is that the patient should receive an unspecified dose of one of a number of drugs in the beta-blocker class. A physician can be “compliant” with the SCIP measure by prescribing only a tiny dose of esmolol. It doesn’t take a cardiologist to determine that this is ridiculous. What’s worse is that aggressive beta-blocker use around the time of surgery may be hazardous. A 2013 review and meta-analysis of the medical literature concluded that starting beta-blockers before surgery caused a 27% increase in 30-day all-cause mortality.  Beta-blockade reduced the incidence of heart attacks, but increased the incidence of stroke and significant hypotension.

The authors of SCIP-CARD-2 cite scientific evidence to support their recommendations. Unfortunately, much of the science has been discredited. Dr. Dan Poldermans, a prolific Dutch researcher, led the DECREASE-IV study, which is no longer considered credible. The authors of the previously cited meta-analysis remarked about DECREASE-IV that “the key data required to judge outcomes were missing and the adjudication committee was fabricated.”

Larry Husten of Forbes.com has been reporting the damage caused by the Poldermans research misconduct with the overwrought headline: “Medicine or Mass Murder? Guideline Based on Discredited Research May Have Cause 800,000 Deaths in Europe Over the Last 5 Years.” Even if Husten exaggerates, the SCIP beta-blocker measure certainly warrants critical reexamination. Good control of a patient’s blood pressure and heart rate is common sense, and another medication might be better than a beta-blocker for a specific patient’s perioperative needs.

It’s complicated: Lessons from primary care

Improving quality in primary care seems to be just as tough as it is in the perioperative world. A major article in JAMA recently reported on the three-year experience of trying to transform primary care practices in Pennsylvania, with the goals of improving quality and containing costs. The pilot practices received technical assistance to establish registries to identify patients overdue for chronic disease services, create monthly quality indicator reports, and implement electronic prescribing. They were encouraged to establish a “team-based model” of primary care practice, and received annual bonus payments per full-time equivalent clinician ranging from $28,000 to $95,000.

Electronic prescribing increased from 38% to 86%, and the use of registries to identify patients overdue for chronic disease services increased from 30% to 85%. Of note, six of the pilot practices were “nurse-managed health centers.” Pilot participation was significantly associated with greater performance improvement on only one quality measure: nephropathy monitoring in diabetes. On 10 other quality measures — including screening for cervical cancer and colon cancer, and monitoring LDL and Hba1c — the pilot practices showed no statistically significant performance improvement relative to comparison practices. Perhaps most important, the results demonstrated noreduction in hospital admissions, emergency department visits, ambulatory care services, or total costs.

The results were a disappointment to the Obama administration, which has heavily promoted the “patient-centered” primary care model, as the Wall Street Journal pointed out on in a February 25 article.  The WSJ quoted Dr. Mark W. Friedberg, a researcher at RAND Corp. and lead author of the study, who said people have been convinced that the new model “is a proven intervention that doesn’t even need to be tested or refined. Our findings will hopefully change those views.” An accompanying editorial in JAMA noted that the model “has been promoted for widespread adoption, using a fairly generic and fixed set of structural practice features, even before being fully developed in targeted high-risk populations or before clearly understanding which features or combination of features are most effective with which patients. It is time to replace enthusiasm and promotion with scientific rigor and prudence.”

What should we conclude?

Scientific rigor and prudence are standards that have been sadly lacking in the evolution of a number of “quality measures,” from SSI prevention to perioperative beta-blockade and tight glucose control. One notable exception is the reduction of infection rates in central venous catheters placed in the ICU, with consistent use of a simple standardized protocol of hand-washing, full barrier precautions, skin cleansing with chlorhexidine, avoidance of the femoral site, and prompt removal of unnecessary catheters. These were common-sense measures that resulted in immediate and sustained reduction in ICU central line infections and the enormous cost of treating them.

Dr. Peter Pronovost’s success in reducing catheter-related infections is a good illustration of how standardizing routine, well-defined processes can improve quality and reduce errors. This is a basic business principle with broad application to any process that can be broken down into components, analyzed, modified, studied, and improved. It is also a guiding principle behind the Perioperative Surgical Home concept, which aims to take the same kind of rigorous process analysis, led by anesthesiologists, and apply it to the care of patients before, during, and after surgery.

Unfortunately, the direction CMS is taking — led by a nurse (Marilyn Tavenner) rather than a physician — lacks any semblance of scientific rigor and prudence. The “value-based” system will increasingly link payments made to physicians and hospitals with patient satisfaction scores. While this may appeal to consumer advocates, the fact is that the delivery of medical care is not like running a restaurant, and the customer is not always right. A 2013 study in JAMA Internal Medicine reported that patient preferences for shared decision-making were associated with longer inpatient stays and with 6% higher total hospitalization costs. Dr. Frederick Greene, a nationally recognized general surgeon, wrote recently, “We have all experienced patients who demand certain drugs, imaging studies, surgical procedures and home care strategies that may be unwarranted, unhelpful or downright wrong!” Physician pay should not be tied to Press Ganey scores, Dr. Greene argued. “The only winners in this game of linking pay to patient satisfaction are the entities created to measure and promulgate these highly suspect data.”

CMS also plans to put higher percentages of hospital payment at risk in the event of poor clinical outcomes: hospital-acquired complications of care, and excessive 30-day mortality rates for acute MI, heart failure and pneumonia. The list of linked outcomes is certain to expand and include 30-day mortality, morbidity, and hospital readmission rates surrounding specific surgeries. This trend will incentivize community hospitals to funnel every complicated, frail, or high-risk patient to the nearest tertiary care center which won’t be able to turn them away. How CMS expects tertiary care public and teaching hospitals to remain financially viable is unclear.

Perhaps the only positive note to less reliance on fee-for-service payment is that we might see more appropriate clinical decision-making for high-risk patients before they are subjected to major invasive procedures. Too often these patients spend their last weeks of life enduring a series of painful interventions in the ICU, when conservative medical management might have been more appropriate than surgery. The time to have that conversation with a patient and family is before the episode of care begins, not in the preoperative holding area.

Otherwise, the future of the clinical anesthesiologist appears to include increasing attention turned away from the patient and toward a computer screen so that adherence to “quality measures” is documented to the satisfaction of CMS, NQF, SCIP, and any other acronym that hospital administrators want to appease. The “rules” will become increasingly arbitrary and less rational, as they seem to do with every Joint Commission visit. Humanistic, individualized medical care? We’ll see that, maybe, in old movies.

Karen S. Sibert is an associate professor of anesthesiology, Cedars-Sinai Medical Center, Los Angeles, CA.  She blogs at A Penned PointThis article originally appeared in the Anesthesia Business Consultants Spring 2014 Communique.

View 5 Comments >