Advances in Addressing Patient Heterogeneity in Economic Evaluation: A Review of the Methods Literature.

Cost-effectiveness analyses commonly use population or sample averages, which can mask key differences across subgroups and may lead to suboptimal resource allocation. Despite there being several new methods developed over the last decade, there is no recent summary of what methods are available to researchers. This review sought to identify advances in methods for addressing patient heterogeneity in economic evaluations and to provide an overview of these methods. A literature search was conducted using the Econlit, Embase and MEDLINE databases to identify studies published after 2011 (date of a previous review on this topic). Eligible studies needed to have an explicit methodological focus, related to how patient heterogeneity can be accounted for within a full economic evaluation. Sixteen studies were included in the review. Methodologies were varied and included regression techniques, model design and value of information analysis. Recent publications have applied methodologies more commonly used in other fields, such as machine learning and causal forests. Commonly noted challenges associated with considering patient heterogeneity included data availability (e.g., sample size), statistical issues (e.g., risk of false positives) and practical factors (e.g., computation time). A range of methods are available to address patient heterogeneity in economic evaluation, with relevant methods differing according to research question, scope of the economic evaluation and data availability. Researchers need to be aware of the challenges associated with addressing patient heterogeneity (e.g., data availability) to ensure findings are meaningful and robust. Future research is needed to assess whether and how methods are being applied in practice.

Risk Prediction Using Bayesian Networks: An Immunotherapy Case Study in Patients With Metastatic Renal Cell Carcinoma.

PURPOSE: To address the need for more accurate risk stratification models for cancer immuno-oncology, this study aimed to develop a machine-learned Bayesian network model (BNM) for predicting outcomes in patients with metastatic renal cell carcinoma (mRCC) being treated with immunotherapy. METHODS: Patient-level data from the randomized, phase III CheckMate 025 clinical trial comparing nivolumab with everolimus for second-line treatment in patients with mRCC were used to develop the BNM. Outcomes of interest were overall survival (OS), all-cause adverse events, and treatment-related adverse events (TRAE) over 36 months after treatment initiation. External validation of the model's predictions for OS was conducted using data from select centers from the International Metastatic Renal Cell Carcinoma Database Consortium (IMDC). RESULTS: Areas under the receiver operating characteristic curve (AUCs) for BNM-based classification of OS using baseline data were 0.74, 0.71, and 0.68 over months 12, 24, and 36, respectively. AUC for OS at 12 months increased to 0.86 when treatment response and progression status in year 1 were included as predictors; progression and response at 12 months were highly prognostic of all outcomes over the 36-month period. AUCs for adverse events and treatment-related adverse events were approximately 0.6 at 12 months but increased to approximately 0.7 by 36 months. Sensitivity analysis comparing the BNM with machine learning classifiers showed comparable performance. Test AUC on IMDC data for 12-month OS was 0.71 despite several variable imbalances. Notably, the BNM outperformed the IMDC risk score alone. CONCLUSION: The validated BNM performed well at prediction using baseline data, particularly with the inclusion of response and progression at 12 months. Additionally, the results suggest that 12 months of follow-up data alone may be sufficient to inform long-term survival projections in patients with mRCC.

An empirical investigation of time-varying cost-effectiveness across the product life cycle.

Cost-effectiveness is traditionally treated as a static estimate driven by clinical trial efficacy and drug price at launch. Prior studies suggest that cost-effectiveness varies over the drug's lifetime. We examined the impact of "learning by doing," one of the least studied drivers of changes in cost-effectiveness across the product life cycle. We combined time-series trends in effectiveness over time by cancer regimen using the Surveillance, Epidemiology, and End Results-Medicare database. We estimated the time-varying effects of treatments in colorectal and pancreatic cancer over their life cycle, including FOLFOX (leucovorin, 5-fluorouracil, and oxaliplatin) and gemcitabine, on survival of patients. Mean prices over time by strength and dosage form were calculated using historical wholesale acquisition costs. We found consistent downward trends in the mortality hazard ratios, which suggest that effectiveness improves over time. In the case of first-line FOLFOX for colorectal cancer, the implied incremental cost-effectiveness ratio based on the observational data fell from $610,000 per life year gained in 2004 to $27,000 per life year gained in 2011. Cost-effectiveness estimated at launch is unlikely to be representative of cost-effectiveness over the drug's lifetime. In the drugs studied, the impact of time-varying clinical effectiveness dominated the impact of changing prices overtime.

Maternal and Infant Adverse Outcomes Associated with Mild and Severe Preeclampsia during the First Year after Delivery in the United States.

OBJECTIVE: The burden of preeclampsia severity on the health of mothers and infants during the first year after delivery is unclear, given the lack of population-based longitudinal studies in the United States. STUDY DESIGN: We assessed maternal and infant adverse outcomes during the first year after delivery using population-based hospital discharge information merged with vital statistics and birth certificates of 2,021,013 linked maternal-infant births in California. We calculated sampling weights using the National Center for Health Statistics data to adjust for observed differences in maternal characteristics between California and the rest of the United States. Separately, we estimated the association between preeclampsia and gestational age and examined collider bias in models of preeclampsia and maternal and infant adverse outcomes. RESULTS: Compared with women without preeclampsia, women with mild and severe preeclampsia delivered 0.66 weeks (95% confidence interval [CI]: 0.64, 0.68) and 2.74 weeks (95% CI: 2.72, 2.77) earlier, respectively. Mild preeclampsia was associated with an increased risk of having any maternal adverse outcome (relative risk [RR] = 1.95; 95% CI: 1.93, 1.97), as was severe preeclampsia (RR = 2.80; 95% CI: 2.78, 2.82). The risk of an infant adverse outcome was increased for severe preeclampsia (RR = 2.15; 95% CI: 2.14, 2.17) but only marginally for mild preeclampsia (RR = 0.99; 95% CI: 0.98, 1). Collider bias produced an inverse association for mild preeclampsia and attenuated the association for severe preeclampsia in models for any infant adverse outcome. CONCLUSION: Using multiple datasets, we estimated that severe preeclampsia is associated with a higher risk of maternal and infant adverse outcomes compared with mild preeclampsia, including an earlier preterm delivery.

One size does not always fit all in value assessment.

Laying a clear path for incorporating reliable evidence on heterogeneity in value assessments could improve their applicability for healthcare decision making.

An empirical tool for estimating the share of unmet need due to healthcare inefficiencies, suboptimal access, and lack of effective technologies.

BACKGROUND: Although there has been growing attention to the measurement of unmet need, which is the overall epidemiological burden of disease, current measures ignore the burden that could be eliminated from technological advances or more effective use of current technologies. METHODS: We developed a conceptual framework and empirical tool that separates unmet need from met need and subcategorizes the causes of unmet need into suboptimal access to and ineffective use of current technologies and lack of current technologies. Statistical models were used to model the relationship between health-related quality of life (HR-QOL) and treatment utilization using data from the National Health and Wellness Survey (NHWS). Predicted HR-QOL was combined with prevalence data from the Global Burden of Disease Study (GBD) to estimate met need and the causes of unmet need due to morbidity in the US and EU5 for five diseases: rheumatoid arthritis, breast cancer, Parkinson's disease, hepatitis C, and chronic obstructive pulmonary disease (COPD). RESULTS: HR-QOL was positively correlated with adherence to medication and patient-perceived quality and negatively correlated with financial barriers. Met need was substantial across all disease and regions, although significant unmet need remains. While the majority of unmet need was driven by lack of technologies rather than ineffective use of current technologies, there was considerable variation across diseases and regions. Overall unmet need was largest for COPD, which had the highest prevalence of all diseases in this study. CONCLUSION: We developed a methodology that can inform decisions about which diseases to invest in and whether those investments should focus on improving access to currently available technologies or inventing new technologies.

The Lifetime Health Burden of Delayed Graft Function in Kidney Transplant Recipients in the United States.

Background. Although delayed graft function (DGF) is associated with an increased risk of acute rejection and decreased graft survival, there are no estimates of the long-term or lifetime health burden of DGF. Objectives. To estimate the long-term and lifetime health burden of DGF, defined as the need for at least one dialysis session within the first week after transplantation, for a cohort representative of patients who had their first kidney transplant in 2014. Methods. Data from the United States Renal Data System (USRDS; 2001-2014) were used to estimate a semi-Markov parametric multi-state model with three disease states. Maximum length of follow-up was 13.7 years, and a microsimulation model was used to extrapolate results over a lifetime. The impact of DGF was assessed by simulating the model for each patient in the cohort with and without DGF. Results. At the end of 13.7 years of follow-up, DGF reduces the probability of having a functioning graft from 52% to 32%, increases the probability of being on dialysis from 10% to 19%, and increases the probability of death from 38% to 50% relative to transplant recipients who do not experience DGF. A typical transplant recipient with DGF (median age = 53) is observed to lose 0.87 quality-adjusted life-years (QALYs). Extrapolated over a lifetime, the same 53-year-old DGF patient is projected to lose 3.01 (95% confidence interval: 2.33, 3.70) QALYs relative to a transplant recipient with the same characteristics who does not experience DGF. Conclusions. The lifetime health burden of DGF is substantial. Understanding these consequences will help health care providers weigh kidney transplant decisions and inform policies for patients in the context of varying risks of DGF.

The Economic and Social Burden of the 2014 Ebola Outbreak in West Africa.

Background: The 2014 Ebola virus disease (EVD) outbreak in West Africa was more devastating than all previous EVD outbreaks combined; however, many estimates of its economic burden did not capture its significant social costs. This study aimed to review all currently available estimates, identify social components missed by these estimates, and generate a comprehensive cost of the 2014 EVD outbreak. Methods: We conducted a systematic review of the grey (reports produced by nonprofit or nongovernmental organizations, government, or industry) and published literature to identify current estimates of the burden of the outbreak. Based on the findings of this review, we identified 6 key areas absent or underestimated from previous estimates and calculated the underestimated social costs using estimates from the literature and extrapolation. Results: Current estimates of the economic burden of the outbreak range from $2.8 to $32.6 billion in lost gross domestic product. We estimated the comprehensive economic and social burden from the 2014 EVD outbreak to be $53.19 billion (2014 USD). The most significant component, $18.8 billion, was deaths from non-Ebola causes. Conclusions: A more complete understanding of the burden of the 2014 EVD outbreak underscores the value of interventions that can mitigate or reduce the severity of future outbreaks.

Assessing Variation in the Cost of Palivizumab for Respiratory Syncytial Virus Prevention in Preterm Infants.

BACKGROUND: The variability in cost of palivizumab treatment, indicated for prevention of respiratory syncytial virus (RSV) infections in high-risk infants, has not been robustly estimated in prior studies. This study aimed to determine the cost variations of palivizumab from a US payer perspective for otherwise healthy preterm infants born 29-35 weeks gestational age (wGA) using infant characteristics and applied dosing regimens. METHODS: Fenton Growth Charts were merged with World Health Organization Child Growth Standards to estimate preterm infant growth patterns. The merged growth chart was applied to infants who received palivizumab from a prospective, observational registry to determine future body weight using each infant's wGA and birth weight. Using quarter 3 (Q3) 2016-Q2 2017 vial cost, treatment costs at monthly dosing intervals were estimated using expected weights and averaged by age to derive expected mean 2016-2017 RSV seasonal costs per infant under various dosing scenarios. RESULTS: Given different dosing scenarios (two to five doses), birth month, and growth patterns for preterm infants 29-35 wGA, the estimated average 2016-2017 seasonal cost of palivizumab treatment ranged from $3221 to $12,568. Outpatient-only cost (excluding first dose at hospital discharge) ranged from $1733 to $11,862. The main drivers of costs were dosing regimen (74% of variance), dosing interacted with birth month (17%), and wGA (6%). CONCLUSION: The considerable variability in the average cost of palivizumab treatment for preterm infants is driven by choice of dosing regimen, wGA, and birth month. Therefore, when estimating the cost of palivizumab, it is important to consider both infant characteristics at each dose and potential dosing regimens.

Amazon plant diversity revealed by a taxonomically verified species list.

Recent debates on the number of plant species in the vast lowland rain forests of the Amazon have been based largely on model estimates, neglecting published checklists based on verified voucher data. Here we collate taxonomically verified checklists to present a list of seed plant species from lowland Amazon rain forests. Our list comprises 14,003 species, of which 6,727 are trees. These figures are similar to estimates derived from nonparametric ecological models, but they contrast strongly with predictions of much higher tree diversity derived from parametric models. Based on the known proportion of tree species in neotropical lowland rain forest communities as measured in complete plot censuses, and on overall estimates of seed plant diversity in Brazil and in the neotropics in general, it is more likely that tree diversity in the Amazon is closer to the lower estimates derived from nonparametric models. Much remains unknown about Amazonian plant diversity, but this taxonomically verified dataset provides a valid starting point for macroecological and evolutionary studies aimed at understanding the origin, evolution, and ecology of the exceptional biodiversity of Amazonian forests.

Short-term costs of preeclampsia to the United States health care system.

BACKGROUND: Preeclampsia is a leading cause of maternal morbidity and mortality and adverse neonatal outcomes. Little is known about the extent of the health and cost burden of preeclampsia in the United States. OBJECTIVE: This study sought to quantify the annual epidemiological and health care cost burden of preeclampsia to both mothers and infants in the United States in 2012. STUDY DESIGN: We used epidemiological and econometric methods to assess the annual cost of preeclampsia in the United States using a combination of population-based and administrative data sets: the National Center for Health Statistics Vital Statistics on Births, the California Perinatal Quality Care Collaborative Databases, the US Health Care Cost and Utilization Project database, and a commercial claims data set. RESULTS: Preeclampsia increased the probability of an adverse event from 4.6% to 10.1% for mothers and from 7.8% to 15.4% for infants while lowering gestational age by 1.7 weeks (P < .001). Overall, the total cost burden of preeclampsia during the first 12 months after birth was $1.03 billion for mothers and $1.15 billion for infants. The cost burden per infant is dependent on gestational age, ranging from $150,000 at 26 weeks gestational age to $1311 at 36 weeks gestational age. CONCLUSION: In 2012, the cost of preeclampsia within the first 12 months of delivery was $2.18 billion in the United States ($1.03 billion for mothers and $1.15 billion for infants), and was disproportionately borne by births of low gestational age.

Value of improved lipid control in patients at high risk for adverse cardiac events.

OBJECTIVES: Lipid-lowering therapy (LLT) is suboptimally used in patients with hyperlipidemia in the 2 highest statin benefit groups (SBGs), as categorized by the American College of Cardiology and the American Heart Association. This study estimated the social value of reducing low-density lipoprotein cholesterol (LDL-C) levels by 50% for patients in SBGs 1 and 2 who have been treated with standard LLT but have not reached LDL-C goal, as well as the potential value of PCSK9 inhibitors for patients in these groups. STUDY DESIGN: Simulation model. METHODS: We used National Health and Nutrition Examination Surveys (NHANES) and US Census data to project the population of SBGs 1 and 2 in the time period 2015 to 2035. We used insurance claims data to estimate incidence rates of major adverse cardiac events (MACEs), and NHANES with National Vital Statistics data to estimate cardiovascular disease mortality rates. Using established associations between LDL-C and MACE risk, we estimated the value of reducing LDL-C levels by 50%. We incorporated results from a meta-analysis to estimate the value of PSCK9 inhibitors. RESULTS: Among those treated with LLT with LDL-C > 70 mg/dL in SBGs 1 and 2, the cumulative value of reducing LDL-C levels by 50% would be $2.9 trillion from 2015 to 2035, resulting primarily from 1.6 million deaths averted. The cumulative value of PCSK9 inhibitors would range from $3.4 trillion to $5.1 trillion (1.9-2.8 million deaths averted), or $12,000 to $17,000 per patient-year of treatment. CONCLUSIONS: Lowering LDL-C in high-risk patients with hyperlipidemia has enormous potential social value. For patients in these high-risk groups, PCSK9 inhibitors may have considerable net value depending on the final prices payers ultimately select.

The wider public health value of HCV treatment accrued by liver transplant recipients.

OBJECTIVES: Organs for transplantation are scarce, but new medical therapies can prevent organ failure and the need for transplants. We sought to describe the unique value created by treatments that spare organs from failure and thus conserve donated organs for transplant into others, using hepatitis C virus (HCV) as a case study. STUDY DESIGN: Epidemiologic-economic model. METHODS: Using data on trends in chronic liver disease, liver disease progression, and liver transplant allocation models, as well as the effectiveness of new HCV treatments, we estimate the potential effects of systematic HCV screening and treatment on the demand for liver transplants in the United States. We estimate the spillover benefits to patients with all-cause liver disease in terms of increased availability of transplants and life-years gained. RESULTS: We estimated that systematic HCV screening and treatment could spare 10,490 liver transplants to HCV-infected patients from 2015 to 2035. An estimated 7321 transplants would accrue to patients with end-stage liver disease without HCV and 3169 transplants to those with uncured HCV, providing approximately 52,700 and 22,800 additional life-years, respectively. CONCLUSIONS: Treatment advances for HCV have the potential to generate considerable spillover benefits to patients awaiting transplants for non-HCV-mediated liver failure. For other diseases in which organ transplants are in short supply, our study provides a novel pathway by which positive spillovers may accrue from treatments that prevent end-stage organ disease.

Estimating the future burden of cardiovascular disease and the value of lipid and blood pressure control therapies in China.

BACKGROUND: Lifestyle and dietary changes reflect an ongoing epidemiological transition in China, with cardiovascular disease (CVD) playing an ever-increasing role in China's disease burden. This study assessed the burden of CVD and the potential value of lipid and blood pressure control strategies in China. METHODS: We estimated the likely burden of CVD between 2016 and 2030 and how expanded use of lipid lowering and blood pressure control medication would impact that burden in the next 15 years. Accounting for the costs of drug use, we assessed the net social value of a policy that expands the utilization of lipid and blood pressure lowering therapies in China. RESULTS: Rises in prevalence of CVD risk and population aging would likely increase the incidence of acute myocardial infarctions (AMIs) by 75 million and strokes by 118 million, while the number of CVD deaths would rise by 39 million in total between 2016 and 2030. Universal treatment of hypertension and dyslipidemia patients with lipid and blood pressure lowering therapies could avert between 10 and 20 million AMIs, between 8 and 30 million strokes, and between 3 and 10 million CVD deaths during the 2016-2030 period, producing a positive social value net of health care costs as high as $932 billion. CONCLUSIONS: In light of its aging population and epidemiological transition, China faces near-certain increases in CVD morbidity and mortality. Preventative measures such as effective lipid and blood pressure management may reduce CVD burden substantially and provide large social value. While the Chinese government is implementing more systematic approaches to health care delivery, prevention of CVD should be high on the agenda.

The Rising Burden of Preeclampsia in the United States Impacts Both Maternal and Child Health.

OBJECTIVE: Preeclampsia is one of the top six causes of maternal mortality in the United States (US) and is associated with considerable perinatal morbidity and mortality. Evidence suggests the US incidence of preeclampsia has increased dramatically over the past two decades. This study aims to compile, summarize, and critique the literature on the health and economic burden of preeclampsia and early-onset preeclampsia. STUDY DESIGN: We reviewed the literature for estimates of burden of preeclampsia and early-onset preeclampsia to both mother and child, summarized the evidence on economic and social burden, and highlighted current gaps in the literature. RESULTS: No recent studies comprehensively assess the costs and health consequences of preeclampsia or early-onset preeclampsia for both mother and child. Where it exists, the literature suggests preeclampsia and early-onset preeclampsia cause numerous adverse health consequences, but these conditions currently lack effective treatment. The need for preterm delivery from early-onset preeclampsia suggests its costs are substantial: very (28-31 weeks) and extremely (<28 weeks) preterm birth cost approximately 40 and 100 times a term pregnancy, respectively. CONCLUSION: Given the severity of outcomes from preeclampsia, further research on its health and economic consequences is essential to inform policy and resource allocation decisions in health care.

Cancer mortality reductions were greatest among countries where cancer care spending rose the most, 1995-2007.

Health care spending and health outcomes vary markedly across countries, but the association between spending and outcomes remains unclear. This inevitably raises questions as to whether continuing growth in spending is justified, especially relative to the rising cost of cancer care. We compared cancer care across sixteen countries over time, examining changes in cancer spending and two measures of cancer mortality (amenable and excess mortality). We found that compared to low-spending health systems, high-spending systems had consistently lower cancer mortality in the period 1995-2007. Similarly, we found that the countries that increased spending the most had a 17 percent decrease in amenable mortality, compared to 8 percent in the countries with the lowest growth in cancer spending. For excess mortality, the corresponding decreases were 13 percent and 9 percent. Additionally, the rate of decrease for the countries with the highest spending growth was faster than the all-country trend. These findings are consistent with the existence of a link between higher cancer spending and lower cancer mortality. However, further work is needed to investigate the mechanisms that underlie this correlation.

Turning evidence into practice under payment reform: the new frontier of translational science.

Momentum is building to replace the current fee-for-service payment system with value-based reimbursement models that aim to deliver high quality care at lower costs. Although the goals of payment and delivery system reforms to improve quality and reduce costs are clear, the actual path by which provider groups can achieve these goals is not well understood, in large part because the role of identifying and discouraging the use of low-value, high-cost services and encouraging the use of high-value, low-cost services has traditionally fallen to health plans, not provider groups. The shifting focus towards provider accountability for costs and quality promises to expand the role of provider organizations from mainly delivering care to both delivering and prioritizing it based on costs and quality. We discuss how progress on two important but challenging fronts will be needed for provider groups to successfully translate evidence into value. First, robust evidence on the costs and benefits of treatments will need to be developed and made easily accessible to provider groups. Second, provider groups will need to translate that evidence into systems that support cost-effective clinical decisions.

A Cost-Benefit Analysis of Using Evidence of Effectiveness in Terms of Progression Free Survival in Making Reimbursement Decisions on New Cancer Therapies.

Payers increasingly require evidence of a statistically significant difference in overall survival (OS) for reimbursement of new cancer therapies. At the same time, it becomes increasingly costly to design clinical trials that measure OS endpoints instead of progression-free survival (PFS) endpoints. While PFS is often an imperfect proxy for OS effects, it is also faster and cheaper to measure accurately. This study develops a general cost-benefit framework that quantifies the competing trade-offs of the use of PFS versus that of OS in oncology reimbursement. We then apply this general framework to the illustrative case of metastatic renal cell carcinoma (mRCC). In the particular case of mRCC, the framework demonstrates that the net benefit to society from basing reimbursement decisions on PFS endpoints could be between $271 and $1271 million in the United States, or between €171 and €1128 million in Europe. In longevity terms, waiting for OS data in this case would result in a net loss of 3549-14,557 life-years among US patients, or 6785-27,993 life-years for European patients. While more stringent standards for medical evidence improve accuracy, they also impose countervailing costs on patients in terms of foregone health gains. These costs must be weighed against the benefits of greater accuracy. The magnitudes of the costs and benefits may vary across tumor types and need to be quantified systematically.