A Comprehensive Value Equation for a Comprehensive Cancer Center
Over the past decade, many forms of the value equation have emerged from the original “Value = Outcomes/Cost” framework promulgated by Professors Michael Porter, Elizabeth Teisberg, and Robert Kaplan of the Harvard Business School. Health care systems, hospital alliances, government health care funders, and even individual providers and health care administrators within the same institution have developed their own versions of the value equation. For example, many health systems have emphasized the concept of “high reliability,” interpreting the value equation as “Value = Safety/Cost.” However, by focusing the numerator only on harm, this form of the equation disregards the value of the positive outcomes that are sought from health care encounters. Alternative proposals have emphasized patient experience in the denominator, to the exclusion of either quality or safety. However, other versions combine these concepts:
“Value = (Quality + Outcomes)/Cost”
“Value = (Outcomes + Patient Experience)/(Direct + Indirect Cost)”
“Value = (Quality + Service)/Cost”
Each of the above variations of the value equation has fundamental flaws that prevent meaningful measurement of value. The major problem with each version is that the terms are nebulous, leaving too much room for interpretation when precision is needed. What is meant by safety? What is quality? Which costs should be included? Likewise, these equations do not always measure value from a patient’s perspective. Frequently, the denominator of costs has been interpreted as institutional costs of provisioning care or third-party reimbursement when patient-centered value focuses only on out-of-pocket patient expenditures for health care, including copays and deductibles, but also on travel expenses and lost wages.
Adding to the already challenging task of developing measurement for a multidimensional equation is the complexity of cancer care and the heterogeneity of the patient population. Each type of cancer is not a single disease but a conglomerate of thousands of diseases with distinct characteristics that carry varying risks and require different treatments. Even within the same cancer disease type, there is variability in the genomic pathways that require individual considerations. Overlying the oncologic variability is variability in our patients’ social determinants of health and their subjective outcome priorities. Faced with this compounded variability, our challenge was to develop a value equation that is inclusive of all the important domains while remaining scalable across a variety of diseases, perspectives, and priorities.
Defining Quality and Safety: Outcomes That Matter to Patients
To achieve this objective, we first expanded the terms in the numerator to contrast the outcome of “Quality” (defined as the achievement of a positive outcome) against the outcome of “Safety” (defined as the avoidance of a negative outcome). Here, “Safety” is mathematically equivalent to 1/“Harm.” Although interrelated, the terms quality and safety are very different. Quality is the indication for the treatment, the reason we perform therapeutic intervention, while safety is what we do during the intervention to avoid harm. For almost every medical intervention (including drug therapies), at the time the intervention is initiated, the patient and provider cannot see the quality of the intervention (e.g., long-term survival after cancer surgery) ( Fig. 53.1A ). However, they are definitely exposed to the risk of harm (e.g., postoperative thromboembolism).
As previously discussed, a commonly derived transformation of the value equation has been tailored for high reliability, expressed as “Value = Safety/Cost.” However, this version of the formula does not fully capture the impact of high reliability. The reason it is laudable to focus on a high-reliability environment is not only that it emphasizes the need to limit the risk for harm, but equally or more importantly, a narrowed risk for harm through high-reliability processes allows the patient and the provider to see the quality objective sooner and to a greater extent ( Fig. 53.1B ). In an error-prone, low-reliability environment, or in the case of chance complications, it is well understood that even one adverse event can completely erase the possibility of experiencing the intended quality of the intervention ( Fig. 53.1C ). The fact that 90% of “Quality measures” in health care are actually metrics describing harm, and we seldom measure and report on the positive quality outcomes of treatment, is a primary driver of provider burnout in medicine. Based on these and other arguments, it is imperative that the value equation speak to both positive qualities and potential harm.
To account for both sides of this coin, we operationalized the numerator of the value equation as “Quality minus Harm.” With these two domains defined, the patient can start to appreciate both the positive and negative attributes of the proposed intervention and to compare one therapy to another or one providing institution to another for the same therapy. This configuration likewise feels natural to providers, as it is the embodiment of the modern consent process, where we describe the benefits, risks, and alternatives of an intervention to the patient and their associated caregivers, allowing all parties to determine whether the balance favors (or not) the proposed procedure or treatment.
The Full Value Equation
With the foundation of “Value = Quality minus Harm over Cost,” we sought to define and measure the individual components of each of these three terms. To accomplish this, we interviewed patients, caregivers, and providers. We expected to find a vast array of qualities and harms that spoke to patients and providers alike. However, ultimately (and conveniently) all of the feedback we received could be folded under three primary qualities that patients seek from cancer care and three specific areas of harm that they seek to avoid ( Fig. 53.2 ). In the next section, we define each of these subcategories of value.
Similar to the traditional basic form of the value equation, the University of Texas M.D. Anderson Cancer Center (UTMDACC) value equation is patient-centric, focusing on the value of care provided from the patient’s perspective. However, unlike other forms of the value equation, this equation considers the factors that impact decisions from the perspective of all stakeholders, defines measurement for each component, and identifies their respective data sources to enable actual calculations and comparisons of care value. Additionally, the universality of the equation allows value calculations to easily transition between different cancers while still maintaining meaningful insights to inform decisions. The analyses are also easily dissectible to concentrate on specific patient populations, disease types, and outcome sets. The equation can be similarly utilized in noncancer-related care settings.
Defining the Value Equation
The Numerator: Quality
Based on our qualitative interviews and quantitative feedback, it is clear that patients seek three qualities from their cancer care: survivals, functional recovery, and a positive experience (PE) .
Ninety percent of surveyed cancer providers reported that survival is the number one priority of our patients. Interestingly, only 50% of patients ranked survival as their first priority, as many more patients than might be expected were focused on returning to pressing life responsibilities that cancer had “interrupted” and/or maintaining their independence in the vein of “Quality of life over Quantity of life.” Even patients who rank survival first, when pressed why their priority would be the achievement of a milestone birthday (e.g., So your first priority is to make it to 80 years old?). Most reply that the desire for longer survival is actually only driven by the observation that they “have more to do in life.” In other words, functional recovery from illness (discussed later) is clearly the primary goal of cancer care for the vast majority of our patients, and perhaps the most underaddressed value proposition in health care . This having been said, cancer survival remains a crucial and highly measurable Quality metric. It can be calculated from when a patient is diagnosed or starts treatment for a specific cancer to the date of recurrence, progression, or death. Leveraging our Tumor Registry, which contains patient follow up data on over 1.5 million patients beginning in 1944, analyses are performed to determine long-term cancer survival after care at UTMDACC, which can then be benchmarked against equivalent populations in national datasets, as demonstrated in Fig. 53.3 .
Almost all patients come to medical care with some level of disease-induced disability. Patients with malignancy are no exception, as they are frequently diagnosed in a debilitated state with cachexia, malnutrition, infection, immunosuppression, anxiety, and cancer-related pain. Unfortunately, most cancer treatments further impair the patient’s ability to perform desired activities by inducing symptoms and side effects. This sinister “double-hit” leaves patients with substantial levels of dysfunction. In part, this explains why functional recovery is the primary goal of cancer patients, as described earlier. It also implies that in addition to wearing badges that read “Surgical Oncology,” “Medical Oncology,” or “Radiation Oncology,” all cancer providers are also rehabilitation specialists. In fact, functional recovery is so preeminent in patient-centric value that a health care enterprise’s ultimate true value could be summed to the degree to which it recovers patients from illness . Rapid recovery also speaks to the value employers place on their employees rapidly returning to the workforce. Despite the preeminence of this value proposition, it is remarkable that functional recovery data are rarely found in our historical written medical records or our current electronic billing records (EHRs). This is because the only mechanism to assess functional recovery is to ask the patients how functional they are in their daily lives. Not only have we paternalistically never asked, we poorly recorded our own impressions of recovery via simple tools such as the Karnofsky status or Eastern Cooperative Oncology Group (ECOG) performance status. , Fortunately, we are evolving, and the field of patient-reported outcomes (PROs) has exploded into medical research and care, allowing patients to voice and numerically score their degree of recovery. , Only by using PRO surveys at strategic time points (e.g., preintervention and 14, 30, and 90 days postintervention) in clinical care can we measure our ability to achieve short-term functional recovery after interventions. Validated PRO instruments tend to be dominated by measures of symptom burden but the best tools for measurement of functional recovery contain measures of life interference from symptoms. Our research shows that symptoms are variable, difficult to impact, and frequently affected by psychologic overtones. The goal of enhanced recovery programs is to facilitate function with symptoms. Therefore PRO instruments that include questions about walking, work, ability to care for self, ability to care for others, and enjoyment of life serve as actionable dependent variables for hospitals to measure the value of their function recovery efforts. Armed with these measures of success, enhanced recovery programs have entered the cancer care environment with ubiquitous benefits that impact multiple areas of the value equation.
The third quality patients seek from cancer care is a positive experience (PE). PE is sometimes conflated with PROs, but symptoms and function are distinct from feelings of respect, courtesy, and open communication, as measured by patient experience and satisfaction surveys. An increasing number of institutions are collecting patient feedback through patient satisfaction reports, feedback surveys, and patient and family advisory councils. These data points are assessed to rate an important part of the success of the service provided to patients along their cancer care journey.
The Numerator: Safety/Harm
Every cancer intervention has the potential to harm the patient before the benefits of treatment are ever realized. These unintended negative results, which may be preventable, include acute complications of care, pain, and long-term disutility . Both mathematically and qualitatively, if the harm encountered exceeds the quality, the numerator crosses into the negative and value is extinguished.
Complications of Care
The medical profession is awash in short-term complication data that are produced locally, administratively, and even nationally by regulatory bodies. Surgeons, medical oncologists, radiation therapists, and all other care specialists collect and obsess over short-term harm data such as postoperative complications, toxicity grades, readmissions, and hospital-acquired infections. Ideally, these data are risk-adjusted for fair comparison, but frequently we have found even raw complication rate data to be actionable. Despite the availability of these data, it has only recently been reported more transparently to patients and caregivers.
Patient pain scores are ubiquitously collected throughout all facets of cancer care. Despite the availability of these data, substantial progress, and the development of medical specialties in pain management, pain continues to be a principle negative outcome for patients. More than 50% of patients receiving cancer treatment report severe pain during their continuum of care. For this reason, we believe that pain warrants a unique place in the value equation and hope that calling it out will bring more attention to its profound impact on patients’ quality of life. There is no doubt that intentional programs that successfully reduce pain to levels that permit function improve the value of cancer care.
Much like short-term functional recovery, the measurement of long-term disutility from cancer care has been an underaddressed issue. As cancer treatments improve, patients live longer and rightfully demand durable functional recovery. Therefore the collection of long-term functional recovery data, again via PROs, has become a requisite to measure value. Urologic oncology providers who track long-term bowel, bladder, and sexual function after prostate surgery and radiation have demonstrated the feasibility of capturing these data and having them inform treatment decisions. The widespread collection of these data will be a heavy lift for patients and providers but is required to completely factor the numerator of the value equation. To accomplish this, institutions are encouraged to form PRO Governance Committees to oversee the build of validated PROs into the EHR and to ensure that they take into consideration patient survey fatigue, changes to clinical workflow, and other factors that are critical to successful PRO implementation and data collection.
The Numerator: Accounting for All Three Tiers of Value
As originally proposed by Porter and Teisberg, value should be measured across time using the concept of three tiers of outcomes measures hierarchy. , Adhering to these principles, our detailed value formula incorporates the dimension of time by including short-term complications with long-term disutility, and both early and late survival with functional recovery.
Like the term “Outcomes,” the value denominator term of “Cost” has been equally opaque. In one dimension, we must grapple with determining if we are describing charges, prices, collections, acquisition costs, costs of provisioning care, and/or amortization of indirect costs to overall costs. In another dimension, we are compelled to assume the perspective of the patient (who measures out-of-pocket payments and largely disregards third-party payer contributions or the institutional costs of provisioning care). At the same time, population-level value measurement requires the inclusion of third-party payor/society financial contributions and the institutions’ expense side of the balance sheet. In a pure patient-driven value equation, we only concern ourselves with patient-borne cost, but holistically, we cannot effect positive change in the health care landscape without including the payor and provider perspectives. Given these seemingly conflicting priorities, we quarantined the denominator of our value equation to describe real dollars in the form of collections from the patient (patient-borne cost) and third-party payors (including private, nonprofit, and governmental sources), as well as the institutional cost for the hospital system to provide care . The institutional cost of care is optimally captured using time-driven activity-based costing (TDABC) methodologies but in the absence of this capability, financial data may also be acquired from traditional hospital costing systems. Importantly, charges are not included in any component of the equation, as these are not real dollars; they reflect inflated pricing of services and are subject to deep contractual discounts. The avoidance of the “monopoly money” of charges allows the Value equation to rigidly report on real dollars in the form of collections, and to subtract the institutional Cost to provision care yielding the margin. Deep knowledge of margin data is critically important for the future establishment of fair bundled payment pricing strategies.
Due to multiple factors, including the emergence of high-deductible insurance products and narrow network policies, the percentage and total amount of out-of-pocket health care costs that are being paid directly by patients are rising. Although every state, demographic, and socioeconomic group has been impacted, the low- and middle-class wage earners have been disproportionately impacted by these trends, explaining the fact that medical costs account for over 60% of bankruptcies in the United States and multitudes of patients are avoiding necessary care in order to avoid financial toxicity. Due to the complex nature of cancer, and extremely high-priced treatments (including robotic surgery, proton radiation therapy, and immunotherapy) financial toxicity for cancer patients is a serious concern. A portion of the patients’ burden of out-of-pocket expenses can be determined by analyzing payments received directly from patients for care in the form of copays, deductibles, and/or coinsurance. However, travel costs and reduced income from lost work time cannot currently be factored into the equation. In the future, transparent disclosure of out-of-pocket estimates within the full framework of the value formula prior to the initiation of therapy is required to avoid catastrophic surprise billing and to allow for comprehensive shared decision-making. Previously, hospital systems have avoided these disclosures by hiding behind the perceived complexity of accurate determination of out-of-pocket cost estimates; however, with modern EHRs, this capability is now available ubiquitously. The time is now for every patient to know what out-of-pocket expenses they can anticipate from medical care.
Third-Party Payment Cost
Reimbursement for care from third parties often involves complex contracts that outline discounts, plan maximums, stop-loss provisions, and nontransparent accounting. Given that most institutions accept a vast array of plans, the simplest method to calculate third-party payor contributions is to determine the actual payments received for care based on collections via the institutional accounting system. Modern systems can define these collections to the penny.
Determining the actual cost of providing care to an individual patient has historically been elusive for all hospitals and remains a challenge. Fully loaded costs encompass goods, drugs, equipment, personnel, services, time, information technologies, capital purchases, and a multitude of indirect costs that require attribution and allocation. In the absence of a definitive strategy to accurately assess the true costs of medical care, many bodies, including quality reporters such as Vizient, are forced to impute costs of care by applying a ratio of costs to charges based on each institution’s Medicare Cost report. Although this methodology provides a rough surrogate estimate of resource utilization, it is an inaccurate way of representing the health care costs associated with cancer care because it neither factors in overhead costs associated with the facilities, administration, nonbillable items, and so on, nor considers the actual flow of expenses and revenue. TDABC is a more accurate accounting method for measuring the cost of care. It utilizes process maps to capture the resources actually attributable to patient care. Fully loaded cost, which accounts for direct and indirect costs of care, is then applied to each process to calculate the cost for the cycle of care. Although historically tedious to fully implement, electronic tools have substantially increased the pace and ability to accurately measure costs of care using the TDABC methodology.
The Value Equation: Summary
Fully realized, the Value equation is represented by three major domains as “Quality minus Harm over Costs.” Each of the three main groups has three subcategories that are easily understood and measurable. The six numerator values account for all three tiers of outcome measures hierarchy, highlighting critically important factors for patients such as functional recovery and patient experience. Notably, two of the six factors (short-term functional recovery and long-term disutility) are dependent on PROs. For isolated patient-centric value, third-party payor contributions and institutional costs of care are dropped from the denominator, leaving seven total factors, as the denominator is limited only to the patient-borne expenses. When used for population health, payment reform, and institutional quality improvement is initiated, all nine factors are entered into the equation ( Fig. 53.2 ).
Many of the data sources for these nine elements are currently available in most health care systems (e.g., collections), whereas others require a build (e.g., PRO platforms). For the most part, the heavy institutional lift is simply the channeling of existing data streams to a central value group to assimilate them across patient types, diseases, stages, and interventions. The goal of this work is to integrate all of the data streams into the formula for transparent visualization at the patient encounter. Investing in and resourcing the QA capabilities for collecting and reporting these data can add expense, but has a tremendous return on investment (ROI). We predict that institutions that achieve this capability will have a substantial competitive advantage both in the ability to market their value proposition and negotiate premium reimbursement rates that reflect the full spectrum of the value they provide ( Table 53.1 ).